Inadequate safety reporting in the publications of randomised clinical trials in irritable bowel syndrome: drug versus probiotic interventions

Abstract

Randomised controlled clinical trials (RCTs) offer a unique opportunity to obtain controlled efficacy and safety data to support clinical decisions. However, most RCT reporting has a stronger focus on efficacy rather than safety. This study aimed to identify the safety profile of both probiotic and drug interventions in irritable bowel syndrome (IBS). In connection to this paper, an accompanying paper was published in which a meta-analysis was conducted to evaluate the efficacy of probiotic interventions compared to that of drug interventions in IBS. Together, these two studies provide a first assessment regarding the feasibility to determine a burden to benefit ratio for both probiotic and drug interventions in IBS. RCTs including participants (>18 years old) with IBS and comparing probiotic or drugs interventions with control groups were identified by a systematic search of MEDLINE (January 2015 – Jan 2021). Reported safety profiles in drug studies were completer and more detailed as compared with studies on probiotics. Several inconsistencies in safety reporting were identified between and within drug and probiotic studies, such as: didn’t report on safety; only reported adverse reactions (ARs) or adverse events (AEs) with a certain severity; didn’t report the total number of AEs; didn’t split in the control- or experimental arm; didn’t specify AEs; and used different thresholds for ‘common’ AEs. Hence, it is difficult to compare safety data from drug and probiotic RCTs across and between different studies. On the current approaches to safety reporting, we could not establish an unambiguous safety profile for neither probiotic and drug interventions in IBS. These shortcomings hamper a critical comparison of the burden to benefit ratio for IBS intervention.

Comparing probiotic and drug interventions in irritable bowel syndrome: a meta-analysis of randomised controlled trials

“…Of all included studies 14 entailed probiotic-and 18 drug interventions. Supplementary Tables S2 and S3 of the companion paper in Beneficial Microbes ( Van der Geest et al, 2022) show the characteristics of all studies included in this meta-analysis.…”

“…Interestingly, some mechanisms might be widespread among commonly studied probiotic genera; but most often, a particular strain is needed to intervene with a specific pathophysiological process (Hill et al 2014). As shown in Van der Geest et al (2022: see Table 2), a wide variety of mono-or multi-strain probiotics are used in order to treat IBS symptoms. Since the distribution of mechanisms of probiotics can be very specific even up to strain level, and probiotics work via different pathophysiologic mechanisms, efficacy can be compared between types of probiotics, however, efficacy with exception of general effects such as…”

“…Although very little knowledge exists on doseeffect relationship for probiotics (Ouwehand, 2017), this study suggests that studying dose-effect relationship would be worthwhile. Table 2 in the companion paper in Beneficial Microbes ( Van der Geest et al, 2022) shows that several clinical trials included in this study used different dosages; hence the effect size might be influenced due to these differences.…”

“…As shown in Van der Geest et al (2022: see Table 2 and 3), there is considerable variation between and within the articles regarding the distribution of IBS subtypes. However, most articles do not report their outcomes per IBS subtype but only make the distinction between control-and experimental group.…”

“…To our knowledge, no clinical trials have compared the efficacy of drug and probiotic interventions in patients with IBS. To nonetheless harness acquired insights into the comparable efficacy of both treatments, we here determined the weighted average efficacy of both probiotic and drug interventions by conducting a meta-analysis, also we determined the safety profile for both probiotic-and drug interventions in a companion article (Van der Geest et al, 2022)..…”