Two Organs, One Prescription
Residual sleep apnea despite good PAP adherence, and early diabetic kidney disease, sit in the same patient at the same visit. Tirzepatide, semaglutide, and an SGLT2 inhibitor each have their strongest evidence pointed at a different one of his problems.
H.O. checks his PAP machine's data app most mornings the way other people check the weather — a habit his sleep clinic actually encouraged, and one that shows, unambiguously, that he wears the device more than seven hours most nights. His AHI has come down from a pretreatment 55 to 18 on therapy, real and meaningful improvement, but not the near-normalization his sleep physician was hoping adherence this good would produce. He is 58, has had type 2 diabetes for nine years, and a routine urine study last month came back with a mildly elevated albumin-to-creatinine ratio — his first sign of diabetic kidney involvement, caught early, while it's still a trajectory rather than an established diagnosis.
He is asking, reasonably, whether one of the newer weight-loss medications could help with "all of it at once." The honest answer is that the three real candidates don't have the same evidence pointed at the same problem. Tirzepatide's SURMOUNT-OSA trial studied AHI reduction directly as a primary outcome in a cohort whose mean baseline AHI sat near 50 and mean BMI near 39, evidence built for the exact question his sleep physician is asking, and at a BMI of 34 with an AHI of 18 on therapy he sits inside its entry criteria rather than beside them. The other two are a different story. Semaglutide's headline cardiovascular result comes from SELECT, which required established atherosclerotic disease — a prior myocardial infarction, stroke, or symptomatic peripheral arterial disease — and excluded anyone with a history of diabetes or an HbA1c at or above 6.5%. H.O. has neither the qualifying event nor the permitted glycemic profile: no prior MI or stroke, and nine years of type 2 diabetes at 7.4%. He is outside that trial twice over, in opposite directions. Empagliflozin's dedicated kidney-outcome evidence runs into the same wall from the other side. CREDENCE enrolled at a UACR above 300, DAPA-CKD at 200 or more, and EMPA-KIDNEY required a UACR of at least 200 in anyone whose eGFR was 45 or higher — and his eGFR is 78 with a UACR that has only just crossed into the abnormal range. FLOW, the semaglutide kidney trial, drew its line at an eGFR of 75 with a UACR above 300; he clears neither. The uncomfortable shape of this visit is that the very thing making his nephropathy worth treating — that it was caught early — is what puts him outside every trial that proved these drugs protect kidneys at all.
Joint sleep medicine / endocrinology visit
He's doing everything right on PAP and his AHI is still 18. SURMOUNT-OSA studied exactly this — AHI reduction as a primary endpoint in patients who look like him. Of the three options in front of us, tirzepatide is the one with trial evidence actually built to answer the question the sleep clinic is asking.
I'd frame the urgency differently. His AHI has already come down from 55 to 18 on PAP alone — real, substantial improvement. His UACR just turned abnormal for the first time. That's the finding that's early enough to actually change trajectory. I want to be careful about how I say this, though: the dedicated renal-outcome trials don't cover him. CREDENCE started at a UACR above 300, DAPA-CKD at 200, EMPA-KIDNEY at 200 for anyone with an eGFR of 45 or better. His eGFR is 78 and his UACR has barely crossed the line. What I'm relying on is the guideline recommendation for SGLT2 inhibition in albuminuric type 2 diabetes, which extends below where those trials enrolled, plus a renal mechanism that is direct and not a weight-loss side effect — and that is an extrapolation, not a population match.
Tirzepatide might help his AHI further, but "already substantially improved" and "just starting to decline" aren't the same kind of urgency — I'd rather treat the trajectory that's still early enough to fully arrest.
Before we get to which organ wins, I want to close off the option someone always raises at this point, because I nearly raised it myself: semaglutide as the neutral middle, on SELECT's cardiovascular result. It doesn't survive contact with his chart. SELECT required an established atherosclerotic event — MI, stroke, or symptomatic peripheral arterial disease — and excluded anyone with a history of diabetes or an HbA1c of 6.5% or above. He has no qualifying event and nine years of diabetes at 7.4%. He fails the inclusion and meets the exclusion. FLOW is the semaglutide trial that did enroll diabetic kidney disease, and it drew its line at an eGFR of 75 with a UACR over 300; his numbers are 78 and barely abnormal.
Which leaves me disagreeing with the framing more than with either of you. You've each argued as though one drug has to be chosen — but the nephrologist has just conceded his own recommendation is an extrapolation rather than a trial match, and that cuts against treating this as a contest between two equally-evidenced claims. Tirzepatide has a population match for the AHI question; empagliflozin has a guideline recommendation and no matching trial for the kidney question. Those are different kinds of warrant, not competing ones, and they don't have to be traded against each other at a single visit.
Agreed: tirzepatide and empagliflozin both started today, UACR and AHI both rechecked at four months, PAP therapy unchanged and continued.
Not fully agreed: the nephrologist would have preferred to see empagliflozin's renal effect in isolation first before adding a second new agent at the same visit, given his eGFR is still preserved and not urgently declining; the endocrinologist and primary care physician judged starting both together reasonable given his overall risk profile and the low likelihood of the two agents' effects being difficult to distinguish clinically. The nephrologist's preference for a more staged approach was heard but not adopted this visit.