Associations between lipid-lowering drugs and pregnancy and perinatal outcomes: a Mendelian randomization study

Abstract

Introduction: Mounting evidence has indicated that maternal dyslipidemia is associated with adverse obstetric outcomes, and the actions of lipid-lowering drugs in pregnant women remain controversial. Hence, this study aimed to appraise the causal relationship of lipid-lowering drugs [hydroxymethylglutaryl-coenzyme reductase (HMGCR) inhibitors, PCSK9 inhibitors, and NPC1L1 inhibitors] with pregnancy and perinatal outcomes using drug-targeting Mendelian randomization analysis. Methods: As a proxy for lipid-lowering drug exposure, two genetic instruments were used: genetic variants within or near the gene linked to low-density lipoprotein cholesterol (LDL-C) and the expression of quantitative trait loci of the drug target gene. Effect estimates were calculated using the inverse variance weighting (IVW) method and summary data-based Mendelian randomization (SMR) method. Heterogeneity and pleiotropy were assessed by Mendelian randomization–Egger regression, the Cochran Q test, and MR-PRESSO analysis. Results: HMGCR inhibitors were ascribed to a reduced risk of preeclampsia in both the IVW-MR method [odds ratio (OR) 0.583; 95% confidence interval (CI) 0.418–0.812; P = 0.001] and SMR analysis (OR 0.816; 95% CI 0.675–0.986; P = 0.036). The causal link between HMGCR inhibitors and offspring birthweight was statistically significant only in the analysis using the IVW method (OR, 0.879; 95% CI, 0.788–0.980; P = 0.020), and the combined results of the OR values supported the potential inhibitory effect of HMGCR inhibitors on offspring birthweight. Causal associations between lipid-lowering drugs and gestational diabetes, preterm birth, and congenital anomalies were not detected in either analysis. Conclusion: No causal associations were observed between lipid-lowering drugs and gestational diabetes, preterm birth or congenital anomalies, whereas genetically predicted HMGCR inhibition dramatically reduced the risk of preeclampsia but attenuated offspring birthweight.

Accuracy in drug target Mendelian randomization of maternal and foetal health

“…We utilized identical methodologies for single nucleotide polymorphisms (SNPs) selection as described in a previously published study that identified gene targets for lipid-lowering drugs. We conducted LD pruning by using the clump_data() function implemented in TwoSampleMR (https://github.com/MRCIEU/TwoSam-pleMR) R package [1,5,6]. The SNPs employed as instrumental variables in our study were consistent with those used in other studies (Huang et al, Zhao et al, Qin et al), thus mitigating the risk of reducing the standard error of the observed association or narrowing the confidence interval beyond the true value [1,3,4].…”

“…First, in this study, we referred to the published literature and set the threshold of linkage disequilibrium (r 2 < 0.3), in order to avoid the influence of strong linkage disequilibrium on the results [1][2][3][4]. We utilized identical methodologies for single nucleotide polymorphisms (SNPs) selection as described in a previously published study that identified gene targets for lipid-lowering drugs.…”

“…We conducted LD pruning by using the clump_data() function implemented in TwoSampleMR (https://github.com/MRCIEU/TwoSam-pleMR) R package [1,5,6]. The SNPs employed as instrumental variables in our study were consistent with those used in other studies (Huang et al, Zhao et al, Qin et al), thus mitigating the risk of reducing the standard error of the observed association or narrowing the confidence interval beyond the true value [1,3,4]. The code could also be available in the Code availability.…”

“…e have concerns regarding the accuracy of results and conclusions in 'Associations between lipidlowering drugs and pregnancy and perinatal outcomes: a Mendelian randomization study' [1]. This study used Mendelian randomization to explore effects of inhibiting targets of lipid-lowering medications [3hydroxy-3-methylglutaryl-CoA reductase (HMGCR), proprotein convertase subtilisin/kexin type 9 (PCSK9) and Niemann-Pick C1-Like 1 (NPC1L1)] on pregnancy outcomes.…”