Novartis's Modeling Flaw Threatens Cardiovascular Outcome Trials Across Industry

A recent failure by Novartis in its Lp(a) outcomes study has exposed a modeling flaw that now threatens Amgen and Lilly's cardiovascular outcome trials. According to biotech analyst Brian Skorney, when event rates in an outcomes trial are mismodeled, the study almost always fails.

Novartis had flagged that cardiovascular events in its Horizon study were accruing more slowly than expected. Skorney argued that this slow-accrual pattern is a red flag and has now hung over Amgen's olpasiran (data possibly 2027) and Lilly's lepodisiran (data possibly 2029), both of which reduce Lp(a) more deeply than pelacarsen's roughly 80% but rest on the same genetic hypothesis.

The Horizon trial enrolled roughly 8,000 patients and came up flat. The drug, pelacarsen, did exactly what it was designed to do: it crushed Lp(a) levels by about 80%. However, it did not prevent heart attacks, strokes, cardiovascular deaths, or urgent revascularizations.

The failure has forced a wholesale re-examination of how biotech investors model these trials. Skorney's heuristic holds that when event rates are mismodeled, the study usually fails.