Corticosteroids for septic arthritis in children
Abstract
Background
Septic arthritis is an acute infection of the joints characterised by erosive disruption of the articular space. It is the most common non‐degenerative articular disease in developing countries. The most vulnerable population for septic arthritis includes infants and preschoolers, especially boys. Septic arthritis disproportionately affects populations of low socioeconomic status. Systemic corticosteroids and antibiotic therapy may be beneficial for treatment of septic arthritis. Even if the joint infection is eradicated by antibiotic treatment, the inflammatory process may produce residual joint damage and sequelae.
Objectives
To determine the benefits and harms of corticosteroids as adjunctive therapy in children with a diagnosis of septic arthritis.
Search methods
We searched MEDLINE, Embase, the Cochrane Central Register of Controlled Trials (CENTRAL), in the Cochrane Library, Latin American Caribbean Health Sciences Literature (LILACS), the World Health Organization (WHO) trials portal (www.who.int/ictrp/en/), ClinicalTrials.gov (www.ClinicalTrials.gov), and Google Scholar. We searched all databases from their inception to 17 April 2018, with no restrictions on language of publication.
Selection criteria
We included randomised controlled trials (RCTs) with patients from two months to 18 years of age with a diagnosis of septic arthritis who were receiving corticosteroids in addition to antibiotic therapy or as an adjuvant to other therapies such as surgical drainage, intra‐articular puncture, arthroscopic irrigation, or debridement.
Data collection and analysis
Two review authors independently assessed eligibility, data extraction, and evaluation of risk of bias. We considered as major outcomes the presence of pain, activities of daily living, normal physical joint function, days of antibiotic treatment, length of hospital stay, and numbers of total and serious adverse events. We used standard methodological procedures expected by Cochrane. We assessed the evidence using GRADE (Grading of Recommendations Assessment, Development and Evaluation) and created a 'Summary of findings' table.
Main results
We included two RCTs involving a total of 149 children between three months and 18 years of age who were receiving antibiotics for septic arthritis. The most commonly affected joints were hips and knees. These studies were performed in Costa Rica and Israel. In both studies, dexamethasone administered intravenously (ranging from 0.15 to 0.2 mg/kg/dose every six to eight hours) during four days was the corticosteroid, and the comparator was placebo. Trials excluded patients with any degree of immunodeficiency or immunosuppression. The longest follow‐up was one year. Trials did not report activities of daily living nor length of hospital stay. Both studies used adequate processes for randomisation, allocation concealment, and blinding, and review authors judged them to have low risk of selection and performance bias. Losses to follow‐up were substantive in both studies, and we judged them to have high risk of attrition bias and of selective outcome reporting. We graded all outcomes as low quality due to concerns about study limitations and imprecision.
The risk ratio (RR) for absence of pain at 12 months of follow‐up was 1.33, favouring corticosteroids (95% confidence interval (CI) 1.03 to 1.72; P = 0.03; number needed to treat for an additional beneficial outcome (NNTB) = 13, 95% CI 6 to 139; absolute risk difference 24%, 95% CI 5% to 43%).
The RR for normal function of the affected joint at 12 months of follow‐up was 1.32, favouring corticosteroids (95% CI 1.12 to 1.57; P = 0.001; NNTB = 13, 95% CI 7 to 33; absolute risk difference 24%, 95% CI 11% to 37%).
We found a reduction in the number of days of intravenous antibiotic treatment favouring corticosteroids (mean difference (MD) ‐2.77, 95% CI ‐4.16 to ‐1.39) based on two trials with 149 participants.
Researchers did not report length of hospital stay. One trial (49 participants) reported that treatment with dexamethasone was associated with a shorter duration of IV antibiotic treatment, leading to a shorter hospital stay, and although duration of hospitalisation was a primary outcome of the study, study authors did not provide data on the duration of hospitalisation. We downgraded the quality by one level for concerns about study limitations (high risk of attrition bias and selective reporting), and by another level for imprecision.
In one trial of 49 participants, researchers followed 29 children for 12 months, and parents reported that no children demonstrated adverse effects of the intervention.
Authors' conclusions
Evidence for corticosteroids as adjunctive therapy in children with a diagnosis of septic arthritis is of low quality and is derived from the findings of two trials (N = 149). Corticosteroids may increase the proportion of patients without pain and the proportion of patients with normal function of the affected joint at 12 months, and may also reduce the number of days of antibiotic treatment. However, we cannot draw strong conclusions based upon these trial results. Additional randomised clinical trials in children with relevant outcomes are needed.
PICOs
Plain language summary
Corticosteroids for children with septic arthritis
Researchers conducted a review of the effects of corticosteroids given in addition to antibiotics to children with septic arthritis. Evidence was sought until April 2018. After searching for all relevant studies, reviewers found two studies with 149 children. These studies were conducted in hospitalised children with a normal immune system between the ages of three months and 18 years living in Costa Rica and Israel. The longest follow‐up was one year. Reviewer findings are summarised below.
What is septic arthritis and what are corticosteroids?
Septic arthritis, which is more frequent in children, is a serious disease caused by bacteria that infect the joints. Patients are usually treated with antibiotics, but secondary inflammation can destroy the joint and can reduce the ability of the joint to function normally. Corticosteroids are a group of medications with anti‐inflammatory properties. Corticosteroids may reduce the consequences of inflammation in the joints.
For children with septic arthritis who are taking antibiotics compared to placebo (fake medication)
1. Corticosteroids may reduce pain in affected joints at one year of follow‐up
2. Corticosteroids may improve normal function of affected joints at one year of follow‐up
3. Corticosteroids may reduce days of intravenous antibiotic treatment needed
4. Corticosteroids may have little or no effect on total or serious adverse effects
We do not have information about the effects of corticosteroids on activities of daily living.
What happens to children with septic arthritis who take corticosteroids in addition to antibiotics?
Absence of pain
1. 24 more of 100 children experienced absence of pain after 12 months with corticosteroids (24% absolute improvement)
2. 96 of 100 children experienced absence of pain compared to 72 of 100 children who took a placebo
Activities of daily living
Included studies did not report this outcome.
Normal physical joint function
1. 24 more of 100 children who received corticosteroids had normal function of the joint after 12 months (24% absolute improvement)
2. 98 of 100 children experienced absence of pain compared to 74 of 100 children who received a placebo
Number of days of intravenous antibiotic treatment
1. Children who received corticosteroids compared with placebo had 2.77 fewer days of intravenous antibiotic treatment
2. Children who received corticosteroids had 8.09 days of intravenous antibiotic treatment
3. Children who received placebo had 10.86 days of intravenous antibiotic treatment
Length of hospital stay
1. We are uncertain whether corticosteroids had an effect on the length of hospital stay because the evidence was of very low quality
Total or serious adverse events
1. None of the patients treated with corticosteroids reported adverse effects at 12 months
Quality of the evidence
Overall, these studies provided low‐quality evidence due to small numbers of study participants and concerns about study design. Evidence on length of hospital stay was of very low quality, as this was not clearly reported.
Authors' conclusions
Summary of findings
| Corticosteroids compared to placebo for septic arthritis in children | ||||||
| Patient or population: septic arthritis in children taking antibiotics | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | No. of participants | Certainty of the evidence | Comments | |
| Risk with placebo | Risk with corticosteroids | |||||
| Pain ‐ absence of pain at 12 months | 72 per 100 | 96 per 100 | RR 1.33 | 49 | ⊕⊕⊝⊝ | Absolute risk difference 24% (95% CI 5% to 43%) NNTB = 13 (95% CI 6 to 139) |
| Activities of daily living | See comments | ‐ | (0 RCTs) | ‐ | This outcome was not reported in any of the included trials | |
| Number of participants with normal physical joint function ‐ normal function at 12 months of follow‐up (long term) | 74 per 100 | 98 per 100 | RR 1.32 | 100 | ⊕⊕⊝⊝ | Absolute risk difference 24% (95% CI 11% to 37%) NNTB = 13 (95% CI 7 to 33) |
| Number of days of antibiotic treatment ‐ number of days of intravenous antibiotic treatment | Number of days of intravenous antibiotic treatment with placebo was 10.86 | Number of days of intravenous antibiotic treatment with corticosteroids was 8.09 | MD 2.77 lower | 149 | ⊕⊕⊝⊝ | Almost 3 days lower (95% CI 4 days to 1.5 days lower) |
| Length of hospital stay | See comments | ‐ | (0 RCTs) | ‐ | This outcome was not reported in any of the included trials. Study authors report that treatment with dexamethasone was associated with a shorter duration of IV antibiotic treatment, leading to a shorter hospital stay | |
| Total adverse events at 12 months | See comments | ‐ | 49 | ⊕⊕⊝⊝ | Trial reported that none of the participants showed adverse effects | |
| Serious adverse events at 12 months | See comments | ‐ | 49 | ⊕⊕⊝⊝ | Trial reported that none of the participants showed serious adverse effects | |
| *The risk in the intervention group (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). | ||||||
| GRADE Working Group grades of evidence. | ||||||
| aWe downgraded by one level for concerns about study limitations (high risk of attrition bias and selective reporting). We downgraded by another level for imprecision. bWe downgraded by one level for concerns about study limitations (unclear risk of detection and reporting bias and high risk of attrition bias). We downgraded by another level for imprecision. cWe downgraded by one level for concerns about study limitations (detection, reporting, and attrition bias). We downgraded by another level for imprecision. | ||||||
Background
Description of the condition
Septic arthritis is an acute infection of the joints characterised by erosive disruption of the articular space. It is the most common non‐degenerative articular disease in developing countries. Although septic arthritis can be diagnosed at any age, children are most often affected. The incidence of septic arthritis in children from developed countries ranges from five to 12 cases for every 100,000 persons per year (Forero 2013; García‐Arias 2011). Infants and preschoolers, especially boys, are most vulnerable to septic arthritis. Septic arthritis disproportionately affects populations of low socioeconomic status (García‐Arias 2011).
Septic arthritis often affects the hip, knee, elbow, and ankle joints (Young 2011). Signs and symptoms include pain, oedema, erythema, and functional limitation, as well as shivering, sickness, emesis, and fever. Symptoms may vary in very young or breastfeeding children (Harel 2011; Odio 2003).
The bacteria responsible for septic arthritis generally vary with age, comorbid conditions, socioeconomic status, and immunogenic and vaccination status of the patient (Dodwell 2013). Staphylococcus aureus is isolated in 37% to 56% of all cases (García‐Arias 2011). Studies have shown that gram‐negative bacteria such as Escherichia coli, Proteus mirabilis, Klebsiella spp, and Enterobacteriaceae are also responsible for septic arthritis, specifically among children younger than four years. Septic arthritis due to Streptococcus has decreased as a result of mass vaccination (Mathews 2008a).
Recent studies have found that emerging bacteria play a critical role in septic arthritis. Methicillin‐resistant Staphylococcus aureus (MRSA) has been found to be responsible for 30% to 40% of osteoarticular infections in the United States (Dodwell 2013). Kingella kingae, which causes unusual clinical symptoms, has also been isolated. Other pathogens such as Haemophilus influenzae type B,Neisseria gonorrhoeae, andCandida albicans have been involved in development of the disease (Dodwell 2013).
Bacterial colonisation in septic arthritis most frequently occurs via the hematogenous route. Researchers have also described direct inoculation through an injury or infection in surrounding tissue (Mathews 2008a). The risk of complications is higher among children under two years of age because the epiphysis is permeated by blood vessels, and this facilitates bacterial colonisation through the growth plates (Kang 2009; Mathews 2008b).
Characteristics of the synovial membrane foster bacterial reproduction 24 to 48 hours after infection. The synovial membrane reacts with hyperplasia, and neutrophils and monocytes release cytokines and proteases, producing exudation (Kang 2009). Consequently, neutrophils inhibit synthesis of the cartilage, causing destruction of the articulation in approximately seven days. Infiltration of neutrophils can damage adjacent bone and metaphyseal growth cartilage, producing osteoarthritis (Mathews 2010).
Clinicians evaluate C‐reactive protein (CRP), erythrocyte sedimentation rate (ESR), procalcitonin (PCT), and blood cultures (Choe 2013; Harel 2011; Odio 2003), as well as synovial liquid (Al Saadi 2009), in diagnosing septic arthritis. They also perform other examinations such as X‐ray, echography, magnetic resonance imaging, gammagraphy, and evaluation of the synovial liquid (Al Saadi 2009). Accuracy of the diagnosis depends on bacterial isolation from the articular exudate or the blood culture (Harel 2011).
Description of the intervention
Staphylococus aureus hampers the autoimmune response via factors related to virulence, as this organism triggers numerous signals that exacerbate the humoral immune response. Staphylococus aureus suppresses the innate immune system, increasing humoral immune answer and facilitating the development of septic arthritis. This in turn promotes progressive destruction of the cartilage and of the chondral bone, generating the clinical manifestations associated with septic arthritis (Bremell 1992).
Systemic corticosteroids and antibiotics may be beneficial for treatment of septic arthritis (Al Saadi 2009). In septic arthritis, even if the joint infection is eradicated by antibiotic treatment, the inflammatory process produces residual joint damage and sequelae. Corticosteroids are thought to have a therapeutic role in attenuating the inflammatory response mediated by cytokines and interleukins, and in avoiding sequelae or complications (i.e. restricted articulation movements, articular dysfunction) (Baghdadi 2012; Brouwer 2015).
How the intervention might work
During acute joint infection, Toll‐like receptors are activated, with substantial release of pro‐inflammatory cytokines such as Interleukin (IL)‐1‐beta, IL‐17, IL‐6, and tumour necrosis factor (TNF)‐alpha (Colavite 2014; Farrow 2015). This inflammatory process promotes osteoclast differentiation, bone reabsorption, and matrix metalloproteinase (MMP) release, leading to bone and cartilage destruction (Farrow 2015; Kwan 2004). Intra‐articular pressure is increased (secondary to joint effusion), which mechanically reduces blood/nutrient supply to the joint and increases cellular damage (Farrow 2015; Shirtliff 2002).
Pharmacodynamic effects of corticosteroids are mediated through binding to glucocorticoid receptors. Phospholipase A2 (a potent intracellular producer of prostaglandins, free radicals, and leukotrienes) is inhibited, and deleterious effects on several pro‐inflammatory cytokines such as IL‐1, interferon (IFN)‐γ, and TNF‐alpha, are evident. As a result, corticosteroids could attenuate the destruction of cartilage and of the synovial membrane and capsule.
Why it is important to do this review
Septic arthritis is a limiting and relatively common condition, especially in children. Corticosteroid therapy may have potential benefits for septic arthritis, as well as a positive impact on the patient's quality of life. Therefore, it is important to conduct a systematic review to determine the benefits and harms of such treatment, to guide clinicians when treating paediatric patients with this diagnosis.
We conducted this review according to guidelines provided by the Cochrane Musculoskeletal Group Editorial Board (Ghogomu 2014).
Objectives
To determine the benefits and harms of corticosteroids as adjunctive therapy in children with a diagnosis of septic arthritis.
Methods
Criteria for considering studies for this review
Types of studies
We included randomised controlled trials (RCTs) reported as full text, those published as abstract only, and unpublished data. We applied no language restrictions.
Types of participants
We included patients from two months to 18 years of age with the diagnosis of septic arthritis, as defined by the authors of each study, who were receiving corticosteroids in addition to antibiotic therapy or as adjuvant to other therapies.
The diagnosis of septic arthritis is based mainly on the presence of clinical symptoms, a detailed history, and a careful examination (acute onset of swelling pain, local warmth, and severe limitation of motion in any joint), along with findings of laboratory tests (elevated levels of acute phase reactants, erythrocyte sedimentation rate (ESR), and C‐reactive protein (CRP), and/or increase in white blood cell (WBC) count). A definitive diagnosis of septic arthritis is made by direct demonstration of bacteria from synovial fluid or blood (gram‐stained smear or direct cultures).
Types of interventions
We included trials in which researchers treated participants with antibiotics and/or other co‐interventions plus corticosteroids administered by any route (oral, intramuscular, intravenous, or intra‐articular), at any dose. As controls, we considered patients treated only with antibiotics and/or any therapy other than corticosteroids, such as surgical drainage, intra‐articular puncture (using a needle to release articular pressure or to extract intra‐articular fluid), arthroscopic irrigation, or debridement.
We allowed the following co‐interventions, provided they were not part of the randomised treatment and were given to participants in all treatment arms of the trial: drainage: intra‐articular puncture, arthroscopic irrigation, and debridement.
Types of outcome measures
Major outcomes
-
Pain measured by different scales (visual analogue scales, numerical rating scales, or facial drawings scales)
-
Activities of daily living (measured by parent report, self‐report, or any assessment instrument appropriate to participant age (e.g. Activities of Daily Living (ADL) Scale, Klein‐Bell Activities of Daily Living Scale (KB ADL))
-
Number of participants with normal physical joint function (normal angle of joint movement according to age and sex, measured with a goniometer/other instrument and/or compared with that of the contralateral healthy joint)
-
Number of days of antibiotic treatment
-
Length of hospital stay
-
Total number of adverse events
-
Number of serious adverse events
Outcomes were measured at short term (one day to one month), medium term (one month to six months), and long term (beyond six months).
Scales used to measure paediatric pain are subjective; therefore, numerical scales and scales of facial drawings are used more frequently.
We created a 'Summary of findings' table using the same outcomes as above to report long‐term measurements (beyond six months).
Search methods for identification of studies
Electronic searches
We searched the following databases: MEDLINE, Embase, the Cochrane Central Register of Controlled Trials (CENTRAL), in the Cochrane Library, Latin American Caribbean Health Sciences Literature (LILACS), the World Health Organization (WHO) trials portal (www.who.int/ictrp/en/), ClinicalTrials.gov (www.ClinicalTrials.gov), and Google Schoolar. We searched all databases from their inception to 17 April 2018, with no restrictions on language of publication.
See Appendix 1, Appendix 2, Appendix 3, Appendix 4, Appendix 5, Appendix 6, and Appendix 7, for details of each search strategy.
Searching other resources
We checked the reference lists of all primary studies and reviewed articles for additional references. We searched relevant manufacturers' websites for trial information.
We searched for errata or retractions from included studies published in full text on PubMed (www.ncbi.nlm.nih.gov/pubmed) and reported within the review the date this was done.
We sent emails to the authors of included clinical trials to ask if they were aware of other similar clinical trials.
Data collection and analysis
Selection of studies
Two review authors (AF, JF) independently screened titles and abstracts for inclusion of all potentially relevant studies identified as a result of the search, and coded them as 'retrieve' (eligible or potentially eligible/unclear) or 'do not retrieve'. We retrieved the full‐text study reports/publications, and two review authors independently screened these and identified studies for inclusion, then identified and recorded the reasons for exclusion of ineligible studies. We resolved disagreements through discussion or, if required, by consultation with a third review author (MD, JMC). We identified and excluded duplicates and collated multiple reports of the same study, so that each study, rather than each report, was the unit of interest in the review. We recorded the selection process in sufficient detail to complete a PRISMA flow diagram and a Characteristics of excluded studies table.
Data extraction and management
We used a data collection form that had been piloted on at least one study in the review when extracting study characteristics and outcome data. One review author (AF or JF) extracted study characteristics from included studies. A second review author (JAC) spot‐checked study characteristics for accuracy against the trial report. We extracted the following study characteristics.
-
Methods: study design, total duration of study, number of study centres and locations, study setting, withdrawals, and date of study.
-
Participants: N, age mean and range, sex, disease duration, severity of condition, diagnostic criteria, important baseline data; inclusion criteria, and exclusion criteria.
-
Interventions: intervention of interest, comparison, concomitant medications, and excluded medications.
-
Outcomes: primary and secondary outcomes specified and collected, and time points reported.
-
Characteristics of the design of the trial, as outlined below in the Assessment of risk of bias in included studies section.
-
Notes: funding for trial and declarations of interest of all trial authors.
Two review authors (AF, JF) independently extracted outcome data from included studies. We extracted numbers of events and numbers of participants per treatment group for dichotomous outcomes, and means and standard deviations and numbers of participants per treatment group for continuous outcomes. We noted in the Characteristics of included studies table if outcome data were not reported in a usable way, and if data had been transformed or estimated from a graph. We resolved disagreements by consensus or by consultation with a third person (JAC or MD). One review author transferred data into the Review Manager file (RevMan 2014).
We double‐checked that data had been entered correctly by comparing data presented in the systematic review against the study reports.
For data extraction, we analysed data as presented by study authors for each outcome considered. Regarding outcomes, we extracted final values data, adjusted values if reported, intention‐to‐treat (ITT) data, and the final time point if available.
Assessment of risk of bias in included studies
Two review authors (AF, JF) independently assessed the risk of bias for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We resolved disagreements by discussion or by consultation with another review author (MD, JAC). We assessed the risk of bias according to the following domains.
-
Random sequence generation.
-
Allocation concealment.
-
Blinding of participants and personnel.
-
Blinding of outcome assessment.
-
Incomplete outcome data.
-
Selective outcome reporting.
-
Other bias.
We graded each potential source of bias as having high, low, or unclear risk, and we provided a quote from the study report, as well as a justification for our judgement, in the 'Risk of bias' table. We summarised 'Risk of bias' judgements across different studies for each of the domains listed. We considered the impact of missing data for key outcomes.
When information on risk of bias relates to unpublished data or corresponds to a trial list, we noted this in the 'Risk of bias' table.
When considering treatment effects, we took into account the risk of bias for studies contributing to that outcome.
We have presented the figures generated by the 'Risk of bias' tool to provide summary assessments of risk of bias.
Assessment of bias in conducting the systematic review
We conducted this review according to the published protocol and reported any deviations from it in the Differences between protocol and review section.
Measures of treatment effect
We analysed dichotomous data as risk ratios (RRs) or Peto odds ratios when the outcome was a rare event (approximately < 10%), along with 95% confidence intervals (CIs).
We analysed continuous data (i.e. continuous pain scales, visual analogue scales, numerical rating scales, or facial drawings scales) as mean differences (MDs) or standardised mean differences (SMDs) (depending on whether the same scale was used to measure an outcome), along with 95% CIs. We transformed outcome data to a continuous scale depending on each situation (e.g. facial drawings scales 0 to 6). We entered data presented as a scale with a consistent direction of effect across studies.
When researchers used different scales to measure the same conceptual outcome (e.g. disability), we calculated SMDs with corresponding 95% CIs. We back‐translated the SMD to a typical scale (i.e. 0 to 10 for pain) by multiplying the SMD by a typical among‐person standard deviation (e.g. the standard deviation of the control group at baseline from the most representative trial).
In the Effects of interventions results section and in the 'Comments' column of the 'Summary of findings' table, we provided the absolute per cent difference, the relative per cent change from baseline, and the number needed to treat for an additional beneficial outcome (NNTB) (we provided the NNTB only when the outcome showed a statistically significant difference).
For dichotomous outcomes, such as serious adverse events, we calculated NNTB from the control group event rate and the RR using the Visual Rx NNTB calculator (Cates 2008). We calculated the NNTB for continuous measures using the Wells calculator.
For dichotomous outcomes, we calculated the absolute risk difference using the risk difference statistic in RevMan and expressed this result as a percentage. For continuous outcomes, we calculated the absolute benefit as improvement in the intervention group minus improvement in the control group, expressed in original units.
Unit of analysis issues
When a single trial reported multiple trial arms, we included only the relevant arms. If two comparisons (e.g. drug A vs placebo and drug B vs placebo) were combined in the same meta‐analysis, we halved the control group to avoid double‐counting.
Dealing with missing data
We contacted investigators or study sponsors to verify key study characteristics and to obtain missing numerical outcome data when possible (e.g. when a study was identified as abstract only, when data are not available for all participants). When it was not possible, and when missing data were thought to introduce serious bias, we explored the impact of including such studies in the overall assessment of results by performing a sensitivity analysis. We clearly described assumptions and imputations applied to handle missing data, and we explored the effect of imputation by performing sensitivity analyses.
For dichotomous outcomes (e.g. number of withdrawals due to adverse events), we calculated the withdrawal rate by using the number of randomised participants in the group as the denominator.
For continuous outcomes (e.g. mean change in pain score), we calculated the MD or the SMD based on the number of participants analysed at a given time point. If researchers did not present the number of participants analysed for each time point, we used the number of randomised participants in each group at baseline.
When possible, we computed missing standard deviations from other statistics such as standard errors, CIs, or P values, according to the methods recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). If we could not calculate standard deviations, we imputed them (e.g. from other studies in the meta‐analysis).
Assessment of heterogeneity
We assessed clinical and methodological heterogeneity for the included studies in terms of participants, interventions, outcomes, and study characteristics to determine whether a meta‐analysis was appropriate. We did this by observing data included in the data extraction tables. We assessed statistical heterogeneity by visually inspecting the forest plot to assess obvious differences in results between studies, and by using I² and Chi² statistical testing.
As recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011), an I² value of 0% to 40% might 'not be important'; 30% to 60% may represent 'moderate' heterogeneity; 50% to 90% may represent 'substantial' heterogeneity; and 75% to 100% represents 'considerable' heterogeneity. As noted in the Cochrane Handbook for Systematic Reviews of Interventions, we kept in mind that the importance of I² depends on (1) magnitude and direction of effects and (2) strength of evidence for heterogeneity.
For the Chi² test, a P value ≤ 0.10 indicated evidence of statistical heterogeneity.
When we identified substantial heterogeneity, we reported it and investigated possible causes by following the recommendations provided in Section 9.6 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
Assessment of reporting biases
If we needed to evaluate reporting bias, we created and examined a funnel plot to explore a possible small‐study bias. To interpret funnel plots appropriately, we looked at different possible reasons for funnel plot asymmetry, as outlined in Section 10.4 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011), and we related this to review results (Sterne 2011).
To assess outcome reporting bias, we checked the trial protocols against published reports. For studies published after 1 July 2005, we screened the Clinical Trials Register at the International Clinical Trials Registry Platform of the World Health Organization (http://apps.who.int/trialssearch) for the a priori trial protocol. We evaluated whether selective reporting of outcomes was present.
Data synthesis
We conducted meta‐analyses only when this was meaningful (i.e. if treatments, participants, and the underlying clinical question were similar enough for pooling). We used a random‐effects model.
If pooling was not possible, we summarised results of the included trials by using a descriptive approach. We made a summary table under each study to report participants, interventions, controls, and outcomes. Then, we discussed study findings, strengths, and weaknesses from a clinical perspective.
'Summary of findings' table
We created a 'Summary of findings' table using the outcomes listed at Types of outcome measures.
Two review authors(AF, JF) independently assessed evidence quality. We used the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness, and publication bias) to assess the quality of a body of evidence as it relates to studies that contributed data to the meta‐analyses for prespecified outcomes. We used methods and recommendations described in Sections 8.5 and 8.7, and in Chapters 11 and 13, Section 13.5, of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011; Schünemann 2011), as well as GRADEpro 2015. We justified all decisions to downgrade or upgrade the quality of studies by using footnotes, and we made comments when necessary to aid the reader in understanding the review.
In the Comments column of the 'Summary of findings' table, we provided the absolute per cent difference, the relative per cent change from baseline, and the number needed to treat for an additional beneficial outcome (NNTB) (we provided the NNTB only when the outcome showed a statistically significant difference).
Subgroup analysis and investigation of heterogeneity
We planned to carry out the following subgroup analyses.
-
Age (infants (up to two years), preschoolers (two to five years), and school children (older than six years)).
-
Route of administration (intra‐articular vs intramuscular vs oral vs intravenous).·
-
Co‐intervention type: drainage, intra‐articular puncture, arthroscopic irrigation, or debridement.
We planned to use the formal test for subgroup interactions in Review Manager (RevMan 2014), and we intended to apply caution in interpreting subgroup analyses, as advised in Section 9.6 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
Sensitivity analysis
We planned to carry out the following sensitivity analyses.
-
By risk of bias of included studies: we will classify a study as having high risk of bias when one or more domains of the 'Risk of bias' assessment tool is classified as being at high risk of bias.
-
By diagnosis: we will classify a study based on clinical criteria without bacterial isolation.
Interpreting results and reaching conclusions
We followed the guidelines for interpreting results as provided in Chapter 12 of the Cochrane Handbook for Systematic Reviews of Interventions, and we were aware of distinguishing lack of evidence of effect from lack of effect (Higgins 2011). We based our conclusions only on findings from the quantitative or narrative synthesis of included studies prepared for this review. We avoided making recommendations for practice, and our implications for research suggest priorities for future research and outline remaining uncertainties in the area.
Results
Description of studies
Results of the search
Through our search strategy, we identified the following numbers of studies until 13 April 2018: MEDLINE 12, Embase 0, CENTRAL 10, LILACS 4, WHO 48, ClinicalTrials.gov 15, and Google Scholar 388, for a total of 477 records. We have presented the search process in a flow chart (Figure 1).
Study flow diagram.
Included studies
We included two studies ‐ Odio 2003 and Harel 2011 (total 149 children) ‐ conducted to evaluate the clinical effects of corticosteroids as adjunctive therapy for children with septic arthritis. Both studies were randomised controlled trials, and both used placebo in the control group. Study authors stated no funding sources. We recorded details of these papers under Included studies and in the Characteristics of included studies tables. We found no other references to these studies.
The author of one of the included studies was a co‐author of a cohort study that we excluded (Fogel 2015).
We sent emails to the authors of included clinical trials to ask if they were aware of other similar clinical trials, but we received no response.
Participants
Odio 2003 was a clinical trial conducted in Costa Rica in 123 children, and Harel 2011 was conducted in Israel in 46 patients. Children in Odio 2003 were between three months and 13 years of age. In Harel 2011, the age of participants was between six months and 18 years. The most commonly affected joints were lower limbs, hips, and knees in both trials. Trials excluded patients with any degree of immunodeficiency or immunosuppression.
Interventions
Both trials used dexamethasone as the experimental intervention. In Odio 2003, investigators administered treatment by the intravenous route, at 0.2 mg/kg/dose IV every eight hours for 12 consecutive doses. In Harel 2011, study authors provided 0.15 mg/kg/dose every six hours for 16 consecutive doses.
Outcomes
The main outcome in Odio 2003 was the normal function of the affected joint at the end of therapy (short term), at six months (medium term), and at 12 months (long term). Trialists measured outcomes as dichotomous variables. Secondary endpoints included time to clinical and laboratory normalisation in terms of the following: days of resolution of symptoms, absence of fever, absence of spontaneous pain or pain to passive or active movement, absence of warmth and oedema, normal range of movement of the joint, and days of normalisation of serum C‐reactive protein. Another outcome was duration (number of days) of intravenous and oral antibiotics.
The main outcome in Harel 2011 was the number of days to clinical and laboratory normalisation (number of days until no fever, number of days until no local heat, number of days until no redness of the joint, number of days until pain free, number of days until full range of movement of the joint, number of days until normal function of the joint, last day ESR +25 mm/h, last day WBC +15.000, and last day CRP +0.5 mg/dL). Additionally, study authors reported duration (number of days) of intravenous and oral administration of antibiotics. They reported the presence of pain as dichotomous data over the long term and as assessed by phone call at 12 months. Finally, they reported numbers of total and serious adverse events as assessed by parental reporting at 12 months.
Excluded studies
We excluded three studies: van Oosterhout 2006, Arti 2014, and Fogel 2015. Reasons for exclusion included type of intervention and type of study design. We have provided specific reasons for exclusion in the Excluded studies table.
Risk of bias in included studies
We assessed the risk of bias in included studies using details provided in Chapter 8.7a of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
We have included a summary for 'Risk of bias' in the Characteristics of included studies table, in the methodological quality graph (Figure 2), and in the methodological quality summary (Figure 3).
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Random sequence generation
We judged both Odio 2003 and Harel 2011 as having low risk of bias. These studies explicitly reported the randomisation process. Odio 2003 explicitly reported the randomisation sequence assignment.
Allocation
We judged both Odio 2003 and Harel 2011 as having low risk of bias, as they explicitly reported the allocation concealment process.
Blinding
The two included clinical trials reported an appropriate blinding procedure for both participants and study personnel; thus, we judged them as having low risk of bias (Harel 2011; Odio 2003).
Blinding of outcome assessment
When we assessed blinding of outcome assessment, we judged Odio 2003 as having unclear risk of detection bias because study authors did not mention how they assessed outcomes. We judged Harel 2011 as having low risk of detection bias.
Incomplete outcome data
Both studies reported losses to follow‐up (Harel 2011; Odio 2003). Odio 2003 described loss to follow‐up of 18.7%, with similar data for experimental and control groups. Harel 2011 indicated that loss to follow‐up was 40.9%, and that this finding was not similar between groups (intervention 30%, placebo 52%). We judged both studies as having high risk of bias for incomplete outcome data.
Selective reporting
Harel 2011 did not clearly describe how study authors measured sequelae as a long‐term outcome. In addition, study authors did not report the primary prespecified outcome of the trial, as mentioned in the Methods section of the trial report (length of hospital stay). Therefore, we judged this study as having high risk of bias for selective reporting. Odio 2003 did not present in a clear way data for the outcome of range of movement of the joint.
Effects of interventions
1. Pain
1.1. Absence of pain
Only Harel 2011 with 49 participants considered this outcome at long term (12 months). Researchers reported data as a categorical outcome (presence or absence of pain). The risk ratio (RR) for absence of pain at 12 months of follow‐up (long‐term) was 1.33 (95% confidence interval (CI) 1.03 to 1.72; P = 0.03; number needed to treat for an additional beneficial outcome (NNTB) = 13, 95% CI 6 to 139)). We downgraded quality by one level for concerns about study limitations (high risk of bias for attrition and selective reporting). We downgraded by another level for imprecision due to small sample size, which resulted in a large confidence interval.
1.2. Number of days until pain free
Only Harel 2011 with 49 participants considered this outcome. Compared with placebo, the number of days until pain free was 3.58 days lower with corticosteroids (mean difference (MD) ‐3.58, 95% CI ‐7.41 to 0.25). Data show no statistically significant differences (P = 0.07). The quality of the evidence was low due to concerns about attrition bias, selective reporting, and imprecision (due to a large confidence interval and the small number of study participants).
2. Activities of daily living
We found no studies reporting this outcome.
3. Number of participants with normal physical joint function
3.1. Number of days until normal function of the joint
Harel 2011 with 49 participants considered this outcome. Compared with placebo, the number of days until normal function of the joint was 2.07 days lower with corticosteroids (MD ‐2.07, 95% CI ‐5.47 to 1.33). Data show no statistically significant differences (P = 0.23). The quality of the evidence was low due to concerns about attrition bias, selective reporting, and imprecision.
3.2. Number of days until full range of movement
Harel 2011 with 49 participants considered this outcome. Compared with placebo, the number of days until full range of movement was 5.24 days lower with corticosteroids (MD ‐5.24, 95% CI ‐10.37 to ‐0.11). Data show statistically significant differences (P = 0.05). The quality of the evidence was low due to concerns about attrition bias, selective reporting, and imprecision, because the confidence interval was very large for number of days until full range of movement.
3.3. Normal function at the end of treatment (short term)
Odio 2003 with 100 participants considered this outcome in the short term at the end of treatment. Study authors reported data as categorical outcomes. The risk ratio for normal function at the end of treatment (short term) was 1.41 (95% CI 1.16 to 1.72; P = 0.0007; NNTB = 8, 95% CI 5 to 20). We downgraded quality by one level for concerns about study limitations (unclear risk of detection and reporting bias and high risk of attrition bias). We downgraded by another level for imprecision (due to few participants and a wide confidence interval).
3.4. Normal function at six months of follow‐up (medium term)
Odio 2003 with 100 participants considered this outcome in the medium term. Researchers reported data as categorical outcomes. The estimated risk ratio for normal function at six months of follow‐up (medium term) was 1.58 (95% CI 1.27 to 1.97; P < 0.0001; NNTB = 5, 95% CI 3 to 11). We downgraded quality by one level for concerns about study limitations (unclear risk of detection and reporting bias and high risk of attrition bias). We downgraded by another level for imprecision (due to few participants and a wide confidence interval).
3.5. Normal function at 12 months of follow‐up (long term)
Odio 2003 with 100 participants considered this outcome in the long term. Investigators reported data as categorical outcomes. The risk ratio for normal function at 12 months of follow‐up (long term) was 1.32 (95% CI 1.12 to 1.57; P = 0.001; NNTB = 13, 95% CI 7 to 33). We downgraded quality by one level for concerns about study limitations (unclear risk of detection and reporting bias and high risk of attrition bias). We downgraded by another level for imprecision (due to few participants and a wide confidence interval).
3.6. Number of days until resolution of symptoms
Only Odio 2003 with 100 participants considered this outcome. Compared with placebo, the number of days until resolution of symptoms was 5.47 days lower with corticosteroids (MD ‐5.47, 95% CI ‐6.98 to ‐3.96). Data show statistically significant differences (P < 0.00001). We downgraded the quality of the evidence because of unclear risk of detection bias and selective reporting and high risk of attrition bias. This result may be clinically relevant from the patient perspective.
4. Number of days of antibiotic treatment
4.1. Number of days of intravenous antibiotic treatment
Two studies with a total of 149 participants measured this outcome (Harel 2011; Odio 2003). Study authors reported this as a continuous variable. Compared with placebo, the number of days of intravenous antibiotic treatment was 2.77 days lower with corticosteroids (MD ‐2.77, 95% CI ‐4.16 to ‐1.39). Data show statistically significant differences (P < 0.0001), and heterogeneity between studies was 0%. We downgraded quality by one level for concerns about study limitations (detection, reporting, and attrition bias). We downgraded by another level for imprecision (due to few participants and a wide confidence interval).
4.2. Number of total days of antibiotic treatment
Two studies with a total of 149 participants measured this outcome and reported it as a continuous variable (Harel 2011; Odio 2003).
Compared with placebo, the total number of days of antibiotic treatment was 3.11 days lower with corticosteroids (MD ‐3.11, 95% CI ‐6.94 to 0.73). Data show no statistically significant differences (P = 0.11), and heterogeneity between studies was 60%. We downgraded quality by one level for concerns about study limitations (detection, reporting, and attrition bias). We downgraded by another level for imprecision (due to few participants and a wide confidence interval).
5. Length of hospital stay
We found no studies reporting this outcome.
Harel 2011 with 49 participants indirectly reported this outcome. Researchers found that treatment with dexamethasone was associated with a shorter duration of IV antibiotic treatment, leading to a shorter hospital stay. They did not provide numerical data and did not clearly report the outcome. We downgraded quality by one level for concerns about study limitations (high risk of bias of attrition and selective reporting). We downgraded by another level for imprecision due to small sample size.
6. Total adverse events
Harel 2011 with 49 participants reported this outcome. At long term, investigators followed up with 29 children, and none demonstrated adverse effects of the intervention. We downgraded quality by one level for concerns about study limitations (high risk of attrition bias and selective reporting). We downgraded by another level for imprecision due to small sample size, which resulted in a large confidence interval.
7. Serious adverse events
Harel 2011 with 49 participants reported this outcome as presented above. No other studies reported additional serious adverse events.
We conducted no planned subgroup or sensitivity analyses due to lack of data.
Discussion
Summary of main results
For septic arthritis in children, even though bacterial infection is eradicated by antibiotic treatment, the inflammatory process continues, generating residual articular damage with irreversible loss of joint function. The inflammatory process is characterised by rapid arrival of polymorphonuclear cells, activated macrophages, and T cells, as well as by the production of cytokines, which destroy the articular cartilage (Colavite 2014). It is thought that corticosteroids could have a therapeutic role in reducing the inflammatory response as mediated by cytokines and interleukins, thus avoiding residual sequelae or improving function with septic arthritis (Kang 2009; Kwan 2004).
In our review, we aimed to determine the effect of corticosteroids as adjunctive therapy for children with septic arthritis and found that findings of our comprehensive literature search were scarce. Only two randomised clinical trials conducted in Costa Rica and Israel, respectively, met our inclusion criteria (Harel 2011; Odio 2003); thus, researchers have tested very few interventions with corticosteroids for this condition during childhood. Outcomes reported in those trials, such as pain, symptom resolution, joint function, joint range of movement, and the presence of sequelae, are clinically important. None of the included studies followed up with children beyond 12 months. Trials did not report activities of daily living nor length of hospital stay.
Regarding measured outcomes, the risk ratio (RR) for absence of pain at 12 months of follow‐up was 1.33, favouring corticosteroids, with the number needed to treat for an additional beneficial outcome (NNTB) of 13. The RR for normal function of the affected joint at 12 months of follow‐up was 1.32, favouring corticosteroids, with an NNTB of 13. We conducted a meta‐analysis including the two studies and found a reduction in the number of days of intravenous antibiotic treatment favouring corticosteroids, with a mean difference (MD) of ‐2.77 days (Harel 2011; Odio 2003). In one trial of 49 participants followed to 12 months (Harel 2011), parents reported that no children demonstrated adverse effects of the intervention (summary of findings Table for the main comparison).
Both studies described an adequate randomisation process and adequate allocation concealment and blinding, and we judged them to have low risk of selection and performance bias. The number of losses to follow‐up was substantive in both studies, and we judged these studies as having high risk of attrition bias and selective outcome reporting. We graded all outcomes as low quality due to concerns about study limitations and imprecision (summary of findings Table for the main comparison).
Both studies assessed outcomes not prespecified in our protocol, such as return to normal values of parameters used during clinical follow‐up of infectious processes (e.g. C‐reactive protein). Both studies reported a shorter time to reaching normal values.
Overall completeness and applicability of evidence
Because of the potential benefit of an inexpensive adjuvant corticosteroid therapy, our review question remains important. Additional studies including larger numbers of participants would increase our confidence in these results. Lack of studies was a major limitation of this review. Participants and interventions included in these trials are representative of usual clinical practice. Included studies were published in Costa Rica and Israel and included children between three months and 18 years of age, with lower limbs reported as the most commonly affected joints.
Both trials used dexamethasone as the experimental intervention. Dexamethasone led to a reduction in pain at 12 months, an increase in the number of participants with normal function of the joint at 12 months, and a reduced number of days of IV antibiotic treatment. However, the included studies did not assess important outcomes from the patient perspective, such as activities of daily living. Current data are limited to 12 months of follow‐up. We conducted no planned subgroup or sensitivity analyses due to lack of data.
We classified quality of the evidence as low due to limitations in risk of bias and to the number of studies with imprecision. Nevertheless, results have led us to believe that this is an acceptable intervention for children with septic arthritis. Corticosteroids are an accessible, available, and inexpensive intervention.
Quality of the evidence
Studies identified for this review remain insufficient to provide robust support for use of this intervention. We identified two trials for inclusion in this review, and both used dexamethasone as the corticosteroid and placebo as the comparative intervention (Harel 2011; Odio 2003). Investigators studied a total of 149 children with a follow‐up period of one year. Both studies described adequate processes for randomisation, allocation concealment, and blinding, and we judged them as having low risk of selection and performance bias. However, the number of losses to follow‐up was substantive in both studies, and we judged them as having high risk of attrition bias. In addition, we judged Harel 2011 as having high risk of selective outcome reporting. One trial described serious adverse events and length of hospital stay as narrative statements without providing additional data.
We used the GRADE approach to evaluate the body of evidence, and we prepared summary of findings Table for the main comparison. In general, we assessed the quality of evidence for the most important outcomes as low. We downgraded quality of the evidence for outcomes because of concerns about attrition bias, selective reporting, and imprecision.
Potential biases in the review process
The possibility that some studies may not have been identified during the search conducted for this review is small. We sent emails to the authors of included clinical trials to ask if they were aware of other similar clinical trials, but we received no response.
The small number of trials included in analyses reduced the robustness of these findings, introducing imprecision to the estimates. In our review process, we attempted to control for bias by strictly adhering to Cochrane methods (Higgins 2011). We applied no language restrictions, and we updated searches several times using multiple databases and handsearching. By searching clinical trial registries (e.g. clinicaltrials.gov), we enhanced the opportunity to identify unpublished trials as well as selective reporting of outcomes.
We used a standardised procedure to determine selection and inclusion of studies in this review, and review authors were trained in data extraction.
Agreements and disagreements with other studies or reviews
Some studies support a potentially harmful effect when corticosteroids are used to treat infectious disease (Aljeab 2016). However, some systematic reviews of interventions show a beneficial effect of corticosteroids in such conditions. Results obtained with these interventions include a reduction in sequelae and mortality when used for child meningitis (Brouwer 2015). Researchers have examined effects of corticosteroids on neonatal meningitis (Ogunlese 2015), on sepsis (Annane 2009), and during acute sinusitis in children (Head 2016; Zalmanovici 2013). However, regarding septic arthritis in children, and even in adults, a small number of clinical trials have published research on this topic. Therefore, the main limitation of this systematic review is the lack of clinical trials; this insufficient evidence makes difficult the routine use of corticosteroids for treating children with septic arthritis.
The findings of this systematic review are in accordance with the findings of another recent review (Farrow 2015). The Farrow review identified the same clinical trials included here and also included studies based on different designs and studies conducted in animal models. However, only one author conducted the former systematic review, and that review author limited the search to articles published in the English language.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Comparison 1 Corticosteroids versus placebo, Outcome 1 Pain ‐ absence of pain.
Comparison 1 Corticosteroids versus placebo, Outcome 2 Pain ‐ number of days until pain free.
Comparison 1 Corticosteroids versus placebo, Outcome 3 Number of participants with normal physical joint function ‐ number of days until normal function of the joint.
Comparison 1 Corticosteroids versus placebo, Outcome 4 Number of participants with normal physical joint function ‐ number of days until full range of movement.
Comparison 1 Corticosteroids versus placebo, Outcome 5 Number of participants with normal physical joint function ‐ normal function at the end of treatment (short term).
Comparison 1 Corticosteroids versus placebo, Outcome 6 Number of participants with normal physical joint function ‐ normal function at 6 months of follow‐up (medium term).
Comparison 1 Corticosteroids versus placebo, Outcome 7 Number of participants with normal physical joint function ‐ normal function at 12 months of follow‐up (long term).
Comparison 1 Corticosteroids versus placebo, Outcome 8 Number of participants with normal physical joint function ‐ number of days until resolution of symptoms.
Comparison 1 Corticosteroids versus placebo, Outcome 9 Number of days of antibiotic treatment ‐ number of days of intravenous antibiotic treatment.
Comparison 1 Corticosteroids versus placebo, Outcome 10 Number of days of antibiotic treatment ‐ number of total days of antibiotic treatment.
| Corticosteroids compared to placebo for septic arthritis in children | ||||||
| Patient or population: septic arthritis in children taking antibiotics | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | No. of participants | Certainty of the evidence | Comments | |
| Risk with placebo | Risk with corticosteroids | |||||
| Pain ‐ absence of pain at 12 months | 72 per 100 | 96 per 100 | RR 1.33 | 49 | ⊕⊕⊝⊝ | Absolute risk difference 24% (95% CI 5% to 43%) NNTB = 13 (95% CI 6 to 139) |
| Activities of daily living | See comments | ‐ | (0 RCTs) | ‐ | This outcome was not reported in any of the included trials | |
| Number of participants with normal physical joint function ‐ normal function at 12 months of follow‐up (long term) | 74 per 100 | 98 per 100 | RR 1.32 | 100 | ⊕⊕⊝⊝ | Absolute risk difference 24% (95% CI 11% to 37%) NNTB = 13 (95% CI 7 to 33) |
| Number of days of antibiotic treatment ‐ number of days of intravenous antibiotic treatment | Number of days of intravenous antibiotic treatment with placebo was 10.86 | Number of days of intravenous antibiotic treatment with corticosteroids was 8.09 | MD 2.77 lower | 149 | ⊕⊕⊝⊝ | Almost 3 days lower (95% CI 4 days to 1.5 days lower) |
| Length of hospital stay | See comments | ‐ | (0 RCTs) | ‐ | This outcome was not reported in any of the included trials. Study authors report that treatment with dexamethasone was associated with a shorter duration of IV antibiotic treatment, leading to a shorter hospital stay | |
| Total adverse events at 12 months | See comments | ‐ | 49 | ⊕⊕⊝⊝ | Trial reported that none of the participants showed adverse effects | |
| Serious adverse events at 12 months | See comments | ‐ | 49 | ⊕⊕⊝⊝ | Trial reported that none of the participants showed serious adverse effects | |
| *The risk in the intervention group (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). | ||||||
| GRADE Working Group grades of evidence. | ||||||
| aWe downgraded by one level for concerns about study limitations (high risk of attrition bias and selective reporting). We downgraded by another level for imprecision. bWe downgraded by one level for concerns about study limitations (unclear risk of detection and reporting bias and high risk of attrition bias). We downgraded by another level for imprecision. cWe downgraded by one level for concerns about study limitations (detection, reporting, and attrition bias). We downgraded by another level for imprecision. | ||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Pain ‐ absence of pain Show forest plot | 1 | 49 | Risk Ratio (M‐H, Random, 95% CI) | 1.33 [1.03, 1.72] |
| 2 Pain ‐ number of days until pain free Show forest plot | 1 | 49 | Mean Difference (IV, Random, 95% CI) | ‐3.58 [‐7.41, 0.25] |
| 3 Number of participants with normal physical joint function ‐ number of days until normal function of the joint Show forest plot | 1 | 49 | Mean Difference (IV, Random, 95% CI) | ‐2.07 [‐5.47, 1.33] |
| 4 Number of participants with normal physical joint function ‐ number of days until full range of movement Show forest plot | 1 | 49 | Mean Difference (IV, Random, 95% CI) | ‐5.24 [‐10.37, ‐0.11] |
| 5 Number of participants with normal physical joint function ‐ normal function at the end of treatment (short term) Show forest plot | 1 | 100 | Risk Ratio (M‐H, Random, 95% CI) | 1.41 [1.16, 1.72] |
| 6 Number of participants with normal physical joint function ‐ normal function at 6 months of follow‐up (medium term) Show forest plot | 1 | 100 | Risk Ratio (M‐H, Random, 95% CI) | 1.58 [1.27, 1.97] |
| 7 Number of participants with normal physical joint function ‐ normal function at 12 months of follow‐up (long term) Show forest plot | 1 | 100 | Risk Ratio (M‐H, Random, 95% CI) | 1.32 [1.12, 1.57] |
| 8 Number of participants with normal physical joint function ‐ number of days until resolution of symptoms Show forest plot | 1 | 100 | Mean Difference (IV, Random, 95% CI) | ‐5.47 [‐6.98, ‐3.96] |
| 9 Number of days of antibiotic treatment ‐ number of days of intravenous antibiotic treatment Show forest plot | 2 | 149 | Mean Difference (IV, Random, 95% CI) | ‐2.77 [‐4.16, ‐1.39] |
| 10 Number of days of antibiotic treatment ‐ number of total days of antibiotic treatment Show forest plot | 2 | 149 | Mean Difference (IV, Random, 95% CI) | ‐3.11 [‐6.94, 0.73] |
