SGLT2i or GLP-1 RA: Two Patients, Two Different Tipping Factors
Both patients face the same guideline-neutral fork between two cardioprotective drug classes. The disagreement isn't about which class is generally better — it's about which patient-specific detail should be allowed to override guideline neutrality, and the two cases pull in opposite directions.
Denise K., a 57-year-old woman, has spent the last two years training for and finishing half-marathons — a hobby she picked up after her husband's own heart attack scared them both into overhauling how they live, and one she's proud enough of that she keeps every race bib pinned to a corkboard in the garage. She has type 2 diabetes of nine years' duration and established coronary artery disease, having had a non-ST-elevation MI and a single drug-eluting stent placed two years ago; her ejection fraction is normal and her creatinine has never been abnormal. She's on metformin and a statin, with an A1c that has crept up to 8.1% despite consistent adherence, and her cardiologist and endocrinologist agree she needs a second agent with proven cardiovascular benefit — the guideline-favored move for a patient with her history. What she volunteers, unprompted, when the options are named: she is genuinely afraid of needles, has fainted twice getting blood drawn as an adult, and would strongly prefer something she can swallow.
What she doesn't volunteer, and what the chart actually shows on review, is three documented episodes of vulvovaginal candidiasis in the past year alone, each treated with a short antifungal course and each, by her own account at the time, "worse than the last." That detail changes the read on this decision more than her needle aversion does. SGLT2 inhibitors work by inducing glucosuria — deliberately spilling glucose into the urine to lower blood sugar — and that same glucose-rich urinary environment is the documented mechanism behind the class's elevated genital mycotic infection rate. She isn't a theoretical candidate for that side effect; she is, by her own recent history, already someone the drug's own mechanism would predictably make worse.
Either class is a defensible second-line choice for her — established ASCVD, no heart failure, no CKD, both SGLT2 inhibitors and GLP-1 receptor agonists carry proven cardiovascular benefit in that population, and the guidelines don't hand one a preference over the other here. Given that, I'd start with an SGLT2 inhibitor. It's one pill, once a day, and she's told us directly she'd rather not deal with an injection if she doesn't have to.
I'd weigh her own history more heavily than her stated preference. Genital mycotic infection is a real, mechanistically explained SGLT2i class effect — the induced glucosuria feeds exactly the kind of overgrowth she's already had three times this year, each one by her own report worse than the last. That's not a hypothetical side effect for her; it's a documented pattern we'd be knowingly reproducing.
The needle aversion is real, but it's a preference, and a once-weekly GLP-1 injection is a genuinely small ask compared to a recurring infection she's already told us is getting worse. And at a BMI of 38, she stands to gain meaningfully more weight benefit from a GLP-1 receptor agonist than from an SGLT2 inhibitor — that's not a tiebreaker here, it's a second reason pointing the same direction.
I'd still lean SGLT2 inhibitor, on the strength of the trial record specifically — EMPA-REG OUTCOME, CANVAS, and DECLARE-TIMI 58 all independently showed cardiovascular benefit for this class, while the GLP-1 RA record is more mixed: LEADER and SUSTAIN-6 were positive, but EXSCEL and ELIXA came back neutral. That consistency matters to me in a patient with her exact risk profile.
But I'll concede the point on her yeast infections directly — that's a real, already-demonstrated harm in front of us, not a trial-level abstraction, and three episodes in a year with each one worse than the last isn't something I'd knowingly walk her back into for a difference in trial consistency this narrow. If it comes down to her actual history versus the tidiness of the outcome-trial record, her history should win.
Semaglutide started, with counseling on injection technique to address her needle aversion directly rather than let it silently drive the drug choice. All three voices agreed this was the right call once her own infection history was weighed against the trial-consistency argument.
Harold M., a 61-year-old man, has driven a long-haul furniture delivery route for over two decades, a job built around gas-station meals and long stretches without a real bathroom break — habits he knows haven't done him any favors, and ones his wife has been after him to change since his own cardiologist put a stent in his right coronary artery eighteen months ago after a STEMI. His type 2 diabetes, diagnosed around the same time, has been managed on metformin with an A1c holding at 7.9%; his ejection fraction is normal, and his kidney function has stayed unremarkable on every check since the stent. Like Denise, he needs a second agent with proven cardiovascular benefit for established ASCVD without heart failure or CKD — the same guideline-neutral fork she faced. Unlike her, he has no needle aversion at all; he takes his own blood pressure readings with a home cuff without complaint and says an injection "doesn't bother him one way or another."
What does bother him, and what surfaces only when asked directly, is a pattern of early fullness and intermittent bloating after meals that he's had for years and always attributed to "just how his stomach is" — never formally worked up, never enough to stop him eating a full route-day meal, but present most days. A GLP-1 receptor agonist's dominant tolerability issue is exactly this territory: the drug slows gastric emptying by design, which is part of how it lowers glucose and suppresses appetite, and in a patient who already runs mildly delayed doesn't add a generic side-effect risk so much as compound a mechanism he's already living with. The two cases are not quite mirror images, though, and the difference is evidentiary rather than clinical: Denise’s tipping factor is a harm three times documented in her own chart, while his has never been worked up at all, which means the team is weighing a proven recurrence against a symptom nobody has yet confirmed the cause of.
I'd start with a GLP-1 receptor agonist here, same guideline-neutral logic as Denise's case — proven ASCVD benefit, no HF or CKD pulling toward either class specifically, and he has none of her needle hesitation to work around.
I'd go the other way for him specifically, and it's the mirror of Denise's argument. His years of early satiety and bloating aren't a random comorbidity sitting next to this decision — they're the exact symptom territory a GLP-1 RA's delayed-gastric-emptying mechanism would predictably worsen. He's not a patient with an elevated theoretical risk of GI intolerance; he's a patient already living with mild features of the thing the drug is most likely to intensify.
And unlike Denise, he genuinely has no meaningful preference standing in the way of the alternative — an SGLT2 inhibitor is a once-daily pill with no gastric-emptying effect at all, in a patient who's told us directly he doesn't mind either route.
Before this settles as a clean swap to SGLT2 inhibitor, his chart has a detail that changes the risk profile on that side too: he's on chronic furosemide for longstanding leg edema, not heart failure. SGLT2 inhibitors cause their own osmotic diuresis on top of that, and starting one on an already-diuresed patient raises real volume-depletion and acute kidney injury risk, especially given his irregular meal and fluid intake on the road.
That's not an argument for reversing course back to a GLP-1 RA — his GI history still argues against that class specifically. It's an argument for sequencing: reduce or hold the furosemide dose first, confirm his edema tolerates it, then start the SGLT2 inhibitor with a volume-status recheck at two weeks rather than starting both changes at once.
Furosemide reduced to 10mg daily, edema reassessed at one week and unchanged, empagliflozin started with a two-week renal-function and volume-status recheck scheduled. All three voices treated the diuretic detail as a real modification to the plan, not a reason to reverse it.
Not agreed: whether his undiagnosed early-satiety pattern deserves its own workup independent of this decision. The endocrinologist raised it as possibly worth a gastric-emptying study on its own merits; the primary care physician saw no urgency once the GLP-1 RA route was already avoided. Left as an open recommendation for a future visit, not resolved here.