Esophageal Candidiasis: Empiric Treatment or Confirmatory Endoscopy First
New odynophagia in an immunocompromised patient with visible oral thrush is classic enough to treat empirically by most standard practice — but a genuine minority of these presentations turn out to be something else entirely, and the cost of being wrong is a real part of the argument.
Terrence O., 44, was diagnosed with HIV eleven years ago and has been consistently on antiretroviral therapy since, a routine he keeps as carefully as he keeps the community garden plot he's tended every growing season for the past five years — something he mentions almost as a way of establishing that his life is, by his own account, well-managed and stable. He missed several clinic visits over the past year during a stretch of unstable housing, and his most recent labs, drawn today, show a CD4 count of 94, down from 310 eighteen months ago, with a detectable viral load — a real, unambiguous immunologic decline, not incidental. He presents today with five days of pain on swallowing, worse with solids, and visible white plaques on his tongue and buccal mucosa consistent with oral thrush.
Odynophagia with a CD4 count this low and visible oral candidiasis is classic enough that most standard practice treats empirically without endoscopy first, and that's a reasonable default in most cases like his. But a real minority of presentations that look exactly like this — oral thrush plus odynophagia in advanced immunosuppression — turn out on endoscopy to be something else: CMV esophagitis, HSV esophagitis, or idiopathic HIV-associated esophageal ulceration, none of which fluconazole treats, and each of which has a genuinely different regimen and, in CMV's case, a substantially different urgency. The actual disagreement isn't whether empiric therapy is ever reasonable — it clearly is, most of the time — it's whether Terrence's specific presentation has anything that should raise the threshold for confirming the diagnosis before treating rather than after.
That minority has a size. The NIH HIV opportunistic-infection guidelines and multiple endoscopy-based series place CMV, HSV, and idiopathic esophageal ulceration squarely in the standard differential once CD4 falls under 100 — his is 94 — with CMV specifically flagged as the concern once counts drop under 50, a threshold he is approaching, not one he has safely cleared — which is exactly why the infectious disease physician wanted his specific CD4 trajectory weighed rather than treated as generically "low."
How classic is classic enough to skip the scope
I'd treat empirically with fluconazole and reserve endoscopy for non-response. This is a classic presentation — oral thrush, new odynophagia, advanced immunosuppression — and he's had a prior episode that responded to exactly this regimen. Standard practice treats empirically here specifically because the substantial majority of patients who look like this respond, and an invasive procedure isn't warranted as a first step for the majority who don't need one.
I don't disagree that candida remains the single most likely cause. What I want weighed more heavily is how far his CD4 count has fallen — 94, down from 310, with a detectable viral load after a real gap in care. At this level of immunosuppression, CMV esophagitis becomes a genuinely more likely alternative than it would be at a higher CD4 count, and it needs a different, more urgent treatment.
I'm not arguing against empiric therapy as the first move — I'm arguing that his specific trajectory should shorten how long we're willing to wait before escalating if it doesn't work.
I think you're both right about different parts of this. Empiric fluconazole is a reasonable first step even at his CD4 count — candida is still the most likely single cause. But I'd set an explicit, short reassessment point now, at prescribing, rather than a vague 'follow up if it's not better.' Given his risk profile, a defined 72-hour check-in with a low threshold to proceed straight to endoscopy resolves both of your concerns without either of you having to fully win the probability argument today.
Agreed: fluconazole started today as empiric therapy, with a specific, dated 72-hour reassessment rather than an open-ended trial, and a documented low threshold to proceed directly to endoscopy with biopsy if symptoms haven't meaningfully improved by then. Terrence was also connected with case management the same visit to address the housing instability behind his gap in HIV care.