Background
Etiology of polycystic ovary syndrome (PCOS) is attributed to genetic and environmental factors. One environmental factor is oxidative stress. Paraoxonase 1 (PON1) is an antioxidant high-density lipoprotein-associated enzyme encoded by the PON1 gene. The PON1 gene has been implicated in the risk for PCOS, the influence of which appears to come from single nucleotide variants (SNVs) at multiple genetic loci. However, association study reports have been inconsistent which compels a meta-analysis to obtain more precise estimates.
Methods
From 12 publications, extracted genotype data were used in two genetic procedures. First, linkage disequilibrium (LD) was used to group eight PON SNVs into three: LD1, LD2 and LD3. Second, frequencies of the variant (var), wild-type (wt) and heterozygous (het) genotypes were used for genetic modeling (allele-genotype for LD1 and standard for LD2 and LD3). Risk associations were expressed in terms of pooled odds ratios (ORs), 95% confidence intervals (CIs) and Pa-values. Evidence was considered strong when significance was high (Pa < 0.0001) and heterogeneity absent (I2 = 0%). Pooled effects were subjected to modifier (power), subgroup (Asian/Caucasian), outlier, sensitivity and publication bias treatments. Multiple comparisons were Bonferroni-corrected.
Results
This meta-analysis generated 11 significant outcomes, five in LD1, six in LD2 and none in LD3. All six LD2 outcomes did not survive the Bonferroni-correction but two of the five in LD1 did. These two core LD1 findings conferred greater odds of PCOS to the var allele in the highly significant (Pa < 0.0001) overall (OR 1.44, 95% CI 1.24–1.67) and Asian (OR 1.41, 95% CI 1.20–1.65) outcomes. Of these two core outcomes, the Asian effect was homogeneous (I2 = 0%) but not the overall (I2 = 29%).
Conclusions
Of the eight PON SNVs examined, two (rs854560 and rs662) were associated with PCOS risk. These 1.4-fold increased risk effects rendered Asians susceptible to PCOS. High statistical power, high significance, zero to low-level heterogeneity, robustness and lack of bias in the core outcomes underpinned the strong evidence for association.