The Diabetes Risk Nobody Argues Should Stop the Statin, and Why That's Not the Same as Ignoring It
A woman with elevated cardiovascular risk and pre-diabetic features weighs a real, dose-dependent statin-associated diabetes signal against a cardiovascular benefit both physicians agree clearly outweighs it — and debates what that agreement should still change about how she's monitored.
Sandra O., a 57-year-old hospital billing coordinator, came in for a cholesterol recheck she'd been putting off since her last physical flagged both her LDL and her fasting glucose as 'borderline, recheck in six months' — a phrase that, six months later, had quietly become two separate borderline results instead of one resolved one. Her mother had a stroke at 66 and has been on dialysis for the past four years, a fact Sandra brought up unprompted twice during the visit, each time followed by a version of the same question: is this going to be me?
Her numbers today make the cardiovascular case straightforward on their own: LDL 168mg/dL, a calculated 10-year ASCVD risk of 8.2%, and a blood pressure that's crept up to 138/86 over the past two years. What complicates the conversation isn't whether to treat that risk — it's that her fasting glucose, at 108mg/dL, sits inside the impaired-fasting-glucose range, and her BMI of 31 places her closer to the population in which statin therapy's own dose-dependent diabetes signal was first characterized than the average patient starting one. Both facts are true about her at once: she is a strong candidate for statin therapy, and she is also a patient in whom that therapy's best-known non-lipid side effect is more likely to show up than average.
The mechanism behind that signal is worth naming rather than treating as an unexplained statistical quirk: statins appear to modestly impair pancreatic beta-cell insulin secretion and reduce GLUT4-mediated glucose uptake in skeletal muscle, effects downstream of the same HMG-CoA reductase inhibition that lowers LDL, which is part of why the risk scales with dose and intensity rather than appearing as an all-or-nothing side effect. For a patient already sitting at the edge of impaired fasting glucose, that mechanism isn't hypothetical background — it's a real, if modest, additional push in a direction she's already drifting.
In clinic, two borderline results and one decision
I want this named plainly before we start her: statins carry a real, dose-dependent increase in incident diabetes, though the magnitude is routinely overstated. JUPITER, the trial that first raised the signal, reported a 26% relative increase in physician-reported diabetes; the pooled picture is much smaller — Sattar and colleagues' 2010 collaborative meta-analysis of 13 trials and 91,140 participants found a 9% relative increase, and Preiss's 2011 analysis of intensive versus moderate dosing found 12%. What matters for her is where that small excess concentrates: she's sitting with a fasting glucose of 108 and a BMI of 31.
None of that is an argument against starting the statin. It's an argument for saying it out loud, checking an A1c before we start rather than after, and not being surprised if her glucose drifts a little once we do.
Agreed on all of that, and I'd add the part of the same trials that gets underweighted in this conversation: the Cholesterol Treatment Trialists' 2024 individual-participant analysis found the excess is real and dose-dependent but lands overwhelmingly in people whose glycemic markers already sit near the diagnostic threshold — and that any cardiovascular risk those small glucose shifts might carry is already counted inside the event reduction the same trials show. Sattar's own arithmetic put it at roughly one extra case of diabetes per 255 patients treated for four years. Her 10-year risk is 8.2%, above the borderline band this discipline usually debates — this isn't a marginal call.
I'm not disputing the net benefit calculus — I'd be making the same recommendation in her position. What I'd add is that 'the trials show net benefit' and 'this specific patient's glucose trajectory needs watching' aren't in tension with each other; they're two true things that both belong in the same visit.
A statin-induced glucose rise, when it happens, is generally modest and doesn't reliably reverse on discontinuation the way some patients assume it will, which is exactly why catching it early with a follow-up A1c matters more than treating this as a one-time consent conversation.
Agreed, without real friction: start atorvastatin 40mg, check a baseline A1c today, and recheck at six months rather than waiting for her next annual physical to catch a glycemic shift late. Both physicians told Sandra directly that her diabetes risk was real and worth watching, while being equally direct that it did not change the recommendation to treat her cholesterol now.
The one open item, left for the six-month recheck rather than decided today: if her A1c does cross into the diabetic range, whether that should prompt a statin dose reduction or a switch to a lower-intensity agent, versus treating the new diagnosis on its own separate track while keeping her statin dose exactly where it is. Both physicians expect to lean toward the latter, but neither wanted to commit to it before there's an actual number to react to.