A Safari Fever at 24 Weeks: Doxycycline Despite the Pregnancy
A pregnant woman develops fever, an eschar, and a rash ten days after a South African safari — a textbook presentation of African tick bite fever. The disagreement isn't about the diagnosis — it's about whether the tetracycline caution generations of clinicians absorbed from a rickettsiosis that kills should govern one that, on the record, never has.
C.F., a 31-year-old woman, is 24 weeks into her first pregnancy and still glowing, by her own description, from the two-week "babymoon" safari she and her husband took through Kruger National Park before she got too far along to travel comfortably. They'd been warned about malaria and taken their prophylaxis faithfully; nobody had specifically mentioned ticks, and she doesn't recall any single bite standing out among the general itch of a week spent walking through bush grass at dawn. Ten days after returning home, a fever started, climbing to 39.2°C over two days, alongside a spreading maculopapular rash and body aches severe enough that she initially wondered whether something was wrong with the pregnancy itself rather than with her.
What actually settled the diagnosis, before any lab result came back, was a small blackened scab on her left calf that neither she nor her husband had noticed until the exam — a tick-bite eschar, the single most useful bedside sign for African tick bite fever. Rickettsia africae is a far milder organism than the Rocky Mountain spotted fever the tetracycline-avoidance teaching was built around: no death has ever been attributed to it, and treated patients usually defervesce within about two days. Doxycycline is the drug of choice for both regardless, so the caution generations of clinicians absorbed about tetracyclines and fetal tooth and bone effects arrives here just as reflexively as it would in a domestic RMSF case. Her 24 weeks is the number that actually sorts that caution. She is well past the first-trimester organogenesis window most drug-safety anxiety attaches to, and squarely inside the second- and third-trimester window in which tetracycline binding to developing fetal dentition and bone was described in the first place — which is to say the caution is aimed at precisely where she is. It has softened since: Todd and colleagues found in 2015 that children given short doxycycline courses for suspected RMSF showed no visible dental staining, reassurance that changed pediatric practice directly. The pregnancy-specific data never moved with it.
The drug everyone was taught to avoid here
Start doxycycline today — and let me be careful about why, because the easy version of this argument is wrong. This will not kill her. African tick bite fever has never killed anyone on record, and I'm not going to inflate that to win a point. What I'm treating is a 39.2°C fever with thrombocytopenia and transaminitis in a 24-week pregnancy, running for an unknown number of further days, when a five-day course reliably ends it in about two. The teratogenic concern that built the tetracycline-avoidance teaching came from prolonged, high-dose courses of the older tetracyclines, not from a short course of doxycycline. Todd and colleagues found in 2015 that children given short doxycycline courses for suspected Rocky Mountain spotted fever showed no visible dental staining; that reassurance changed how pediatrics treats this drug directly.
Waiting for serology buys us nothing here — it won't turn positive for one to two weeks, well past the window where treatment delay actually matters clinically.
I take the pediatric data seriously, and I'm not arguing to withhold treatment. What I'd push back on is treating that reassurance as if it directly answers the pregnancy question, when it doesn't — it's a different exposed population, at a different developmental stage of risk, studied for a different outcome. There is no comparable prospective pregnancy safety dataset for doxycycline, full stop.
And I'd want azithromycin genuinely on the table, not waved past. It has real use in spotted-fever-group rickettsiosis in pregnancy and in young children, it carries none of doxycycline's fetal dentition question, and you have just told me this organism has never killed anyone — which is exactly the setting where accepting a possibly weaker drug costs least. I am not asking for chloramphenicol. Its marrow toxicity is worse than the problem we are solving.
Azithromycin is the fairer comparator and I'm glad it was raised rather than chloramphenicol, which I'd take off the table entirely — marrow suppression in the mother, gray baby syndrome in a neonate exposed near term, and no practical modern US oral formulation. But I'd hold azithromycin to the same standard we just held doxycycline to. Its record in this genus is a handful of Mediterranean spotted fever reports; against R. africae specifically, nobody has tested it. Swapping a fetal-safety question her gestational age has partly answered for an efficacy question nothing has answered isn't reducing risk, it's moving it somewhere we can't see it.
Given how clinically apparent this presentation already is — the eschar, the exposure history, the rash — I'd treat empirically today with doxycycline rather than wait on serology that can't help us for another one to two weeks regardless of which drug we're debating.
Agreed: doxycycline started today at standard dosing, with close outpatient follow-up and fetal monitoring arranged before discharge.
Not agreed: the maternal-fetal medicine specialist's underlying view, stated plainly even after accepting doxycycline as the reasonable choice for this specific, clinically obvious presentation, is that azithromycin should be tried first in a patient earlier in pregnancy than C.F. is — where the fetal dentition window her 24 weeks sits inside has not yet opened, and an untested drug against a never-fatal organism is a cheaper bet than it is today. The infectious disease physician's own position, unmoved, is that an untested drug is untested at every gestational age, and that trading known efficacy for unknown efficacy gets no safer earlier in a pregnancy. Both left agreeing on today's patient and disagreeing on the next one.