The Drug With the Best Liver Evidence and a Cardiac Asterisk
Pioglitazone reverses liver fibrosis better than anything else in the diabetes formulary. It also raises heart failure risk, and his echocardiogram just gave that risk a face.
Roland F., a 59-year-old man, restores old radios in a workshop behind his house, a habit he picked up from his father and has kept going for over thirty years despite, as he puts it, "the parts getting harder to find every decade." He has had type 2 diabetes for eight years, managed on metformin with an A1c of 7.6%, and a liver ultrasound ordered for persistently elevated transaminases led to an elastography study confirming metabolic dysfunction-associated steatohepatitis with moderate fibrosis — stage F2, not yet cirrhosis, but past the point where lifestyle change alone reliably reverses the disease. A recent echocardiogram, ordered after his primary care physician noted a soft S4 on exam, showed mild diastolic dysfunction with preserved ejection fraction — not heart failure, but an early structural finding that changes how the team reads what comes next.
Pioglitazone has more evidence behind it for actually reversing NASH fibrosis on repeat biopsy than any other diabetes medication, which is exactly why it's on the table for a patient whose liver disease has already progressed past the simple-steatosis stage. But that same drug carries two separate historical shadows worth reading carefully rather than reflexively: the PROactive trial's real, replicated signal for increased heart failure hospitalization, which lands directly on a patient who already has an echo finding in that direction, and an older bladder-cancer association that larger, more recent epidemiologic studies have substantially narrowed, though not entirely erased from the conversation.
A drug with the best liver evidence and a real cardiac question mark
I'd avoid pioglitazone here. PROactive showed a real, replicated increase in heart failure hospitalization with this drug, and that's not a population he's approaching from a neutral starting point — his echo already shows diastolic dysfunction. Newer agents now offer real liver benefit without this specific risk sitting on top of an existing structural finding.
I'd weigh the fibrosis data more heavily than that framing allows. Pioglitazone remains the diabetes medication with the strongest repeat-biopsy evidence for actual fibrosis regression, not just normalized enzymes — and he's already at stage F2, past the point where I'm comfortable just watching and hoping lifestyle change catches up.
The bladder-cancer signal you might also be thinking of has been substantially narrowed by larger, more recent epidemiologic data — it's not the strong concern it was made out to be a decade ago. And diastolic dysfunction without diagnosed heart failure is a reason to monitor him closely, not a reason to withhold the drug with the best liver evidence outright.
There's a way to get real liver benefit without opening the cardiac question at all, at least not yet. ESSENCE, the phase 3 semaglutide trial reported by Sanyal and colleagues, enrolled exactly his population — biopsy-confirmed MASH at fibrosis stage F2 or F3 — and found resolution of steatohepatitis without worsening fibrosis in 62.9% on semaglutide against 34.3% on placebo at 72 weeks, which is what carried the drug to FDA approval for this indication in 2025 alongside its established glycemic and weight effects — not as strong an evidence base as pioglitazone's on fibrosis specifically, but real, and without the heart-failure signal.
I'd start there, with a repeat elastography at six to twelve months. If his fibrosis hasn't meaningfully improved, pioglitazone is still on the table then, with a fresh echo to see whether his diastolic function has changed in the meantime either way.
Semaglutide started, with repeat liver elastography and echocardiogram both scheduled at six months. All three voices agreed this sequencing was reasonable, though for different underlying reasons — the cardiologist saw it as avoiding an unnecessary risk, the hepatologist as a reasonable trial before reaching for the stronger agent.
Not agreed: how much fibrosis improvement at six months would count as "enough" to avoid pioglitazone entirely. The hepatologist set a higher bar, wanting to see genuine stage regression before calling semaglutide sufficient; the endocrinologist would accept stable, non-progressive fibrosis as a reasonable outcome to continue watching rather than escalating. Left for the six-month visit to resolve with real data in hand.