Antibiotic-Refractory Lyme Arthritis: DMARD Therapy After a Negative Synovial PCR
A hiker's knee synovitis persists two months after a second, IV course of antibiotics for Lyme arthritis, with synovial fluid now PCR-negative for Borrelia burgdorferi. The disagreement is whether a negative PCR genuinely closes the infectious question, or whether it's reasonable to try antibiotics once more before committing to anti-inflammatory therapy.
Aaron K., a 52-year-old man, has hiked the same ridge trails with his local trail club nearly every Saturday for eleven years, and was three miles into one of them six months ago when he noticed the expanding red rash on his thigh that turned out to be erythema migrans. No prior tick-borne illness, no chronic joint disease, nothing else in his history — his one real risk factor is that he spends every weekend in tick habitat. Two-tier serology confirmed Lyme disease and he completed twenty-one days of oral doxycycline.
Six weeks after finishing that course, his right knee swelled — monoarticular, warm, a large effusion — and repeat serology plus synovial fluid analysis confirmed Lyme arthritis rather than a new tick exposure. He completed a second, twenty-eight day course of oral doxycycline per guideline-standard therapy for Lyme arthritis specifically. When the knee failed to improve at follow-up, he moved to a twenty-eight day course of IV ceftriaxone, the next guideline-recommended step for incomplete response. That course ended two months ago. Today his knee remains swollen and warm, with a persistent effusion on exam and synovial fluid still showing a markedly elevated, lymphocyte-predominant white count.
The finding that actually settles what happens next isn't in his exam — it's in the lab result his rheumatologist ordered specifically to answer this question. Synovial fluid PCR for Borrelia burgdorferi DNA came back negative. The 2020 IDSA/AAN/ACR guideline is careful about what that result can carry: it recommends serum antibody testing over PCR for diagnosing Lyme arthritis at all, and reserves synovial PCR for seropositive patients in whom a treatment decision needs more than serology settles — which is exactly why his rheumatologist sent it. What actually triggers the guideline's move away from antibiotics is not the PCR but his treatment history: one failed oral course and one failed IV course, the threshold at which it advises referral for non-antibiotic therapy. Steere and colleagues, who first characterized this entity, describe it as an HLA-DR-associated autoimmune synovitis triggered by the original infection rather than sustained by it — and the guideline built on their work now calls it post-antibiotic Lyme arthritis rather than antibiotic-refractory, precisely so the older name stops implying an organism still waiting to be killed. He has the treatment history that threshold describes, and a joint with nothing left in it to find.
Rheumatology clinic, two months after IV therapy ended
The negative PCR is the finding that should end this conversation, not restart it. Steere and colleagues, who first described antibiotic-refractory Lyme arthritis, characterized it as an HLA-DR-associated autoimmune synovitis that the original infection triggers but no longer sustains. The 2020 IDSA/AAN/ACR Lyme disease guideline sets its threshold on treatment history rather than on any single test: once a patient has failed one oral and one IV course, it advises referral for consideration of DMARDs, biologics, intraarticular steroids or arthroscopic synovectomy. He has failed both. I'd start hydroxychloroquine now, given his mild presentation, and move to methotrexate if he doesn't respond.
You're right that a third antibiotic course isn't what the guideline supports, and I'm not going to argue he needs one on principle. But listen to how you just used that PCR. The guideline you're citing recommends serology over PCR for this diagnosis and treats a synovial PCR as a tiebreaker, not a proof of sterility — and you've turned one negative result on an imperfectly sensitive assay into the thing that closes the question.
Here is where that actually bites, and it isn't about antibiotics. The same recommendation lists four options side by side — DMARDs, biologics, intraarticular steroids, arthroscopic synovectomy — and ranks none of them. You've read it as endorsing your drug. It endorses a category. This is one knee. An intraarticular corticosteroid treats that knee without committing an otherwise healthy fifty-two-year-old to months of systemic immunosuppression on the strength of a test the guideline itself declines to lean on. I want that on the table before hydroxychloroquine, not after it.
I'd go further, and I'd point at the same sentence. Synovectomy sits in that list too, with no more hedging attached to it than hydroxychloroquine gets. The sequence the two of you are describing — antimalarial, then methotrexate, then surgery if those fail — is a ranking the guideline doesn't make. It's a habit. And it isn't free: it's potentially half a year of a knee that stays swollen while we work down a list, in a man whose stated reason for caring about any of this is that he walks ridge trails every Saturday. I'll accept starting with a drug, because you're both right that it's the less invasive first move and he's never had one. What I won't accept is an open-ended trial. Put a date on it.
Agreed: start hydroxychloroquine rather than a further antibiotic course, on the strength of his failed oral and IV courses rather than the PCR alone, with a hard reassessment at eight weeks rather than an open-ended trial — the one point the Orthopedic Surgeon made a condition of his agreeing to a drug at all. If the effusion and synovitis have not meaningfully improved by then, escalate to methotrexate.
Not agreed: whether systemic therapy was the right thing to reach for first at all. The Infectious Disease Physician's position — that the guideline lists intraarticular steroids, biologics, DMARDs and synovectomy without ranking them, so a single knee does not obviously call for a systemic drug — was not answered, only outvoted by the other two preferring the least invasive systemic option in a man who has never had one. The Orthopedic Surgeon shares the textual reading but not the conclusion, and both flagged that the customary drug-first sequence is practice rather than evidence. An intraarticular injection remains available at the eight-week visit and was not treated as foreclosed by today's decision.