Interventions for treating people with symptoms of bladder pain syndrome: a network meta-analysis

Abstract

Background: Bladder pain syndrome (BPS), which includes the condition of interstitial cystitis, is a poorly understood clinical condition for which patients present with varying symptoms. Management of BPS is challenging for both patients and practitioners. At present, there is no universally accepted diagnosis and diverse causes have been proposed. This is reflected in wide-ranging treatment options, used alone or in combination, with limited evidence. A network meta-analysis (NMA) simultaneously comparing multiple treatments may help to determine the best treatment options for patients with BPS. Objectives: To conduct a network meta-analysis to assess the effects of interventions for treating people with symptoms of bladder pain syndrome (BPS). Search methods: We searched the Cochrane Incontinence Specialised Register, which contains trials identified from the Cochrane Central Register of Controlled Trials (CENTRAL, in the Cochrane Library), MEDLINE, MEDLINE In-Process, MEDLINE Epub Ahead of Print, ClinicalTrials.gov, the World Health Organization International Clinical Trials Registry Platform (WHO ICTRP) and handsearched journals and conference proceedings (searched 11 May 2018) and the reference lists of relevant articles. We conducted a further search on 5 June 2019, which yielded four small studies that were screened for eligibility but were not incorporated into the review. Selection criteria: We included randomised controlled trials (RCTs) and quasi-RCTs of interventions for treating adults with BPS. All types of interventions (including conservative, pharmacological and surgical) were eligible. Data collection and analysis: We assessed the risk of bias of included studies using Cochrane's 'Risk of bias' tool. Primary outcomes were the number of people cured or improved, pain, frequency and nocturia. For each outcome, random-effects NMA models were fitted using WinBUGS 1.4. We monitored median odds ratios (ORs) for binary outcomes and mean differences (MDs) for continuous outcomes with 95% credible intervals (Crls). We compared results of the NMA with direct evidence from pairwise meta-analysis of head-to-head trials. We used the CINeMA tool to assess the certainty of evidence for selected treatment categories. Main results: We included 81 RCTs involving 4674 people with a median of 38 participants (range 10 to 369) per RCT. Most trials compared treatment against control; few trials compared two active treatments. There were 65 different active treatments, and some comparisons were informed by direct evidence from only one trial. To simplify, treatments were grouped into 31 treatment categories by mode of action. Most studies were judged to have unclear or high risk of bias for most domains, particularly for selection and detection bias. Overall, the NMA suggested that six (proportion cured/improved), one (pain), one (frequency) and zero (nocturia) treatment categories were effective compared with control, but there was great uncertainty around estimates of effect. Due to the large number of intervention comparisons in this review, we focus on three interventions: antidepressants, pentosan polysulfate (PPS) and neuromuscular blockade. We selected these interventions on the basis that they are given 'strong recommendations' in the EAU Guidelines for management of BPS (EAU Guidelines 2019). We found very low-certainty evidence suggesting that antidepressants were associated with greater likelihood of cure or improvement compared with control (OR 5.91, 95% CrI 1.12 to 37.56), but it was uncertain whether they reduced pain (MD -1.27, 95% CrI -3.25 to 0.71; low-certainty evidence), daytime frequency (MD -2.41, 95% CrI -6.85 to 2.05; very low-certainty evidence) or nocturia (MD 0.01, 95% CrI -2.53 to 2.50; very low-certainty evidence). There was no evidence that PPS had improved cure/improvement rates (OR 0.14, 95% CrI 0.40 to 3.35; very low-certainty evidence) or reduced pain (MD 0.42, 95% CrI -1.04 to 1.91; low-certainty evidence), frequency (MD -0.37, 95% CrI -5.00 to 3.44; very low-certainty evidence) or nocturia (MD -1.20, 95% CrI -3.62 to 1.28; very low-certainty evidence). There was evidence that neuromuscular blockade resulted in greater cure or improvement (OR 5.80, 95% CrI 2.08 to 18.30) but no evidence that it improved pain (MD -0.33, 95% CrI -1.71 to 1.03), frequency (MD -0.91, 95% CrI -3.24, 1.29) or nocturia (MD -0.04, 95% CrI -1.35 to 1.27). The certainty of this evidence was always very low. Authors' conclusions: We are uncertain whether some treatments may be effective in treating patients with BPS because the certainty of evidence was generally low or very low. Data were available for a relatively large number of trials, but most had small sample sizes and effects of treatments often could not be estimated with precision. An NMA was successfully conducted, but limited numbers of small trials for each treatment category hampered our ability to fully exploit the advantages of this analysis. Larger, more focused trials are needed to improve the current evidence base.