PCSK9 and HMGCR genetic variants are associated to risk of type 2 diabetes

Background

Low LDL-c levels either due to genetic variants in the 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGCR) gene or to statin treatment, are associated with the risk of coronary heart disease (CHD) but also with modest hyperglycaemia, increased body weight and a modest increased risk of type 2 diabetes (T2DM)1-3. To investigate whether low LDL-c due to low levels of PCSK9 also exert this effect, 2 mendelian studies have been performed in which genetic variants of PCSK9, but also of HMGCR, were related to the incidence of diabetes and cardiovascular events.

This relates to a study of Ference et al and of Schmidt et al that very recently appeared. Ference used 112,772 participants of which 14,120 and 10,635 experienced a cardiovascular event or developed diabetes, respectively. In this study, genetic scores were generated for PCSK9 and HMGCR. Scores were dichotomized and used to divide participants into 2 groups of approximately equal sizes on the basis of whether the genetic score was above or below the median. The primary composite endpoint was first occurrence of myocardial infarction (MI) or death from CHD and the primary safety outcome was diabetes.

In the study of Schmidt, 4 variants of PCSK9 (rs11591147, rs2479409, rs11206510 and rs11583680) were selected based on several criteria. Participating studies executed a common analysis script on their own data and submitted this to a central analysis centre. Subsequently, data were meta-analysed. The composite endpoint was prevalent or incident T2DM.

Main results

In the study of Ference and colleagues, 7 variants of PCSK9 and 6 variants of HMGCR were included.

• Individuals in the group with higher PCSK9 genetic scores had lower mean LDL-c levels than those with lower PCSK9 scores (difference of -4.2 mg/dL/-0.11 mmol/L, P=5.6*10-16).
• Similarly for non-HDL (difference of -4.5 mg/dL/-0.12 mmol/L, P=1.8*10-16), triglycerides (difference of -5.3 mg/dL/-0.06 mmol/L, P=6.8*10-10) and HDL-c (difference of 0.5 mg/dL/0.01 mmol/L, P=5.4*10-5).
• Participants in the group with higher PCSK9 scores had an 8.4% lower risk of primary endpoint, as well as lower risks of major coronary events, major vascular events, MI and death from CHD. They also had a 6.1% higher risk of diabetes.
• Individuals in the group with higher HMGCR genetic scores had lower mean LDL-c levels than those with lower HMGCR scores (difference 3.2 mg/dL/0.08 mmol/L, P=2.9*10-15).
• Participants in the group with higher HMGCR scores had and 6.6% lower risk of MI or death from CHD, which was consistent for all secondary endpoints. They also had a 5% higher risk of diabetes.
• PCSK9 and HMGCR genetic scores had additive effects on LDL-c levels and corresponding risk of cardiovascular events.
• A genetic score consisting of variants in the LDL receptor gene had a very similar effect on the risk of diabetes per unit decrease in the LDL receptor levels.

Combining 50 studies including 245,942 individuals and data from repositories including another 322,506 individuals, Schmidt and colleagues observed 51,623 cases of incident or prevalent T2DM.

• The 4 selected PCSK9 variants were associated with reductions in LDL-c; -0.02 mmol/L (95% CI: -0.03 to -0.02) for rs11583680, -0.04 (95% CI: -0.05 to -0.04) for rs2479409, -0.06 (95% CI: -0.07 to -0.05) for rs11206510 and -0.34 mmol/L (95% CI: -0.36 to -0.32) for rs11591147.
• 1 mmol/L lower LDL-c was associated with an increase in body weight of 1.03 kg (95% CI: 0.24 to 1.82), an increase of 0.006 (95% CI: 0.003 to 0.010) in waist-to-hip ratio but a potentially neutral association with BMI (0.11 kg/m2, 95% CI: -0.09 to 0.30).
• 1 mmol/L lower LDL-c was associated with 0.09 mmol/L (95% CI: 0.02 to 0.15) higher fasting plasma glucose, HbA1c of 0.03% (95% CI: -0.01 to 0.08) but no change in fasting insulin.
• Surprisingly, there was an effect on height with a mean difference of 0.008 m (95% CI: 0.0008 to 0.015).
• 1 mmol/L lower LDL-c was associated with an increased risk of T2DM (OR 1.29, 95% CI: 1.11-1.50). Separate analysis showed an OR of 1.15 (95% CI: 0.76-1.72) for incident T2DM and 1.26 (95% CI: 0.88-1.80) for prevalent T2DM.

Conclusion

References

Find the article of Ference and colleagues online at NEJMFind the article of Schmidt and colleagues online at Lancet Diabetes Endocrinol