Comment on de Carvalho et al. Proprotein Convertase Subtilisin/Kexin Type 9 (PCSK9) Inhibitors and Incident Type 2 Diabetes: A Systematic Review and Meta-analysis With Over 96,000 Patient-Years. Diabetes Care 2018;41:364–367

We read with interest the article by de Carvalho et al. (1). This meta-analysis reported the results from 68,123 individuals (20 randomized clinical trials) during 78 weeks of follow-up and evaluated the effects of proprotein convertase subtilisin/kexin type 9 inhibitors (PCSK9i) on glucose metabolism. They revealed that PCSK9i increased fasting blood glucose (FBG) and HbA_1c (P < 0.001, respectively) compared with placebo. Satisfactorily, PCSK9i had no impact on the incidence of new-onset diabetes but showed a positive association between diabetes risks and the potency (P = 0.029) and duration (P = 0.026) of PCSK9i therapy. The research appears informative clinically. However, we would like to add several comments on this article.

First, in their analysis they included SPIRE trials when calculating the change of FBG and HbA_1c posttreatment from baseline. In fact, unlike evolocumab and alirocumab, bococizumab is not a fully humanized monoclonal antibody; it contains ∼3% of the murine sequence in the antigen-binding complementarity-determining region (2). During bococizumab treatment, nearly half the patients developed antidrug antibodies, and 29% of patients had neutralizing antibodies that reduced the clinical efficacy, which never occurred in evolocumab and alirocumab therapy. Furthermore, no significant benefit was observed with respect to the primary end point. Thus, this PCSK9 monoclonal antibody was withdrawn from future use. Hence, the analysis covering the SPIRE trials might be improper for current reference to the clinical PCSK9i application. Importantly, we performed similar meta-analysis without SPIRE trials. The data suggested that PCSK9i had no effect on circulating FBG or HbA_1c levels and that treatment duration or percent change of LDL cholesterol did not influence diabetes risk. Although the analyses with or without SPIRE trials may be a small difference point, the conclusion is quite meaningful. Originally, there was a small increase in FBG and no change in plasma HbA_1c levels or new-onset diabetes after bococizumab treatment when compared with placebo. Thereby, the results of the meta-analysis by de Carvalho et al. (1) might be cautiously regarded due to the possibility of misunderstandings.

Second, new-onset diabetes is defined as patients without diabetes at study entry and initiation of antidiabetes medication, FPG ≥7.0 mmol/L, or HbA_1c ≥ 6.5% posttreatment. To investigate the risk of development of diabetes in patients given statins, Sattar et al. (3) and Preiss et al. (4) evaluated the development of diabetes in participants without diabetes at baseline. Unfortunately, de Carvalho et al. (1) covered patients with diabetes at baseline, which might result in serious bias during the meta-analysis according to our knowledge. Logically, the value of their analysis might be limited by this methodological imperfection.

Evidence suggested that statins moderately increased new-onset diabetes and HbA_1c (3,4). PCSK9i are the most powerful lipid-lowering drugs currently (5) and naturally aroused the same concerns. Therefore, we should be more prudent to evaluate the data to make a conclusion. Given the experience with statins, evidence from a larger number of participants and longer follow-up clinical trials concerning PCSK9i is required in the future.