Statin Use and Cancer Incidence in Patients with Type 2 Diabetes Mellitus: A Network Meta-Analysis

Abstract

Background: Type 2 diabetes mellitus (T2DM) patients are involved closely with cancer. This work aims to conduct a systematic review and network meta-analysis (NMA) to examine the effect of different types of statins on cancer incidence in patients with T2DM.

Methods: We systematically searched the Cochrane Library, PubMed, Embase, and Wanfang databases from January 1999 to March 2017. We performed a pairwise meta-analysis to estimate the pooled ratios (ORs) and 95% confidence intervals (CIs). A NMA was performed to compare different types of statins.

Results: Seven publications were included. In pairwise meta-analysis, the incidence of cancer in T2DM patients was reduced when simvastatin, atorvastatin, pravastatin, fluvastatin, lovastatin, rosuvastatin, and pitavastatin were used. In the result of NMA, the usage of simvastatin (RR 0.30 and 95% CI 0.16-0.56), atorvastatin (RR 0.29 and 95% CI 0.09-0.88), pravastatin (RR 0.34 and 95% CI 0.12-0.93), fluvastatin (RR 0.27 and 95% CI 0.09-0.83), rosuvastatin (RR 0.22 and 95% CI 0.10-0.49), and pitavastatin (RR 0.33 and 95% CI 0.20-0.57) was superior to the nonstatin groups. When compared with six other statins, rosuvastatin appeared to be the best one.

Conclusions: Different statins can reduce the risk of cancer in patients with T2DM. Our analyses suggest that rosuvastatin may be more effective than others.

Figure 1

Flow chart of studies selected for this meta-analysis.

Figure 2

Network of comparisons for the multiple-treatment meta-analysis for cancer occurrence in diabetic patients. The width of the solid lines presents proportional to the number of trials that compares each pairwise treatment, and the size of each node presents proportional to the number of participants.

Figure 3

Direct meta-analysis results for pairwise comparisons were represented by ORs with 95% CI. ORs which were higher than 1 favored the column-defining treatment and ORs which were lower than 1 supported horizontal treatment. Statistically significant differences between interventions are shown in bold, underlined font. OR: odds ratio; CI: credibility interval.

Figure 4

Consistency of direct evidence and indirect evidence was based on the value of ROR. The letters in the figure represented different treatment measures: a: simvastatin; b: atorvastatin; c: pravastatin; d: fluvastatin; e: lovastatin; f: rosuvastatin; g: pitavastatin; and h: nonstatin.

Figure 5

The pairwise comparison contributes to merged comparison. The letters in the figure represented different treatment measures: a: simvastatin; b: atorvastatin; c: pravastatin; d: fluvastatin; e: lovastatin; f: rosuvastatin; g: pitavastatin; and h: nonstatin.

Figure 6

A complete summary of estimates from the network meta-analysis for occurrence of cancer in diabetes patients with statins or not (rate ratios with 95% CI). Statistically significant differences between interventions are shown in bold, underlined font.

Figure 7

Distribution of probabilities of each intervention is ranked at each of the possible 8 positions. The horizontal axis is the ranking order, and the vertical axis is the probability of a certain order in a certain order.

Figure 8

Funnel plots of the meta-analysis. The ordinate (Y) presents the effect of each study, and abscissa (X) is the total effect on each control group. Scatter represents a study in the comparison of the two treatment measures. The letters in the figure represent different treatment measures: a: simvastatin; b: atorvastatin; c: pravastatin; d: fluvastatin; e: lovastatin; f: rosuvastatin; g: pitavastatin; and h: nonstatin.