복통

Background: Peppermint oil (PO) has intrinsic properties that may benefit patients with irritable bowel syndrome (IBS) symptoms. The study objective was to determine the effect of peppermint oil in the treatment of the IBS.

Methods: We systematically searched MEDLINE (PubMed), Cochrane Central Register of Controlled Trials (Cochrane CENTRAL), ClinicalTrials.gov, EMBASE (Ovid), and Web of Science for randomized controlled trials (RCTs) of PO for IBS. We appraised the eligible studies by the Cochrane risk of bias tool. We performed random-effects meta-analysis on primary outcomes including global improvement in IBS symptoms and abdominal pain. A PRISMA-compliant study protocol is registered in PROSPERO Register [2016, CRD42016050917].

Results: Twelve randomized trials with 835 patients were included. For global symptom improvement, the risk ratio (RR) from seven RCTs for the effect of PO (n = 253) versus placebo (n = 254) on global symptoms was 2.39 [95% confidence interval (CI): 1.93, 2.97], I2 = 0%, z = 7.93 (p < 0.00001). Regarding abdominal pain, the RR from six RCTs for the effect of PO (n = 278) versus placebo (n = 278) was 1.78 [95% CI: 1.43, 2.20], I2 = 0%, z = 5.23 (p < 0.00001). Overall, there were no differences in the reported adverse effects: PO (32 events, 344 total, 9.3%) versus placebo (20 events, 327 total, 6.1%) for eight RCTs; RR 1.40 [95% CI: 0.87, 2.26] I2 = 0%, z = 1.39 (p = 0.16). The number needed to treat with PO to prevent one patient from having persistent symptoms was three for global symptoms and four for abdominal pain.

Conclusions: In the most comprehensive meta-analysis to date, PO was shown to be a safe and effective therapy for pain and global symptoms in adults with IBS.

Electronic supplementary material: The online version of this article (10.1186/s12906-018-2409-0) contains supplementary material, which is available to authorized users.

Study selection: A literature search conducted from inception to April 11, 2018, identified 759 studies. After duplicates were removed, a total of 427 studies remained for a review of titles and abstracts, from which 22 trials were identified that underwent full text screening. A total of ten studies were excluded (Additional file 2: Table of Excluded Studies), and twelve randomized studies (835 patients) that met the inclusion criteria were identified and underwent systematic review and data synthesis. A flow diagram of the study selection is summarized in Fig. 1.

Study characteristics: The included studies were published over five decades, from 1979 to 2016. Patients from Asia, Europe, and North America were recruited. Studies were of varied sample sizes, from 18 to 178. The settings of most trials were teaching hospitals. Most studies were double-blind parallel group RCTs with follow-up durations ranging from 3 weeks to 12 weeks. Cross-over design were observed in three studies. Table 2 summarizes the characteristics of the included studies. Of note, among the twelve included studies, there was significant variation in the use of validated tools for the diagnosis of IBS. Alam et al., Cappello et al., Capanni et al., and Merat et al. utilized the Rome II criteria. Cash et al. used the Rome III criteria and found a statistically significant benefit for PO relative to placebos for the global improvement of IBS symptoms. Dew et al. conducted a double-blind cross-over study with a washout period defined by the recurrence of active symptoms, however, they failed to utilize any validated inclusion criteria. Lech et al. failed to utilize any validated inclusion criteria, though they did find a significant benefit for PO relative to placebo for improvements in abdominal pain. Liu et al. failed to utilize any validated inclusion criteria, though they did find a significant benefit for PO relative to placebo for the improvement of abdominal pain. Rees et al., Schneider et al., Weiss et al., and Carling et al. also failed to mention the use of validated inclusion criteria.

Risk of bias assessment: Incomplete outcome data was the most concerning problem observed in the included studies (Fig. 2). Six out of 12 studies were assessed as having high risk of attrition bias as these studies had over 10% loss to follow-up and dealt with missing data by excluding those patients from final analyses. Two studies were funded by industry and were assessed as having high risk of bias due to conflicts of interest. In addition, random sequence generation and allocation concealment were not reported (unclear risk of selection bias) in many studies. In contrast, the blinding of participants and personnel were well performed in all studies (low risk of performance bias). The selective reporting was not observed in any studies (low risk of reporting bias).

Meta-analysis: Seven studies reported treatment outcomes for the global improvement of IBS symptoms (Fig. 3a). The risk ratio (RR) for seven RCTs for the effect of PO (n = 253) versus placebo (n = 254) on global symptoms was 2.39 [95% confidence interval (CI): 1.93, 2.97], I2 = 0%, z = 7.93 (p < 0.00001) (Fig. 3a). The number of patients needed to be treated with peppermint oil versus the placebo to induce a global improvement of IBS symptoms was three (Table 3). No statistically significant heterogeneity was detected for this comparison (τ2 = 0.00, X2 = 5.29, P = 0.51, I2 = 0%) (Table 3).

Six studies reported treatment outcomes of improvement of abdominal pain (Fig. 3b). The RR for six RCTs for the effect of PO (n = 278) versus placebo (n = 278) on abdominal pain was 1.78 [95% CI: 1.43, 2.20], I2 = 0%, z = 5.23 (p < 0.00001) (Fig. 2b). The number of patients that needed to be treated with peppermint oil versus the placebo to improve abdominal pain was four (Table 3). No statistically significant heterogeneity was detected for this comparison (τ2 = 0.00, X2 = 3.01, P = 0.70, I2 = 0%) (Table 3).

Of the studies included, overall, only a few adverse events were reported. One study reported no adverse events. Most of the reported adverse events were mild and transient. The adverse events included: heartburn, dry mouth, belching, peppermint taste, rash, dizziness, headache, and less frequent increased appetite. There was no statistically significant difference in reported adverse effects in IBS subjects using PO (32 events, 344 total, 9.3%) versus placebo (20 events, 327 total, 6.1%); RR 1.40 [95% CI: 0.87, 2.26], I2 = 0%, z = 1.39 (p = 0.16) (Fig. 3c).

The planned funnel plots and Egger’s test for publication bias was not applicable as the number of studies for each outcome is less than ten because test power is usually too low to distinguish chance from real asymmetry. The sensitivity analyses excluding studies with high attrition bias provided similar results to our primary analyses.

The GRADE criteria were used to assess the overall quality of the evidence reported (Table 3). The global improvement in IBS symptoms outcomes was upgraded to high because of the large magnitude of the effect. The outcome of improvement of abdominal pain was downgraded to moderate because of the risk of bias. The outcome “adverse events” was downgraded to “low quality” because of the risk of bias and imprecision (Table 3).