Comparative efficacy and safety of treatments for secondary Raynaud's phenomenon: a systematic review and network meta-analysis of randomised trials

Several pharmacological treatments are available for secondary Raynaud's phenomenon, but there is uncertainty regarding the best options. We aimed to assess and compare the benefits and harms of treatments available for secondary Raynaud's phenomenon.We did a systematic review and network meta-analysis of randomised controlled trials (RCTs) of pharmacological treatments. We searched for systematic reviews published in MEDLINE and the Cochrane Database of Systematic Reviews up to Jan 31, 2017, and for RCTs published from inception to Sept 24, 2019 in MEDLINE, Embase, and ClinicalTrials.gov. We included double-blind RCTs (parallel or crossover) that compared two or more pharmacological treatments or placebo in patients with secondary Raynaud's phenomenon. Individual patient data were obtained for one unpublished RCT. Three researchers independently screened the texts and extracted the data. Efficacy outcomes included severity (on a ten-point scale), daily frequency, and mean duration of Raynaud's phenomenon attacks. We also examined tolerability and acceptability. Pairwise meta-analyses and Bayesian random-effects network meta-analyses were used to synthesise data. This study is registered with PROSPERO (CRD42017057518).We included 58 RCTs in the analysis, comprising 3867 patients (3540 [91·5%] with secondary Raynaud's phenomenon) and 15 classes of drugs. Phosphodiesterase 5 (PDE5) inhibitors were more effective than placebo for frequency (mean difference -0·36 [95% credibility interval -0·69 to -0·04]), severity (-0·34 [-0·66 to -0·03]), and duration (-3·42 [-6·62 to -0·29]) of attacks (low to moderate level of evidence). Calcium channel blockers (CCBs) were superior to placebo for frequency (-0·35 [-0·67 to -0·02]) and severity (-0·84 [-1·25 to -0·45]) of attacks (low level of evidence). For severity of attacks, selective serotonin-reuptake inhibitors (-1·54 [-2·68 to -0·41]; very low level of evidence) and oral prostacyclin receptor agonists (-0·48 [-0·80 to -0·16]; low level of evidence) were superior to placebo. No other drug classes were significantly superior to placebo with regard to efficacy outcomes. Compared with placebo, tolerability was lower for PDE5 inhibitors (incidence rate ratio for serious adverse events or early study exit due to adverse events 3·30 [95% CrI 1·49 to 7·55]) and CCBs (3·13 [1·33 to 7·04]). For all outcomes, global heterogeneity and between-study variance ranged from low (I2=0% and τ2=0·0 for attack severity and duration) to moderate (I2=41% and τ2=0·2 for tolerability). The overall risk of bias was judged to be low in 22 (38%), high in ten (17%), and unclear in 26 (45%) RCTs.PDE5 inhibitors and CCBs are the most effective pharmacological options, albeit with moderate efficacy and a low level of evidence. Current evidence does not support the use of any other drug in secondary Raynaud's phenomenon.None.