Suctioning of clear amniotic fluid at birth: A systematic review

Context Upper airway suctioning at birth was considered standard procedure and is still commonly practiced. Negative effects could exceed benefits of suction. Question In infants born through clear amniotic fluid (P) does suctioning of the mouth and nose (I) vs no suctioning (C) improve outcomes (O). Data sources Information specialist conducted literature search (12th September 2021, re-run 17th June 2022) using Medline, Embase, Cochrane Databases, Database of Abstracts of Reviews of Effects, and CINAHL. RCTs, non-RCTs and observational studies with a defined selection strategy were included. Unpublished studies, reviews, editorials, animal and manikin studies were excluded. Data extraction Two authors independently extracted data, risk of bias was assessed using the Cochrane ROB2 and ROBINS-I tools. Certainty of evidence was assed using the GRADE framework. Review Manager was used to analyse data and GRADEPro to develop summary of evidence tables. Meta-analyses were performed if ≥2 RCTs were available. Outcomes Primary: assisted ventilation. Secondary: advanced resuscitation, oxygen supplementation, adverse effects of suctioning, unanticipated NICU admission. Results Nine RCTs (n = 1096) and 2 observational studies (n = 418) were identified. Two RCTs (n = 280) with data concerns were excluded post-hoc. Meta-analysis of 3 RCTs, (n = 702) showed no difference in primary outcome. Two RCTs (n = 200) and 2 prospective observational studies (n = 418) found lower oxygen saturations in first 10 minutes of life with suctioning. Two RCTs (n = 200) showed suctioned newborns took longer to achieve target saturations. Limitations Certainty of evidence was low or very low for all outcomes. Most studies selected healthy newborns limiting generalisability and insufficient data was available for planned subgroup analyses. Conclusions Despite low certainty evidence, this review suggests no clinical benefit from suctioning clear amniotic fluid from infants following birth, with some evidence suggesting a resulting desaturation. These finding support current guideline recommendations that this practice is not used as a routine step in birth. Funding The International Liaison Committee on Resuscitation provided access to software platforms, an information specialist and teleconferencing. Clinical Trial Registration This systematic review was registered with the Prospective Register of Systematic Reviews (https://www.crd.york.ac.uk/prospero/) (identifier: CRD42021286258).


Introduction
At birth, all infants have fluid-filled lungs and upper airways. Lung fluid is absorbed within the lungs. Healthy infants may clear upper airway fluid by some combination of swallowing, inhalation and sometimes, sneezing. Despite this, longstanding practice was to routinely provide oro/nasopharyngeal suctioning at birth in many parts of the world. There have been increasing concerns that this practice may not confer benefit and may have undesirable consequences.
ILCOR prepared an evidence worksheet in 2010 and concluded that: "Routine intrapartum oropharyngeal and nasopharyngeal suctioning for newborn infants with clear or meconium-stained amniotic fluid is no longer recommended". 1 The World Health Organisation (WHO) reviewed 3 studies 3-5 in a 2017 systematic review 6 and recommended that: "In neonates born through clear amniotic fluid who start breathing on their own after birth, suctioning of the mouth and nose should not be performed.". The WHO guideline authors made a further consensusbased recommendation that: "In neonates born through clear amniotic fluid who do not start breathing after thorough drying and rubbing the back 2-3 times, suctioning of the mouth and nose should not be done routinely before initiating positive pressure ventilation. Suctioning should be done only if the mouth or nose is full of secretions.".
In addition to no benefit, both ILCOR and WHO found literature suggesting possible adverse effects of suctioning, including lower oxygen saturations over the first 10 minutes of life and lower likelihood of Apgar score of 10 at 10 minutes. Other reported associations include increased risk for bradycardia 4,7-8 apnea, 8 hypoxemia and arterial oxygen desaturation, 3,[9][10] hypercapnia, 11 impaired cerebral blood flow regulation, 12 increased intracranial pressure 13 and infection. 8 One study reported that suctioning was commonly applied despite opposing recommendations in resuscitation guidelines. 14 This question was prioritized by the ILCOR Neonatal Life Support Task Force because an ILCOR scoping review in 2019 found sufficient new studies to justify updating the systematic review, 2 and to assess the certainty of evidence using Grading of Recommendations, Assessment, Development and Evaluation (GRADE) methodology. 15 The aim of the review was to assess the role of routine suctioning of clear fluid in the upper airway, compared to no routine suctioning in newborn infants.

Protocol
This systematic review (SR) was completed as part of the ILCOR NLS Task Force continuous evidence review process based on knowledge gaps identified in the 2020 ILCOR NLS Consensus on the Science of Resuscitation with Treatment Recommendations. 2 The SR and meta-analysis were guided by the Cochrane Handbook for Systematic Reviews of Interventions 15 and reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement for meta-analysis of health care interventions. 16 The protocol was registered with the Prospective Register of Systematic Reviews (PROSPERO; CRD42021286258) on 22nd October 2021. The study was conducted in the a priori planned way included in the Prospero registration, except for updated literature search dates where database access was subtly different.

Inclusion and exclusion criteria
Randomized controlled trials (RCTs) and non-randomized studies (non-randomized controlled trials, interrupted time series, controlled before-and-after studies, cohort studies) are eligible for inclusion. Unpublished studies (e.g., conference abstracts, trial protocols), review articles, editorials, comments, case reports, animal studies, and manikin studies were excluded. All years were included without language restrictions if an English abstract was available.
For this review, observational studies were cohort studies eligible for inclusion if they used a defined strategy to ensure that the participants were either all of those who received an exposure of interest in a defined population (e.g., infants born at a hospital between specified dates), or they were sampled in such a way as to be representative of such a population. Otherwise, the study was an (ineligible) case series.
Population, Intervention, Comparator, Outcome, Study Design, Time Frame (PICOST) question Among neonates who are born through clear amniotic fluid in the delivery room (population) does initial suctioning of the mouth and nose (intervention) compared with no initial suctioning (comparison) change outcome?
The PICOST question was developed by the authors in collaboration with the ILCOR NLS Task Force and approved by the ILCOR Scientific Advisory Committee.
Outcome ratings using the GRADE certainty of evidence (COE) classifications 17 of critical or important outcomes were based on a consensus for international neonatal resuscitation guidelines (range 1-3 low importance, 4-6 important but not critical, 7-9 critical for decision-making).
The primary outcome was receipt of assisted ventilation (important). Secondary outcomes were advanced resuscitation (critical), receipt and duration of oxygen supplementation (important), adverse effects of intervention (important) and unanticipated admission to the Neonatal Intensive Care Unit (NICU) (important). Appendix A defines these outcomes.

Search strategy
Literature searches in Medline, Embase, the Cochrane Database of Systematic Reviews, the Cochrane Central Register of Controlled Trials, the Cochrane Methodology Register, the Database of Abstracts of Reviews of Effects, and Cumulative Index to Nursing and Allied Health Literature (CINAHL) were developed by an information specialist (MDW) iteratively, in consultation with the review team. The subject headings and keywords were adapted for the respective databases. The search was completed on 12th September 2021 and updated on 17th June 2022. For the updated literature search the EBM Review suite of databases was no longer available through the Information Specialist's institution. In order to recreate the original search, the Cochrane Library (online through Wiley) was searched for CDSR and CCRCT (Trials). Covidence Systematic Review software 18 was used for management of the search results.

Study selection
Authors independently screened titles and abstracts, studies required agreement from two authors to be excluded or included for full text review. Full text reviews were conducted independently by authors and two authors need to agree on inclusion. Disagreements were resolved by consensus of the full review team. The process was conducted using Covidence software (Veritas Health Innovation, Melbourne, Australia).

Data extraction, bias, and quality assessment
The study review group worked collaboratively to extract data from included studies. Study investigators were emailed if data queries arose. All data for pre-specified outcomes were included where studies reported on these outcomes. Studies were assessed for risk of bias (RoB) using the Cochrane ROB2 tool 19 for RCTs and the Risk of Bias in Non-Randomized Studies of Interventions (ROBINS-I) 20 for observational studies, using templates constructed in Covidence systematic review software. 21 RoB was defined at a study level, and where studies contributed data to an individual outcome, their RoB for that outcome was assessed. All RoB assessments were decided by consensus of the full review group.
Certainty of evidence (confidence in the estimate of effect) for each outcome was decided by consensus among the review group using the GRADE framework.
Review team members were excluded from assessing inclusion or RoB for any study in which they had participated as an investigator. The evidence profile tables were presented and discussed with the ILCOR NLS Task Force and content experts.

Data analysis
Review Manager 22 was used to analyse data and GRADEPro 23 to develop summary of evidence tables. Meta-analyses were performed if !2 RCTs were available. Observational studies were analysed and reported if fewer than 2 RCTs were available. For dichotomous outcomes, pooled unadjusted risk ratios (RRs) and corresponding 95% confidence intervals (CIs) were reported using the Mantel-Haenszel fixed effect method. The pooled risk difference (RD) and the absolute risk difference (ARD) were calculated. Pooled continuous variables were reported as mean differences (MDs) and corresponding 95% CIs using the Mantel-Haenszel fixed effect method.
Forest plots were created for graphical representation of RRs and MDs. Heterogeneity was measured using the I 2 statistic. Significant heterogeneity was considered present if the I 2 statistic was >50%. We explored statistical heterogeneity using post-hoc sensitivity analyses. Subgroup analyses were planned according to gestational age (term vs late preterm infants), mode of delivery and type of suctioning device (bulb vs catheter).
Communication of the findings of the review was based on GRADE guidelines with wording decided by the ILCOR NLS Task Force through consensus.

Literature search and study selection
The search strategies identified 2453 unique records, for which titles and abstracts were screened, and 2411 studies were excluded (Fig. 1). From them, 42 full-text articles were assessed for eligibility and 11 were included in the final review.

Study Characteristics
The SR included 9 RCTs 3-5,9,24-28 and 2 14,29 observational studies enrolling a total of 1514 newborn infants (1096 in RCTs, 418 in observational studies) ( Table 1). All the RCTs only recruited term newborn infants except for one 26 that recruited newborn infants >35 weeks. One observational study 29 recruited only term newborn infants whilst another 14 recruited term and preterm newborns.
For two of the RCTs 3,4 enrolling 280 participants, the task force had concerns about the reliability of the oxygen saturation and heart rate data. The reported standard deviations were unusually small in comparison to other published studies and the data in each study were remarkably similar. The author was contacted to provide clarification; however, at the time of publication the task force had not received a reply. Therefore, the results of these studies have been excluded from the meta-analysis. Exclusion of these studies did not change the conclusion of this systematic review but in the interests of transparency, analyses were repeated including these two studies and the results are shown in an online supplement.

Risk of bias
RoB was increased for all studies because blinding of those performing the intervention to group assignment was not considered feasible ( Table 2). Some concerns about selective reporting of outcomes were present for two studies.

Certainty of evidence
Evidence for the primary and all but one of the secondary outcomes was rated as low or very low certainty because of high RoB and indirectness ( Table 2). As the studies predominantly recruited healthy term newborn infants, they were downgraded for indirectness for all outcomes because they were not considered representative of all newborn infants, including those at high risk of need for assisted ventilation or other adverse outcomes.

Outcomes
Primary outcome -Assisted ventilation: Three RCTS 24,26,27 , including 702 participants found that for suctioning compared to no suctioning, clinical benefit or harm could not be excluded (RR 0.72; 95% CI 0.40, 1.31 p = 0.28; absolute risk difference (ARD) 18 fewer per 1000 95% CI, 39 fewer to 20 more per 1000). Two of these RCTs 24,27 recruited healthy infants and reported assisted ventilation was not required so the event rate was zero in both groups. Evidence was of very low certainty (downgraded for very serious risk of bias, seri-ous inconsistency, very serious indirectness and very serious imprecision).

Secondary outcomes
Advanced resuscitation and stabilization interventions (intubation, chest compressions/epinephrine (adrenaline) in DR) Very low certainty evidence from three RCTS 24,26,27 including 702 participants found that for suctioning vs no suctioning, clinical benefit or harm could not be excluded (RR 0.72; 95% CI, 0.40, 1.31p = 0.28; ARD 18 fewer per 1000 95% CI, 39 fewer to 20 more patients per 1000). Two of these RCTs 24,27 recruited healthy infants and reported advanced resuscitation was not required so the event rate was zero in both groups. Evidence was downgraded for very serious risk of bias, serious inconsistency, very serious indirectness and very serious imprecision.
Receipt and duration of oxygen supplementation: Two RCTs 24,27 included 254 healthy term infants and reported all newborns were born in good clinical condition and did not need supplemental oxygen. Clinical benefit or harm could not be excluded as the event rate was zero in both groups so a relative risk could not be calculated.
Oxygen saturations over the first 10 minutes from birth: Data were presented in different ways in different studies, precluding a comprehensive meta-analysis of all studies that reported data on this outcome. Two RCTs 9,27 (200 participants) and 2 observational studies 14,29 (418 participants) found lower oxygen saturations in those receiving suctioning within first 10 minutes, while two other RCTs 5,24 did not find significant differences. All evidence was of very low certainty.
Time to reach target oxygen saturations of 86% or 92%: One RCT 9 found time to reach 86% SaO 2 and two RCTs 9,27 found time to 92% SaO 2 was quicker in the non-suctioned group. One RC showed  (continued on next page) Observations up to 10 minutes after birth, showed no statistically significant difference, but the nonsuction group tended to stabilize both SpO2 and HR earlier than the suction group. that 90.6% of those suctioned had achieved 92% saturations at 10 minutes vs 100% of those not suctioned. The oxygenation targets were those selected by the authors. Other oxygenation outcomes: One prospective observational study 14 including 346 participants reported 1 episode of severe desaturation to <75% following suctioning. One prospective observational study 29 enrolled 138 infants born at term by elective caesarean section to examine cerebral and peripheral muscle tissue oxygenation. Between groups of 36 infants who received oropharyngeal suctioning and 36 controls, there was no difference in heart rate, oxygen saturations, cerebral and peripheral muscle tissue oxygenation.
Respiratory rate >60 in the first 24 hours: Moderate certainty evidence from one RCT with 488 participants (not restricted to healthy infants and including those !35 weeks' gestation), showed clinical benefit or harm could not be excluded (RR 0.99; 95% CI, 0.82, 1.20); ARD 5 fewer per 1000 with those receiving suctioning vs no suctioning (95% CI, 83 fewer to 92 more per 1000 patients receiving suctioning).
Heart rate at 5 minutes: Very low certainty evidence from one RCT 24 including 84 participants found clinical benefit or harm could not be excluded [MD À1.00 (95%CI, À7.96, 5.96)] however both groups had a heart rate in the normal range and no bradycardias were reported in either group. Evidence was downgraded for inconsistency and indirectness.
Apgar scores: Insufficient data on the secondary outcome of low Apgar scores (<7) was available for analysis. For the outcome of Apgar score of 10 at 5 minutes very low certainty evidence from one RCT 27  Other secondary outcomes: Insufficient data were available to be able to report on the important secondary outcomes of soft tissue injury, infection and bradycardia.

Subgroup analyses
Gestational age: Insufficient data were available for this subgroup analysis. Only one prospective observational study 14 and one RCT 26 included both preterm and term infants although most babies in both studies were born at term.
Vaginal vs Caesarean section: insufficient data were available for a subgroup analysis of the following outcomes: receipt of assisted ventilation, advanced resuscitation, receipt of supplemental oxygen, unanticipated NICU admission.
For the outcome of oxygen saturations at 5 minutes there was a difference favouring no suctioning in both vaginal delivery and caesarean section subgroups with high heterogeneity within subgroups (I 2 = 97%) and evidence of an interaction by delivery type (test for subgroup differences 0.03) also with high heterogeneity between subgroups (I 2 = 78.6%). Given the very high heterogeneity, despite almost identical results in two studiesp 3,4 a sensitivity analysis was carried out. With the two Gungor studies 3,4 removed from both subgroups there was no difference in saturations in either subgroup with no interaction (p = 0.86) and heterogeneity reduced (I 2 = 0%).  Among the two methodologically identical RCTs by Gungor 3,4 one studied vaginally born infants and the other those born by caesarean section, each included 140 participants and found identical times to achieve saturations of 86% or 92%.

Suction device used (Bulb vs Catheter Suction)
Two RCTs 5,26 studied infants receiving bulb suction vs no suction or wiping but no studies compared bulb suction to catheter suction.

Discussion
This systematic review (SR) analysed 9 RCTs 3-5,9,24-28 and 2 prospective observational studies 14,29 all of which noted that suctioning of clear amniotic fluid from the mouth and/or nose has been a common or routine historical practice in many parts of the world. The procedure is still used frequently, and suctioning can take a long time, 14 thereby potentially delaying the start of necessary critical interventions such as positive pressure ventilation. Most international guidelines recommend that if aeration of the lungs is difficult and airway obstruction is suspected then positioning to improve airway patency and if necessary, suctioning should be performed.
This systematic review found no evidence of benefit of suctioning the upper airway (compared to no suctioning) although evidence was very low certainty. Several studies reported lower oxygen satura-tions in infants receiving suctioning. However, combining the data for a meta-analysis was not possible due to differences in the presentation of data in the included studies. Some studies reported continuous measurements over time, others reported time to achieve a certain saturation.
Two RCTs 3,4 enrolling 280 participants, were originally selected for inclusion but were excluded post-hoc. The studies, which enrolled distinct groups of newborn infants (one enrolled infants born by caesarean section and the other, vaginal births) reported almost identical results for oxygen saturation levels, with much smaller standard deviations than those seen in other studies. Because a data reporting error was considered possible, a decision was made to omit the studies from the review. For transparency, analyses including them are shown in Appendix A. Their inclusion would have made little difference to the overall findings of the systematic review.
There are case reports in the literature of rare potential side effects of upper airway suctioning including cardiac arrest in one case. 7 The studies included in the review did not report any instances of severe bradycardia, but they are of insufficient size to assess low frequency adverse events. In the absence of evidence of benefit, it seems unjustified to expose large numbers of newborn infants to any risk of harm by using upper airway suctioning.
The review could not exclude the possibility that there are subgroups of newborn infants who could benefit from upper airway suctioning. The focus of this review was infants with clear amniotic fluid, Both excluded Gungor studies showed maximum time to SaO2 !92% (6 vs 11 min) and !86% (5 vs 8 min) were shorter in the no suction group (P < 0.001). MD: mean difference, CI: confidence interval, SD: standard deviation, RCT: randomised controlled trial, SaO 2 arterial oxygen saturation.
so the results cannot be considered to apply to those with blood clots, meconium or other particulate material in the amniotic fluid. The included studies included mostly healthy term infants, limiting the generalisability to preterm babies or those requiring resuscitation. We found no studies that targeted recruitment of depressed or very preterm infants. Strengths of this review include that it was conducted rigorously and in accord with a pre-registered protocol that was developed in collaboration with the combined expert opinion of the ILCOR NLS Task Force. It used a search strategy developed by an expert information specialist and was performed in adherence with established guidelines for systematic reviews. Limitations include the difficulties of obtaining additional information from authors and the differences in presentation of study results in the included studies, which precluded some of the intended meta-analyses, as well as pre-planned subgroup analyses. This may have prevented recognition of important subgroups of infants in whom the balance of risks and benefits differs.

Conclusion
This systematic review found no evidence of benefit for routine suctioning of clear amniotic fluid, compared to no suctioning, although the evidence is of low to very low certainty. There was also very low certainty evidence of a temporary adverse effect on oxygen saturation levels, of uncertain clinical significance. The review supports current guidelines which advise against routine suctioning of the upper airway in infants with clear amniotic fluid.

Contributor's statement
Drs. Fawke, Wyllie, Udaeta, Rü diger and Ersdal prepared the protocol, screened studies, completed full text reviews, abstracted data, completed risk-of-bias and GRADE evaluations, completed the analysis, and prepared the manuscript. Mary-Doug Wright developed the search strategy with the review group and conducted the initial and updated literature searches.
Drs. Liley, Weiner, Wyckoff, and Rabi reviewed the protocol, abstracted data, reviewed the analysis and edited the manuscript.
All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

Publication statement
This systematic review and meta-analysis was performed under the umbrella of the 2022 Neonatal Consensus on Science with Treatment Recommendations (CoSTR) and evidence to Decision (EtD) framework. Whilst a summary of this systematic review and metaanalysis will be included in the 2022 CoSTR paper, the submitted systematic review and meta-analysis is a more detailed version which includes all related data, figures and tables. It has not been published previously and the manuscript is not under consideration elsewhere.