Dose-response meta-analysis (DRMA) is one of the branches of meta-analysis, which has provided important evidence for clinical research. Since introducing into China, it has gained great attention. In order to improve the reporting quality of DRMA, Dr. Chang Xu et al. developed the reporting guideline for DAMA——G-Dose Checklist. It was published in Chinese Jouranl of Evidence-based Medicine in 2016. This paper interprets the checklists so as to promote the understanding and use of it.
ObjectiveTo further evaluate the relation between usage of proton pump inhibitor (PPI) and the risk of pancreatic cancer. MethodThe observational studies were systematically searched in the databases of PubMed, Embase, Web of Science, Cochrane Library, ClinicalTrials.gov, CNKI, Wanfang, and VIP. The combined odds ratio (OR) and 95% confidence interval (CI) of pancreatic cancer risk were estimated by the corresponding effect model according to the heterogeneous results, and the subgroup analysis, meta-regression, and sensitivity analysis were performed. In addition, the relation between the defined daily dose (DDD) and usage time of PPI and the pancreatic cancer risk were studied by using restricted cubic spline. ResultsA total of 14 studies were included, including 1 601 430 subjects. The meta-analysis result showed that usage of PPI was positively correlated with the risk of pancreatic cancer [I2=98.9%, OR (95%CI)=1.60 (1.21, 2.11), P<0.001]. The subgroup analysis results showed that usage of PPI would increase the risk of pancreatic cancer in the subgroups of literature published before 2018 [OR (95%CI)=1.88 (1.05, 3.38), P=0.034], non-Asian regions [OR (95%CI)=1.37 (1.04, 1.82), P=0.028], case-control studies [OR (95%CI)=1.59 (1.16, 2.18), P=0.004], cohort studies [OR (95%CI)=1.65 (1.13, 2.39), P=0.009], and high-quality studies [OR (95%CI)=1.62 (1.19, 2.20), P=0.002]. The dose-response curve showed that there was a nonlinear relation between the usage of PPI and the risk of pancreatic cancer (χ2linear=2.27, P=0.132; Pnonlinear=0.039). When the usage of PPI was 800 DDD or less, usage of PPI would increase the risk of pancreatic cancer, but there was no statistical significance when the usage of PPI was more than 800 DDD. The time-effect curve showed that there was a linear relation between the usage time of PPI and the risk of pancreatic cancer (χ2linear=6.92, P=0.009), and the risk of pancreatic cancer would increase by 2.3% if the usage of PPI increased by one month [OR=1.02, 95%CI (1.01, 1.04), P=0.009]. The sensitivity analysis confirmed that the results were stable by gradually eliminating each study, the OR (95%CI) of the risk of pancreatic cancer was 1.37 (1.08, 1.74) to 1.66 (1.22, 2.27), and the publication bias was not found by Egger test (P=0.594).ConclusionsFrom the results of this meta-analysis, usage of PPI will increase the risk of pancreatic cancer, and the dosage of PPI and usage time of PPI may be related to the risk of pancreatic cancer. The clinical usage of PPI should be strictly controlled, and the dosage and usage time should also be carefully considered.