Low Lp(a): a FOURIER analysis offers reassurance on several risks but finds a link with diabetes

A safety analysis in patients with existing disease

Lp(a) is considered a causal factor in atherosclerosis. As treatments designed to lower it undergo clinical evaluation, researchers examined whether low concentrations were associated with certain adverse events.

The FOURIER trial randomly assigned 27,564 patients with stable atherosclerotic cardiovascular disease to evolocumab or placebo, in addition to statin therapy. For this analysis, a baseline Lp(a) measurement was available for 25,090 participants.

The important distinction: treatment was randomly assigned, but baseline Lp(a) concentration was not. Links between this concentration and the events observed therefore remain associations, despite statistical adjustments.

No association detected with several adverse events

The researchers found no association between Lp(a) concentration and the occurrence of hemorrhagic stroke, serious bleeding, neurocognitive events, cancer or atrial fibrillation. This finding also applied to participants whose Lp(a) was at or below 13 nmol/L.

These results are reassuring for the events studied in this population. They demonstrate neither a complete absence of risk nor the long-term safety of all treatments capable of lowering Lp(a).

The diabetes signal does not prove a treatment effect

Lower baseline Lp(a) was associated with both existing diabetes and the development of diabetes during follow-up. For each 50 nmol/L lower baseline concentration, the adjusted hazard ratio for new-onset diabetes was 1.05, with a 95% confidence interval of 1.02 to 1.08.

This is a relative association, not a five-percentage-point increase in absolute risk. The abstract does not provide the case counts needed to express this difference in terms of absolute risk.

Meanwhile, evolocumab did not increase diabetes risk, regardless of baseline Lp(a). A naturally low concentration and a reduction achieved through treatment therefore cannot be considered equivalent on the basis of these findings.

What we can take away

This analysis does not justify considering low Lp(a) inherently dangerous. The observed link with diabetes needs to be distinguished from treatment effects: no increase in diabetes was detected with evolocumab.