SSRIs for Illness Anxiety Disorder: Does OCD-Style Dosing Apply?
A single patient whose illness anxiety looks phenomenologically close to OCD — checking, reassurance-seeking, intrusive health fears — but whose own hypersensitivity to bodily sensation makes the OCD playbook's aggressive dosing a genuinely double-edged recommendation.
R.P., a 52-year-old man, has ridden a century — a hundred miles in a day — every September for the past eleven years with the same group of cycling friends, a streak he is determined not to break despite what the last two years have done to his relationship with his own pulse. It started after a training ride when he felt his heart pounding on a hill climb and convinced himself, in the emergency department that night, that he was having a heart attack. The troponins were normal. The stress test three weeks later was normal. The echocardiogram after that was normal. None of it stuck. He now checks his pulse by hand several times a day, has an app that logs his heart rate continuously, and has been to the emergency department four times this year, each visit ending the same way — reassurance that lasts perhaps a week before the next sensation restarts the cycle. His PHQ-9 is unremarkable; he isn't depressed, isn't avoiding his life, still rides and still works. What he has is a fear that returns no matter how many times the data contradicts it, and a checking ritual that has quietly become the actual organizing structure of his days.
The overlap with OCD isn't just a loose analogy here — it's the actual clinical question the team has to answer before writing a dose. Greeven and colleagues randomized 112 patients with hypochondriasis to sixteen weeks of paroxetine, CBT, or placebo and found both active treatments beat placebo, with no real difference between them. The detail that complicates the dosing question is buried in that trial's own methods: paroxetine was titrated flexibly toward 60 mg/day, and the mean final dose was 40 mg — which is not a depression-range dose at all. It is precisely the target dose paroxetine's own label specifies for OCD. The trial usually described as the standard-dose evidence in this diagnosis was, in practice, already dosing into OCD territory. What remains genuinely unsettled is what nothing in that trial speaks to: his pulse-checking, the part of his presentation that behaves less like generalized worry and more like a compulsion with its own reinforcement cycle, and whether it needs pushing past the 40 mg Greeven's patients averaged toward the 60 mg ceiling.
Deciding on a starting dose
Greeven and colleagues ran the actual randomized trial in this exact diagnosis — 112 patients, sixteen weeks, paroxetine against CBT and placebo — and found real separation from placebo. I want to be precise about where that trial landed on dose, because it's often described loosely: the paroxetine arm titrated flexibly toward 60 mg and averaged 40 mg at endpoint. So the evidence base in hypochondriasis is not a low-dose evidence base. It already sits at what the label calls the OCD target. Forty milligrams is where I'd aim, and I'd want a specific reason before going past the dose the only on-point trial actually tested.
You're right that Greeven's trial is the direct evidence in this diagnosis, and I'm not arguing to ignore it. But look at what he actually does several times a day — the pulse-checking, the app, four ED visits chasing the same reassurance that never holds. That's not generalized worry, that's a compulsion with a checking ritual attached to it, phenomenologically closer to OCD than to the diffuse anxious rumination Greeven's trial population likely centered on.
And I'd turn your own correction back on you. You just told us Greeven's patients averaged 40 mg — the OCD target — which means the trial you're citing as a reason not to reach for OCD-style dosing is the trial that quietly used it. A mean of 40 also means a real share of that arm sat above 40, closer to the 60 mg ceiling, and the trial was never powered to tell us which patients needed which end of that range. If the checking is the part actually running his life, 40 mg is the floor that argument supports, not the ceiling.
I've fielded his calls for two years, including the one ten days into his last SSRI trial when an ordinary bout of jitteriness convinced him something was cardiac and he stopped the drug himself. Whichever target dose you two eventually agree is right, that history says the titration pace matters as much as the destination for this specific patient — a hypervigilant patient can turn a routine side effect into the next health fear before the drug ever gets the chance to work.
I'd start low, well under either of your proposed targets, and move slowly enough that any new sensation is small enough not to become the next crisis — genuinely agnostic about whether Greeven's 40 mg average or something nearer the 60 mg ceiling is ultimately correct for him, because at the rate his last trial ended, neither answer matters if he doesn't stay on the drug long enough to find out.
Agreed: start paroxetine at half the usual starting dose, with explicit instructions to call before stopping rather than after, and a two-week check-in specifically to talk through any new sensation before it becomes its own crisis.
Not agreed, and left open rather than resolved by today's cautious start:
Reassess with Greeven's endpoint in mind — if checking behavior hasn't meaningfully improved at the dose that trial actually averaged, that becomes the pharmacist's argument for pushing on toward the 60 mg ceiling.
The dosing-philosophy question becomes moot until a tolerated baseline is established — CBT specifically targeting the checking ritual moves up as the nearer-term plan.
The psychiatrist and pharmacist still disagree about which target dose is ultimately right for him; both accepted the primary care physician's read that the disagreement is premature until he's actually still taking the drug in six weeks.