Diagnosed Two Years After the Delivery That Caused It
A single patient, two years past a postpartum hemorrhage severe enough to require transfusion, now presenting with a hormonal picture consistent with Sheehan syndrome. The disagreement is whether to complete full dynamic testing before treating, or start replacement now and test around it.
Q.A., a 34-year-old woman, is two years and one toddler out from a delivery complicated by severe postpartum hemorrhage that required emergency transfusion of four units of blood — an event she survived and, until this visit, had mostly filed away as a frightening but resolved complication rather than the plausible cause of everything that's followed. She never successfully breastfed despite trying for weeks, her periods never fully returned to normal, and a fatigue she initially attributed to new-parent exhaustion has never lifted the way she expected it eventually would. Labs drawn by her primary care physician found a low-normal free T4 with an inappropriately non-elevated TSH, a random cortisol of 3.2 mcg/dL drawn at 9 a.m. — low enough to be concerning on its own even without provocative testing — and low estradiol with inappropriately low-normal gonadotropins, a combined pattern of anterior pituitary hormone loss that, together with her delivery history, points strongly toward Sheehan syndrome: ischemic pituitary necrosis from severe peripartum hemorrhage, occurring in a gland that undergoes substantial physiologic enlargement during pregnancy and becomes correspondingly more vulnerable to hemorrhage-related infarction.
The diagnostic-versus-treatment tension here is genuinely time-sensitive rather than academic. Formal confirmation of her adrenal axis would traditionally involve dynamic testing — an ACTH stimulation test at minimum, an insulin tolerance test in some protocols, though the latter is relatively contraindicated in a patient with this degree of suspected hypocortisolism given its own risk of provoking a crisis it's meant to diagnose. Her random morning cortisol of 3.2 mcg/dL, however, already sits low enough that most endocrine society guidance treats a value in this range as sufficient grounds for presumptive treatment without waiting for further dynamic confirmation, precisely because the risk of an undiagnosed, untreated adrenal insufficiency progressing to a genuine crisis during the additional days or weeks formal testing would take outweighs the value of a more precise number before starting glucocorticoid replacement.
Two years of unexplained fatigue, and a diagnosis that can't wait for perfect data
I'd start hydrocortisone today rather than wait for formal dynamic testing. Her morning cortisol of 3.2 mcg/dL already clears the threshold most guidance treats as sufficient for presumptive treatment on its own, and an insulin tolerance test — the more definitive dynamic study — carries real risk of provoking the exact crisis we're trying to diagnose in a patient this likely to actually be deficient.
I agree with starting hydrocortisone now, and I'd add one sequencing point that matters as much as the decision to treat itself: levothyroxine has to wait until after glucocorticoid coverage is established, not started alongside it or before it. Thyroid hormone replacement increases the metabolic clearance of cortisol, and starting it first in a patient with unrecognized or undertreated adrenal insufficiency can precipitate the adrenal crisis we're specifically trying to prevent by treating early.
That sequencing risk is easy to lose track of once the decision to treat both deficiencies has been made — the urgency of her cortisol shouldn't collapse into treating both hormones as a package started on the same day.
Right, and I'd build that sequencing into the plan explicitly rather than trust it to be obvious later — hydrocortisone starting today, levothyroxine held for at least a week to ten days once cortisol coverage is clearly established and tolerated, with an ACTH stimulation test still obtained in the interim to formally confirm the diagnosis, but as a confirmatory step running alongside treatment, not as a gate she has to clear before treatment starts.
Agreed: start hydrocortisone today on a presumptive basis, given her morning cortisol already meeting the threshold most guidance treats as sufficient without further dynamic testing. Levothyroxine explicitly delayed for 1-1.5 weeks until glucocorticoid coverage is established and tolerated, given the real risk of precipitating adrenal crisis by starting thyroid replacement first. ACTH stimulation testing obtained in parallel to formally confirm the diagnosis, without functioning as a precondition for starting treatment.
Both physicians converged fully on this plan; the pharmacologist's contribution was making the glucocorticoid-before-thyroid sequencing an explicit, dated step in the plan rather than an assumption that could get lost once both deficiencies were confirmed and "just needed treating."