Pathogen‐reduced platelets for the prevention of bleeding

Summary of findings for the main comparison. Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding

Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding in thrombocytopenia
Patient or population: thrombocytopenia Settings: hospital Intervention: pathogen‐reduced platelets
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No. of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Control	Pathogen‐reduced platelets
Number of participants with 'any bleeding' event(s) (WHO grade 1 to 4 or equivalent) ‐ follow‐up > 7 days Follow‐up: 8 to 70 days	Study population		RR 1.09 (1.02 to 1.15)	1085 (5 studies)	⊕⊕⊝⊝ low^1,2	Random‐effects analysis: RR 1.14, 95% CI 0.93 to 1.38; I² = 59%
	696 per 1000	758 per 1000 (710 to 800)
	Moderate
	800 per 1000	872 per 1000 (816 to 920)
Number of participants with 'clinically significant' bleeding event(s) (WHO grade ≥ 2 or equivalent) ‐ follow‐up > 7 days Follow‐up: 8 to 70 days	Low		RR 1.10 (0.97 to 1.25)	1392 (5 studies)	⊕⊕⊕⊝ moderate²	Assumed risks from published data³
	570 per 1000	627 per 1000 (553 to 712)
	Moderate
	700 per 1000	770 per 1000 (679 to 875)
	High
	790 per 1000	869 per 1000 (766 to 988)
Number of participants with 'severe' bleeding event(s) (WHO grade ≥ 3 or equivalent) ‐ follow‐up > 7 days Follow‐up: 8 to 70 days	Study population		RR 1.24 (0.76 to 2.02)	1495 (6 studies)	⊕⊕⊕⊝ moderate⁴	Medium‐risk data taken from PLADO trial (Slichter 2010).
	36 per 1000	44 per 1000 (27 to 72)
	Moderate
	100 per 1000	124 per 1000 (76 to 202)
All‐cause mortality Follow‐up: 4 to 12 weeks	Study population		RR 0.81 (0.50 to 1.29)	1509 (6 studies)	⊕⊕⊕⊝ moderate⁵
	54 per 1000	43 per 1000 (27 to 69)
	Moderate
	25 per 1000	20 per 1000 (12 to 32)
Number of participants with a serious adverse event Follow‐up: 15 to 84 days	Study population		RR 1.09 (0.88 to 1.35)	1340 (7 studies)	⊕⊕⊕⊝ moderate⁵
	179 per 1000	196 per 1000 (158 to 242)
	Moderate
	204 per 1000	222 per 1000 (180 to 275)
Number of participants experiencing platelet refractoriness Follow‐up: 0 to 24 hours	Study population		RR 2.94 (2.08 to 4.16)	1525 (7 studies)	⊕⊕⊕⊕ high	4 studies defined refractoriness as 2 successive 1‐hour CCIs below 5 x 10³ (Cazenave 2010; Janetzko 2005; McCullough 2004; van Rhenen 2003), while Kerkhoffs 2010 defined refractoriness as 2 successive 1‐hour CCIs below 7.5 x 10³ or 24‐hour CCIs below 4.5 x 10³ and presence of antibodies against platelets.
	51 per 1000	149 per 1000 (106 to 212)
Number of platelet transfusions per participant Multiple platelet transfusion trials	The mean number of platelet transfusions per participant was 4.7 to 8.4 platelet transfusions.	MD 1.23 platelet transfusions higher (0.86 to 1.61 higher)	‐	1509 (6 studies)	⊕⊕⊕⊕ high
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CCI: corrected count increment; CI: confidence interval; MD: mean difference; RR: risk ratio; WHO: World Health Organization.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹Downgraded one point due to inconsistency. ²Downgraded one point due to risk of other bias (differences in the assessment and grading of bleeding) ³Low risk = data taken from autologous transplantation participants from PLADO (PLAtelet DOse) trial (Slichter 2010). Medium risk = overall bleeding rate for all participants in the PLADO trial (Slichter 2010). High risk = data taken from allogeneic transplantation participants in PLADO trial (Slichter 2010). ⁴Downgraded one point due to inconsistency. One large study (McCullough 2004) had proportionately more events in the control group than the other five studies. Although the studies within this subgroup had similar follow‐up periods and to our knowledge there are no other major differences between them, the difference in the event rate may nevertheless be due to undetected methodological differences between the studies. ⁵Downgraded one point due to imprecision. Wide confidence intervals that include the risk of significant harm or benefit.

Background

Platelet transfusions are used to prevent and treat bleeding in people who are thrombocytopenic. Despite improvements in donor screening and laboratory testing, a small risk of transfusion‐transmitted infections (TTIs) remains. One solution to minimise the risk of TTIs from platelet transfusion is pathogen reduction, a process by which pathogens are either inactivated or significantly depleted in number, thereby reducing the chance of transmission. A number of published clinical studies have raised concerns about the effectiveness of pathogen‐reduced platelets for post‐transfusion platelet recovery and the prevention of bleeding when compared with standard platelets. The objective of this review was to evaluate the evidence from randomised controlled trials (RCTs) assessing the effectiveness of pathogen‐reduced platelets compared with standard platelets for the prevention of bleeding in people requiring platelet transfusion(s).

See Published notes for explanation of some technical terms and abbreviations.

Description of the condition

Significant progress has been made in recent years to improve the safety of blood transfusion with regard to transfusion‐transmitted infections (TTIs). Consequently the risk of viral TTIs (e.g. human immunodeficiency virus (HIV), hepatitis B (HBV), hepatitis C (HCV), and human T‐cell lymphotropic virus (HTLV) type I and II) is now considered to be low in high‐income countries (Zou 2012). Nevertheless, there remains the threat both of newly emerging blood‐transmitted infections (e.g. from viruses such as Zika, prions, and protozoa) and of TTIs caused by bacterial contamination of blood products. Platelet concentrates in particular are more vulnerable to bacterial contamination owing to their higher storage temperature (22°C) in comparison to other blood components (Blajchman 2006). The UK's Serious Hazards of Blood Transfusion (SHOT) programme reported that overall a total of 36 out of 43 bacterial transfusion‐transmissions to individual recipients (33 incidents) were caused by the transfusion of platelets between 1996 and 2014 (Bolton‐Maggs 2015). In the USA between January 2007 and December 2011, 38 episodes of probable or definite post‐transfusion sepsis including four fatalities were reported after transfusion of an apheresis platelet component to their haemovigilance programme (Benjamin 2014). In 2011, over 90% of all platelet components issued in the USA were apheresis platelet components (Whitaker 2011). For pools of whole blood‐derived platelet components prepared by the platelet‐rich plasma method and constituted just prior to transfusion, the rate for sepsis is approximately five times higher than in apheresis platelets (Jacobs 2008).

Both the UK and the USA use passive reporting systems, which may under‐report the number of cases. Using active reporting (based on prospective follow‐up of transfused recipients), single‐centre studies have identified higher rates of transfusion‐associated septic reactions that are up to 10 times that identified by passive reporting (Hong 2015; Jacobs 2008). The World Health Organization has therefore actively issued a recommendation of haemovigilance programmes to monitor and prevent unwanted transfusion events and increase safety and efficacy of blood transfusion (WHO 2016).

Description of the intervention

There are a number of current interventions that aim to reduce the risk of TTIs from platelet transfusion. These include improved donor screening, registries of previously deferred donors, single‐donor apheresis platelets, improved laboratory testing including the use of nucleic acid amplification testing, optimised skin preparation, removal of the first 10 to 30 mL of donated blood, and limited storage duration of platelet units. Some countries also carry out routine bacterial screening (Benjamin 2014; Pearce 2011). These methods have reduced but not eliminated the risk of both bacterial and viral TTIs from platelet transfusion, for example viral contamination can still occur during the 'window period' when the pathogen load is low enough to escape detection (Bolton‐Maggs 2015). The appearance of new blood‐borne viruses, such as West Nile virus (WNV), severe acute respiratory syndrome (SARS), and most recently the Zika virus, has highlighted the ongoing need to prevent the transmission of emerging pathogens.

It is against this background that photochemical pathogen‐reduction technologies have been developed and tested. Although solvent‐detergent treatment of pooled plasma for fractionation and transfusion has effectively eliminated the transmission of HIV, HBV, and HCV (Goodnough 2003; Pelletier 2006), these methods are cytocidal and therefore incompatible with cellular blood components. Pathogen‐reduction technology uses photochemical processes either entirely to inactivate or significantly reduce a broad spectrum of infectious agents such as viruses, fungi, bacteria and parasites within cellular blood components, including platelet concentrates. In addition, pathogen reduction should increase the availability of platelet concentrates by retarding bacterial growth, thereby extending platelet storage life from five to seven days and reducing costs.

How the intervention might work

Pathogen reduction uses photochemical treatment (PCT) to prevent replication of nucleic acid‐containing microbes whilst aiming to preserve cellular function and minimising toxicity. Two pathogen‐reduction systems for the treatment of platelet concentrates are currently available commercially. These utilise either a synthetic psoralen compound (amotosalen) or riboflavin (vitamin B₂), both in the presence of ultraviolet (UV) light.

Amotosalen

Amotosalen hydrochloride acts by selectively binding to nucleic acids within DNA and RNA and forming cross‐links upon photoactivation by exposure to long‐wave ultraviolet A (UVA) light. The UVA light induces a photochemical reaction that transforms the pre‐existing noncovalent link between pyrimidic base residues in DNA and RNA chains into irreversible. Such cross‐links render nucleic acids, and thus pathogens and white blood cells, incapable of replication (Grass 1998; Pelletier 2006). Amotosalen additionally binds to ribosomal RNA, inhibiting synthesis of proteins such as cytokines, and thus may also reduce the incidence of platelet transfusion reactions (McCullough 2007). The Intercept blood system (Cerus Corporation, Concord, CA, USA) achieves pathogen reduction through a combination of amotosalen and UVA light. It also employs a compound absorption device to scavenge unwanted postproduction residual amotosalen and photoproducts.

Amotosalen has been shown to inactivate 10³ to 10⁶ of enveloped viruses (HIV, WNV, HBV, HCV), both gram‐negative (Escherichia coli, Klebsiella pneumoniae) and gram‐positive bacteria (Staphylococcus aureus, Staphylococcus epidermidis), plus protozoa commonly implicated in TTIs (Lin 2004; Lin 2005). However, it has variable efficacy against non‐enveloped viruses, parvovirus is partially resistant, and hepatitis A and E viruses are resistant to inactivation by amotosalen (Hauser 2014; Kwon 2014).

In vitro data have demonstrated that in platelets treated with amotosalen plus UVA light, platelet function is comparable following storage for five and seven days despite transient platelet activation (Janetzko 2002; Lin 1998; Moog 2004; Van Rhenen 2000). Two phase II studies reported that although platelet viability was found to be significantly reduced in vivo in terms of recovery, survival, and corrected count increments, there was no significant difference in correction of bleeding time and transfusion interval (Slichter 2006; Snyder 2004). The Intercept® pathogen‐reduced products have demonstrated sufficient safety in terms of genotoxicity, carcinogenicity, and phototoxicity in single and repeated doses in both in vivo and in vitro studies (Ciaravino 2003; Tice 2007). This was also shown from combined seven‐year prospectively collected data from a cohort of 4067 participants receiving 19,175 transfusions as reported by the haemovigilance programme across 11 countries (Knutson 2015).

Riboflavin

Mirasol® pathogen‐reduction technology (CaridianBCT, Lakewood, CO, USA) reduces pathogens in platelet concentrates by utilising UVA light in the presence of riboflavin (vitamin B₂). This combination introduces intercalations with nucleic acids which, through photolysis upon exposure to UV light, promotes guanine oxidation and single‐strand breaks of nucleic acids (Kumar 2004). The riboflavin in conjunction with UV light associates with nucleic acids and mediates an oxygen‐independent electron transfer process leading to the oxidation of guanine residues and strand breaks. Each treated platelet unit with Mirasol Pathogen Reduction Technology has a fixed amount of riboflavin (35 mL) and can be used without the removal of residual riboflavin or its photoproduct. The Mirasol method causes irreversible damage to the RNA and DNA of of viruses, bacteria, and parasites, stopping them from replicating and causing infection.

Pathogens and white blood cells are thereby rendered inactive and effectively incapable of replication (Goodrich 2006; Kumar 2004). Riboflavin is a naturally occurring compound that does not require postproduction removal, and the breakdown products resulting from photolysis have been demonstrated to be satisfactorily safe (Hardwick 2004; Reddy 2008).

In vitro studies have shown that riboflavin inactivates viruses including HIV, WNV, and a number of bacteria commonly implicated in causing TTIs, including S epidermidis, S aureus, and E coli by 10³ to 10⁶ (Goodrich 2006; Ruane 2004). It may be effective against hepatitis E (Owada 2014), however it has variable efficacy against other viruses, including parvovirus, dengue virus, and hepatitis A virus (Faddy 2016; Kwon 2014).

Why it is important to do this review

Both systems are widely used in Europe, although there is a greater uptake of amotosalen, which is routinely used in 18 countries, and riboflavin, which is routinely used in 15 countries. For example, from 2011, Intercept was implemented nationwide in Switzerland, and amotosalen is widely used in Belgium. Advances in haematological treatments have led to an increased need for safer platelets transfusion strategies, and therefore motivated the development of pathogen inactivation products in order to reduce the risk of contamination. Functional and biochemical studies have revealed that pathogen inactivation may have some impact on the platelets' function, generating concerns about the haemostatic abilities of the inactivated platelets. Although the haemostatic function of the pathogen‐reduced platelets appeared to be preserved (AuBuchon 2005; Goodrich 2006; Lin 2004; Perez‐Pujol 2005; Picker 2009), some reduction in post‐transfusion recovery and survival in vivo were suggested (AuBuchon 2005; Slichter 2006; Snyder 2004).

Although current pathogen‐reduction technologies are very effective in reducing viruses and bacteria in platelet concentrates, concern remains about the haemostatic effectiveness of platelets treated in this way. While some studies have suggested that cell quality and platelet function are largely comparable between pathogen‐reduced and standard platelets (AuBuchon 2005; Goodrich 2006; Lin 2004; Perez‐Pujol 2005; Picker 2009), other studies have reported that current methods of pathogen reduction appear to reduce significantly post‐transfusion platelet recovery and survival in vivo when compared with standard platelets (AuBuchon 2005; Slichter 2006; Snyder 2004). Moreover, a recent RCT found Intercept® PCT pooled platelets to be less clinically effective at preventing bleeding than standard pooled platelets (Kerkhoffs 2010). As pathogen‐reduced platelets must be found to be safe and effective in achieving haemostasis prior to their introduction into routine clinical use, a systematic review of RCTs in this area was required.

Subsequently, results from an active haemovigilance programme across 11 countries involving 19,175 Intercept® transfusions, predominantly in a haemato‐oncology population, confirmed the Intercept® safety profile (Knutson 2015). However, the efficacy and safety of pathogen‐reduced platelets methods have not been systematically compared to standard platelets in people with thrombocytopenia.

Objectives

To assess the effectiveness of pathogen‐reduced platelets for the prevention of bleeding in people of any age requiring platelet transfusions.

Methods

Criteria for considering studies for this review

Types of studies

RCTs only (including cross‐over and multicentre, but excluding quasi‐randomised trials), irrespective of language or publication status.

Types of participants

People of any age requiring platelet transfusion(s).

Types of interventions

We included the following comparisons:

Pathogen‐reduced platelets (either Intercept® pathogen‐reduction technology (PCT) or Mirasol® PCT or other pathogen‐reduction technologies) versus 'standard platelets' in people requiring platelet transfusion(s) (note that for the purposes of this review, we have used each trial's definition of 'standard platelets', and have described these differences in detail in the Characteristics of included studies section).
Intercept® PCT versus Mirasol® PCT versus other pathogen‐reduced platelets in people requiring platelet transfusion(s). (We found no trials including this comparison.)

Types of outcome measures

Primary outcomes

Number, type, and severity of bleeding episodes. Bleeding was measured at up to 48 hours, 48 hours to 7 days, and more than 7 days.
All‐cause mortality. Mortality was measured at up to 4 weeks, 4 to 12 weeks, and more than 12 weeks. Of particular interest were deaths from platelet transfusion‐related complications such as bleeding, infection, thromboembolism, or transfusion reactions.

Secondary outcomes

Adverse events: recognised complications of platelet transfusions, e.g. mild transfusion reactions (rigors, fever, skin rash, urticaria) and platelet refractoriness.
Other adverse events, e.g. anaphylaxis, transfusion‐related acute lung injury (TRALI), infection, arterial or venous thromboembolism.
Laboratory assessment of response to platelet transfusion by post‐transfusion platelet count increment and corrected count increment (CCI).
Platelet and red cell transfusion requirement and interval.
Quality of life measured using validated scales.

Search methods for identification of studies

Electronic searches

The Systematic Review Initiative Information Specialist (CD) formulated updated search strategies in collaboration with the Cochrane Haematological Malignancies Review Group based on those used in previous versions of this review. In this update we searched the following databases for RCTs.

Cochrane Central Register of Controlled Trials on 24 October 2016 (CENTRAL, the Cochrane Library 2016, Issue 9) (www.cochranelibrary.com/) (Appendix 1)
MEDLINE (OvidSP Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE Daily & Ovid MEDLINE 1946 to 24 October 2016) (Appendix 2)
Embase (OvidSP 1974 to 24 October 2016) (Appendix 3)
CINAHL (Cumulative Index to Nursing and Allied Health Literature) (EBSCOhost 1937 to 24 October 2016) (Appendix 4)
PubMed (for e‐publications ahead of print, in‐process and other non‐indexed citations only on 24 October 2016) (www.ncbi.nlm.nih.gov/pubmed) (Appendix 5)
Transfusion Evidence Library (1950 to 24 October 2016) (www.transfusionevidencelibrary.com) (Appendix 6)
LILACS (Latin American and Caribbean Health Sciences Literature) (1982 to 24 October 2016) (lilacs.bvsalud.org/en/) (Appendix 7)
KoreaMed (1958 to 24 October 2016) (koreamed.org/) (Appendix 8)
IndMED (1986 to 24 October 2016) (indmed.nic.in/indmed.html) (Appendix 9)
Web of Science Conference Proceedings Citation Index‐Science (CPCI‐S) (1990 to 24 October 2016) (Appendix 10)

We searched the following ongoing trial databases (all years to 24 October 2016).

World Health Organization International Clinical Trials Registry Platform (WHO ICTRP) (apps.who.int/trialsearch/) (Appendix 11)
ClinicalTrials.gov (www.clinicaltrials.gov/) (Appendix 12)

We combined the search strategies in MEDLINE and Embase with adaptations of the Cochrane sensitive RCT search filter as detailed in Chapter 6 of the Cochrane Handbook for Systematic Reviews of Interventions (Lefebvre 2011), and in CINAHL with the RCT filter designed by the Scottish Intercollegiate Guidelines Network (SIGN) (www.sign.ac.uk/methodology/filters.html). We conducted our original search on 18 February 2013 (Butler 2013); search strategies for each database from the original review are available in Appendix 13, Appendix 14, Appendix 15, Appendix 16, Appendix 17, and Appendix 18.

Searching other resources

Handsearching of reference lists

We checked and citation‐tracked the reference lists of all identified trials, relevant review articles, and current treatment guidelines for further literature, but limited these searches to the 'first‐generation' reference lists.

Personal contacts

We contacted lead authors of relevant studies, study groups, and international experts known to be active in the field to request any unpublished material, missing data, or further information on ongoing trials. This included making contact with Dr Laurence Corash (Intercept, Cerus Corporation), Dr Raymond Goodrich (Mirasol®, CaridianBCT), and Dr Paolo Rebulla.

Data collection and analysis

Selection of studies

The Systematic Review Initiative Information Specialist (CD) initially removed all duplicate records. Two review authors (LE, RM) then independently screened all the remaining references for relevance against the full eligibility criteria using Covidence software. We obtained full‐text papers of all references for which eligibility could not be verified based on title and abstract alone in addition to records meeting the review inclusion criteria. It was necessary to seek further information from several authors, as most articles of interest contained insufficient data to make a decision about eligibility. Differences of opinion were resolved through discussion and consensus within the review team. Details of the studies that did not meet our eligibility criteria are recorded in the Characteristics of excluded studies table.

Data extraction and management

We updated the data extraction performed for the previous version of this review (Butler 2013), including data extraction for all new studies included since the previous review, all new review outcomes that were not part of the previous review (e.g. quality of life), and information on trial registration.

Two review authors (LE, RM) conducted the data extraction according to the guidelines proposed by The Cochrane Collaboration (Higgins 2011a). Any disagreements between the review authors were resolved by consensus. The review authors were not blinded to names of authors, institutions, journals, or the outcomes of the trials.

The agreed data extraction form included the following.

General information: study ID, first author of study, author's contact address (if available), citation of paper, objectives of the trial.
Trial details: trial design, location, setting, sample size, power calculation, trial registration, methods of treatment allocation, randomisation and blinding, inclusion and exclusion criteria, reasons for exclusion, comparability of groups, length of follow‐up, stratification, stopping rules described, method of statistical analysis used, and funding.
Characteristics of participants: age, gender, ethnicity, total number recruited, total number randomised, total number analysed, types of underlying condition(s), losses to follow‐up, dropouts (percentage in each arm) with reasons, protocol violations, current treatment, previous treatments, prognostic factors.
Interventions: experimental and control interventions, type of platelet given, method of platelet production (e.g. apheresis or pooled buffy‐coat derived), type and dosage of platelet given, timing of intervention, compliance with interventions, additional interventions given (especially with regard to red cell transfusions), any other differences between interventions.
Outcomes: mortality due to bleeding, long‐term morbidity due to bleeding (e.g. stroke, myocardial infarction, pulmonary embolism), number and severity of bleeding episodes, laboratory assessment of response to platelet transfusion by post‐transfusion platelet count increment, platelet and red cell transfusion interval, platelet and red cell transfusion requirement, recognised complications of platelet transfusion (e.g. transfusion reaction, TTI, platelet refractoriness, thromboembolism, development of platelet antibodies), and any adverse or other serious adverse event (e.g. anaphylaxis, acute lung injury).

Assessment of risk of bias in included studies

Two review authors (LE, RM) independently applied the Cochrane 'Risk of bias' tool per Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions to assess potential biases across the included studies (Higgins 2011b). We rated each domain of the tool as 'high', 'low', or 'unclear' risk of bias, and provided a brief description to support our judgement for each rating. We compared our statements, involving a third review author when agreement could not be reached. We used the assessment of risk of bias in each trial to conduct sensitivity analyses where appropriate as part of the exploration of heterogeneity among included trials. Details of the potential bias for each individual study are presented in the Characteristics of included studies tables.

The Cochrane 'Risk of bias' tool addresses the following domains.

Selection bias: random sequence generation and allocation
concealment.
Performance bias: blinding of participants and personnel.
Detection bias: blinding of outcome assessment.
Attrition bias: incomplete outcome data.
Reporting bias: selective reporting.
Other bias.

Measures of treatment effect

For continuous outcomes, we extracted the mean, standard deviation, and total number of participants in both the treatment and control groups. For continuous outcomes using the same scale, we performed analyses using the mean difference (MD) with 95% confidence intervals (CI). When different scales were used to measure the same outcomes, we applied the standardised mean difference (SMD) when data were pooled.

For dichotomous outcomes, we recorded the number of events and the total number of participants in both the treatment and control groups. We reported the pooled risk ratios (RR) with a 95% CI using the Mantel‐Haenszel method. Where the number of observed events was small (less than 5% of sample per group) and trials had balanced treatment groups, we applied the Peto odds ratio (OR) with 95% CI (Deeks 2011).

If available, we planned to extract and report hazard ratios (HR) for mortality data. If HRs were not available, we made every effort to estimate as accurately as possible the HR using the available data and a purpose‐built method based on the Parmar and Tierney tool (Parmar 1998; Tierney 2007). However, we were unable to estimate any hazard ratios.

We planned to report the number needed to treat for an additional beneficial outcome (NNTB) and the number needed to treat for an additional harmful outcome (NNTH) with CIs for outcomes that showed a benefit or harm for transfusion‐related adverse events.

We also reported data in a narrative synthesis when data were not available in a usable format.

We applied the generic inverse variance method when pooling data from cross‐over design trials if period effects had been appropriately adjusted for.

Unit of analysis issues

We included no cluster‐randomised trials in this review.

For cross‐over trials, we included data from both trial periods if period effects had been appropriately adjusted for (more than 24‐hour washout period), using the generic inverse variance method. Otherwise we used data only from the first trial period, if data from the two trial periods were reported separately. If they were not reported separately, we reported data narratively. We included three cross‐over trials (Johansson 2013; Simonsen 2006; Slichter 2006). We were unable to use data provided by Johansson 2013 in any of the meta‐analyses as there was no clear washout period between the two trial phases. In Simonsen 2006 and Slichter 2006, we assessed dichotomous outcomes that included data from one or both of these trials using the generic inverse variance model. It should therefore be noted that we could not include outcomes with a zero value result in this type of analysis. Furthermore, for continuous outcomes for which data were available from the cross‐over trials, it was necessary to include data from the first period of each trial only. As Slichter 2006 reported only combined results for both periods of the two trial arms, we could not include these data in any of the meta‐analyses of continuous outcomes and have therefore reported them narratively throughout.

We identified duplication of participants through inclusion in more than one study cycle in three Intercept trials (McCullough 2004; Simonsen 2006; van Rhenen 2003). McCullough 2004 reported data from only the first cycle of the study, whilst we extracted the data for van Rhenen 2003 from cycle 1 only, either from the published literature or obtained through direct author contact. In Simonsen 2006, 1 participant out of 20 received both study transfusions on 2 separate occasions. We did not consider this as having a significant impact on the accuracy of this cross‐over trial's outcomes, despite its small size, and we included data for this participant in the analyses.

Dealing with missing data

We dealt with missing data according to the recommendations in Chapter 16 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011c). We attempted to contact the authors of all 12 included trials via email in order to request additional information and data that were either missing or unclear in the published papers. We received additional data for 10 of the 12 included trials (Cazenave 2010; Janetzko 2005; Johansson 2013; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; Simonsen 2006; Slichter 2006; van Rhenen 2003). We recorded the number of participants lost to follow‐up for each trial. Where possible, we used intention‐to‐treat data, but if not available, we reported per‐protocol data.

Assessment of heterogeneity

The decision to combine the results of individual trials depended on an assessment of heterogeneity, which we performed according to the recommendations of The Cochrane Collaboration (Deeks 2011). If we considered trials to be sufficiently homogeneous in terms of their clinical and methodological characteristics, we carried out a meta‐analysis and assessed the statistical heterogeneity of treatment effects between trials using a Chi² test with a significance level at P < 0.1. We used the I² statistic to quantify the degree of possible heterogeneity, considering an I² > 30% as moderate heterogeneity and I² > 75% as considerable heterogeneity. Where there was moderate heterogeneity, we conducted both fixed‐effect and random‐effects analyses, presenting the former but reporting results for both in the text.

Assessment of reporting biases

We had intended in meta‐analyses of more than 10 trials to explore potential publication bias (small‐trial bias) by generating a funnel plot and using a linear regression test. However, the number of trials included in any of our meta‐analyses was insufficient to perform this analysis, per Chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions (Sterne 2011).

Data synthesis

We performed analyses according to the recommendations of The Cochrane Collaboration using aggregated data (Deeks 2011). For statistical analysis, we entered data into the Cochrane statistical package Review Manager 5 (Review Manager 2014). One review author performed data entry, and another review author checked the data entry for accuracy. We performed meta‐analyses using the Mantel‐Haenszel method with the fixed‐effect model. Where data did not allow for meta‐analysis, we reported outcome data descriptively within the Results section of the review. For the three included cross‐over trials (Johansson 2013; Simonsen 2006; Slichter 2006), if they appropriately adjusted for period effects, for dichotomous outcomes we included data from the trial periods using the generic inverse variance method, as described above. For continuous outcomes, we included data only from the first period, if available, as these data cannot be analysed using the generic inverse variance method. Where data were not reported separately, we reported these narratively. However, we used no data from Johansson 2013, as the washout period between the two phases was less than 24 hours, and data from phase one were not available.

For each outcome, we combined data to compare pathogen‐reduced platelets with standard platelet transfusion in the meta‐analyses. We made no comparisons between pathogen‐reduced products, as no completed or ongoing trials assessed this.

'Summary of findings' table

We used the GRADE profiler to create a 'Summary of findings' table, as suggested in Chapter 12 of the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2011), based on our primary outcomes and adverse events. These included the following.

Number of participants with any bleeding
Number of participants with clinically significant bleeding
Severe bleeding
All‐cause mortality (4 to 12 weeks)
Serious adverse events
Platelet refractoriness
Number of platelet transfusions per participant

The outcomes within the 'Summary of findings' table had not been prespecified in the protocol.

Subgroup analysis and investigation of heterogeneity

Sufficient data were not available to perform our prespecified subgroup analyses relating to length of platelet storage and the underlying diagnosis of participants.

In the previous version of this review, Butler 2013, it became apparent upon the completion of searches that two different trial designs were used for the Intercept PCT versus standard platelet trials: single or multiple study platelet transfusions. In the previous version of this review a post hoc division of trials was performed, separating the trials according to the pathogen‐reduced platelet technique employed (as predefined) and transfusion exposure.

In this version of the review we prespecified these subgroups:

a single transfusion of Intercept PCT versus standard platelets;
multiple transfusions of Intercept PCT versus standard platelets;
a single transfusion of Mirasol PCT versus standard platelets;
multiple transfusions of Mirasol PCT versus standard platelets.

Sensitivity analysis

We intended to assess the robustness of our findings by performing the following two sensitivity analyses:

including only those trials at low risk of bias;
including only those trials in which 25% participants or less were lost to follow‐up.

However, we only carried out the first analysis, as the loss to follow‐up was not significantly high for any of the 10 completed trials included in the meta‐analyses.

Results

Description of studies

See Characteristics of included studies; Characteristics of ongoing studies; and Characteristics of excluded studies.

Results of the search

The original search (conducted February 2013) identified a total of 1138 potentially relevant records. After removal of duplicates 863 records remained, of which 812 records were excluded on the basis of the abstract. The original systematic review identified 51 records that appeared relevant on the basis of their full text or abstract using the original inclusion/exclusion criteria (Butler 2013). Ten studies were identified and included in the previous version of the review (Agliastro 2006; Cazenave 2010; De Francisci 2004; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Simonsen 2006; Slichter 2006; van Rhenen 2003).

The latest update of the search (conducted October 2016) identified a total of 1407 potentially relevant records. After the removal of duplicates 474 records remained. Two review authors (LE, RM) excluded 441 records on the basis of the abstract using Covidence software. Two review authors (LE, RM) retrieved for relevance and reviewed 33 full‐text articles. We identified two new completed studies, Johansson 2013 and Rebulla 2016, and three ongoing studies (EUCTR2015‐001035‐20‐DE; Kerkhoffs 2013; NCT01789762) (see the PRISMA study flow diagram in Figure 1).

Figure 1

Updated review study flow diagram.

For this 2017 update, we have included the previous two ongoing trials in the review (Johansson 2013; Rebulla 2016), for a total of 12 included completed studies (Agliastro 2006; De Francisci 2004; Janetzko 2005; Johansson 2013; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; Simonsen 2006; Slichter 2006; van Rhenen 2003), as well as three ongoing trials (EUCTR2015‐001035‐20‐DE; Kerkhoffs 2013; NCT01789762).

Included studies

All 12 trials were published in English between 2003 and 2016. The included trials were grouped by the type of the PCT product type: Intercept PCT and Mirasol PCT. Of the 12 included trials, no trials compared different types of pathogen‐reduced platelets with each other.

Ten of the 12 included trials compared the Intercept PCT product with standard platelets (Agliastro 2006; De Francisci 2004; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; Simonsen 2006; Slichter 2006; van Rhenen 2003). Of these 10 trials, eight were published as full papers (Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; Simonsen 2006; Slichter 2006; van Rhenen 2003), and two as abstracts only (Agliastro 2006; De Francisci 2004).

Three of the 12 included trials compared the Mirasol PCT product with standard platelets (Cazenave 2010; Johansson 2013; Rebulla 2016). All three trials were reported in full (Cazenave 2010; Johansson 2013; Rebulla 2016).

In one trial (Rebulla 2016), three centres compared Intercept PCT to standard platelets, and another three centres compared Mirasol PCT to standard platelets. Consequently, in this review, we have reported data from the Rebulla 2016 trial separately for Intercept PCT versus standard platelets (three centres) and Mirasol PCT versus standard platelets (three centres).

Trial settings

All included studies were set in high‐income countries according to the World Bank classification (WB 2016).

Intercept PCT platelet trials

The 10 trials comparing Intercept PCT with standard platelets were conducted across 10 countries: Denmark, the Netherlands, Italy, Spain, Germany, France, Switzerland, Sweden, the UK, and the USA. Three trials were multinational: Janetzko 2005 (Germany, France, and Switzerland); Lozano 2011 (Spain, the UK, France, and Sweden); and van Rhenen 2003 (the Netherlands, Sweden, France, and the UK). Five were multicentre trials conducted in a single country: Cazenave 2010 (six centres in France); Kerkhoffs 2010 (eight centres in the Netherlands); McCullough 2004 (12 centres in the USA); Rebulla 2016 (three centres in Italy); and Slichter 2006 (two centres in the USA). One included trial was conducted in a single centre in Denmark (Simonsen 2006). Two trials conducted in Italy did not report the number of participating centres (Agliastro 2006; De Francisci 2004).

Mirasol PCT platelet trials

Of the three trials comparing Mirasol PCT with standard platelets, two were multicentre trials (Cazenave 2010 (six centres in France); Rebulla 2016 (three centres in Italy)), and the third was a single‐centre trial in Denmark (Johansson 2013).

Trial design

Three included trials had a cross‐over design (Johansson 2013; Simonsen 2006; Slichter 2006), while the other nine trials were parallel‐group RCTs (Agliastro 2006; Cazenave 2010; De Francisci 2004; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; van Rhenen 2003).

None of the three cross‐over studies reported the pre‐defined washout period to minimise the carry‐over effects due to transfusion sequences. However, there was at least a 24‐hour period between study transfusions in Simonsen 2006 and Slichter 2006. In Johansson 2013 no information with regard to the period between the two transfusion types was provided.

Intercept PCT platelet trials

Of the 10 RCTs comparing Intercept PCT with standard platelets (Agliastro 2006; De Francisci 2004; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Simonsen 2006; Slichter 2006; Rebulla 2016; van Rhenen 2003), two were cross‐over design trials (Simonsen 2006; Slichter 2006), and the rest were parallel design trials. Three trials were two‐arm non‐inferiority design trials (Lozano 2011; McCullough 2004; Rebulla 2016). Two trials were two‐arm equivalence trials (Janetzko 2005; van Rhenen 2003). One trial was a three‐arm non‐inferiority trial (Kerkhoffs 2010), in which Intercept PCT platelets were compared with platelets stored in both plasma and InterSol platelet additive solution. We selected the plasma arm as the control group (i.e. as the 'standard' platelet product) for comparison in this review. Two trials were parallel design trials that were only reported as an abstract, with no further details given (Agliastro 2006; De Francisci 2004).

Mirasol PCT platelet trials

Of the three trials comparing Mirasol PCT with standard platelets, one was a cross‐over equivalence trial (Johansson 2013), and two were parallel design non‐inferiority trials (Cazenave 2010; Rebulla 2016).

Trial size

The size of included trials varied from 16 participants in Johansson 2013 to 671 participants in McCullough 2004. Six trials included more than 100 participants (Cazenave 2010; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; van Rhenen 2003).

Two trials were stopped early (Kerkhoffs 2010; Rebulla 2016). In Rebulla 2016, the initial target was to enrol 828 participants, but the study was stopped early for administrative reasons with only 424 participants included. The independent data safety monitoring board for the Kerkhoffs 2010 study recommended early stopping of the study, as significantly higher bleeding rates and lower one‐hour corrected count increments (CCIs) were identified in the Intercept PCT platelets arm.

Intercept PCT platelet trials

The size of the included trials varied from 28 participants in Simonsen 2006 to 671 participants in McCullough 2004. Five trials included more than 100 participants (Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; van Rhenen 2003). Two studies were stopped early (Kerkhoffs 2010; Rebulla 2016).

Mirasol PCT platelet trials

The size of the included trials varied from 16 participants in Johansson 2013 to 194 participants in Rebulla 2016. Two trials included more than 100 participants (Cazenave 2010; Rebulla 2016). One study was stopped early (Rebulla 2016).

Participant

Overall, 2075 participants were randomised in the 12 included trials, with only 1981 participants receiving at least one study platelet transfusion. However, we only included data from the 1887 participants in two arms of the three‐arm Kerkhoffs 2010 trial.

All participants, apart from those recruited to one trial (De Francisci 2004), were thrombocytopenic individuals who had a haematological or oncological diagnosis and were either in need of, or expected to be in need of, platelet transfusion. The exception, De Francisci 2004, explored two different groups: one adult group awaiting liver transplantation requiring platelet support before invasive procedures, and one paediatric group with congenital cyanotic cardiopathy undergoing cardiac surgery requiring per‐protocol platelet transfusions during extracorporeal circulation due to platelet consumption.

The age of participants ranged from 4 to 85 years. Two trials recruited only children (Agliastro 2006 and the cardiac group of De Francisci 2004), whilst two other trials included children from the age of 6 or 12 years (McCullough 2004; van Rhenen 2003). The remaining eight trials recruited adults aged over 16 or 18 years (Cazenave 2010; Janetzko 2005; Johansson 2013; Kerkhoffs 2010; Lozano 2011; Rebulla 2016; Simonsen 2006; Slichter 2006).

All trials had comparable proportions of males and females, except for Johansson 2013, in which almost all included participants were male.

Seven trials excluded individuals with a previous history of alloimmunisation, antibodies formed against human leukocyte antigens or human platelet antigens, or positive lymphocytotoxic antibody panels with high reactivity (Cazenave 2010; Johansson 2013; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; Slichter 2006; van Rhenen 2003). Nine trials excluded individuals with a previous history of clinical refractoriness to platelet transfusions (Cazenave 2010; Janetzko 2005; Johansson 2013; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; Simonsen 2006; van Rhenen 2003). Five of the 12 included trials excluded from enrolment people who were actively bleeding (Cazenave 2010; Janetzko 2005; Johansson 2013; Simonsen 2006; Slichter 2006), whilst another study postponed the study transfusion if the person had evidence of active bleeding within the last 12 hours (Lozano 2011). In Rebulla 2016, participants with active bleeding were not excluded.

Intercept PCT platelet trials

A total of 1740 participants were enrolled in the Intercept PCT versus standard platelets trials, of which 1662 received at least one study platelet transfusion, and 1568 were included in our analyses.

Four trials included children (Agliastro 2006; De Francisci 2004; McCullough 2004; van Rhenen 2003), and six trials included only adults (Janetzko 2005; Kerkhoffs 2010; Lozano 2011; Rebulla 2016; Simonsen 2006; Slichter 2006). All participants, apart from those recruited to one trial (De Francisci 2004), were thrombocytopenic individuals who had a haematological or oncological diagnosis and were either in need of, or expected to be in need of, platelet transfusion.

Mirasol PCT platelet trials

A total of 335 participants were enrolled in the Mirasol PCT versus standard platelets trials (Cazenave 2010; Johansson 2013; Rebulla 2016), of which 319 received at least one study platelet transfusion. All three trials included adults (aged 18 years or older) with thrombocytopenic haematology or oncology conditions that required platelet transfusions.

Interventions

For full details of the source (apheresis or buffy coat), storage solution, mean storage duration and mean platelet dose of the study and standard platelet products for each of the included trials see the Characteristics of included studies section and Table 1. The transfusion triggers used were either prophylactic (generally a platelet count of 10 to 20 x 10⁹/L or prior to invasive procedures) or therapeutic due to bleeding, the details of which are given (where reported) in the Characteristics of included studies section and Table 2.

Table 1. Trial characteristics

Study

(publication)

Study design

(analysis)

Total no. of participants

(PCT plts/std plts)

Underlying disease

Type of treatment received for underlying

disease (%)

Type of platelet component

Follow‐up periods for mortality

(days)

Mean

plt storage duration

(days)

Platelet dose^a

(Mean plt transfusion dose) (10¹¹/unit)

PCT plts

Std plts

Single‐transfusion studies ‐ Intercept

De Francisci 2004

(A)

(NR)

(22/22)

Adult liver transplant/

paediatric cardiac surgery

Liver: 14

Cardiac: 8

Liver: 14

Cardiac: 8

Lozano 2011

(F)

(Non‐I)

211

(105/106)

Adult haemato‐

oncological disease

Auto: 42

Allo: 29

Ch: 27

Other: 3

Auto:40

Allo: 25

Ch: 34

Other: 2

86% BC

14% Aph

6.8

Intermediate

(4.2)

Simonsen 2006

(F)

C‐O

(Non‐I)

PCT‐std: 9

std‐PCT: 11

Adult haemato‐

oncological disease

PCT‐std

Std‐PCT

Low/intermediate

(2.9)

Allo: 33

Ch: 67

Allo: 64

Ch: 36

Slichter 2006

(F)

C‐O

(E)

PCT‐std: NR

std‐PCT: NR

Adult haemato‐

oncological disease

Aph

42^b

High

(7.5)

Multiple‐transfusion studies – Intercept

Agliastro 2006

(A)

(NR)

(19/11)

Paediatric haemato‐

oncological disease

PCT plts: BC

Std plts: Aph

Low/intermediate

(2.9)

Janetzko 2005

(F)

(E)

(22/21)

Adult haemato‐

oncological disease

SCT: 68

Ch: 27

Other: 5

SCT: 62

Ch: 38

Aph

3.1

Intermediate

(4.0)

Kerkhoffs 2010

(F)

(Non‐I)

184

(85/99)

Adult haemato‐

oncological disease

Auto: 39

Allo: 7

Ch: 49

Other: 5

Auto:32

Allo: 12

Ch: 53

Other: 3

Intermediate

(3.7^c)

McCullough 2004

(F)

(Non‐I)

645

(318/327)

Paediatric + adult haemato‐oncological

disease

Auto: 48

Allo: 28

Ch: 21

Other: 3

Auto: 52

Allo: 28

Ch: 18

Other: 2

Aph

3.5^d

Intermediate

(3.9^c)

(F)

(Non‐I)

228

(113/115)

Adult haemato‐

oncological disease

Ch: 101

Allo: 12

Ch: 102

Allo: 13

1% Aph

99% BC

1.29 PR

1.48 C

Low/intermediate

(2.9)

van Rhenen 2003

(F)

(E)

103

(52/51)

Paediatric + adult haemato‐oncological

disease

SCT: 37

Ch: 63

SCT: 37

Ch: 63

3.5

Intermediate

(4.1^c)

Single‐transfusion studies ‐ Mirasol

Johansson 2013

(F)

C‐O

(NR)

PCT‐std: 8

std‐PCT: 7

Adult haemato‐

oncological disease

2.8 PR

2.3 C

Low/intermediate

(3.0^c)

Multiple‐transfusion studies – Mirasol

Cazenave 2010

(F)

(Non‐I)

110

(56/54)

Adult haemato‐

oncological disease

25% BC

75% Aph

2.7

Intermediate

(5.2)

(F)

(Non‐I)

196

(99/97)

Adult haemato‐

oncological disease

Ch: 85

Allo: 14

Ch: 83

Allo: 14

49% Aph

51% BC

1.66 PR

1.73 C

Low/intermediate

(3.3)

Key:
A: abstract only
Allo: treated with allogeneic stem cell transplantation
Aph: apheresis
Auto: treated with autologous stem cell transplantation
BC: buffy coats
C: control
Ch: treated with chemotherapy, but without stem cell transplantation
C‐O: cross‐over trial
E: equivalence trial
F: full paper
Non‐I: non‐inferiority
NR: not reported
P2: parallel, 2 arms
P3: parallel, 3 arms
PCT: photochemically treated
Plt: platelet
PR: pathogen‐reduced
SCT: treated with stem cell transplantation (undifferentiated)
Std: standard

^aPlatelet dose has been categorised according to the low‐, intermediate‐, and high‐dose categories in the PLADO study (Slichter 2010). Low/intermediate means that the dose was between the low and intermediate categories of PLADO.
^bCross‐over design means adverse events (including mortality) were not specifically attributed to either PCT or standard transfusion.
^cStatistically significant lower mean platelet doses were issued for PCT versus standard platelet transfusions (P < 0.001 for Kerkhoffs 2010, McCullough 2004, and van Rhenen 2003). However, the doses in both arms were within the intermediate dose category of the PLADO study in all of these studies (Slichter 2010). Although these differences were statistically significant overall there was less than a 10% difference in dose between the study arms in the three trials (range 8% in McCullough 2004 to 15% in Kerkhoffs 2010).
^dStatistically significant difference between treatment arms for duration of platelet storage in McCullough 2004 (P = 0.03).

See: Summary of findings for the main comparison Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding

Table 2. Platelet transfusion data

Study	Platelet transfusion protocol	No. of plt transfusions/participant (mean ± SD)		Total % (no.) of off‐protocol transfusions
Study	Platelet transfusion protocol	PCT plts	Std plts	PCT plts	Std plts
Single‐transfusion studies ‐ Intercept
De Francisci 2004	P (NR)	1	1	NR	NR
Lozano 2011	P (10 to 20)	1	1	NA	NA
Simonsen 2006	P (10 to 20)	1	1	(2/25)	(1/25)
Slichter 2006	P (20)	1	1	NA	NA
Multiple‐transfusion studies ‐ Intercept
Agliastro 2006	T	NR	NR	NR	NR
Janetzko 2005	P (20); PP (NR) or T	4.7 ± 3.3	5.5 ± 4.7	16.5% (17/103)	7% (8/115)
Kerkhoffs 2010	P (10 to 40); PP (40 to 100) or T	5 ± 2^a	4 ± 2^a	34% (134/391)	18% (65/357)
McCullough 2004	P (NR) or T	8.4 ± 8.6^a	6.2 ± 7.0^a	8.5% (232/2715)	4.8% (101/2092)
Rebulla 2016	P (10 to 20) or T	5.9 ± 5.8^a	3.8 ± 3.4^a	2.8% (19/667)	2.0% (9/441)
van Rhenen 2003	P (20) or T	7.5 ± 5.8	5.6 ± 5.5	20.3% (79/390)	10.5% (30/286)
Single‐transfusion studies ‐ Mirasol
Johansson 2013	P (10 to 50)	1	1	NA	NA
Multiple‐transfusion studies ‐ Mirasol
Cazenave 2010	P (10 to 20); PP (50) or T	5.4 ± 3.4^b 1.2 ± 2.3^c	4.4 ± 3.4^b 1.3 ± 3.6^c	17.7% (65/368)	23.2% (72/310)
Rebulla 2016	P (10 to 20) or T	4.6 ± 4.0^a	3.4 ± 2.1^a	6.3% (29/457)	0% (0/334)

Key:

NA: not applicable
NR: not reported
P: prophylactic transfusion (threshold plt count x 10⁹/L)
PCT: photochemically treated
plt: platelet
PP: pre‐procedure transfusion (threshold plt count x 10⁹/L)
SD: standard deviation
Std: standard
T: therapeutic transfusion

^aStatistically significant difference in the number of platelet transfusions received per participant.
^bOn‐protocol transfusions only.
^cOff‐protocol transfusions only.

Intercept PCT platelet trials

Single transfusion

Three trials involved a single study platelet transfusion exposure (Lozano 2011; Simonsen 2006; Slichter 2006). De Francisci 2004 did not report the number of study platelets received, but as these were given to cover a single exposure period (surgery/pre‐procedure), we included them in the single study platelet transfusion exposure group (see Table 1).

Lozano 2011 and Simonsen 2006 stored their standard platelet products in platelet additive solution (PAS) rather than plasma. Slichter 2006 stored both Intercept platelets and standard platelets in plasma. De Francisci 2004 was only available as an abstract, and did not provide information related to the storage solution.

Multiple transfusion

Six trials involved multiple study platelet transfusion exposures (Agliastro 2006; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). In Kerkhoffs 2010, Intercept PCT platelets were compared with PASIII (platelets stored in PAS) and plasma (standard platelets suspended in plasma). van Rhenen 2003 and Rebulla 2016 stored standard platelet products in PAS rather than in plasma. Janetzko 2005 and McCullough 2004 stored standard platelet products in plasma. Agliastro 2006 was only available as an abstract, and did not provide information related to the storage solution.

Mirasol PCT platelet trials

Single transfusion

In Johansson 2013, a cross‐over design study, each participant received two prophylactic transfusions of buffy‐coat derived platelets, one Mirasol and one control platelets, stored in PAS solution for two to four days.

Multiple transfusion

In Rebulla 2016, participants were assigned to receive multiple platelets transfusions of Mirasol or standard platelets derived by buffy‐coat method or apheresis and stored in PAS for a maximum of five days. Cazenave 2010 stored standard platelet products in plasma.

Off‐protocol transfusions

Off‐protocol transfusions were permitted in all trials, including two of the cross‐over trials (Simonsen 2006; Slichter 2006), with a defined, 24‐hour washout period between study transfusions in Simonsen 2006. No information was given with regard to the proportion of participants who received off‐protocol transfusions in three trials (Agliastro 2006; De Francisci 2004; Johansson 2013).

Co‐interventions

Red cell transfusions

Participants in all trials received red cell transfusions as required. Only two trials reported red cell transfusion protocols (Rebulla 2016; Slichter 2006). In Slichter 2006, red blood cell transfusions were given to maintain a minimum haematocrit of 25% across both trial arms. In Rebulla 2016, red blood cell transfusions were given when the haemoglobin dropped below 80 g/L.

Outcomes

Of the 12 included trials, only three trials reported bleeding as a primary outcome measure (McCullough 2004; Rebulla 2016; Slichter 2006).

On the other hand, six included trials used CCI as a primary outcome measure (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; Simonsen 2006; van Rhenen 2003). In two trials the primary outcome measures were not clear (Agliastro 2006; De Francisci 2004), although both trials reported bleeding and CCI.

Intercept PCT platelet trials

All 10 included trials comparing Intercept PCT to standard platelets stated bleeding and severity of bleeding complications as outcomes. However, in the two trials reported as abstracts (Agliastro 2006; De Francisci 2004), information relevant to participants' haemostatic status, bleeding scale assessment, and the timing of bleeding observation was not provided. The World Health Organization (WHO) and the Common Terminology Criteria for Adverse Events (CTCAE) were used to assess bleeding across the remaining eight trials (Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Simonsen 2006; Slichter 2006; Rebulla 2016; van Rhenen 2003). Based on the WHO bleeding classification, four trials reported on bleeding grade 2 and above (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016), and five trials reported bleeding grade 3 and above (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003).

The period during which bleeding assessments were performed varied between trials. In four trials (Janetzko 2005; Lozano 2011; McCullough 2004; Slichter 2006), bleeding assessments were made six hours before and six hours after platelet transfusions. The longest monitoring period was reported in the Kerkhoffs 2010 trial, in which the bleeding assessment was made daily over the 42‐day study period.

Mirasol PCT platelet trials

All three trials comparing Mirasol PCT to standard platelets reported bleeding (Cazenave 2010; Johansson 2013; Rebulla 2016). Bleeding was a primary outcome measure in one trial (Rebulla 2016). The primary outcome measure in the other two trials was the platelet CCI measured one hour post‐transfusion (Cazenave 2010; Johansson 2013).

In all three trials bleeding was reported according to WHO bleeding grades. The period during which bleeding assessments were performed varied between trials. In Cazenave 2010, bleeding assessments were performed 1 hour and 24 hours after each study transfusion and on the final follow‐up visit. In Johansson 2013, bleeding assessments were performed 12 hours before and after the platelet transfusion. In Rebulla 2016, a daily bleeding assessment was made during the 28‐day study period.

Authorship and funding sources

Intercept PCT platelet trials

Only one Intercept trial, Kerkhoffs 2010, was established to be independently financed and published without connection to either Cerus Corporation (manufacturers of the Intercept PCT platelet product) or Baxter Healthcare Corporation (manufacturer of InterSol PAS), or both. Six Intercept trials were supported by either one or both of these companies, with the data held by Cerus Corporation (Janetzko 2005; Lozano 2011; McCullough 2004; Simonsen 2006; Slichter 2006; van Rhenen 2003). The author and critical reviewer of many of these trials, Dr Larry Corash, an employee of Cerus Corporation, provided additional data for all six trials. The funding source was not reported for the two abstracts (Agliastro 2006; De Francisci 2004).

The Italian Platelet Technology Assessment Study (IPTAS) study comparing both Intercept and Mirasol was sponsored by the Italian Ministry of Health (Grant Ricerca Finalizzata 2009) and cofunded by Cerus and Terumo BCT (Rebulla 2016).

Mirasol PCT platelet trials

One Mirasol trial, Cazenave 2010, was supported by CaridianBCT (manufacturers of Mirasol PCT platelets), of which the corresponding author, Dr Raymond Goodrich, is an employee. Additional information was again provided through direct contact with the Mirasol study group to supplement the published data.

Four authors in Johansson 2013 were employees of Terumo BCT. The IPTAS study comparing both Intercept and Mirasol was sponsored by the Italian Ministry of Health (Grant Ricerca Finalizzata 2009) and cofunded by Cerus and Terumo BCT (Rebulla 2016)

Ongoing studies

We identified three ongoing RCTs, two have been completed but has not yet published (NCT01789762; Kerkhoffs 2013). Both trials were non‐inferiority, parallel‐group, multicentre RCTs, one comparing Mirasol and the other comparing Intercept. The third RCT, EUCTR2015‐001035‐20‐DE, will compare Theraflex pathogen inactivation technology to standard platelets. See Characteristics of ongoing studies for details.

Intercept PCT platelet trials

NCT01789762 was conducted in France and recruited 842 adults hospitalised for bone marrow aplasia from May 2013 to January 2016.The study compared Intercept platelets with platelets prepared in PAS, with bleeding as the primary outcome. This study has not yet been published.

Mirasol PCT platelet trials

Kerkhoffs 2013 was conducted in the Netherlands and recruited 567 participants with acute myeloid leukaemia (personal communication from study author). The study is comparing pooled buffy coat‐derived Mirasol pathogen‐reduced platelets with standard plasma platelets. The trial stopped enrolling participants in May 2016 and has not yet been published (personal communication from study authors).

Theraflex PCT platelet trials

EUCTR2015‐001035‐20‐DE will be conducted in Germany in adults with haematological malignancies. The study will compare Theraflex ultraviolet platelets with plasma‐reduced platelet concentrates stored in SSP+ additive solution.

Excluded studies

We excluded 21 studies from the review following full‐text eligibility assessment (see Characteristics of excluded studies).

The reasons for exclusion were as follows.

Nine were not RCTs (Ayupova 2014; Berger 2013; Corash 2012; Drawz 2015; Infanti 2011; NCT00742001; NCT02330081; Rummler 2013; Sigle 2013).

Four were studies in healthy donors (AuBuchon 2005; Corash 2000; Dumont 2010; NCT02653443).

Five studies used the wrong intervention (Andreu 1993; Blundell 1996; MacLennan 2015; NCT01615146; TRAP 1997). Three studies compared treatment with ultraviolet B irradiation only (Andreu 1993; Blundell 1996; TRAP 1997). Ultraviolet B irradiation has been used to prevent the formation of antiplatelet alloantibodies and refractoriness to platelet transfusions (Blundell 1996), but not to reduce or remove pathogens. MacLennan 2015 compared platelets stored two to five days with platelets stored for six or seven days. NCT01615146 compared a therapeutic versus prophylactic platelet transfusion policy.
Three studies were reviews (Cid 2012; Cook 2013; Vamvakas 2012).

Risk of bias in included studies

See Figure 2 and Figure 3 for visual representations of the ’Risk of bias’ assessments across all included trials and for each individual item in the included trials. See the Characteristics of included studies section ’Risk of bias’ table for further information about the bias identified within the individual trial.

Figure 2

'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

Figure 3

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

The overall risk of bias was generally low, with much of the methodological information confirmed through direct contact with the trial authors or study group in addition to the published data. Two of the included trials were available only as abstracts (Agliastro 2006; De Francisci 2004), and therefore were often rated as at unclear risk of bias due to insufficient information.

Allocation

Intercept PCT platelet trials

Eight trials were at low risk of selection bias due to adequate methods of sequence generation and allocation concealment (Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; Simonsen 2006; Slichter 2006; van Rhenen 2003). We sought additional information from the authors of five trials in order to support the published information and confirm the method of randomisation or allocation concealment, or both (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Simonsen 2006; van Rhenen 2003).

The two trials published as an abstract did not report their method of randomisation or allocation concealment and were therefore judged to be at unclear risk of selection bias (Agliastro 2006; De Francisci 2004).

Mirasol PCT platelet trials

All three trials were at low risk of selection bias due to adequate methods of sequence generation and allocation concealment (Cazenave 2010; Johansson 2013; Rebulla 2016).

Blinding

Intercept PCT platelet trials

Blinding of participants and personnel (performance bias)

Six trials were at low risk of performance bias because both treatment providers and participants were blinded to the intervention status (Lozano 2011; McCullough 2004; Rebulla 2016; Simonsen 2006; Slichter 2006; van Rhenen 2003).

Two trials were at unclear risk of performance bias because they did not provide information on blinding (Agliastro 2006; De Francisci 2004)

Two trials were at high risk of performance bias. Janetzko 2005 was at high risk due to the lack of uniformity in the handling and blinding of the study transfusions, which could have had an impact on self reporting of bleeding, for example. Kerkhoffs 2010 was at high risk of performance and detection bias because no blinding of any kind was implemented.

Blinding of outcome assessors (detection bias)

Seven trials were at low risk of detection bias because outcome assessors were blinded (Lozano 2011; Janetzko 2005; McCullough 2004; Rebulla 2016; Simonsen 2006; Slichter 2006; van Rhenen 2003).

Two trials were at unclear risk of detection bias because they did not provide information on blinding (Agliastro 2006; De Francisci 2004).

One trial was at high risk of detection bias because no blinding of any kind was implemented (Kerkhoffs 2010).

Mirasol PCT platelet trials

Blinding of participants and personnel (performance bias)

One trial was at low risk of performance bias because both trial participants and personnel were unaware of the intervention status (Rebulla 2016).

One trial was at unclear risk of performance bias because no clear information was provided (Johansson 2013).

One trial was at high risk of performance bias because the lack of blinding of the pathogen‐reduced platelet appearance could have affected ordering of additional platelets transfusions by the attending physician (Cazenave 2010).

Blinding of outcome assessors

All three trials were at low risk of detection bias because outcome assessors were blinded (Cazenave 2010; Johansson 2013; Rebulla 2016).

Incomplete outcome data

Intercept PCT platelet trials

Seven trials were at low risk of attrition bias because no participants were lost to follow‐up (De Francisci 2004); all participants who received a platelet transfusion were included in the analysis (Janetzko 2005; Kerkhoffs 2010; Rebulla 2016); all participants who received a platelet transfusion using the modified method of production were included in the analysis (Slichter 2006); or there was balanced loss to follow‐up between treatment arms, and the numbers lost to follow‐up were low (Lozano 2011; McCullough 2004).

Three trials were at unclear risk of attrition bias. Agliastro 2006 was only reported as an abstract, and no information was provided about whether any randomised participants were lost to follow‐up or withdrawn from treatment. In van Rhenen 2003, more participants receiving the intervention completed cycle 1 of the study than those receiving standard platelet transfusions (83% versus 67% completed cycle 1; P = 0.06). In Simonsen 2006, 28 participants were randomised, but only 20 were included in the per‐protocol analysis (3 did not receive a platelet transfusion; 3 received only 1 of the 2 types of platelet transfusion; and there were 2 major protocol violations). The number of participants who received a platelet transfusion that were lost to follow‐up was balanced between treatment arms (3 out of 12 randomised to receive an Intercept PCT platelet transfusion first and 2 out of 13 randomised to receive a standard platelet transfusion first).

Mirasol PCT platelet trials

Two trials were at low risk of attrition bias because all participants who received a platelet transfusion were included in the analysis (Cazenave 2010; Rebulla 2016).

One trial was at unclear risk of attrition bias. Johansson 2013 enrolled 16 participants, but only 10 participants were included in the study's primary outcome; 12 participants were included in other efficacy outcomes; and 15 participants were included in safety outcomes.

Selective reporting

Intercept PCT platelet trials

Seven trials were at low risk of reporting bias because all prespecified outcomes were reported in the publications or provided via direct author contact (Janetzko 2005; Kerkhoffs 2010; Rebulla 2016; Simonsen 2006; Slichter 2006; van Rhenen 2003).

One trial reported only as an abstract was at unclear risk of reporting bias (Agliastro 2006).

One trial was at high risk of reporting bias. In De Francisci 2004, the 1‐hour CCI was not reported for the adult cirrhotic group; no standard deviations were reported for the mean 1‐ and 24‐hour CCIs for either group, and pre‐ and post‐transfusion platelet counts for both groups were missing despite their inclusion as outcomes in the trial's methods section.

Mirasol PCT platelet trials

Two trials were at low risk of reporting bias because all prespecified outcomes were reported in the publications or provided via direct author contact (Cazenave 2010; Rebulla 2016).

One trial was at unclear risk of reporting bias because the trial protocol was not available, and trial registration occurred after the trial had been completed (Johansson 2013).

Other potential sources of bias

Intercept PCT platelet trials

Eight trials were at low risk of other potential sources of bias apart from the outcome of bleeding (Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; Simonsen 2006; Slichter 2006).

Six trials reported on red cell transfusions received as an indirect measure of bleeding (Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; van Rhenen 2003; Slichter 2006), but only one of these reported a red cell transfusion policy of maintaining a haematocrit of at least 25% across both trial arms (Slichter 2006). Variations in red cell transfusion policies across centres within a trial could affect the assessment of bleeding grade and therefore lead to bias. Also, variations in the use of red cell transfusions between trials could affect the results of any meta‐analysis (Estcourt 2012).

Duplication of participants through inclusion in more than one study cycle was identified in three Intercept trials (McCullough 2004; Simonsen 2006; van Rhenen 2003). One participant out of 20 received both study transfusions on two separate occasions in Simonsen 2006. This was not thought to pose a high risk of bias, as a single participant included twice should not have a significant impact on the accuracy of the trial's outcomes, despite the small size of the study. McCullough 2004 reported data from only the first cycle of study, whilst the data for van Rhenen 2003 were extracted from cycle 1 only, either from the published literature or obtained through direct author contact.

Two trials were at unclear risk of other bias because they were only reported as abstracts (Agliastro 2006; De Francisci 2004).

Mirasol PCT platelet trials

Two trials were at low risk of other potential sources of bias (Cazenave 2010; Rebulla 2016).

Johansson 2013 was at unclear risk of other potential sources of bias because the trial was conducted mainly on male participants, and the study registration was made after trial completion.

Effects of interventions

For an overview of the main findings of the review, please see summary of findings Table for the main comparison.

We included three cross‐over trials (Johansson 2013; Simonsen 2006; Slichter 2006). We were unable to use data provided by Johansson 2013 in any of the meta‐analyses as there was no clear washout period between the two trial phases.

Where possible, for all outcomes we reported on overall differences between pathogen‐reduced platelet transfusions compared with standard platelet transfusions, then afterward reported the same outcomes for Intercept platelets and Mirasol platelets individually.

Primary outcomes

Number, type, timing, and severity of bleeding episodes

All 12 trials stated bleeding as an outcome. Bleeding was the primary outcome measure in two trials (McCullough 2004; Rebulla 2016). The duration of bleeding assessment varied between trials. Five trials reported bleeding events for up to 48 hours (De Francisci 2004; Johansson 2013; Lozano 2011; Simonsen 2006; Slichter 2006), and 7 trials reported bleeding events for more than 7 days (Table 3) (Agliastro 2006; Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). No trials reported bleeding between 48 hours and 7 days.

Table 3. Bleeding assessment

Study	Bleeding scale used	Duration of bleeding assessment	Timing of bleeding assessment(s)	Maximum number of days of on‐study plt transfusion support	Bleeding results reported in meta‐analysis
Single‐transfusion studies ‐ Intercept
De Francisci 2004	NR	S	NR	1	NR
Lozano 2011	WHO¹	S	6 hrs pre‐ and 6 hrs post‐Tx	1	Post‐Tx bleeding score
Simonsen 2006	WHO¹ (m)	S	≤ 12 hrs post‐Tx	1	Post‐Tx bleeding score
Slichter 2006	WHO¹ (m)	S	6 hrs pre‐ and 6 hrs post‐Tx	1	Post‐Tx bleeding score
Multiple‐transfusion studies ‐ Intercept
Agliastro 2006	NR	L	NR	NR	NR
Janetzko 2005	WHO¹	L	6 hrs pre‐ and post‐ each plt Tx + daily	35	Max bleeding score/participant
Kerkhoffs 2010	CTCAE²	L	Post‐1st Tx + daily	42	Max bleeding score/participant
McCullough 2004	WHO¹ (e)	L	12 hrs post‐Tx + daily	28	Max bleeding score/participant
Rebulla 2016	WHO¹	L	Post‐1st Tx + daily	28	Max bleeding score/participant
van Rhenen 2003	WHO¹ (m)	L	6 hrs pre‐ and 6 hrs post‐Tx	56	Max bleeding score/participant
Single‐transfusion studies ‐ Mirasol
Johansson 2013	WHO¹	S	12 hours pre‐ and post‐Tx	1	NR
Multiple‐transfusion studies ‐ Mirasol
Cazenave 2010	WHO¹	L	1 hr pre‐ and post‐Tx + 24 hrs post‐Tx + final follow‐up visit^a	28	Max bleeding score/participant
Rebulla 2016	WHO¹	L	Post‐1st Tx + daily	28	Max bleeding score/participant

Key:

(e): scale expanded (more specifically defined WHO grades, including sites of bleeding)
L: long‐term bleeding assessment (> 7 days) post‐transfusion
(m): scale modified (only 3 scores: none = 0, minor = 1 (equivalent to WHO grades 1 and 2), major = 2 (equivalent to WHO grades 3 and 4))
NA: not applicable
NR: not reported
plt: platelet
S: short‐term bleeding assessment (up to 48 hours) post‐transfusion
M: medium‐term bleeding (2 to 7 days) post‐transfusion
Tx: transfusion

^aOn‐protocol platelet transfusions only.

References:

1. WHO scale – WHO. WHO Handbook for Reporting Results of Cancer Treatment. Geneva: World Health Organization; 1979.
2. U.S. Dept of Health and Human Services (National Institutes of Health and National Cancer Institute). Common Terminology Criteria for Adverse Events (CTCAE) Version 3.0, September 2006.

For trials that reported bleeding for up to 48 hours, we extracted the post‐transfusion bleeding assessment data from the trial reports.

For trials that reported bleeding for more than seven days, we extracted the number of participants with bleeding documented during a specific bleeding assessment. These bleeding assessments were reported to be undertaken on at least a daily basis in four trials (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016); pre‐ and post‐platelet transfusion in two trials (Cazenave 2010; van Rhenen 2003); and were not described in two trials (Agliastro 2006; De Francisci 2004). We did not extract bleeding data reported within a trial's adverse event data.

We undertook meta‐analyses for 'any bleeding' (WHO grades 1 to 4, CTCAE 1 to 5 or equivalent), 'clinically significant bleeding' (WHO/CTCAE grade 2 and above or equivalent), and 'severe bleeding' (WHO/CTCAE grade 3 and above or equivalent), with trials grouped according to the duration of bleeding assessment.

As might be expected, the single platelet transfusion trials assessed bleeding for up to 48 hours, while the multiple platelet transfusion trials assessed bleeding over a period of more than seven days.

Any bleeding (WHO) grades 1 to 4 ‐ assessed at up to 48 hours

Four trials assessed the risk of any bleeding at less than 48 hours (Johansson 2013; Lozano 2011; Simonsen 2006; Slichter 2006). One cross‐over trial did not report a washout period (Johansson 2013). Post‐transfusion, there were three bleeding events (petechiae or purpura) after both pathogen‐reduced platelet transfusions and standard platelet transfusions. The three remaining trials all compared Intercept and standard platelets (Lozano 2011; Simonsen 2006; Slichter 2006).

There was no evidence of a difference in the risk of any bleeding between the Intercept and standard platelets (3 trials, 309 participants; risk ratio (RR) 0.86, 95% confidence interval (CI) 0.63 to 1.19; P = 0.38; I² = 0%) (Analysis 1.1).

Any bleeding (WHO) grades 1 to 4 ‐ assessed at more than 7 days

Five trials reported this outcome (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; van Rhenen 2003).

We are very uncertain as to whether recipients of pathogen‐reduced platelets have a higher risk of any bleeding. Using a fixed‐effect analysis, recipients of pathogen‐reduced platelets were at a higher risk of any bleeding when compared with standard platelets (5 trials, 1085 participants; RR 1.09, 95% CI 1.02 to 1.15; P = 0.007; I² = 59%) (Analysis 1.2). However, using a random‐effects model, there was no longer any evidence of a difference between treatment arms (5 trials, 1085 participants; RR 1.14, 95% CI 0.93 to 1.38; P = 0.20; I² = 59%) (analysis not shown).

Intercept PCT platelet trials

Four Intercept trials reported this outcome (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; van Rhenen 2003).

We are very uncertain as to whether recipients of Intercept platelet transfusions have a higher risk of any bleeding. Using a fixed‐effect analysis, recipients of Intercept platelet transfusions were at a higher risk of any bleeding when compared with standard platelets (4 trials, 975 participants; RR 1.07, 95% CI 1.01 to 1.13; P = 0.03; I² = 47%) (Analysis 1.2). However, using a random‐effects model, there was no longer any evidence of a difference between treatment arms (4 trials, 975 participants; RR 1.08, 95% CI 0.91 to 1.30; P = 0.37; I² = 47%) (analysis not shown).

Mirasol PCT platelet trials

One Mirasol trial reported the risk of any bleeding (Cazenave 2010). There was no evidence of a difference between the Mirasol and control platelets (1 trial, 110 participants; RR 1.38, 95% CI 0.95 to 2.02; P = 0.09) (Analysis 1.2).

There was no evidence of subgroup differences between studies of Intercept and Mirasol platelets (test for subgroup differences: Chi² = 1.78, df = 1, P = 0.18, I² = 43.8%).

Clinically significant bleeding WHO grade 2 or above ‐ assessed at up to 48 hours

We could not incorporate data from three trials into an analysis. Slichter 2006 reported bleeding events as either minor or major and, as minor bleeding could not be split into the equivalent WHO/CTCAE grades 1 and 2, these results were not included in this outcome. In Johansson 2013 there was one bleeding event (WHO grade 2 petechiae or purpura) after a standard platelet transfusion. In Simonsen 2006 there was one bleeding event (WHO grade 2) in the Intercept PCT platelets arm and no bleeding events in the standard platelet transfusion arm. We could not include data from these two trials in a meta‐analysis using the generic inverse variance method because no events occurred in one of the study arms (Johansson 2013; Simonsen 2006).

In Lozano 2011 there was no evidence of a difference in clinically significant bleeding between treatment arms (1 trial, 211 participants; RR 1.01, 95% CI 0.34 to 3.03; P = 0.99) (Analysis 1.3).

Clinically significant bleeding WHO grade 2 or above ‐ assessed at more than 7 days

van Rhenen 2003 reported bleeding events as either minor or major and, as minor bleeding could not be split into the equivalent WHO/CTCAE grades 1 and 2, these results could not be included in this outcome.

Five trials reported this outcome (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016). There was no evidence of a difference in the risk of clinically significant bleeding between recipients of pathogen‐reduced platelets and recipients of standard platelets (5 trials, 1392 participants; RR 1.10, 95% CI 0.97 to 1.25; P = 0.15; I² = 0%) (Analysis 1.4).

Intercept PCT platelet trials

Four Intercept trials reported this outcome (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016). There was no evidence of a difference in the risk of clinically significant bleeding between recipients of Intercept platelets and standard platelets (4 trials, 1088 participants; RR 1.06, 95% CI 0.94 to 1.21; P = 0.34; I² = 0%) (Analysis 1.4). However, this analysis was dominated by the large McCullough 2004 study (over 80% of weighting).

Mirasol PCT platelet trials

Two trials reported this outcome (Cazenave 2010; Rebulla 2016).

There was no evidence of a difference in the risk of clinically significant bleeding between recipients of Mirasol platelets and standard platelets (2 trials, 304 participants; RR 1.54, 95% CI 0.86 to 2.76; P = 0.15; I² = 0%) (Analysis 1.4).

There was no evidence of subgroup differences between studies of Intercept and Mirasol platelets (test for subgroup differences: Chi² = 1.44, df = 1, P = 0.23, I² = 30.8%).

Severe bleeding (WHO) grade 3 or above ‐ assessed at up to 48 hours

Four trials reported this outcome (Johansson 2013; Lozano 2011; Simonsen 2006; Slichter 2006). No severe bleeding events occurred in three trials (Johansson 2013; Simonsen 2006; Slichter 2006).

In Lozano 2011 there was no evidence of a difference in severe bleeding outcomes between treatment arms (1 trial, 211 participants; RR 2.02, 95% CI 0.19 to 21.93; P = 0.56) (Analysis 1.5).

Severe bleeding (WHO) grade 3 or above ‐ assessed at more than 7 days

Six trials reported this outcome (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003).

There was no evidence of a difference in the risk of developing clinically severe bleeding between recipients of pathogen‐reduced platelets and standard platelets (6 trials, 1495 participants; RR 1.24, 95% CI 0.76 to 2.02; P = 0.39; I² = 32%) (Analysis 1.6). This analysis was dominated by the large McCullough 2004 study (over 70% of weighting).

We defined an I² value higher than 30% as moderate statistical heterogeneity. Applying the random‐effects model did not alter the results (6 trials, 1495 participants; RR 1.51, 95% CI 0.67 to 3.40; P = 0.32; I² = 32%) (analysis not shown).

In McCullough 2004 more participants in the standard platelet transfusion arm had severe bleeding, whereas in the other three trials the results favoured PCT platelets. Removing McCullough 2004 data from the analysis removed all statistical heterogeneity, and suggested that participants receiving PCT platelets in the other three trials may be at a higher risk of developing severe bleeding (5 trials, 850 participants; RR 2.64, 95% CI 1.19 to 5.86; P = 0.02; I² = 0%) (analysis not shown).

Intercept PCT platelet trials

Five trials reported this outcome (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). There was no evidence of a difference in the risk of developing clinically severe bleeding between recipients of Intercept and standard platelets (5 trials, 1191 participants; RR 1.21, 95% CI 0.70 to 2.09; P = 0.49; I² = 48%) (Analysis 1.6). This analysis was dominated by the large McCullough 2004 study (over 85% of weighting).

We defined an I² value higher than 30% as moderate statistical heterogeneity. Applying the random‐effects model did not alter the results (5 trials, 1191 participants; RR 2.08, 95% CI 0.60 to 7.20; P = 0.25; I² = 48%) (analysis not shown).

Removing McCullough 2004 data from the analysis removed all statistical heterogeneity, and suggested that participants receiving Intercept platelets in the remaining trials may be at a higher risk of developing severe bleeding (4 trials, 546 participants; RR 4.84, 95% CI 1.41 to 16.60; P = 0.01; I² = 0%) (analysis not shown).

Mirasol PCT platelet trials

Two multiple‐transfusion trials reported this outcome (Cazenave 2010;Rebulla 2016). There was no evidence of a difference in the risk of developing clinically severe bleeding between recipients of Mirasol and standard platelets (2 trials, 304 participants; RR 1.36, 95% CI 0.45 to 4.16; P = 0.59; I² = 0%) (Analysis 1.6).

There was no evidence of subgroup differences between studies of Intercept and Mirasol platelets (test for subgroup differences: Chi² = 0.03, df = 1, P = 0.85, I² = 0%).

Type of bleeding

Six trials reported the sites of bleeding with very variable methods of assessment and reporting, and so we have reported them narratively here (Janetzko 2005; Lozano 2011; McCullough 2004; Rebulla 2016; van Rhenen 2003).

Where reported, the majority of bleeding occurred at mucocutaneous sites (Janetzko 2005; Johansson 2013; Lozano 2011; McCullough 2004; Rebulla 2016; van Rhenen 2003). Invasive, gastrointestinal, and genitourinary WHO grade 2 or more bleeding events were seen quite frequently in McCullough 2004 (18.9% to 32.7%), although few such episodes were reported by the three other trials. However, McCullough 2004 was a larger trial that reported on bleeding daily and did not exclude people with active bleeding.

Mortality

Ten trials reported mortality over a variety of time periods (Table 1) (Cazenave 2010; Janetzko 2005; Johansson 2013; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; Simonsen 2006; Slichter 2006; van Rhenen 2003). Two trials did not report this outcome (Agliastro 2006; De Francisci 2004).

As Johansson 2013 (no deaths), Simonsen 2006 (no deaths), and Slichter 2006 (three deaths) were cross‐over trials, it was not possible to assign deaths to a single trial arm, therefore we could not include these data in any meta‐analysis. We have not reported data from these cross‐over trials further in this section on mortality.

We defined short‐term mortality as up to 4 weeks, medium‐term mortality as 4 weeks to 12 weeks, and long‐term mortality as greater than 12 weeks. One trial (a single‐transfusion trial) reported all‐cause mortality for up to 4 weeks after the start of the study period (Lozano 2011), and six trials (all multiple‐transfusion trials) reported all‐cause mortality for 4 to 12 weeks (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003).

All‐cause mortality (up to 4 weeks)

One single‐transfusion trial reported this outcome (Lozano 2011).

There was no evidence of a difference in all‐cause mortality between the two study arms (1 trial, 211 participants; RR 0.14, 95% CI 0.01 to 2.76) (Analysis 1.7).

Mortality due to bleeding, thromboembolism, or transfusion reactions (up to 4 weeks)

One single‐transfusion trial reported these outcomes (Lozano 2011). No deaths occurred due to bleeding, thromboembolism, or transfusion reactions.

Mortality due to infection (up to 4 weeks)

One single‐transfusion trial reported these outcomes (Lozano 2011). There was no evidence of a difference in mortality due to infection between the two study arms (1 trial, 211 participants; RR 0.20, 95% CI 0.01 to 4.16) (Analysis 1.7).

All‐cause mortality (4 weeks to 12 weeks)

Six multiple‐transfusion trials reported this outcome (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003).

There was no evidence of a difference in all‐cause mortality between participants receiving pathogen‐reduced platelets and participants receiving standard platelets (6 trials, 1509 participants; RR 0.81, 95% CI 0.50 to 1.29; P = 0.37; I² = 26%) (Analysis 1.8).

Intercept PCT platelet trials

There was no evidence of a difference in all‐cause mortality between participants receiving Intercept PCT platelets and participants receiving standard platelets (5 trials, 1203 participants; RR 0.62, 95% CI 0.37 to 1.05; P = 0.07; I² = 0%) (Analysis 1.8).

Mirasol PCT platelet trials

There was no evidence of a difference in all‐cause mortality between participants receiving Mirasol PCT platelets and participants receiving standard platelets (2 trials, 306 participants; RR 2.92, 95% CI 0.81 to 10.59; P = 0.10; I² = 0%) (Analysis 1.8).

There was evidence of subgroup differences between studies of multiple Intercept and Mirasol platelets transfusions (test for subgroup differences: Chi² = 4.93, df = 1, P = 0.03, I² = 79.7%).

Mortality due to bleeding (4 weeks to 12 weeks)

Six multiple‐transfusion trials reported mortality due to bleeding (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003), and there was no evidence of a difference between treatment arms (6 trials, 1509 participants; RR 1.23, 95% CI 0.38 to 3.96; P = 0.73, I² = 0%) (Analysis 1.9).

Intercept PCT platelet trials

There was no evidence of a difference in mortality due to bleeding between participants receiving Intercept PCT platelets and participants receiving standard platelets (5 trials, 1203 participants; RR 1.04, 95% CI 0.29 to 3.75; P = 0.96; I² = 0%) (Analysis 1.9).

Mirasol PCT platelet trials

There was no evidence of a difference in mortality due to bleeding between participants receiving Mirasol PCT platelets and participants receiving standard platelets (2 trials, 306 participants; RR 2.89, 95% CI 0.12 to 69.55; P = 0.51; I² = 0%) (Analysis 1.9).

There was no evidence of subgroup differences between studies of multiple Intercept and Mirasol platelets transfusions (test for subgroup differences: Chi² = 0.34, df = 1, P = 0.56, I² = 0%).

Mortality due to infection (4 weeks to 12 weeks)

Six multiple‐transfusion trials reported mortality due to infection (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003), and there was no evidence of a difference between treatment arms (6 trials, 1509 participants; RR 0.94, 95% CI 0.42 to 2.09; P = 0.88; I² = 0%) (Analysis 1.10).

Intercept PCT platelet trials

There was no evidence of a difference in mortality due to infection between participants receiving Intercept PCT platelets and participants receiving standard platelets (5 trials, 1203 participants; RR 0.84, 95% CI 0.35 to 2.03; P = 0.70; I² = 0%) (Analysis 1.10).

Mirasol PCT platelet trials

There was no evidence of a difference in mortality due to infection between participants receiving Mirasol PCT platelets and participants receiving standard platelets (2 trials, 306 participants; RR 1.62, 95% CI 0.22 to 12.07; P = 0.63; I² = 0%) (Analysis 1.10).

There was no evidence of subgroup differences between studies of multiple Intercept and Mirasol platelets transfusions (test for subgroup differences: Chi² = 0.36, df = 1, P = 0.55, I² = 0%).

Mortality due to thromboembolism (4 weeks to 12 weeks)

Five multiple‐transfusion trials reported on the incidence of thromboembolism (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). No deaths were attributed to thromboembolism in any of the trials.

Mortality due to transfusion reactions (4 weeks to 12 weeks)

Six multiple‐transfusion trials reported anaphylaxis reactions (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). No deaths were attributed to transfusion reactions.

Secondary outcomes

Adverse events

All 12 trials reported adverse events. Six trials included bleeding, platelet refractoriness, and/or acute transfusion reactions within the definition of adverse or serious adverse events (Cazenave 2010; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; van Rhenen 2003).

Due to the difficulties of assigning an adverse event to a product within cross‐over trials, Simonsen 2006 and Slichter 2006 did not report adverse events in detail. Slichter 2006 reported safety outcomes for all 50 randomised participants, although 18 participants received Intercept PCT platelets prepared using a different protocol and were excluded in the clinical efficacy outcomes.

The duration of follow‐up for adverse events varied widely. Participants in Johansson 2013 were observed for acute transfusion reactions and transfusion‐related serious adverse events over the 24 hours after platelet transfusion. De Francisci 2004 reported events up to six months after the single study transfusion episode, whilst Lozano 2011 reported any adverse event until four days post‐transfusion, but serious adverse events up to 15 days post‐transfusion. Rebulla 2016, Janetzko 2005, McCullough 2004, Cazenave 2010, and van Rhenen 2003 reported both any adverse events and serious adverse events that occurred during the study and follow‐up periods (a maximum of 28, 35, 70, and 84 days after the start of the study period).

Frequently, studies reported 'any adverse events' or 'serious adverse events' only if they affected more than a certain proportion of the trial population. In order to extract data on the frequency of specific adverse events, we attempted direct author contact and were successful in all but two cases (Agliastro 2006; De Francisci 2004).

Recognised complications of platelet transfusions

Acute (mild) transfusion reactions

Twelve trials reported on acute transfusion reactions (defined in our protocol as rigors, fever, skin rash, and urticaria) (Agliastro 2006; Cazenave 2010; De Francisci 2004; Janetzko 2005; Johansson 2013; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; Simonsen 2006; Slichter 2006; van Rhenen 2003). Three multiple Intercept transfusion trials reported transfusion reactions within six hours post‐transfusion (Janetzko 2005; McCullough 2004; van Rhenen 2003). Three single‐transfusion trials reported transfusion reactions within 24 hours post‐transfusion (Johansson 2013; Lozano 2011; Simonsen 2006). The remaining six trials did not report the timing for this outcome (Agliastro 2006; Cazenave 2010; De Francisci 2004; Kerkhoffs 2010; Rebulla 2016; Slichter 2006).

We could include data from seven trials in the meta‐analysis for this outcome (Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; Simonsen 2006; van Rhenen 2003), which we combined (due to the inclusion of data from a cross‐over trial) using the generic inverse variance method. There was no difference between treatment groups for the risk of an acute transfusion reaction (7 trials, 1636 participants; RR 0.96, 95% CI 0.75 to 1.24; P = 0.77; I2 = 0%) (Analysis 1.11).

Five trials could not be included in the meta‐analysis, as trials with zero events in one or both treatment arms cannot be included when using the generic inverse variance method of analysis (Agliastro 2006; Cazenave 2010; De Francisci 2004; Johansson 2013; Slichter 2006). Three studies reported no acute transfusion reactions in either trial arm (Agliastro 2006; De Francisci 2004; Johansson 2013). Slichter 2006 reported no events in the PCT platelets arm (although this trial reported acute transfusion reactions with a broader definition than our protocol and included nausea, which was experienced by 5.2% of the Intercept PCT platelets arm) versus five participants (13%) experiencing six events in the standard platelets arm (chills, rash/urticaria, hypertension, bronchospasm, and fever). Cazenave 2010 did not report the total number of participants with acute transfusion reactions. However, it should be noted that Cazenave 2010 did report the number of participants experiencing adverse events that were considered "very likely" to be related to the study platelet transfusions, with no participants in the PCT platelets arm experiencing such an adverse event versus two participants in the standard platelets arm (recorded as anaphylaxis, hypersensitivity, and eyelid oedema), but with no difference between groups (P = 0.28).

Intercept PCT platelet trials

Two single Intercept PCT platelet transfusion trials reported on this outcome with no difference between the two arms (2 trials, 251 participants; RR 1.57, 95% CI 0.71 to 3.48; P = 0.27; I² = 0%) (Analysis 1.11) (Lozano 2011; Simonsen 2006).

Five multiple Intercept PCT platelet transfusion trials reported this outcome (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). There was no evidence of a difference between recipients of Intercept or standard platelets on acute transfusion reactions (5 trials, 1191 participants; RR 0.91, 95% CI 0.69 to 1.18, P = 0.46; I² = 0%) (Analysis 1.11).

Mirasol PCT platelet trials

One multiple‐transfusion trial reported this outcome with no evidence of a difference between the two groups (1 trial, 194 participants; RR 1.99, 95% CI 0.18 to 21.78; P = 0.57) (Analysis 1.11) (Rebulla 2016).

Johansson 2013 reported no adverse transfusion reactions or transfusion‐related serious adverse events within the 24 hours after study transfusions.

There was no evidence of subgroup differences between studies of multiple Intercept and Mirasol platelets transfusions (test for subgroup differences: Chi² = 0.41, df = 1, P = 0.52, I² = 0%).

Platelet refractoriness

Seven trials reported refractoriness to platelet transfusions (Agliastro 2006; Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). Five trials defined refractoriness as two successive one‐hour post‐transfusion CCIs below 5 x 10³ (Cazenave 2010; Janetzko 2005; McCullough 2004; Rebulla 2016; van Rhenen 2003), while Kerkhoffs 2010 defined refractoriness as two successive one‐hour or 24‐hour CCIs below 7.5 x 10³ or 4.5 x 10³, respectively, and the presence of antibodies against platelets. Agliastro 2006 did not define refractoriness.

In six trials people with a previous history of clinical refractoriness to platelet transfusions were not eligible for inclusion (Cazenave 2010; Janetzko 2005; Johansson 2013; Kerkhoffs 2010; McCullough 2004; van Rhenen 2003). Agliastro 2006 was only available as an abstract and did not report the study inclusion and exclusion criteria.

Combining data from seven multiple platelet transfusion trials showed that recipients of pathogen‐reduced platelets were at a higher risk of developing refractoriness compared with recipients of standard platelets (7 trials, 1525 participants; RR 2.94, 95% CI 2.08 to 4.16; P < 0.001; I² = 0%) (Analysis 1.12) (Agliastro 2006; Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). It should be noted that one large study dominated the analysis (57.9% of overall study weighting) (McCullough 2004), and for one other trial an effect could not be estimated (Agliastro 2006).

Intercept PCT platelet trials

Six Intercept multiple‐transfusion trials contributed to this outcome (Agliastro 2006; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). Refractoriness to platelet transfusions was more common in Intercept platelet recipients than standard platelet recipients (6 trials, 1221 participants; RR 2.85, 95% CI 1.96 to 4.15; P < 0.001; I² = 0%) (Analysis 1.12).

Three trials assessed whether amotosalen neoantigens were formed following exposure to Intercept PCT platelets, but no neoantigens were identified (Janetzko 2005; McCullough 2004; van Rhenen 2003).

Mirasol PCT platelet trials

Two Mirasol multiple‐transfusion trials reported this outcome (Cazenave 2010; Rebulla 2016). Refractoriness to platelet transfusions was more common in Mirasol platelet recipients than standard platelet recipients (2 trials, 304 participants; RR 3.47, 95% CI 1.44 to 8.34; P = 0.005; I² = 17%) (Analysis 1.12). Johansson 2013 did not report this outcome.

There was no evidence of subgroup differences between studies of multiple Intercept and Mirasol platelets (test for subgroup differences: Chi² = 0.16, df = 1, P = 0.69, I² = 0%) (Analysis 1.12).

Platelet refractoriness and platelet alloimmunisation

The CCIs formed the basis of most trials' definitions of refractoriness, therefore, in order to explore the development of platelet refractoriness specifically due to alloimmunisation, we performed a further analysis including only data from trials that reported on participants diagnosed both with their trial's definition of clinical refractoriness and with the presence of platelet alloantibodies.

Six multiple‐transfusion trials reported this outcome (Agliastro 2006; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). Five trials excluded people with a previous history of platelet alloimmunisation (Kerkhoffs 2010; Johansson 2013; McCullough 2004; Rebulla 2016; van Rhenen 2003). Janetzko 2005 did not exclude people with a previous history of alloimmunisation, and Agliastro 2006 did not report their inclusion and exclusion criteria.

Participants receiving pathogen‐reduced platelet transfusions were at a higher risk of developing platelet refractoriness due to platelet alloimmunisation (6 trials, 1415 participants; RR 2.35, 95% CI 1.46 to 3.76; P < 0.001; I² = 0%) (Analysis 1.13).

Intercept PCT platelet trials

We combined data from six Intercept multiple‐transfusion trials (Agliastro 2006; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). Participants receiving Intercept platelet transfusions were at a higher risk of developing platelet refractoriness due to platelet alloimmunisation (6 trials, 1221 participants; RR 1.90, 95% CI 1.11 to 3.26; P = 0.02; I² = 0%) (Analysis 1.13).

Mirasol PCT platelet trials

One multiple‐transfusion trial reported this outcome (Rebulla 2016), and participants receiving Mirasol platelet transfusions were at a higher risk of developing platelet refractoriness due to platelet alloimmunisation (1 trial, 194 participants; RR 4.50, 95% CI 1.58 to 12.81; P = 0.005) (Analysis 1.13).

There was no evidence of subgroup differences between studies of multiple Intercept and Mirasol platelets transfusions (test for subgroup differences: Chi² = 2.06, df = 1, P = 0.15, I² = 51.5%).

All adverse events

Nine trials reported on all adverse events, five of which used the CTCAE Version 3.0, National Cancer Institute ‐ Common Toxicity Criteria (NCI‐CTC), and Medical Dictionary for Regulatory Activities (MedDRA) standards to categorise the adverse events (Cazenave 2010; Janetzko 2005; Lozano 2011; McCullough 2004; van Rhenen 2003), and three of which failed to state how adverse events were defined (De Francisci 2004; Simonsen 2006; Slichter 2006). In Rebulla 2016 participants were evaluated for adverse events, which were assessed for relation to the platelet transfusion and graded for clinical severity by the study clinical investigator. Although we had intended to exclude bleeding, transfusion reactions, and mortality data from the adverse events data for each trial, most trials did not present their data in a way that allowed for this, and so we have included all adverse events.

Regarding the three cross‐over trials, Simonsen 2006 reported one adverse event in the standard platelets arm during the 24‐hour post‐transfusion period; Slichter 2006 simply reported adverse events as "common" in both treatment arms during the entire follow‐up period of four weeks, with no additional details given; and in Johansson 2013, adverse transfusion reactions were assessed through physical examination and interviews with the study participants, but no adverse events were reported. Neither Agliastro 2006 nor Kerkhoffs 2010 provided usable data on all adverse events.

When we combined all of the adverse events data that were available from seven trials (Cazenave 2010; De Francisci 2004; Janetzko 2005; Lozano 2011; McCullough 2004; Rebulla 2016; van Rhenen 2003), there was no evidence of a difference between treatment arms (7 trials, 1566 participants; RR 1.01, 95% CI 0.97 to 1.05; P = 0.67; I² = 0%) (Analysis 1.14).

Intercept PCT platelet trials

Data from two single‐transfusion trials showed no evidence of a difference between the two arms (2 trials, 255 participants; RR 0.98, 95% CI 0.87 to 1.09; P = 0.68) (Analysis 1.14) (De Francisci 2004; Lozano 2011).

When we combined data from four multiple‐transfusion trials (Janetzko 2005; McCullough 2004; Rebulla 2016; van Rhenen 2003), there was no evidence of a difference between treatment arms (4 trials, 1007 participants; RR 1.01, 95% CI 0.97 to 1.05; P = 0.61; I² = 0%) (Analysis 1.14).

Mirasol PCT platelet trials

There was no evidence of a difference in the number of adverse events between the recipients of Mirasol and standard platelet transfusions (2 trials, 304 participants; RR 1.04, 95% CI 0.89 to 1.21; P = 0.65; I² = 0%) (Analysis 1.14).

There was no evidence of subgroup differences between studies of multiple Intercept and Mirasol platelets transfusions (test for subgroup differences: Chi² = 0.11, df = 1, P = 0.74, I² = 0%)

Serious adverse events

Nine of the 12 trials reported all serious adverse events (Cazenave 2010; De Francisci 2004; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Simonsen 2006; Slichter 2006; van Rhenen 2003), with six trials providing a definition of a serious adverse event that conformed either to the European Medicines Agency's, EMEA 1995, or the US Food and Drug Administration's, US FDA 2012, definition of a serious adverse event (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; van Rhenen 2003). Although we had intended to exclude bleeding and mortality data from the serious adverse events data for each trial, most trials did not present their data in a way that allowed for this, and so we have included all serious adverse events. In addition, it should be noted that, with the exception of infection, these were all rare events for which the trials were not sufficiently powered to identify differences between treatment arms.

Regarding the two cross‐over trials that reported this outcome (Simonsen 2006; Slichter 2006), Simonsen 2006 reported one serious adverse event (hypotension) in the standard platelet transfusion arm during the 24‐hour period post‐transfusion, and Slichter 2006 reported one serious adverse event (severe hypertension) in the standard platelet transfusion arm.

When we combined data from the remaining seven trials (Cazenave 2010; De Francisci 2004; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; van Rhenen 2003), there was no evidence of a difference in the frequency of serious adverse events between the trial arms (7 trials, 1340 participants; RR 1.09, 95% CI 0.88 to 1.35; P = 0.44; I² = 0%) (Analysis 1.15).

Intercept PCT platelet trials

Data extracted from two single‐transfusion trials showed no evidence of a difference in the frequency of serious adverse events between recipients of Intercept and standard platelets (2 trials, 255 participants; RR 1.21, 95% CI 0.55 to 2.68; P = 0.64; I² = 0%) (Analysis 1.15) (De Francisci 2004; Lozano 2011).

Four multiple‐transfusion trials reported serious adverse events (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; van Rhenen 2003), and combined results showed no evidence of a difference in the frequency of serious adverse events between recipients of Intercept and standard platelets (4 trials, 975 participants; RR 1.07, 95% CI 0.84 to 1.36; P = 0.57; I² = 0%) (Analysis 1.15).

Mirasol PCT platelet trials

Data from one multiple‐transfusion trial did not show any evidence of a difference in the incidence of serious adverse events between recipients of Mirasol and standard platelets (1 trial, 110 participants; RR 1.14, 95% CI 0.56 to 2.32; P = 0.72) (Analysis 1.15).

There was no evidence of subgroup differences between studies of multiple Intercept and Mirasol platelets transfusions (test for subgroup differences: Chi² = 0.03, df = 1, P = 0.87, I² = 0%).

Transfusion‐related acute lung injury

For the purposes of this review, we assumed that transfusion‐related acute lung injury (TRALI) (generally defined as new‐onset acute lung injury within six hours of transfusion with associated hypoxia and bilateral pulmonary infiltrates) was a serious adverse event related to transfusion, therefore if a trial reported that no serious adverse events related to transfusion had occurred, we assumed that no TRALI had occurred.

Eleven trials reported this outcome (Agliastro 2006; Cazenave 2010; De Francisci 2004; Janetzko 2005; Johansson 2013; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; Simonsen 2006; Slichter 2006). van Rhenen 2003 did not report this outcome.

Ten trials reported no cases of TRALI (Agliastro 2006; Cazenave 2010; De Francisci 2004; Janetzko 2005; Johansson 2013; Kerkhoffs 2010; Lozano 2011; Rebulla 2016; Simonsen 2006; Slichter 2006). In McCullough 2004 there were no cases of TRALI with associated cognate antigen‐antibody mismatch.

Transfusion‐transmitted infections

Six trials reported this outcome (Cazenave 2010; Janetzko 2005; Lozano 2011; McCullough 2004; Simonsen 2006; van Rhenen 2003). No episodes of bacterial TTI occurred in any of these six trials.

Anaphylaxis

Seven trials reported anaphylaxis (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Slichter 2006; van Rhenen 2003).

Combining data from six trials showed no evidence of a difference between treatment arms (6 trials, 1296 participants; RR 1.03, 95% CI 0.26 to 4.06; P = 0.97; I² = 0%) (Analysis 1.16) (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; van Rhenen 2003). Slichter 2006, a cross‐over trial, reported no anaphylaxis cases in either group.

Infections

Six trials reported the incidence of post‐transfusion infection (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; van Rhenen 2003). Pathogen‐reduced platelet transfusions increased the risk of developing an infection (6 trials, 1296 participants; RR 1.27, 95% CI 1.09 to 1.48; P = 0.002; I² = 39%) (Analysis 1.17). In the review protocol we stated that we would investigate any statistical heterogeneity over 30%. Removing Cazenave 2010 from the analysis (the only Mirasol PCT trial in the analysis) eliminated the heterogeneity (5 trials, 1186 participants; RR 1.35, 95% CI 1.14 to 1.60; P < 0.001; I² = 0%) (Analysis 1.17).

Intercept PCT platelet trials

One single Intercept PCT platelet transfusion trial showed no evidence of a difference in the number of infections between treatment arms (1 trial, 211 participants; RR 1.18, 95% CI 0.57 to 2.43; P = 0.66) (Analysis 1.17).

Four multiple Intercept PCT platelet transfusion trials showed an increase in the number of infections in the Intercept PCT platelet transfusion arm (4 trials, 975 participants; RR 1.36, 95% CI 1.14 to 1.62; P < 0.001; I² = 28%) (Analysis 1.17) (Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; van Rhenen 2003).

Mirasol PCT platelet trials

One multiple‐transfusion trial reported this outcome (Cazenave 2010), and there was no evidence of a difference between the two treatment arms (1 trial, 110 participants; RR 0.90, 95% CI 0.63 to 1.28; P = 0.56) (Analysis 1.17).

There was evidence of a subgroup difference between studies of multiple Intercept and Mirasol platelets transfusions (test for subgroup differences: Chi² = 4.21, df = 1, P = 0.04, I² = 76.2%).

Venous thromboembolic events

Six trials reported venous thromboembolic events (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; van Rhenen 2003), and they showed no evidence of a difference between treatment arms (6 trials, 1296 participants; RR 1.23, 95% CI 0.67 to 2.26; P = 0.50; I² = 0%) (Analysis 1.18).

Intercept PCT platelet trials

There was no evidence of a difference when we combined data from from the five Intercept platelet transfusion trials (5 trials, 1186 participants; RR 1.33, 95% CI 0.71 to 2.50; P = 0.38; I² = 17%) (Analysis 1.18). There was no change to the combined effect when we performed a sensitivity analysis excluding Lozano 2011 (single‐transfusion trial with no venous thromboembolic events among participants).

Mirasol PCT platelet trials

One multiple‐transfusion trial reported this outcome, finding no evidence of a difference between the two treatment arms (1 trial, 110 participants; RR 0.48, 95% CI 0.05 to 5.16; P = 0.38) (Analysis 1.18) (Cazenave 2010).

There was no evidence of subgroup differences between studies of multiple Intercept and Mirasol platelets transfusions (test for subgroup differences: Chi² = 0.66, df = 1, P = 0.42, I² = 0%).

Arterial thromboembolic events

Seven trials reported arterial thromboembolic events (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; van Rhenen 2003); when we combined the data in a meta‐analysis, we found no difference between treatment arms (7 trials, 1706 participants; RR 0.44, 95% CI 0.14 to 1.42; P = 0.17; I² = 0%) (Analysis 1.19).

Intercept PCT platelet trials

Six Intercept trials contributed to this outcome (Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; van Rhenen 2003), and combining data across the trials resulted in no evidence of a difference between the two treatment arms (6 trials, 1402 participants; RR 0.47, 95% CI 0.13 to 1.64; P = 0.23; I² = 0%) (Analysis 1.19). There was no change to the combined effect when we performed a sensitivity analysis excluding Lozano 2011 (single‐transfusion trial with no arterial thromboembolic events among participants).

Mirasol PCT platelet trials

Two multiple Mirasol transfusion trials reported this outcome (Cazenave 2010; Rebulla 2016), and there was no evidence of a difference between the two treatment arms (2 trials, 304 participants; RR 0.32, 95% CI 0.01 to 7.73; P = 0.48; I² = 0%) (Analysis 1.19).

There was no evidence of a subgroup difference between studies of multiple intercept and Mirasol platelets transfusions (test for subgroup differences: Chi² = 0.04, df = 1, P = 0.83, I² = 0%).

Laboratory assessment of response to platelet transfusion

All trials reported this outcome as a count increment and/or CCI at one hour or 24 hours, or both, post‐transfusion. There was variability between trials in a number of areas, including the mean study platelet transfusion dose (see Table 1), the timings of the platelet dose measurement, and the timings of the pre‐transfusion platelet counts (see Table 4). Furthermore, three of the seven multiple platelet transfusion trials restricted their primary analysis to include only the first five or eight study transfusions (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010), while the other four trials based their analyses on all transfusions (Agliastro 2006; McCullough 2004; Rebulla 2016; van Rhenen 2003).

Table 4. Laboratory data

Study	Timing of pre‐Tx plt count from Tx	Timing of 1‐hr post‐Tx plt count (mins)	Timing of 24‐hr post‐Tx plt count (hrs)	Timing of plt dose measurement	CCI calculation
Single‐transfusion studies ‐ Intercept®
De Francisci 2004	NR	NR	NR	NR	NR
Lozano 2011	Same day	10 to 240	16 to 24	Day 5 of storage	NR
Simonsen 2006	≤ 6 hrs OR same day	10 to 90	NR	Prior to storage	CI x BSA/plt dose
Slichter 2006	NR	60 to 120	18 to 24	NR	CI x BSA/plt dose
Multiple‐transfusion studies ‐ Intercept®
Agliastro 2006	NR	NR	NR	NR	NR
Janetzko 2005	≤ 6 hrs	NR	NR	At issue	CI x BSA/plt dose
Kerkhoffs 2010	≤ 6 hrs	10 to 120	16 to 28	Prior to storage	CI x BSA/plt dose
McCullough 2004	Same day	10 to 240	10 to 24	NR	CI x BSA/plt dose
Rebulla 2016	Same day	10 to 60	16 to 24	At issue	CI x BSA/plt dose
van Rhenen 2003	NR	10 to 240	18 to 24	At issue	CI x BSA/plt dose
Single‐transfusion studies ‐ Mirasol®
Johansson 2013	< 60 min	30 to 60	8 to 26	NR	NR
Multiple‐transfusion studies ‐ Mirasol®
Cazenave 2010	≤ 12 hrs	30 to 90	18 to 26	At issue	CI x BSA/plt dose
Rebulla 2016	Same day	10 to 60	16 to 24	At issue	CI x BSA/plt dos

BSA: body surface area
CCI: corrected count increment
CI: count increment
NR: not reported
plt: platelet
Tx: transfusion

Two of the single study platelet transfusion exposure trials, Lozano 2011 and Simonsen 2006, and four of the multiple study platelet exposure trials, Cazenave 2010, Janetzko 2005, Kerkhoffs 2010, and van Rhenen 2003, used the one‐hour CCI or count increment, or both, as their primary outcome.

Although all trials reported the one‐hour and/or 24‐hour count increments and/or CCIs, we could not include data from four trials in meta‐analyses. Two trials published as abstracts did not report standard deviations (Agliastro 2006; De Francisci 2004), and two cross‐over trials did not report data from the first trial period only (Johansson 2013; Slichter 2006).

One‐hour count increment

When we combined data from the seven trials with data that could be incorporated into a meta‐analysis, there was considerable statistical heterogeneity (I² = 85%) (Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; Simonsen 2006; van Rhenen 2003). We therefore did not combine data for single‐ and multiple‐transfusion trials (test for subgroup differences: Chi² = 23.96, df = 2, P < 0.001, I² = 91.7%).

Single‐transfusion trials

We combined data from two single Intercept platelet transfusion trials (Lozano 2011; and the first period of Simonsen 2006). There was no evidence of a difference between treatment arms for the one‐hour count increment (2 trials, 219 participants; mean difference (MD) ‐1.39 x 10⁹/L, 95% CI ‐4.81 to 2.02; P = 0.42; I² = 0%) (Analysis 1.20).

Multiple‐transfusion trials

When we combined data from the five multiple‐transfusion trials that reported this outcome (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003), there was still considerable statistical heterogeneity (I² = 79%). We noted a significant difference with regard to the platelet doses received by the pathogen‐reduced group and the standard group across three trials (Kerkhoffs 2010; McCullough 2004; van Rhenen 2003), in which lower mean platelet doses were issued to the pathogen‐reduced transfusion group versus the standard platelet transfusion group (P < 0.001) (Table 4). We subsequently decided to pool results from trials with and without a significant difference in platelet doses transfused by the two groups separately.

No difference in platelet dose between treatment arms

We combined data from two multiple‐transfusion trials (Janetzko 2005; Rebulla 2016). Participants who received pathogen‐reduced platelet transfusions had a lower one‐hour count increment (2 trials, 467 participants; MD ‐4.96, 95% CI ‐7.64 to ‐2.28; P < 0.001; I² = 0%) (Analysis 1.20).

Similar results were seen when data for Intercept (2 trials, 271 participants; MD ‐4.63, 95% CI ‐7.42 to ‐1.83; P = 0.001; I² = 0%) and Mirasol (1 trial, 196 participants; MD ‐8.90, 95% CI ‐18.47 to 0.67; P = 0.07) were analysed separately (Analysis 1.20).

There was no evidence of a difference between the two subgroups: Intercept multiple platelet transfusion and Mirasol multiple platelet transfusion (test for subgroup differences: Chi² = 0.71, df = 1, P = 0.40, I² = 0%).

Difference in platelet dose between treatment arms

Combining data from three Intercept PCT multiple‐transfusion trials in which a significant difference in transfused platelet doses was found showed that participants who received Intercept platelet transfusions had a lower one‐hour count increment (3 trials, 932 participants; MD ‐12.72, 95% CI ‐14.66 to ‐10.77; P < 0.001; I² = 20%) (Analysis 1.20) (Kerkhoffs 2010; McCullough 2004; van Rhenen 2003).

One‐hour corrected count increment

When we combined data from eight trials into a meta‐analysis there was considerable statistical heterogeneity (I² = 75%) (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; Simonsen 2006; van Rhenen 2003). We therefore did not combine data for single‐ and multiple‐transfusion trials (test for subgroup differences: Chi² = 15.01, df = 2, P < 0.001; I² = 86.7%).

Single‐transfusion trials

We combined data from two Intercept PCT single‐transfusion trials (Lozano 2011; Simonsen 2006). There was no evidence of a difference in the one‐hour CCI between Intercept platelets and standard platelets (2 trials, 219 participants; MD ‐0.89, 95% CI ‐2.35 to 0.58; P = 0.23; I² = 23%) (Analysis 1.21).

Multiple‐transfusion trials

We combined data from six multiple‐transfusion trials (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). Participants who received pathogen‐reduced platelet transfusions had a lower one‐hour CCI (6 trials, 1495 participants; MD ‐4.11, 95% CI ‐4.83 to ‐3.39; P < 0.001; I² = 61%) (Analysis 1.21).

We defined an I² value higher than 30% as moderate statistical heterogeneity. Applying the random‐effects model did not alter the results (6 trials, 1495 participants; MD ‐3.81, 95% CI ‐5.15 to ‐2.46; P < 0.001; I² = 61%) (analysis not shown). Four trials measured the platelet dose at issue (Cazenave 2010; Janetzko 2005; Rebulla 2016; van Rhenen 2003), and two studies measured the platelet dose either prior to storage, in Kerkhoffs 2010, or did not report when it was measured (McCullough 2004). When we included only the four trials that measured the platelet dose at issue in the analysis there was no evidence of statistical heterogeneity (4 trials, 666 participants; MD ‐2.60, 95% CI ‐3.74 to ‐1.46; P < 0.001; I² = 0%) (analysis not shown).

Intercept PCT platelet trials

We combined data from five Intercept multiple platelet transfusion trials (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). Participants who received Intercept platelet transfusions had a lower one‐hour CCI (5 trials, 1191 participants; MD ‐4.11, 95% CI ‐4.87 to ‐3.35; P < 0.001; I² = 74%) (Analysis 1.21).

When we only included studies that measured platelet dose at issue there was no longer any statistical heterogeneity (3 studies, 362 participants; MD ‐1.99, 95% CI ‐3.34 to ‐0.64; P = 0.004; I² = 0%) (analysis not shown) (Janetzko 2005; Rebulla 2016; van Rhenen 2003).

Mirasol PCT platelet trials

Two Mirasol trials reported this outcome (Cazenave 2010; Rebulla 2016). Participants who received Mirasol platelet transfusions had a lower one‐hour CCI (2 trials, 304 participants; MD ‐4.14, 95% CI ‐6.29 to ‐1.99; P < 0.001; I² = 0%) (Analysis 1.21).

There was no evidence of a difference between studies of multiple Intercept platelet transfusions and Mirasol platelet transfusions (test for subgroup differences: Chi² = 0.00, df = 1, P = 0.98, I² = 0%).

24‐hour count increment

Six studies reported on this outcome: one single platelet transfusion trial, Lozano 2011, and five multiple platelet transfusion trials (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003).

Combining data across six trials showed that participants who received pathogen‐reduced platelet transfusions had a lower 24‐hour count increment (6 trials, 1571 participants; MD ‐7.12, 95% CI ‐8.32 to ‐5.93; P < 0.001; I² = 55%) (Analysis 1.22) (Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; van Rhenen 2003).

Due to the moderate heterogeneity we have also reported single‐ and multiple‐transfusion trials separately.

Single platelet transfusion trials

In Lozano 2011 participants who received Intercept platelet transfusions had a lower 24‐hour count increment (1 trial, 186 participants; MD ‐4.10 x 10⁹/L, 95% CI ‐7.16 to ‐1.04; P = 0.009) (Analysis 1.22).

Multiple platelet transfusion trials

The five multiple‐transfusion studies that reported on this outcome also showed that participants receiving pathogen‐reduced platelet transfusions had a lower 24‐hour count increment (5 trials, 1385 participants; MD ‐7.67, 95% CI ‐8.96 to ‐6.37; P < 0.001; I² = 44%) (Analysis 1.22) (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). Due to the moderate statistical heterogeneity, we also applied the random‐effects model, which did not alter the results (5 trials, 1385 participants; MD ‐7.59, 95% CI ‐9.52 to ‐5.67; P < 0.001; I² = 44%) (analysis not shown). We noticed a significant difference with regard to the platelet doses received by the pathogen‐reduced group and the standard group across three trials (Kerkhoffs 2010; McCullough 2004; van Rhenen 2003), in which lower mean platelet doses were issued to the pathogen‐reduced transfusion group versus the standard platelet transfusion group (P < 0.001) (Table 4). Combining results from the two multiple‐transfusion trials with no significant difference in platelet doses between the two study arms did not alter the results, but did remove any statistical heterogeneity (2 trials, 453 participants; MD ‐5.66, 95% CI ‐7.77 to ‐3.56; P < 0.001; I² = 0%) (analysis not shown) (Janetzko 2005; Rebulla 2016).

Intercept PCT platelet trials

We combined data from five Intercept multiple platelet transfusion trials (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). Participants who received Intercept platelet transfusions had a lower 24‐hour count increment (5 trials, 1191 participants; MD ‐8.39, 95% CI ‐9.82 to ‐6.96; P < 0.001; I² = 0%) (Analysis 1.22).

Mirasol PCT platelet trials

One Mirasol multiple platelet transfusion trial reported this outcome (Rebulla 2016). Participants who received Mirasol platelet transfusions had a lower 24‐hour count increment (1 trial, 194 participants; MD ‐4.30, 95% CI ‐7.38 to ‐1.22; P = 0.006) (Analysis 1.22).

There was evidence of a difference between studies of multiple Intercept platelet transfusions and Mirasol platelet transfusions (test for subgroup differences: Chi² = 5.59, df = 1, P = 0.02, I² = 82.1%).

24‐hour corrected count increment

We combined data from seven trials in a meta‐analysis (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; van Rhenen 2003). Participants who received pathogen‐reduced platelet transfusions had a lower 24‐hour CCI (7 trials, 1681 participants; MD ‐3.02, 95% CI ‐3.57 to ‐2.48; P < 0.001; I² = 15%) (Analysis 1.23). There was evidence of subgroup differences (test for subgroup differences: Chi² = 5.52, df = 2, P = 0.06, I² = 63.8%).

Single platelet transfusion trials

In Lozano 2011 participants who received Intercept platelet transfusions had a lower 24‐hour CCI (1 trial, 186 participants; MD ‐1.96 x 10³/L, 95% CI ‐3.24 to ‐0.68; P = 0.003) (Analysis 1.23).

Multiple platelet transfusion trials

Six multiple platelet transfusion trials reported this outcome (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). Participants who received pathogen‐reduced platelet transfusions had a lower 24‐hour CCI (6 trials, 1495 participants; MD ‐3.26, 95% CI ‐3.86 to ‐2.66; P < 0.001; I² = 0%).

Intercept PCT platelet trials

A meta‐analysis of five Intercept PCT platelets trials showed that participants who received Intercept PCT platelets had a lower 24‐hour CCI (5 trials, 1191 participants; MD ‐3.50, 95% CI ‐4.18 to ‐2.82; P < 0.001; I² = 0%) (Analysis 1.23) (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003).

Mirasol PCT platelet trials

A meta‐analysis of two Mirasol PCT platelets trials showed that participants who received Intercept PCT platelets had a lower 24‐hour CCI (2 trials, 304 participants; MD ‐2.37, 95% CI ‐3.68 to ‐1.06; P < 0.001; I² = 0%) (Analysis 1.23) (Cazenave 2010; Rebulla 2016).

There was no evidence of a difference between studies of multiple Intercept platelet transfusions and Mirasol platelet transfusions (test for subgroup differences: Chi² = 2.26, df = 1, P = 0.13, I² = 55.8%).

Other measures of laboratory assessment of response to platelet transfusion

In addition to count increments, Slichter 2006 reported bleeding time assessments at one to two and 18 to 24 hours' post‐study transfusion as the trial's primary outcome, with paired results available for only 10 of the 32 participants who received both study products after a change in platelet production protocol. There was no evidence of a difference between the bleeding times when Intercept PCT and standard platelet transfusion responses were compared (P = 0.25 for one‐ to two‐hour and P = 0.82 for 18‐ to 24‐hour bleeding time assessments).

Platelet transfusion requirement

Eight studies reported on this outcome, two single study platelet transfusion trials, Lozano 2011 and Simonsen 2006, and six multiple platelet transfusion trials (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003).

Similar prophylactic transfusion thresholds were applied across the trials (platelet count < 10 to 20 x 10⁹/L), although therapeutic platelet transfusion protocols were frequently not reported and follow‐up periods varied (see Table 2).

Number of additional platelet transfusions (single‐transfusion studies)

Two single platelet transfusion studies reported this outcome (Lozano 2011; Simonsen 2006). In Lozano 2011 the number of additional platelet transfusions within the 24 hours post‐study transfusion was similar in both arms (15.2% versus 11.3%, P = 0.398). In Simonsen 2006 6/13 participants in the PCT arm and 4/13 in the standard arm required an additional platelet transfusion between the two study transfusions (unpublished data) (no analysis presented).

Number of platelet transfusions per participant (multiple‐transfusion studies)

Six trials reported this outcome as the number of platelet transfusions received per participant over various follow‐up periods (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003), and three trials also reported the number of platelet transfusions received per participant per day of platelet support (Cazenave 2010; Janetzko 2005; McCullough 2004).

Participants who received pathogen‐reduced platelet transfusions required more platelet transfusions (6 trials, 1509 participants; MD 1.23, 95% CI 0.86 to 1.61; P < 0.001; I² = 27%) (Analysis 1.24). Due to the greater comparability of the data, we combined data from the three trials that reported the number of platelet transfusions received per day of platelet support; this also showed that participants who received pathogen‐reduced platelet transfusions required more platelet transfusions (3 trials, 798 participants; MD 0.07, 95% CI 0.03 to 0.11; P < 0.001; I² = 21%) (analysis not shown).

Intercept PCT platelet trials

Participants who received Intercept platelet transfusions required more platelet transfusions (5 trials, 1203 participants; MD 1.30, 95% CI 0.84 to 1.77; P < 0.001; I² = 49%) (Analysis 1.24). This was also seen for those studies that reported the number of platelet transfusions per day of platelet support (2 trials, 688 participants; MD 0.09, 95% CI 0.04 to 0.13; P < 0.001; I² = 22%) (analysis not shown).

Mirasol PCT platelet trials

Participants who received Mirasol platelet transfusions required more platelet transfusions (2 trials, 306 participants; MD 1.09, 95% CI 0.44 to 1.75; P = 0.001; I² = 0%) (Analysis 1.24). There was insufficient evidence to assess the effect from the only study that reported the number of platelet transfusions per day of platelet support (1 trial, 110 participants; MD 0.04, 95% CI ‐0.03 to 0.11; P = 0.23; I² = 0%) (analysis not shown).

There was no evidence of a subgroup difference between studies of multiple Intercept PCT and Mirasol PCT platelet transfusions (test for subgroup differences: Chi² = 0.27, df = 1, P = 0.60, I² = 0%)

Off‐protocol transfusions

For a variety of reasons, including a shortage of pathogen‐reduced platelet transfusions, five of the six multiple study platelet transfusion trials reported a fairly high use of off‐protocol platelets during their trial periods (see Characteristics of included studies and Table 2). The use of off‐protocol platelet transfusions was uneven between arms, with approximately double the number of off‐protocol transfusions in the PCT platelets arm of four of the five studies (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; van Rhenen 2003). Cazenave 2010 reported a slightly higher use of off‐protocol platelets in the standard platelets arm (23.2%) when compared with the Mirasol PCT platelets arm (17.7%). In Rebulla 2016, off‐protocol transfusion occurred in 10.3% of the Intercept PCT recipients and 6.4% of the Mirasol PCT recipients. Off‐protocol transfusions were not an issue for the single platelet transfusion trials. However, in Simonsen 2006 there were three protocol violations in which two participants did not receive PCT platelets, and one participant did not receive standard platelets. No off‐protocol violation was reported in Johansson 2013.

Platelet transfusion interval

Nine trials measured the interval between platelet transfusions as days to the next platelet transfusion (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; Simonsen 2006; Slichter 2006; van Rhenen 2003). We could incorporate data from eight trials into meta‐analyses (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; van Rhenen 2003; and the first period of Simonsen 2006). Due to moderate statistical heterogeneity (I² = 68%) and considerable differences between subgroups (I² = 91.2%), we did not combine the data, but analysed the subgroups single‐ and multiple‐transfusions separately.

Single platelet transfusion trials

We combined data from two single Intercept platelet transfusion trials (Lozano 2011; and the first period of Simonsen 2006); we found no evidence of a difference in the platelet transfusion interval between treatment arms (2 trials, 194 participants; MD 0.10 days, 95% CI ‐0.13 to 0.34; P = 0.39; I² = 0%) (Analysis 1.25).

Of the data that could not be included in the meta‐analysis, Slichter 2006 reported the mean time to the next platelet transfusion following both sequences of study platelet transfusions as 2.9 ± 1.2 days (N = 26) for the Intercept PCT arm versus 3.4 ± 1.3 days (N = 26) for the standard platelets arm, with no difference between treatment arms (P = 0.18). In the second period of Simonsen 2006, the mean time to the next platelet transfusion following the second single study platelet transfusion was reported as 0.86 ± 0.23 days (N = 11) for the PCT platelets arm versus 1.41 ± 0.22 days (N = 9) for the standard platelets arm, favouring the standard platelets arm (P < 0.001). Nevertheless, despite a minimum washout period of 24 hours, the trial authors reported a significant period‐by‐treatment interaction between study transfusions for the second period of the trial (Simonsen 2006).

Multiple platelet transfusion trials

We combined data from six multiple platelet transfusion trials (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). Participants who received pathogen‐reduced platelet transfusions had a shorter platelet transfusion interval (6 trials, 1489 participants; MD ‐0.42 days, 95% CI ‐0.53 to ‐0.32; P < 0.001; I² = 29%) (Analysis 1.25).

We combined data from the five Intercept multiple platelet transfusion trials (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; Rebulla 2016; van Rhenen 2003). Participants who received Intercept PCT platelet transfusions had a shorter platelet transfusion interval (5 trials, 1197 participants; MD ‐0.50, 95% CI ‐0.61 to ‐0.38; P < 0.001; I² = 0%) (Analysis 1.25).

When we combined data from two Mirasol multiple platelet transfusion trials (Cazenave 2010; Rebulla 2016), there was no evidence of a difference in the platelet transfusion interval between the two treatment arms (2 trials, 306 participants; MD ‐0.16, 95% CI ‐0.38 to 0.07; P = 0.17; I² = 0%) (Analysis 1.25).

There was evidence of a difference between studies of multiple Intercept PCT and Mirasol PCT platelet transfusions (test for subgroup differences: Chi² = 6.78, df = 1, P = 0.009, I² = 85.3%).

Red cell transfusion requirement and interval

Red cell transfusion requirement

Seven trials reported this outcome as the number of red cell transfusions received per participant over various follow‐up periods (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; van Rhenen 2003); five trials also reported the number of red cells received per participant per day of platelet support (Cazenave 2010; Janetzko 2005; Lozano 2011; McCullough 2004; van Rhenen 2003).

There was no evidence of a difference between the pathogen‐reduced platelets arm and the standard platelets arm in the number of red cell transfusions participants received (7 trials, 1720 participants; MD 0.13, 95% CI ‐0.07 to 0.33; P = 0.07; I² = 24%) (Analysis 1.26). Due to the greater comparability of the data, we combined data from the five trials that reported the number of red cell transfusions received per day of platelet support, which also showed that there is probably no difference in the number of red cell transfusions participants received (5 trials, 1112 participants; MD 0.00, 95% CI ‐0.02 to 0.03; P = 0.72; I² = 0%) (analysis not shown).

Intercept PCT platelet trials

Six Intercept trials contributed to this outcome (Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; Rebulla 2016; van Rhenen 2003); combining data across the trials resulted in no evidence of a difference between the two treatment arms (6 trials, 1414 participants; MD 0.07, 95% CI ‐0.15 to 0.28; P = 0.54; I² = 22%) (Analysis 1.26). There was no change to the combined effect when we performed a sensitivity analysis excluding the only single‐transfusion trial, Lozano 2011 (5 trials, 1203 participants; MD 0.38, 95% CI ‐0.08 to 0.85; P = 0.11; I² = 4%) (Analysis 1.26).

Mirasol PCT platelet trials

Two multiple Mirasol transfusion trials reported this outcome (Cazenave 2010; Rebulla 2016), and there was no evidence of a difference between the two treatment arms (2 trials, 306 participants; MD 0.47, 95% CI ‐0.05 to 0.99; P = 0.07; I² = 0%) (Analysis 1.26).

There was no evidence of a difference between studies of multiple Intercept PCT and Mirasol PCT platelet transfusions (test for subgroup differences: Chi² = 0.07, df = 1, P = 0.80, I² = 0%).

Red cell transfusion interval

Kerkhoffs 2010 was the only trial reporting the interval between red cell transfusions, again with no evidence of a difference between intervention arms (P = 0.15).

Sensitivity analyses

Only one of our intended sensitivity analyses was required as, having received additional data from several trial authors, there were no trials in which 25% of participants or more were lost to follow‐up. We identified Cazenave 2010, Kerkhoffs 2010, and Janetzko 2005 as at a high risk of performance or detection bias, or both (Risk of bias in included studies; Figure 2), thus we performed sensitivity analyses. However, upon exclusion of these trials there was no change in the overall effect estimates for any of the outcomes for which they reported data, and thus no impact was detected from these potential sources of bias.

Discussion

The main aims of this review were to comprehensively assess the effectiveness and safety of pathogen‐reduced platelets compared with standard platelets for the prevention of bleeding. We identified 12 completed RCTs involving 2045 participants assessing two commercially available pathogen‐reduction systems (1981 received at least one platelet transfusion): 10 studies compared Intercept PCT platelets with standard platelets (1662 participants), and three studies compared Mirasol PCT platelets with standard platelets (319 participants). All studies were performed in high‐income countries. We also identified three ongoing trials comparing three different types of pathogen‐reduced platelets, Intercept PCT, Mirasol PCT, and Theraflex UV, to standard platelets.

The main rationale for introducing pathogen reduction of platelets is to further reduce the risk of transfusion‐transmitted infection (TTI) from both known and future pathogens. In high‐income countries that perform bacterial screening, the current risks of TTI are exceedingly small, as bacterial screening has proved highly effective in removing platelets contaminated with pathogenic bacteria from the blood supply. In the UK since the introduction of bacterial screening in 2011, there has been only one case of proven bacterial TTI (Bolton‐Maggs 2016).

It is highly unlikely that the sample sizes of any RCTs of pathogen‐reduced platelets would inform these outcomes. However, RCTs of pathogen‐reduced platelets would be expected to address other safety issues and to assess efficacy, including the prevention and treatment of bleeding, which are the major clinical reasons for using platelet transfusions.

Summary of main results

We are very uncertain as to whether pathogen‐reduced platelets increase the risk of any bleeding (WHO grade 1 to 4). There was no evidence of a difference between pathogen‐reduced platelets and standard platelets in the incidence of clinically significant bleeding complications (WHO grade 2 to 4), and there is probably no difference in the incidence of severe or life‐threatening bleeding (WHO grade 3 to 4).
There is probably no difference between pathogen‐reduced platelets and standard platelets in the incidence of all‐cause mortality (4 to 12 weeks). However, there was evidence of a subgroup difference between multiple‐transfusion trials of Intercept and Mirasol (Intercept 5 trials, 1203 participants; RR 0.62, 95% CI 0.37 to 1.05; Mirasol 2 trials, 306 participants; RR 2.92, 95% CI 0.81 to 10.59; test for subgroup differences: Chi² = 4.78, df = 1, P = 0.03, I² = 79.1%).
There is probably no difference between pathogen‐reduced platelets and standard platelets in the incidence of serious adverse events, acute transfusion reactions, or adverse events.
Participants who received pathogen‐reduced platelet transfusions required more platelet transfusions and probably had a shorter time interval between platelet transfusions.
Participants who received pathogen‐reduced platelet transfusions had a lower 24‐hour corrected count increment (CCI).
Participants who received pathogen‐reduced platelet transfusions had an increased risk of developing platelet refractoriness, including an increased risk due to alloimmunisation.
No bacterial TTIs occurred in the six trials that reported this outcome.

Overall completeness and applicability of evidence

This review provides the most up‐to‐date assessment of the effectiveness and safety of pathogen‐reduced platelet transfusions versus standard platelet transfusions.

No studies directly compared different types of pathogen‐reduced platelet transfusions. There are three ongoing studies comparing pathogen‐reduced platelets versus standard platelets; none of these trials will directly compare different pathogen‐reduction technologies (EUCTR2015‐001035‐20‐DE is assessing Theraflex UV platelets (166 participants); Kerkhoffs 2013 is assessing Mirasol PCT platelets (375 participants); and NCT01789762 is assessing Intercept PCT platelets (810 participants)).

There are a number of limitations that may affect the strength of any conclusions in this review. They are as follows.

The majority of the evidence applies to Intercept PCT platelets alone, with only three trials comparing Mirasol PCT platelets with standard platelets (319 participants).
Although platelet transfusions are used in many different clinical settings, the predominantly investigated group in this review were people with haemato‐oncological malignancies. De Francisci 2004 was the only trial including participants with other diagnoses (adults requiring a liver transplant and children requiring cardiac surgery). Although individuals with haemato‐oncological malignancies represent a large proportion of people receiving platelet transfusions, we cannot extrapolate these finding to individuals with other diagnoses (Charlton 2014; Fedele 2016; Fillet 2016; Jones 2013).
Only three trials included children (De Francisci 2004; McCullough 2004; van Rhenen 2003).
It was not possible to extract data on outcomes for all trials, even after contact with the trial authors.
There was considerable variability in the reporting of the methodology of recording and grading bleeding between trials; this included the use of different bleeding scales or bleeding scales that had been modified in different ways (Table 3). Details, such as training of outcome assessors, were often poorly described, which may limit the comparison of bleeding outcomes between trials. Bleeding was assessed in relation to three different bleeding scales in the identified trials: WHO (four‐point scale), CTCAE version 3 (five‐point scale), and a three‐point scale of no bleeding, minor, or major bleeding used in two trials (Slichter 2006; van Rhenen 2003). There remains uncertainty in the literature regarding the clinical importance of detecting, recording, and reporting minor severity of bleeding in trials (such as WHO grade 1), although all types of bleeding may be of concern to patients. Furthermore, the data from the SPRINT trial as reported by McCullough 2004 used an expanded version of the WHO scale, but with the same bleeding events additionally reported and assessed by Snyder 2004 using the CTCAE scale (version 2).
Many participants in these trials received off‐protocol transfusions, with uneven use noted between study arms in four trials (Table 2) (Janetzko 2005; Kerkhoffs 2010; McCullough 2004; van Rhenen 2003). Off‐protocol transfusions could additionally be prescribed during the minimum 24‐hour washout period of two of the cross‐over trials, and hence more than one platelet product type would have been received during the short follow‐up period of these studies (Simonsen 2006; Slichter 2006).
Significantly lower platelet doses were received per transfusion in the pathogen‐reduced platelet transfusion arm of four of the 11 trials that reported this measure (Table 1). As platelet losses are frequently seen during the processing of pathogen‐reduced platelets, one study permitted centres to either increase their apheresis collection dose by 10% or to add an additional buffy coat during the production of the Intercept PCT platelet product (Lozano 2011); despite this adjustment, laboratory responses were still significantly lower following PCT platelet transfusion for both the 24‐hour count increment and CCI in this trial.
Five of the 12 included trials were single‐transfusion trials. The power of these single‐dose trials to identify clinically significant outcomes and safety issues is likely to be considerably reduced compared to trials that evaluated multiple platelet transfusions, in which the follow‐up periods were generally longer.
The comparator standard platelet component in all trials varied by platelet collection method, storage duration, platelet dose, and storage solution (Table 1). Finally, changes in platelet function related to the pathogen‐reduction technologies have been less clearly investigated and were not studied in the trials included in this review.

Quality of the evidence

All included studies were RCTs, of which three were cross‐over trials. Through direct author contact for nine of the 12 included studies we obtained sufficient information to allow us to evaluate the quality of evidence. Due to the limitations of information reported in the two trials published only as abstracts and our inability to contact these trial authors, much of their study methodology remains unclear (Agliastro 2006; De Francisci 2004). The impact on the pooled results of the potentially high risk of performance and or detection bias present in three trials was shown to be insignificant when tested by a series of sensitivity analyses (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010). The usefulness of this method in meta‐analyses involving small numbers of trials is, however, limited. Participants were aware of the intervention status in three trials (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010), and in Kerkhoffs 2010 the outcome assessors were not blinded.

We assessed the quality of the evidence using the GRADE approach, and this ranged from low to high quality (see summary of findings Table for the main comparison).

We assessed the GRADE quality of evidence as high for the following.

Number of participants experiencing platelet refractoriness.
24‐hour CCI.

We assessed the GRADE quality of evidence as moderate for the following.

Number of participants with a clinically significant bleeding event (WHO grade ≥ 2 or equivalent) ‐ follow‐up more than seven days.
Number of participants with a severe bleeding event (WHO grade ≥ 3 or equivalent) ‐ follow‐up more than seven days. We downgraded the evidence because there was inconsistency between the McCullough 2004 trial and the other trials. In McCullough 2004, the number of severe bleeding events in the standard platelet transfusion arm was higher than in the pathogen‐reduced platelet transfusion arm, whereas all of the other studies showed a lower incidence in the standard platelet arm. Also, the incidence of severe bleeding was much higher in the McCullough 2004 study than in the other trials. Although the studies within this subgroup had similar follow‐up periods, and to our knowledge there were no other major differences between them, the difference in the event rate may nevertheless be due to undetected methodological differences between the studies.
All‐cause mortality post‐transfusion. We downgraded the evidence because there was inconsistency between the trials; part of this may have been due to different follow‐up durations between the trials.
Number of participants with a serious adverse event. We downgraded the evidence because there was inconsistency between the trials in the number of serious adverse events; part of this may have been due to different follow‐up durations between the trials.

We assessed the GRADE quality of evidence as low for the following.

Number of participants with any bleeding event(s) (WHO grade 1 to 4 or equivalent) ‐ follow‐up more than seven days. We downgraded the evidence because there was inconsistency between the trials, and results differed depending on whether a fixed‐effect or random‐effects analysis was performed.

Assessing the methodological quality across the included studies was a complex task, as the included studies differed in their design, primary outcomes, transfusion frequency, timing of the assessment, and scales used to document bleeding signs and symptoms. Similar methodological challenges were reported in another review (Cook 2013) that included six of the trials included in this review (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; Lozano 2011; McCullough 2004; van Rhenen 2003). Cook 2013 provided recommendations on how to design future clinical trials of pathogen‐reduced platelets, including: a parallel RCT design should be used for trials with clinical outcomes (e.g. bleeding), whereas a randomised block design should be used for trials with transfusion‐based outcomes (e.g. CCI); "bleeding status should be assessed on a daily basis using a valid and reliable bleeding assessment instrument"; "bleeding data should be analyzed based on the time to the first bleed as well as using the daily bleeding data"; and transfusion‐based outcome data (such as CCIs) should be interpreted with caution "since they are susceptible to selection effects" (Cook 2013). Cook 2013 also highlighted a recently devised universal bleeding assessment tool for use in future platelet transfusion trials (Webert 2012). Meanwhile, the Biomedical Excellence for Safer Transfusion (BEST) Collaborative Study Group has developed "consensus bleeding definitions, a standardized approach to record and grade bleeding, and guidance notes to educate and train bleeding assessors" for use in future platelet transfusion trials.

Potential biases in the review process

There were no obvious biases within the review process. We conducted a wide search that included ongoing trial databases and contact with researchers in the field; we carefully assessed the relevance of each paper identified; and we used no restrictions on the language in which the paper was originally published or its publication status. We performed all screening and data extractions in duplicate. We prespecified all outcomes and subgroups prior to analysis. The numbers of included studies were insufficient for us to combine to complete a funnel plot in order to examine the risk of publication bias.

Although two different methods of platelet pathogen‐reduction were included and their results pooled in this review, subgroups of trials evaluating the different pathogen‐reduction technologies with single or multiple platelet transfusion exposures were described and presented separately. This enabled assessment of differences between study subgroups where identified in addition to assessing the pooled effect.

Agreements and disagreements with other studies or reviews

Comparison with non‐randomised trials

Both pathogen‐reduction technologies have been introduced into routine practice in a number of countries worldwide. A number of surveillance and observational studies have reported on haemovigilance of pathogen‐reduced platelets; there was no increase in the adverse event reporting rate for the Mirasol PCT platelet product, Łętowska 2016, or Intercept PCT platelet products (Cazenave 2011; Osselaer 2008).

An international survey of current practice and concerns or opinions regarding pathogen‐reduced platelets was performed by Reesink 2010. Of the nine responding countries, pathogen‐reduced platelets were already employed by four (Intercept PCT platelets: France, Spain, Kuwait; Mirasol PCT platelets: Norway). No TTIs were identified in these centres, in which approximately 190,000 pathogen‐reduced platelet products had been transfused, with observational reports of fewer febrile, allergic reactions after transfusion of pathogen‐reduced versus standard platelets, assumed to be due to the reduced plasma content of the pathogen‐reduced platelets. Limitations of these studies include non‐standardised reporting by non‐research staff. The findings of these countries support our findings that adverse events are infrequent and not increased by the use of pathogen‐reduced platelets when compared with standard platelets.

Cazenave 2011 showed an increase in the number of platelet components per participant when pathogen‐reduced platelet components were used (mean 5.2 platelet components per participant (platelets in plasma) to mean 6.4 platelet components per participant (Intercept PCT platelets)). This is consistent with the findings of this review, but is inconsistent with another observational study that found no difference in platelet component demand (Osselaer 2009).

Cazenave 2011, conducted in France (control 2050 participants, pathogen‐reduced 2069 participants), was a much bigger study than Osselaer 2009, conducted in Belgium (control 699 participants, pathogen‐reduced 795 participants), but there were no other obvious differences between the two studies.

Comparison with other meta‐analyses

According to the Transfusion Evidence Library, there have been no systematic reviews published since the previous version of this review (Butler 2013). The reviews identified in the previous version of this review, Cid 2012, Vamvakas 2011, and Vamvakas 2012, only included five, Cid 2012 and Vamvakas 2011, or six, Vamvakas 2012, of this review's 12 included trials.

Vamvakas 2011 identified five trials (Cazenave 2010; Janetzko 2005; Kerkhoffs 2010; McCullough 2004; van Rhenen 2003), and, similar to our review, reported a significant reduction of the 1‐ and 24‐hour CCI (1‐hour CCI: MD 3.26, 95% CI 2.45 to 4.79; 24‐hour CCI: MD 3.31, 95% CI 2.03 to 4.60); an increased platelet transfusion requirement (MD 0.93, 95% CI 0.16 to 1.7); and reduced platelet transfusion interval for PCT platelets (MD 0.41, 95% CI 0.13 to 0.67). Unlike our review, Vamvakas 2011 reported an increase in "all bleeding" (odds ratio (OR) 1.58, 95% CI 1.1 to 2.25; estimated to cause a 58% increase) and "clinically significant" bleeding (OR 1.54, 95% CI 1.11 to 2.13; estimated to cause a 54% increase) with the use of pathogen‐reduced platelets. Vamvakas 2011 used the CTCAE‐graded haemorrhagic adverse event data from the largest included study (SPRINT) reported by Snyder 2004 to assess bleeding, whereas our review used the original WHO grading of bleeding rather than data from adverse event reporting. Methodological issues were discussed in the updated review (Vamvakas 2012), which applied different eligibility criteria and included a different additional trial, Lozano 2011 (although again, quantitative results were based on five trials).

Cid 2012 compared Intercept PCT to standard platelet transfusions and included five trials in their meta‐analysis. The authors concluded that the transfusion of Intercept PCT platelets was associated with lower CCIs and a shorter transfusion interval when compared with the transfusion of standard platelets, but there was no difference in the odds ratio of bleeding.

Comparison with other trials

As mentioned above, there are concerns that pathogen‐reduction technologies may affect platelet dose. The question of the effectiveness of different doses of prophylactic platelet transfusions has been addressed by Slichter 2010, which randomly assigned 1272 haematopoietic stem cell transplantation or chemotherapy recipients to receive either low‐, medium‐, or high‐dose platelet transfusions (1.1 x 10¹¹, 2.2 x 10¹¹, or 4.4 x 10¹¹ platelets per square metre of body surface area). Fewer platelets were transfused per participant in the group randomised to low‐dose platelets when compared with medium‐ (P = 0.002) or high‐dose (P < 0.001) platelets, but an increased number of transfusions was required (P < 0.001). The number of platelets within the transfused product did not translate into a difference in the incidence of bleeding. Overall, the platelet dose in the 10 trials included in our review ranged from a mean of 2.8 x 10¹¹ to 7.6 x 10¹¹ per transfusion (see Table 1), which is generally in line with the medium‐ to high‐dose categories identified by the PLADO study (Slichter 2010). Although we did not perform a post hoc analysis to assess our review's outcomes by platelet transfusion dose, participants of three trials (which recruited approximately two‐thirds of the participants included in this review) received a lower platelet dose in the PCT platelets arm (Kerkhoffs 2010; McCullough 2004; van Rhenen 2003). Despite these differences being statistically significant overall, there was less than a 10% difference in dose between the study arms in these three trials (range 8% in McCullough 2004 to 15% in Kerkhoffs 2010). The PLADO study also raised uncertainties about the relationship between the one‐hour count increment and risk of bleeding, reaffirming the need for trials evaluating clinical haemostatic outcomes (Triulzi 2012).

Figure 1

Updated review study flow diagram.

Figure 2

'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

Figure 3

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Analysis 1.1

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 1 Number of participants with 'any bleeding' event(s) (WHO grade 1 to 4 or equivalent) ‐ follow‐up less than 48 hrs.

Analysis 1.2

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 2 Number of participants with 'any bleeding' event(s) (WHO grade 1 to 4 or equivalent) ‐ follow‐up more than 7 days.

Analysis 1.3

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 3 Number of participants with 'clinically significant' bleeding (WHO grade ≥ 2 or equivalent ‐ follow‐up less than 48 hours.

Analysis 1.4

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 4 Number of participants with 'clinically significant' bleeding (WHO grade ≥ 2 or equivalent) ‐ follow‐up more than 7 days.

Analysis 1.5

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 5 Number of participants with 'severe' bleeding (WHO grade ≥ 3 or equivalent) ‐ follow‐up less than 48 hours.

Analysis 1.6

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 6 Number of participants with 'severe' bleeding (WHO grade ≥ 3 or equivalent) ‐ follow‐up more than 7 days.

Analysis 1.7

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 7 All‐cause mortality (follow‐up 0 to < 4 weeks).

Analysis 1.8

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 8 All‐cause mortality (follow‐up 4 to 12 weeks).

Analysis 1.9

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 9 Mortality due to bleeding (follow‐up 4 to 12 weeks).

Analysis 1.10

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 10 Mortality due to infection (follow‐up 4 to 12 weeks).

Analysis 1.11

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 11 Number of participants with an acute transfusion reaction.

$Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 12 Number of participants experiencing platelet refractoriness.$

Analysis 1.12

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 12 Number of participants experiencing platelet refractoriness.

$Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 13 Number of participants experiencing platelet refractoriness and platelet alloimmunisation.$

Analysis 1.13

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 13 Number of participants experiencing platelet refractoriness and platelet alloimmunisation.

Analysis 1.14

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 14 Number of participants with an adverse event.

Analysis 1.15

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 15 Number of participants with a serious adverse event.

Analysis 1.16

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 16 Number of participants with anaphylaxis.

Analysis 1.17

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 17 Number of participants with infection.

Analysis 1.18

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 18 Number of participants with venous thromboembolism (follow‐up 0 to 12 weeks).

Analysis 1.19

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 19 Number of participants with arterial venous thromboembolism (follow‐up 0 to 12 weeks).

Analysis 1.20

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 20 Lab response ‐ 1‐hour count increment [x 10⁹/L].

Analysis 1.21

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 21 Lab response ‐ 1‐hour corrected count increment [x 10³/L].

Analysis 1.22

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 22 Lab response ‐ 24‐hour count increment [x 10⁹/L].

Analysis 1.23

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 23 Lab response ‐ 24‐hour corrected count increment [x 10³/L].

Analysis 1.24

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 24 Number of platelet transfusions per participant.

Analysis 1.25

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 25 Platelet transfusion interval (days to next platelet transfusion).

Analysis 1.26

Comparison 1 Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding, Outcome 26 Number of red cell transfusions per participant.

Summary of findings for the main comparison. Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding

Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding in thrombocytopenia
Patient or population: thrombocytopenia Settings: hospital Intervention: pathogen‐reduced platelets
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No. of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Control	Pathogen‐reduced platelets
Number of participants with 'any bleeding' event(s) (WHO grade 1 to 4 or equivalent) ‐ follow‐up > 7 days Follow‐up: 8 to 70 days	Study population		RR 1.09 (1.02 to 1.15)	1085 (5 studies)	⊕⊕⊝⊝ low^1,2	Random‐effects analysis: RR 1.14, 95% CI 0.93 to 1.38; I² = 59%
	696 per 1000	758 per 1000 (710 to 800)
	Moderate
	800 per 1000	872 per 1000 (816 to 920)
Number of participants with 'clinically significant' bleeding event(s) (WHO grade ≥ 2 or equivalent) ‐ follow‐up > 7 days Follow‐up: 8 to 70 days	Low		RR 1.10 (0.97 to 1.25)	1392 (5 studies)	⊕⊕⊕⊝ moderate²	Assumed risks from published data³
	570 per 1000	627 per 1000 (553 to 712)
	Moderate
	700 per 1000	770 per 1000 (679 to 875)
	High
	790 per 1000	869 per 1000 (766 to 988)
Number of participants with 'severe' bleeding event(s) (WHO grade ≥ 3 or equivalent) ‐ follow‐up > 7 days Follow‐up: 8 to 70 days	Study population		RR 1.24 (0.76 to 2.02)	1495 (6 studies)	⊕⊕⊕⊝ moderate⁴	Medium‐risk data taken from PLADO trial (Slichter 2010).
	36 per 1000	44 per 1000 (27 to 72)
	Moderate
	100 per 1000	124 per 1000 (76 to 202)
All‐cause mortality Follow‐up: 4 to 12 weeks	Study population		RR 0.81 (0.50 to 1.29)	1509 (6 studies)	⊕⊕⊕⊝ moderate⁵
	54 per 1000	43 per 1000 (27 to 69)
	Moderate
	25 per 1000	20 per 1000 (12 to 32)
Number of participants with a serious adverse event Follow‐up: 15 to 84 days	Study population		RR 1.09 (0.88 to 1.35)	1340 (7 studies)	⊕⊕⊕⊝ moderate⁵
	179 per 1000	196 per 1000 (158 to 242)
	Moderate
	204 per 1000	222 per 1000 (180 to 275)
Number of participants experiencing platelet refractoriness Follow‐up: 0 to 24 hours	Study population		RR 2.94 (2.08 to 4.16)	1525 (7 studies)	⊕⊕⊕⊕ high	4 studies defined refractoriness as 2 successive 1‐hour CCIs below 5 x 10³ (Cazenave 2010; Janetzko 2005; McCullough 2004; van Rhenen 2003), while Kerkhoffs 2010 defined refractoriness as 2 successive 1‐hour CCIs below 7.5 x 10³ or 24‐hour CCIs below 4.5 x 10³ and presence of antibodies against platelets.
	51 per 1000	149 per 1000 (106 to 212)
Number of platelet transfusions per participant Multiple platelet transfusion trials	The mean number of platelet transfusions per participant was 4.7 to 8.4 platelet transfusions.	MD 1.23 platelet transfusions higher (0.86 to 1.61 higher)	‐	1509 (6 studies)	⊕⊕⊕⊕ high
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CCI: corrected count increment; CI: confidence interval; MD: mean difference; RR: risk ratio; WHO: World Health Organization.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹Downgraded one point due to inconsistency. ²Downgraded one point due to risk of other bias (differences in the assessment and grading of bleeding) ³Low risk = data taken from autologous transplantation participants from PLADO (PLAtelet DOse) trial (Slichter 2010). Medium risk = overall bleeding rate for all participants in the PLADO trial (Slichter 2010). High risk = data taken from allogeneic transplantation participants in PLADO trial (Slichter 2010). ⁴Downgraded one point due to inconsistency. One large study (McCullough 2004) had proportionately more events in the control group than the other five studies. Although the studies within this subgroup had similar follow‐up periods and to our knowledge there are no other major differences between them, the difference in the event rate may nevertheless be due to undetected methodological differences between the studies. ⁵Downgraded one point due to imprecision. Wide confidence intervals that include the risk of significant harm or benefit.

Summary of findings for the main comparison. Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding

Table 1. Trial characteristics

Study

(publication)

Study design

(analysis)

Total no. of participants

(PCT plts/std plts)

Underlying disease

Type of treatment received for underlying

disease (%)

Type of platelet component

Follow‐up periods for mortality

(days)

Mean

plt storage duration

(days)

Platelet dose^a

(Mean plt transfusion dose) (10¹¹/unit)

PCT plts

Std plts

Single‐transfusion studies ‐ Intercept

De Francisci 2004

(A)

(NR)

(22/22)

Adult liver transplant/

paediatric cardiac surgery

Liver: 14

Cardiac: 8

Liver: 14

Cardiac: 8

Lozano 2011

(F)

(Non‐I)

211

(105/106)

Adult haemato‐

oncological disease

Auto: 42

Allo: 29

Ch: 27

Other: 3

Auto:40

Allo: 25

Ch: 34

Other: 2

86% BC

14% Aph

6.8

Intermediate

(4.2)

Simonsen 2006

(F)

C‐O

(Non‐I)

PCT‐std: 9

std‐PCT: 11

Adult haemato‐

oncological disease

PCT‐std

Std‐PCT

Low/intermediate

(2.9)

Allo: 33

Ch: 67

Allo: 64

Ch: 36

Slichter 2006

(F)

C‐O

(E)

PCT‐std: NR

std‐PCT: NR

Adult haemato‐

oncological disease

Aph

42^b

High

(7.5)

Multiple‐transfusion studies – Intercept

Agliastro 2006

(A)

(NR)

(19/11)

Paediatric haemato‐

oncological disease

PCT plts: BC

Std plts: Aph

Low/intermediate

(2.9)

Janetzko 2005

(F)

(E)

(22/21)

Adult haemato‐

oncological disease

SCT: 68

Ch: 27

Other: 5

SCT: 62

Ch: 38

Aph

3.1

Intermediate

(4.0)

Kerkhoffs 2010

(F)

(Non‐I)

184

(85/99)

Adult haemato‐

oncological disease

Auto: 39

Allo: 7

Ch: 49

Other: 5

Auto:32

Allo: 12

Ch: 53

Other: 3

Intermediate

(3.7^c)

McCullough 2004

(F)

(Non‐I)

645

(318/327)

Paediatric + adult haemato‐oncological

disease

Auto: 48

Allo: 28

Ch: 21

Other: 3

Auto: 52

Allo: 28

Ch: 18

Other: 2

Aph

3.5^d

Intermediate

(3.9^c)

(F)

(Non‐I)

228

(113/115)

Adult haemato‐

oncological disease

Ch: 101

Allo: 12

Ch: 102

Allo: 13

1% Aph

99% BC

1.29 PR

1.48 C

Low/intermediate

(2.9)

van Rhenen 2003

(F)

(E)

103

(52/51)

Paediatric + adult haemato‐oncological

disease

SCT: 37

Ch: 63

SCT: 37

Ch: 63

3.5

Intermediate

(4.1^c)

Single‐transfusion studies ‐ Mirasol

Johansson 2013

(F)

C‐O

(NR)

PCT‐std: 8

std‐PCT: 7

Adult haemato‐

oncological disease

2.8 PR

2.3 C

Low/intermediate

(3.0^c)

Multiple‐transfusion studies – Mirasol

Cazenave 2010

(F)

(Non‐I)

110

(56/54)

Adult haemato‐

oncological disease

25% BC

75% Aph

2.7

Intermediate

(5.2)

(F)

(Non‐I)

196

(99/97)

Adult haemato‐

oncological disease

Ch: 85

Allo: 14

Ch: 83

Allo: 14

49% Aph

51% BC

1.66 PR

1.73 C

Low/intermediate

(3.3)

Table 1. Trial characteristics

Table 2. Platelet transfusion data

Study	Platelet transfusion protocol	No. of plt transfusions/participant (mean ± SD)		Total % (no.) of off‐protocol transfusions
Study	Platelet transfusion protocol	PCT plts	Std plts	PCT plts	Std plts
Single‐transfusion studies ‐ Intercept
De Francisci 2004	P (NR)	1	1	NR	NR
Lozano 2011	P (10 to 20)	1	1	NA	NA
Simonsen 2006	P (10 to 20)	1	1	(2/25)	(1/25)
Slichter 2006	P (20)	1	1	NA	NA
Multiple‐transfusion studies ‐ Intercept
Agliastro 2006	T	NR	NR	NR	NR
Janetzko 2005	P (20); PP (NR) or T	4.7 ± 3.3	5.5 ± 4.7	16.5% (17/103)	7% (8/115)
Kerkhoffs 2010	P (10 to 40); PP (40 to 100) or T	5 ± 2^a	4 ± 2^a	34% (134/391)	18% (65/357)
McCullough 2004	P (NR) or T	8.4 ± 8.6^a	6.2 ± 7.0^a	8.5% (232/2715)	4.8% (101/2092)
Rebulla 2016	P (10 to 20) or T	5.9 ± 5.8^a	3.8 ± 3.4^a	2.8% (19/667)	2.0% (9/441)
van Rhenen 2003	P (20) or T	7.5 ± 5.8	5.6 ± 5.5	20.3% (79/390)	10.5% (30/286)
Single‐transfusion studies ‐ Mirasol
Johansson 2013	P (10 to 50)	1	1	NA	NA
Multiple‐transfusion studies ‐ Mirasol
Cazenave 2010	P (10 to 20); PP (50) or T	5.4 ± 3.4^b 1.2 ± 2.3^c	4.4 ± 3.4^b 1.3 ± 3.6^c	17.7% (65/368)	23.2% (72/310)
Rebulla 2016	P (10 to 20) or T	4.6 ± 4.0^a	3.4 ± 2.1^a	6.3% (29/457)	0% (0/334)
Key: NA: not applicable NR: not reported P: prophylactic transfusion (threshold plt count x 10⁹/L) PCT: photochemically treated plt: platelet PP: pre‐procedure transfusion (threshold plt count x 10⁹/L) SD: standard deviation Std: standard T: therapeutic transfusion ^aStatistically significant difference in the number of platelet transfusions received per participant. ^bOn‐protocol transfusions only. ^cOff‐protocol transfusions only.

Table 2. Platelet transfusion data

Table 3. Bleeding assessment

Study	Bleeding scale used	Duration of bleeding assessment	Timing of bleeding assessment(s)	Maximum number of days of on‐study plt transfusion support	Bleeding results reported in meta‐analysis
Single‐transfusion studies ‐ Intercept
De Francisci 2004	NR	S	NR	1	NR
Lozano 2011	WHO¹	S	6 hrs pre‐ and 6 hrs post‐Tx	1	Post‐Tx bleeding score
Simonsen 2006	WHO¹ (m)	S	≤ 12 hrs post‐Tx	1	Post‐Tx bleeding score
Slichter 2006	WHO¹ (m)	S	6 hrs pre‐ and 6 hrs post‐Tx	1	Post‐Tx bleeding score
Multiple‐transfusion studies ‐ Intercept
Agliastro 2006	NR	L	NR	NR	NR
Janetzko 2005	WHO¹	L	6 hrs pre‐ and post‐ each plt Tx + daily	35	Max bleeding score/participant
Kerkhoffs 2010	CTCAE²	L	Post‐1st Tx + daily	42	Max bleeding score/participant
McCullough 2004	WHO¹ (e)	L	12 hrs post‐Tx + daily	28	Max bleeding score/participant
Rebulla 2016	WHO¹	L	Post‐1st Tx + daily	28	Max bleeding score/participant
van Rhenen 2003	WHO¹ (m)	L	6 hrs pre‐ and 6 hrs post‐Tx	56	Max bleeding score/participant
Single‐transfusion studies ‐ Mirasol
Johansson 2013	WHO¹	S	12 hours pre‐ and post‐Tx	1	NR
Multiple‐transfusion studies ‐ Mirasol
Cazenave 2010	WHO¹	L	1 hr pre‐ and post‐Tx + 24 hrs post‐Tx + final follow‐up visit^a	28	Max bleeding score/participant
Rebulla 2016	WHO¹	L	Post‐1st Tx + daily	28	Max bleeding score/participant
Key: (e): scale expanded (more specifically defined WHO grades, including sites of bleeding) L: long‐term bleeding assessment (> 7 days) post‐transfusion (m): scale modified (only 3 scores: none = 0, minor = 1 (equivalent to WHO grades 1 and 2), major = 2 (equivalent to WHO grades 3 and 4)) NA: not applicable NR: not reported plt: platelet S: short‐term bleeding assessment (up to 48 hours) post‐transfusion M: medium‐term bleeding (2 to 7 days) post‐transfusion Tx: transfusion ^aOn‐protocol platelet transfusions only. References: 1. WHO scale – WHO. WHO Handbook for Reporting Results of Cancer Treatment. Geneva: World Health Organization; 1979. 2. U.S. Dept of Health and Human Services (National Institutes of Health and National Cancer Institute). Common Terminology Criteria for Adverse Events (CTCAE) Version 3.0, September 2006.

Table 3. Bleeding assessment

Table 4. Laboratory data

Study	Timing of pre‐Tx plt count from Tx	Timing of 1‐hr post‐Tx plt count (mins)	Timing of 24‐hr post‐Tx plt count (hrs)	Timing of plt dose measurement	CCI calculation
Single‐transfusion studies ‐ Intercept®
De Francisci 2004	NR	NR	NR	NR	NR
Lozano 2011	Same day	10 to 240	16 to 24	Day 5 of storage	NR
Simonsen 2006	≤ 6 hrs OR same day	10 to 90	NR	Prior to storage	CI x BSA/plt dose
Slichter 2006	NR	60 to 120	18 to 24	NR	CI x BSA/plt dose
Multiple‐transfusion studies ‐ Intercept®
Agliastro 2006	NR	NR	NR	NR	NR
Janetzko 2005	≤ 6 hrs	NR	NR	At issue	CI x BSA/plt dose
Kerkhoffs 2010	≤ 6 hrs	10 to 120	16 to 28	Prior to storage	CI x BSA/plt dose
McCullough 2004	Same day	10 to 240	10 to 24	NR	CI x BSA/plt dose
Rebulla 2016	Same day	10 to 60	16 to 24	At issue	CI x BSA/plt dose
van Rhenen 2003	NR	10 to 240	18 to 24	At issue	CI x BSA/plt dose
Single‐transfusion studies ‐ Mirasol®
Johansson 2013	< 60 min	30 to 60	8 to 26	NR	NR
Multiple‐transfusion studies ‐ Mirasol®
Cazenave 2010	≤ 12 hrs	30 to 90	18 to 26	At issue	CI x BSA/plt dose
Rebulla 2016	Same day	10 to 60	16 to 24	At issue	CI x BSA/plt dos
BSA: body surface area CCI: corrected count increment CI: count increment NR: not reported plt: platelet Tx: transfusion

Table 4. Laboratory data

Comparison 1. Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of participants with 'any bleeding' event(s) (WHO grade 1 to 4 or equivalent) ‐ follow‐up less than 48 hrs Show forest plot	3		Risk Ratio (Fixed, 95% CI)	Subtotals only

1.1 Intercept plts vs standard plts ‐ single platelet transfusion studies	3	309	Risk Ratio (Fixed, 95% CI)	0.86 [0.63, 1.19]
2 Number of participants with 'any bleeding' event(s) (WHO grade 1 to 4 or equivalent) ‐ follow‐up more than 7 days Show forest plot	5	1085	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [1.02, 1.15]

2.1 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	4	975	Risk Ratio (M‐H, Fixed, 95% CI)	1.07 [1.01, 1.13]
2.2 Mirasol plts vs standard plts ‐ multiple platelet transfusion studies	1	110	Risk Ratio (M‐H, Fixed, 95% CI)	1.38 [0.95, 2.02]
3 Number of participants with 'clinically significant' bleeding (WHO grade ≥ 2 or equivalent ‐ follow‐up less than 48 hours Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

3.1 Intercept plts vs standard plts ‐ single platelet transfusion studies	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4 Number of participants with 'clinically significant' bleeding (WHO grade ≥ 2 or equivalent) ‐ follow‐up more than 7 days Show forest plot	5	1392	Risk Ratio (M‐H, Fixed, 95% CI)	1.10 [0.97, 1.25]

4.1 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	4	1088	Risk Ratio (M‐H, Fixed, 95% CI)	1.06 [0.94, 1.21]
4.2 Mirasol plts vs standard plts ‐ multiple platelet transfusion studies	2	304	Risk Ratio (M‐H, Fixed, 95% CI)	1.54 [0.86, 2.76]
5 Number of participants with 'severe' bleeding (WHO grade ≥ 3 or equivalent) ‐ follow‐up less than 48 hours Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

5.1 Intercept plts vs standard plts ‐ single platelet transfusion studies	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6 Number of participants with 'severe' bleeding (WHO grade ≥ 3 or equivalent) ‐ follow‐up more than 7 days Show forest plot	6	1495	Risk Ratio (M‐H, Fixed, 95% CI)	1.24 [0.76, 2.02]

6.1 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	5	1191	Risk Ratio (M‐H, Fixed, 95% CI)	1.21 [0.70, 2.09]
6.2 Mirasol plts vs standard plts ‐ multiple platelet transfusion studies	2	304	Risk Ratio (M‐H, Fixed, 95% CI)	1.36 [0.45, 4.16]
7 All‐cause mortality (follow‐up 0 to < 4 weeks) Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

7.1 All‐cause mortality	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
7.2 Mortality due to bleeding	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
7.3 Mortality due to infection	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
7.4 Mortality due to thromboembolism	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
7.5 Mortality due to transfusion reactions	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
8 All‐cause mortality (follow‐up 4 to 12 weeks) Show forest plot	6	1509	Risk Ratio (M‐H, Fixed, 95% CI)	0.81 [0.50, 1.29]

8.1 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	5	1203	Risk Ratio (M‐H, Fixed, 95% CI)	0.62 [0.37, 1.05]
8.2 Mirasol plts vs standard plts ‐ multiple platelet transfusion studies	2	306	Risk Ratio (M‐H, Fixed, 95% CI)	2.92 [0.81, 10.59]
9 Mortality due to bleeding (follow‐up 4 to 12 weeks) Show forest plot	6	1509	Risk Ratio (M‐H, Fixed, 95% CI)	1.23 [0.38, 3.96]

9.1 Intercept plts vs standard plts‐multiple platelet transfusion studies	5	1203	Risk Ratio (M‐H, Fixed, 95% CI)	1.04 [0.29, 3.75]
9.2 Mirasol plts vs standard plts‐multiple platelet transfusion studies	2	306	Risk Ratio (M‐H, Fixed, 95% CI)	2.89 [0.12, 69.55]
10 Mortality due to infection (follow‐up 4 to 12 weeks) Show forest plot	6	1509	Risk Ratio (M‐H, Fixed, 95% CI)	0.94 [0.42, 2.09]

10.1 Intercept plts vs standard plts‐multiple platelet transfusion studies	5	1203	Risk Ratio (M‐H, Fixed, 95% CI)	0.84 [0.35, 2.03]
10.2 Mirasol plts vs standard plts‐multiple platelet transfusions studies	2	306	Risk Ratio (M‐H, Fixed, 95% CI)	1.62 [0.22, 12.07]
11 Number of participants with an acute transfusion reaction Show forest plot	7	1636	Risk Ratio (Fixed, 95% CI)	0.96 [0.75, 1.24]

11.1 Intercept plts vs standard plts ‐ single platelet transfusion studies	2	251	Risk Ratio (Fixed, 95% CI)	1.57 [0.71, 3.48]
11.2 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	5	1191	Risk Ratio (Fixed, 95% CI)	0.91 [0.69, 1.18]
11.3 Mirasol plts vs standard plts‐multiple platelet	1	194	Risk Ratio (Fixed, 95% CI)	1.99 [0.18, 21.78]
12 Number of participants experiencing platelet refractoriness Show forest plot	7	1525	Risk Ratio (M‐H, Fixed, 95% CI)	2.94 [2.08, 4.16]

12.1 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	6	1221	Risk Ratio (M‐H, Fixed, 95% CI)	2.85 [1.96, 4.15]
12.2 Mirasol plts vs standard plts ‐ multiple platelet transfusion studies	2	304	Risk Ratio (M‐H, Fixed, 95% CI)	3.47 [1.44, 8.34]
13 Number of participants experiencing platelet refractoriness and platelet alloimmunisation Show forest plot	6	1415	Risk Ratio (M‐H, Fixed, 95% CI)	2.35 [1.46, 3.76]

13.1 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	6	1221	Risk Ratio (M‐H, Fixed, 95% CI)	1.90 [1.11, 3.26]
13.2 Mirasol plts vs standard plts‐multiple platelet transfusion studies	1	194	Risk Ratio (M‐H, Fixed, 95% CI)	4.5 [1.58, 12.81]
14 Number of participants with an adverse event Show forest plot	7	1566	Risk Ratio (M‐H, Fixed, 95% CI)	1.01 [0.97, 1.05]

14.1 Intercept plts vs standard plts ‐ single platelet transfusion studies	2	255	Risk Ratio (M‐H, Fixed, 95% CI)	0.98 [0.87, 1.09]
14.2 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	4	1007	Risk Ratio (M‐H, Fixed, 95% CI)	1.01 [0.97, 1.05]
14.3 Mirasol plts vs standard plts ‐ multiple platelet transfusion studies	2	304	Risk Ratio (M‐H, Fixed, 95% CI)	1.04 [0.89, 1.21]
15 Number of participants with a serious adverse event Show forest plot	7	1340	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.88, 1.35]

15.1 Intercept plts vs standard plts ‐ single platelet transfusion studies	2	255	Risk Ratio (M‐H, Fixed, 95% CI)	1.21 [0.55, 2.68]
15.2 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	4	975	Risk Ratio (M‐H, Fixed, 95% CI)	1.07 [0.84, 1.36]
15.3 Mirasol plts vs standard plts ‐ multiple platelet transfusion studies	1	110	Risk Ratio (M‐H, Fixed, 95% CI)	1.14 [0.56, 2.32]
16 Number of participants with anaphylaxis Show forest plot	6	1296	Risk Ratio (M‐H, Fixed, 95% CI)	1.03 [0.26, 4.06]

16.1 Intercept plts vs standard plts‐single platelet transfusion studies	1	211	Risk Ratio (M‐H, Fixed, 95% CI)	1.01 [0.06, 15.93]
16.2 Intercept plts vs standard plts‐ multiple platelet transfusion studies	4	975	Risk Ratio (M‐H, Fixed, 95% CI)	1.79 [0.23, 13.69]
16.3 Mirasol plts vs standard plts‐multiple platelet transfusion studies	1	110	Risk Ratio (M‐H, Fixed, 95% CI)	0.32 [0.01, 7.73]
17 Number of participants with infection Show forest plot	6	1296	Risk Ratio (M‐H, Fixed, 95% CI)	1.27 [1.09, 1.48]

17.1 Intercept plts vs standard plts‐ single platelet transfusion studies	1	211	Risk Ratio (M‐H, Fixed, 95% CI)	1.18 [0.57, 2.43]
17.2 Intercept plts vs standard plts‐multiple platelet transfusion studies	4	975	Risk Ratio (M‐H, Fixed, 95% CI)	1.36 [1.14, 1.62]
17.3 Mirasol plts vs standard plts‐multiple platelet transfusion studies	1	110	Risk Ratio (M‐H, Fixed, 95% CI)	0.9 [0.63, 1.28]
18 Number of participants with venous thromboembolism (follow‐up 0 to 12 weeks) Show forest plot	6	1296	Risk Ratio (M‐H, Fixed, 95% CI)	1.23 [0.67, 2.26]

18.1 Intercept plts vs standard plts‐single platelet transfusion studies	1	211	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
18.2 Intercept plts vs standard plts‐multiple platelet transfusion studies	4	975	Risk Ratio (M‐H, Fixed, 95% CI)	1.33 [0.71, 2.50]
18.3 Mirasol plts vs standard plts‐multiple platelet transfusion studies	1	110	Risk Ratio (M‐H, Fixed, 95% CI)	0.48 [0.05, 5.16]
19 Number of participants with arterial venous thromboembolism (follow‐up 0 to 12 weeks) Show forest plot	7	1706	Risk Ratio (M‐H, Fixed, 95% CI)	0.44 [0.14, 1.42]

19.1 Intercept plts vs standard plts‐single platelet transfusion studies	1	211	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
19.2 Intercept plts vs standard plts‐multiple platelet transfusion studies	5	1191	Risk Ratio (M‐H, Fixed, 95% CI)	0.47 [0.13, 1.64]
19.3 Mirasol plts vs standard plts‐multiple transfusion studies	2	304	Risk Ratio (M‐H, Fixed, 95% CI)	0.32 [0.01, 7.73]
20 Lab response ‐ 1‐hour count increment [x 10⁹/L] Show forest plot	7		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

20.1 Intercept plts vs standard plts ‐ single platelet transfusion studies	2	219	Mean Difference (IV, Fixed, 95% CI)	‐1.39 [‐4.81, 2.02]
20.2 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	5	1203	Mean Difference (IV, Fixed, 95% CI)	‐10.08 [‐11.67, ‐8.48]
20.3 Mirasol plts vs standard plts‐multiple platelet transfusion studies	1	196	Mean Difference (IV, Fixed, 95% CI)	‐8.90 [‐18.47, 0.67]
21 Lab response ‐ 1‐hour corrected count increment [x 10³/L] Show forest plot	8		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

21.1 Intercept plts vs standard plts ‐ single platelet transfusion studies	2	219	Mean Difference (IV, Fixed, 95% CI)	‐0.89 [‐2.35, 0.58]
21.2 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	5	1191	Mean Difference (IV, Fixed, 95% CI)	‐4.11 [‐4.87, ‐3.35]
21.3 Mirasol plts vs standard plts ‐ multiple platelet transfusion studies	2	304	Mean Difference (IV, Fixed, 95% CI)	‐4.14 [‐6.29, ‐1.99]
22 Lab response ‐ 24‐hour count increment [x 10⁹/L] Show forest plot	6	1571	Mean Difference (IV, Fixed, 95% CI)	‐7.12 [‐8.32, ‐5.93]

22.1 Intercept plts vs standard plts ‐ single platelet transfusion studies	1	186	Mean Difference (IV, Fixed, 95% CI)	‐4.1 [‐7.16, ‐1.04]
22.2 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	5	1191	Mean Difference (IV, Fixed, 95% CI)	‐8.39 [‐9.82, ‐6.96]
22.3 Mirasol plts vs standard plts‐multiple platelet transfusion studies	1	194	Mean Difference (IV, Fixed, 95% CI)	‐4.30 [‐7.38, ‐1.22]
23 Lab response ‐ 24‐hour corrected count increment [x 10³/L] Show forest plot	7	1681	Mean Difference (IV, Fixed, 95% CI)	‐3.02 [‐3.57, ‐2.48]

23.1 Intercept plts vs standard plts ‐ single platelet transfusion studies	1	186	Mean Difference (IV, Fixed, 95% CI)	‐1.96 [‐3.24, ‐0.68]
23.2 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	5	1191	Mean Difference (IV, Fixed, 95% CI)	‐3.50 [‐4.18, ‐2.82]
23.3 Mirasol plts vs standard plts ‐ multiple platelet transfusion studies	2	304	Mean Difference (IV, Fixed, 95% CI)	‐2.37 [‐3.68, ‐1.06]
24 Number of platelet transfusions per participant Show forest plot	6	1509	Mean Difference (IV, Fixed, 95% CI)	1.23 [0.86, 1.61]

24.1 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	5	1203	Mean Difference (IV, Fixed, 95% CI)	1.30 [0.84, 1.77]
24.2 Mirasol plts vs standard plts ‐ multiple platelet transfusion studies	2	306	Mean Difference (IV, Fixed, 95% CI)	1.09 [0.44, 1.75]
25 Platelet transfusion interval (days to next platelet transfusion) Show forest plot	8	1697	Mean Difference (IV, Fixed, 95% CI)	‐0.34 [‐0.43, ‐0.24]

25.1 Intercept plts vs standard plts ‐ single platelet transfusion studies	2	194	Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.13, 0.34]
25.2 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	5	1197	Mean Difference (IV, Fixed, 95% CI)	‐0.50 [‐0.61, ‐0.38]
25.3 Mirasol plts vs standard plts ‐ multiple platelet transfusion studies	2	306	Mean Difference (IV, Fixed, 95% CI)	‐0.16 [‐0.38, 0.07]
26 Number of red cell transfusions per participant Show forest plot	7	1720	Mean Difference (IV, Fixed, 95% CI)	0.13 [‐0.07, 0.33]

26.1 Intercept plts vs standard plts ‐ single platelet transfusion studies	1	211	Mean Difference (IV, Fixed, 95% CI)	‐0.02 [‐0.27, 0.23]
26.2 Intercept plts vs standard plts ‐ multiple platelet transfusion studies	5	1203	Mean Difference (IV, Fixed, 95% CI)	0.38 [‐0.08, 0.85]
26.3 Mirasol plts vs standard plts ‐ multiple platelet transfusion studies	2	306	Mean Difference (IV, Fixed, 95% CI)	0.47 [‐0.05, 0.99]

Comparison 1. Pathogen‐reduced platelets versus standard platelets for the prevention of bleeding