A Randomized Placebo-Controlled N-of-1 Trial: The Effect of Proton Pump Inhibitor in the Management of Gastroesophageal Reflux Disease

Abstract

Background. Gastroesophageal reflux disease (GERD) is the most frequent chronic gastrointestinal disorder. It is defined as a condition developed when the reflux of gastric contents causes troublesome symptoms (heartburn and regurgitation). This requires adequate treatment since it can lead to long-term complications including esophagus adenocarcinoma. Proton pump inhibitors (PPI) are generally used to treat GERD due to their high-security profile and efficiency on most patients. However, recurrent reflux despite initial treatment is frequent. N-of-1 trial is a study that allows the identification of the best treatment for each patient. The objective of this study is to compare the efficacy of standard dose with double dosage of esomeprazole, to improve the GERD symptoms in a single patient. Methods. A single-patient trial, placebo-controlled, randomized, double-blind, was carried out from September 25th, 2012, to April 26th, 2013. It included one outpatient at the gastroenterology service in a fourth-level hospital, diagnosed with nonerosive reflux disease (NERD). Yet, his symptoms were heartburn and reflux, and his endoscopic results were normal esophageal mucosa, without hiatal hernia, though pathological pH values. A no-obese male without any tobacco or alcohol usage received esomeprazole 40 mg/day and 40 mg/bid for 24 weeks. A standardized gastroesophageal reflux disease questionnaire (GerdQ) was used weekly to evaluate symptom frequency and severity. The consumption of 90% of the capsules was considered as an adequate treatment adherence. D’agostino–Pearson and Wilcoxon test were used to determine normal or nonnormal distribution and compare both treatments, respectively, both with a significant statistical difference of p<0.05. Results. The patient completed the study with 96% of adherence. The double dosage of esomeprazole did not improve the control of symptoms compared with the standard dosage. Mean symptomatic score was 9.5±0.5 and 10.2±0.6 for each treatment, respectively (p>0.05). Conclusion. There was no significant improvement in the patient GERD symptoms increasing the dose of oral esomeprazole during the 6 months of study. N-of-1 trials in chronic pathologies including GERD are recommended due to their potential value as systematic methods that evaluate therapies without strong scientific evidence.

Comparison of Bayesian Networks, G-estimation and linear models to estimate causal treatment effects in aggregated N-of-1 trials with carry-over effects

“…In addition to research on appropriate statistical models for the analysis of aggregate and individual n-of-1 trials, previous studies have investigated approaches to derive optimal designs regarding sample size and number of cycles [ 6 – 8 ]. For the aggregate statistical analysis of series of N-of-1 trials, popularly used methods include non-parametric methods like the Wilcoxon signed-rank test [ 9 , 10 ], two-sample mean tests [ 11 ], methods that allow for covariate adjustments like linear models [ 12 , 13 ], linear mixed models [ 14 ], and Bayesian approaches [ 15 , 16 ]. Also, autoregressive models to account for time dependencies have been proposed for the analysis [ 17 ].…”

Comparison of Bayesian Networks, G-estimation and linear models to estimate causal treatment effects in aggregated N-of-1 trials with carry-over effects

“…In addition to individual-level analyses for personalized treatment, series of N-of-1 trials can be analyzed jointly Stunnenberg et al [2018], or also combined with results from standard randomized controlled trials (RCTs) to obtain population-level estimates on the effectiveness of treatments with equal or superior efficiency compared to non-crossover RCTs Punja et al [2016], Blackston et al [2019]. For the aggregate statistical analysis of series of N-of-1 trials, popularly used methods include parameter free methods like the Wilcoxon signed-rank test Green et al [2004], Sierra-Arango et al [2019], two-sample mean tests Schmid CH, Duan N., the DEcIDE Methods Center N-of-1 Guidance Panel [2014], methods that allow for covariate adjustments like linear models Odineal et al [2019], Vrinten et al [2015], linear mixed models Herrett et al [2021] and Bayesian approaches Chen and Chen [2014], Samuel et al [2019]. Also, methods for time adjustments such as autoregressive models have been proposed for the analysis Zhou et al [2017].…”