RBC Exchange Transfusion for Babesiosis: We’ve Been Doing This for Decades. Now We Have Data.

The call comes in sometime around midnight. A patient is febrile and jaundiced, their smear shows ring forms in the red cells, and the parasitemia is sitting at 14%. The infectious disease fellow wants to know if we’ll do an exchange. The answer, for most of us, is yes — and it’s not a difficult yes. High parasitemia, signs of hemolysis, organs starting to wobble. The indication feels obvious.

For the better part of forty years, that feeling was the best we had. The first reported case of severe babesiosis treated with RBC exchange was in 1980. Since then, we’ve been doing it based on case reports, small case series, mechanistic logic, and guideline recommendations that were themselves built on case reports, small case series, and mechanistic logic. The data, such as they were, strongly suggested we were doing the right thing. Strongly suggesting and actually demonstrating are different things.

Now we have something closer to an actual answer.

What We’re Dealing With

Babesiosis is a tick-borne illness caused primarily by Babesia microti, an intraerythrocytic parasite — meaning it invades and replicates inside red blood cells. It’s endemic to the northeastern and upper midwestern United States, with incidence increasing steadily in recent years, particularly in New England. The geographic range is expanding, and true incidence may be ten times higher than reported cases.

The disease spectrum is wide. Many people clear infection without knowing they had it. Others end up in the ICU. The patients most at risk for severe disease are immunocompromised — asplenic, actively malignant, post-transplant, on immunosuppression — as well as those at the extremes of age. In hospitalized patients, mortality ranges from about 3% to nearly 9% in the general population and can reach 20% in immunocompromised patients. The parasite’s mechanism of harm is hemolysis: infected red cells rupture, releasing hemoglobin into the plasma, triggering endothelial injury, organ dysfunction, and a proinflammatory cytokine cascade that can take on a life of its own.

Why Exchange Makes Sense

RBC exchange transfusion (ET) is an extracorporeal procedure that removes the patient’s circulating red cells — infected and uninfected alike — while simultaneously replacing them with donor RBCs. The mechanistic case for it in babesiosis is straightforward: fewer infected cells means less hemolysis, less free hemoglobin circulating, and less downstream organ injury. You may also be pulling cytokines out of the circulation, though the contribution of that effect is harder to quantify.

The American Society for Apheresis (ASFA) and the Infectious Diseases Society of America (IDSA) have recommended considering ET for patients with high-grade parasitemia (>10%), severe hemolytic anemia, or acute organ injury for years. What neither guideline could do was point to a study with a real control group and say: we know this changes outcomes.

The Evidence Problem

A 2021 retrospective chart review from Yale (O’Bryan et al., J Clin Apher) was about as rigorous as the pre-existing literature got. Ninety-one patients, single center, 2011–2017. The investigators stratified patients by peak parasitemia — <1%, 1–5%, 5–10%, >10% — and showed that virtually every marker of end-organ dysfunction worsened in a stepwise fashion with increasing parasite burden: hematocrit fell, LDH rose, bilirubin climbed, creatinine drifted up, platelet counts dropped. Nineteen patients received exchange, all with peak parasitemia ≥9% and some degree of organ dysfunction. Parasitemia dropped sharply post-exchange.

The study showed what we already believed: high parasite burden is bad, and exchange reduces parasite burden. What it could not show — because patients who received exchange were sicker at baseline than those who didn’t — is whether exchange changed outcomes. The only prior study that even attempted a comparison had twenty-eight total patients. That is the entire controlled evidence base for a procedure we’ve been doing since the Carter administration.

Enter STOP-BABESIOSIS

In March 2026, Leaf et al. published the STOP-BABESIOSIS study in JAMA Internal Medicine: a multicenter cohort of 3,233 patients hospitalized with babesiosis across 82 sites and 24 medical centers in the northeastern US, spanning 2010 to 2024. Of these, 629 met eligibility criteria for the analysis: parasitemia >10%, or 5–10% with acute organ injury or severe hemolytic anemia.

The investigators used a sequential target trial emulation (TTE) framework — worth pausing on, because it matters for how much you trust the result. Target trial emulation is an approach to observational data analysis that mimics the structure of a randomized controlled trial: you specify eligibility criteria, a treatment strategy, a defined start of follow-up, and outcomes, then apply them to real-world data. The sequential version used here enrolls patients on each of the first 7 days of hospitalization, which eliminates a specific bias problem called immortal time bias and allows confounder adjustment at the actual moment of treatment assignment rather than at a fixed earlier time point. Inverse probability of treatment weighting (IPTW) was applied to balance the groups on measured covariates. It’s rigorous methodology, and it’s increasingly the standard when an RCT is impractical.

The primary endpoint was a composite of in-hospital death or 30-day readmission. Among the 209 patients treated with ET and the 420 who were not: the composite occurred in 3.6% of ET-treated patients versus 9.8% of those not treated. Adjusted odds ratio: 0.22 (95% CI 0.09–0.51). That’s a nearly fivefold reduction in odds, and it held up across eight sensitivity analyses — adjusted for site, year of admission, LDH, restricted to the first three days, restricted to exchanges using at least 10 units of RBCs. The result didn’t move.

What This Changes — and What It Doesn’t

The caveats are real, and the authors don’t hide them. Even after IPTW, the ET group had higher median parasitemia and a greater proportion of immunocompromised patients than the control group. Residual confounding in either direction is possible — meaning the benefit of ET could be an underestimate, but it could theoretically also be overestimated. The total number of deaths was small, so most of the composite endpoint is driven by 30-day readmissions rather than mortality. And the study enrolled only patients in the northeastern US, where B. microti is overwhelmingly the dominant species.

An RCT is almost certainly never coming. The investigators say so directly, and they’re right: the number of sites required, the enrollment duration, the difficulty of maintaining equipoise, and the inevitability of crossover make a randomized trial effectively impossible. This is the best evidence we are likely to have. And the best evidence we are likely to have is a nearly fivefold reduction in death or readmission among severely ill patients who received exchange.

For practice, this means the ASFA/IDSA criteria — parasitemia >10%, or 5–10% with organ injury or severe hemolytic anemia — now have something behind them beyond expert opinion. It also raises questions the study doesn’t answer: which subgroups benefit most? Does exchange help patients who don’t meet the strict eligibility criteria, such as those with milder organ injury? The subgroup analyses showed consistent benefit across age, sex, immunocompromised status, and SOFA score, but the study wasn’t powered for those comparisons.

There is a particular kind of clinical discomfort in practicing in an evidence vacuum — ordering a procedure you believe is right, knowing the belief is built on mechanistic logic and small case series rather than data. Most of transfusion medicine lives in that space. Most of medicine does, if you’re being honest about it.

The STOP-BABESIOSIS investigators enrolled 3,233 patients across 82 sites over fifteen years to give us an answer for 629 of them who met strict eligibility criteria. That is what it takes to generate evidence in a disease this uncommon. The result is about as close to a definitive answer as we’re going to get.

The data support exchange transfusion for severely ill patients with babesiosis. It took forty-six years to say that with numbers behind it.

References

Leaf DE, et al. (STOP-BABESIOSIS Investigators). Red Blood Cell Exchange Transfusion for Severe Babesiosis. JAMA Intern Med. Published online March 30, 2026. doi:10.1001/jamainternmed.2026.0244

O’Bryan J, Gokhale A, Hendrickson JE, Krause PJ. Parasite burden and red blood cell exchange transfusion for babesiosis. J Clin Apher. 2021;36:127–134. doi:10.1002/jca.21853