A Gout Prescription His Primary Care Doctor Almost Wrote Without a Blood Test First
A man with recurrent, worsening gout is a straightforward candidate for allopurinol by every clinical measure except one his primary care doctor nearly overlooked — a specific HLA allele, disproportionately common in his ancestry group, that raises his risk of a rare but frequently fatal drug reaction by more than a hundredfold.
Kevin S., 57, immigrated from southern China in his twenties and has run a small accounting practice out of a strip-mall office for the past eighteen years, work he says he can barely sit through during a gout flare bad enough to keep him off his feet for days. His fourth flare in fourteen months — this one in his left great toe, waking him at 3 a.m. with the pain he now recognizes immediately — brought him back to his primary care physician, who confirmed hyperuricemia on labs drawn during the acute episode and recommended starting allopurinol for long-term urate-lowering therapy once the flare settled, a reasonable and guideline-consistent next step for recurrent gout at his flare frequency.
What the initial prescribing conversation didn't include was HLA-B*58:01 testing — a specific allele disproportionately prevalent in individuals of Han Chinese, Thai, and Korean ancestry, and one that, in carriers, raises the risk of a severe, sometimes fatal cutaneous drug reaction (Stevens-Johnson syndrome/toxic epidermal necrolysis, SJS/TEN) to allopurinol by roughly one to two orders of magnitude above the already-low baseline risk in non-carriers. The Clinical Pharmacogenetics Implementation Consortium's guideline specifically recommends screening before allopurinol initiation in exactly the ancestry groups Kevin's family history places him in — a test that costs little, takes about a week to result, and was nearly bypassed entirely in the interest of getting him started on treatment before his next flare.
Kevin has no personal history of drug reactions of any kind and, by his own account, had never heard of pharmacogenomic testing before today's conversation — his prior experience with new prescriptions has simply been to take the pill and see what happens, an approach that has never caused him trouble with the handful of medications he's tried over the years. He was initially puzzled rather than alarmed when the testing was raised, asking directly why this particular drug and this particular test, at his age, with no prior reaction history to speak of, warranted a step none of his previous prescriptions ever had.
One test, one allele, a risk too large to treat as routine
I'd screen for HLA-B*58:01 before writing this prescription, not after. The CPIC guideline specifically recommends it in populations with elevated allele prevalence, and Han Chinese ancestry is exactly one of those populations — case-control studies have found carriers face a risk of allopurinol-induced Stevens-Johnson syndrome or toxic epidermal necrolysis on the order of one to two hundred times higher than non-carriers. That's not a modest pharmacogenomic signal of the kind we usually shrug off as too small to act on; it's one of the largest, most replicated drug-gene associations in clinical use.
I hear the numbers, and I want to be honest that my instinct going in was to start him now and counsel him carefully — allopurinol hypersensitivity has been managed for decades through patient education on early warning signs and close follow-up through the first eight weeks, before genetic testing was ever part of the conversation. He's in real, recurrent pain, and a week's delay for a test result isn't nothing to a patient who's already missed work four times this year.
But I'll say directly that hearing the actual magnitude of the risk in a carrier changes how I'm weighing that delay — this isn't the kind of small relative-risk bump where vigilance alone reasonably substitutes for knowing his status first.
I'd underline why this case is different from the general skepticism a lot of us reasonably hold about routine pharmacogenomic testing — most single-gene associations shift risk modestly enough that testing everyone rarely changes outcomes meaningfully. HLA-B*58:01 and allopurinol is a genuine outlier among these associations, both in the size of the relative risk increase and in how well it's been replicated across multiple populations since Hung and colleagues' original finding. One week's delay for a test result is a small cost against a reaction that, in a true carrier, has a real mortality rate. If he tests negative, we've lost a week; if he tests positive, we've avoided starting a drug that could kill him.
Agreed: HLA-B*58:01 testing ordered before allopurinol initiation, with low-dose colchicine started as interim flare-prevention therapy during the one-week wait, and febuxostat pre-identified as the fallback plan should the test return positive.
Kevin's HLA-B*58:01 result came back negative five days later, and allopurinol was started that week with standard monitoring; all three clinicians agreed the week's delay was a small, reasonable cost against the risk it was checking for, and the primary care physician noted directly that this case had changed how she would approach the next patient presenting with the same ancestry and the same prescription.