Two Patients Who Cannot Avoid the Drug They React To
Two patients need a drug they are allergic to and have no adequate substitute — one, oral penicillin for neurosyphilis in pregnancy; the other, IV oxaliplatin for stage III colon cancer — and the same rapid desensitization principle produces two very differently structured protocols.
Camille R., 29, is twenty weeks into a pregnancy she and her partner had been trying for since last spring, and was referred to the allergy clinic the same week her routine prenatal RPR screen came back reactive and a confirmatory FTA-ABS established neurosyphilis with early ocular involvement — a diagnosis that, left untreated through the pregnancy, carries a real risk of congenital transmission and fetal harm that no alternative regimen matches. Fifteen years ago, as a teenager, she developed hives and lip swelling within twenty minutes of a dose of amoxicillin for strep throat; she has avoided every penicillin-class antibiotic since, and skin testing performed this week confirmed a positive immediate reaction to both major and minor penicillin determinants.
Penicillin is not simply the preferred treatment for neurosyphilis in pregnancy — it is, per every major guideline, the only treatment with proven efficacy at preventing congenital transmission; the alternative regimens sometimes used in nonpregnant penicillin-allergic patients (doxycycline, ceftriaxone with its own cross-reactivity uncertainty) either cross the placenta poorly or carry unacceptable fetal risk of their own. That leaves exactly one path to treating her safely: desensitize her to penicillin, then treat with it directly, the approach validated in a prospective protocol published by Wendel and colleagues specifically for pregnant women with documented penicillin allergy and a syphilis diagnosis requiring treatment. Her own reaction sets the shape of that protocol, not merely the need for one: hives and lip swelling, no hypotension, nothing respiratory — a purely mucocutaneous history, which is the category the standard graded oral schedule was built around rather than the cardiovascular-involvement history that forces a lower starting dose and a longer ladder.
Camille has spent the days since her diagnosis oscillating between relief that it was caught early enough to treat and a sharper fear than the fetal-risk counseling alone accounts for — her older sister lost a pregnancy to an infection years ago, and the word 'transmission' has been landing differently for her than a straightforward clinical explanation might otherwise land. She has asked, more than once, whether desensitization itself could somehow harm the baby, a question her allergist has answered directly each time: the graded protocol is designed to avoid triggering the kind of systemic reaction that would pose any risk, and decades of use in exactly her situation have not shown otherwise.
There isn't a real second option here, which actually simplifies the conversation: penicillin is the only agent with documented efficacy at preventing congenital syphilis, and every week of delay is a week the fetus remains exposed. The Wendel protocol — a graded oral desensitization over roughly four hours, starting from a dose small enough to be below the threshold that triggered her original reaction — has been used successfully in this exact population for decades.
Agreed on the drug and the protocol family, and I'd add the mechanistic reason oral works well here: graded oral dosing produces a slower rise in serum drug level than IV push, giving mast cells more time to down-regulate IgE receptor signaling at each incremental step rather than being hit with a dose large enough to cross-link enough surface IgE to degranulate. Her reaction history — urticaria and angioedema, no cardiovascular or respiratory involvement — puts her toward the lower-risk end of what desensitization protocols are built to handle, which supports the standard 13-14 dose oral schedule rather than anything more conservative.
I don't think this needs the extended monitoring the chemotherapy protocol below will, precisely because her prior reaction never suggested the kind of severity that argues for a longer, more cautious ladder.
The one addition from my side is fetal monitoring during the desensitization itself, not because the drug threatens the fetus, but because a maternal reaction — even a mild one — can transiently affect uteroplacental perfusion. Continuous fetal heart rate monitoring throughout the four-hour protocol, in a labor and delivery unit rather than an outpatient infusion suite, is the piece I'd add on top of what's already agreed.
Desensitization completed without incident over four hours on the labor and delivery unit; aqueous crystalline penicillin G was started immediately afterward at 18 million units daily by continuous intravenous infusion for fourteen days, with fetal heart tracing reassuring throughout. The weekly benzathine schedule was explicitly not used: benzathine penicillin G does not reliably reach treponemicidal concentrations in cerebrospinal fluid, and the three-dose weekly course belongs to late latent syphilis, not to neurosyphilis or to the ocular involvement that is managed on the same footing.
Walter B., 63, retired eight months ago after thirty years running a small commercial print shop, planning to spend his newly free mornings restoring an old wooden sailboat docked at the marina near his house — a plan interrupted by a stage III colon cancer diagnosis and, more recently, by his sixth cycle of FOLFOX, during which he developed facial flushing, throat tightness, and a blood pressure drop to 78/50 roughly ten minutes into the oxaliplatin infusion, requiring epinephrine and a rapid response team call. He recovered fully within the hour, but oxaliplatin was held pending allergy evaluation.
IgE-mediated hypersensitivity to platinum agents, unlike many chemotherapy reactions, tends to emerge after repeated exposure rather than on the first dose — Walter's reaction on cycle six, not cycle one, is the textbook pattern, consistent with true sensitization rather than a non-immune infusion reaction. Skin testing confirmed a positive immediate reaction to oxaliplatin, and there is no equally effective platinum-free substitute within his current regimen; discontinuing platinum therapy altogether would mean abandoning a component of treatment his oncology team considers meaningfully protective against recurrence at his stage. That leaves rapid desensitization as the only path to completing his planned course — but a reaction that already included hypotension requiring epinephrine argues for a materially more cautious ladder than Camille's did.
The sailboat project has become, in the way these things sometimes do, a stand-in for the larger question he doesn't ask out loud — whether he'll actually get the retirement he planned for, or whether this diagnosis quietly rewrites it. He has been notably practical about the reaction itself, more concerned with the six added hours per infusion day than with the underlying immunology, and has told his oncology team plainly that he'd rather push through a longer, more cautious protocol than lose the platinum component of his treatment on the theory that it might matter less than they think.
I'd like to keep oxaliplatin in his regimen if it's at all feasible, given the added benefit platinum therapy provides at his stage, but I recognize a Grade 3 reaction with documented hypotension is a different animal from a hive-and-lip-swelling history, and I don't think the same protocol used for a milder reaction is the right one here.
Right — this calls for the extended 16-step (or longer) desensitization protocol described by Castells and colleagues for high-risk platinum and taxane reactions, rather than the shorter 12-step version, specifically because his reaction already crossed into a grade where the starting dose needs to be lower and the number of incremental steps higher to keep each dose increment small enough to avoid re-triggering mast cell degranulation. Premedication with H1/H2 blockade and a corticosteroid beforehand, plus the infusion run in a monitored setting with resuscitation equipment immediately at hand, are non-negotiable given the hypotension in his history.
This is the same underlying principle as Camille's protocol — graded re-exposure inducing temporary tolerance — but the number of steps and the premedication burden scale directly with reaction severity, not with the drug class itself; a milder platinum reaction could plausibly use a shorter ladder, and a more severe penicillin reaction would need a longer one.
One more layer worth naming directly: unlike Camille's one-time treatment course, Walter will need this desensitization repeated before every remaining cycle, since the induced tolerance is temporary and doesn't carry forward once the drug clears his system. That's a real burden — roughly six added hours per infusion day for however many cycles remain — and it's worth being explicit with him about that cost now, rather than only after the first one.
The extended 16-step protocol was completed successfully with no reaction, and oxaliplatin resumed as planned; all agreed it would be repeated identically before each remaining cycle rather than shortened once tolerance was demonstrated once, since the induced state does not persist between infusions.
Not fully settled: whether, after several uneventful cycles, the ladder could reasonably be shortened to reduce the six-hour burden on his infusion days. The pharmacist raised it as a legitimate question for later cycles; the allergist was reluctant to modify a protocol that had worked without a clear safety signal justifying the change, and the group left it for reassessment rather than deciding it today.