Does rabbit ATG exposure after transplant affect survival and relapse?

Exposure is modeled, not simply the mg/kg dose

The study asks whether the amount of rATG actually circulating after transplant, rather than the weight-based dose written on the order, tracks with survival, relapse, immune recovery and graft-versus-host disease (GVHD). rATG is given during conditioning to reduce graft rejection and GVHD, but the paper notes that traditional weight-based dosing (mg/kg) produces highly variable post-transplant exposure.

To measure that exposure, the authors did not draw drug levels directly. They took data from patients enrolled in the BMT CTN 1202 registry and fed each patient's transplant date, weight, height, rATG dose and infusion schedule, and absolute lymphocyte count into a previously validated population pharmacokinetic model accessed through the InsightRx platform. The model output is an area under the curve after transplant, expressed in AUxd/mL. This is a simulated exposure derived from dosing records, not a measured drug concentration, which matters when reading every result below.

The authors sorted patients into three previously defined exposure bands: A, below 30; B, 30 to 55; and C, 55 or higher AUxd/mL. They also included 97 patients from the same registry who met eligibility criteria but received no rATG, creating a comparison group that earlier exposure studies lacked. Median post-transplant exposure among treated patients was 44.1 AUxd/mL, with a wide range of 6.2 to 125.0.

Survival falls as modeled exposure rises

Five-year overall survival followed a gradient across exposure bands. It was 67% in group A, 49% in group B and 34% in group C, compared with 51% in the no-ATG group (p = 0.007). In the multivariable Cox model, group C had a hazard ratio of 2.03 (95% CI 1.27 to 3.24) relative to no ATG, while group A's hazard ratio was 0.78 (95% CI 0.43 to 1.40) and group B's was 1.23 (95% CI 0.82 to 1.82).

The confidence intervals are the key detail: only the highest exposure band was statistically distinguishable from no ATG in the adjusted survival model. The low-exposure band's survival advantage over no ATG is visible in the unadjusted five-year rates and in the paper's discussion, but the adjusted hazard ratio for group A crosses 1. The authors also report that the survival difference between low exposure and no ATG was mainly driven by higher GVHD-related deaths in the no-ATG group.

Age at transplant was the other consistent predictor, with an adjusted hazard ratio of 1.02 per year (95% CI 1.01 to 1.03, p < 0.001). An exploratory spline analysis supported roughly 30 AUxd/mL as the exposure threshold most associated with survival, consistent with the authors' prior work.

Relapse and CD4 recovery move in opposite directions

Five-year relapse incidence rose with exposure: 27% in group A, 42% in group B and 58% in group C, versus 31% in the no-ATG group (p < 0.001). Adjusted against no ATG, the hazard ratio was 0.76 for group A (95% CI 0.38 to 1.54), 1.83 for group B (95% CI 1.13 to 2.97) and 2.73 for group C (95% CI 1.58 to 4.73). Here both higher bands separate from no ATG, and the authors highlight this relapse signal as a difference from earlier, largely pediatric studies where the main effect of exposure was on transplant-related mortality.

Early CD4 immune reconstitution, defined as CD4 count above 50 per microliter at day 100, showed the same pattern in reverse: 73% in group A, 51% in group B and 19% in group C, versus 64% with no ATG (p = 0.003). In the logistic model, group C had an odds ratio of 0.13 (95% CI 0.03 to 0.49) relative to no ATG. Peripheral blood as graft source was also associated with better CD4 recovery than bone marrow (56% versus 41%; adjusted odds ratio 3.24, 95% CI 1.16 to 9.97).

The paper links these strands descriptively: patients who reached early CD4 reconstitution had 15% transplant-related mortality versus 22% without it, 60% versus 52% overall survival, and 30% versus 45% relapse. These are unadjusted comparisons across the whole cohort, not a test of whether CD4 recovery causes the difference.

GVHD shows a different and weaker exposure signal

Grade 2 to 4 acute GVHD did not differ significantly across the four groups: 40% with no ATG, 26% in group A, 39% in group B and 35% in group C (p = 0.44). Any rATG exposure was associated with lower moderate or severe chronic GVHD, with a hazard ratio of 0.63 (95% CI 0.42 to 0.94, p = 0.025) compared with no rATG.

Transplant-related mortality did not reach significance across exposure groups (group A 17%, B 21%, C 22%, no ATG 25%; p = 0.48), and the authors attribute the null result to limited power. Among the 164 deaths, the recorded causes were primary disease in 73 (45%), GVHD in 33 (20%) and infection in 25 (15%).

Read together, these results frame the clinical trade-off. Higher exposure appears to buy chronic GVHD protection while costing CD4 recovery and, in this adult cohort, relapse control. The no-ATG group sits at the other extreme, with the highest transplant-related mortality and more GVHD-related deaths, which is why the authors argue against simply omitting the drug.

Why adults may pay a relapse price

The authors offer an explanation for why relapse emerged here when it did not dominate earlier pediatric analyses. In pediatric populations, they write, patients are heavily pretreated and mostly minimal residual disease negative before transplant, so the main observable harm of excess exposure was transplant-related mortality. Adults may depend more on a graft-versus-leukemia effect, which higher post-transplant rATG exposure could compromise.

They also speculate on why low exposure beats no exposure for GVHD control. Because rATG is polyclonal and includes targets on dendritic cells, one proposed mechanism is dendritic cell depletion reducing antigen presentation that initiates GVHD; another is that rATG dampens inflammation in GVHD target organs. The paper presents these as potential mechanisms, not as tested findings.

The authors state they could not analyze minimal residual disease or pre-transplant chemotherapy in this cohort because of limited data, so the relapse explanation remains a hypothesis rather than a demonstrated pathway.

Retrospective modeling limits how far this travels

The study is retrospective and registry-based, pooling de-identified data from multiple institutions, and the authors present this as validating prior single-center and combined-center work rather than as a prospective test. Exposure values come from a pharmacokinetic model, and absolute lymphocyte counts were missing for some patients, in which case the authors substituted expected values according to rules tied to the day of rATG administration.

CD4 data were also limited: 207 patients alive with missing day-100 CD4 values or lost to follow-up before day 100 were excluded from that analysis, while 35 patients who died within 100 days were counted as lacking early reconstitution. The cohort was 93% adult (303 of 325 patients over 18), so the results describe mainly adult T-replete transplant recipients with hematological malignancies, not pediatric or T-cell-depleted settings.

The rATG dosing schedule itself followed institutional practice and its details were not available, which means the exposure bands reflect what centers actually did rather than a controlled dosing strategy.

The practical target is low exposure, not absent exposure

The paper's most decision-relevant comparison is not simply high versus low exposure but low exposure versus no rATG at all. Group A, below 30 AUxd/mL, had the highest five-year survival (67%), the lowest relapse (27%) and the highest CD4 reconstitution (73%), yet the no-ATG group did worse on survival (51%) and had more GVHD-related deaths. That combination is why the authors conclude that targeting lower but non-zero post-transplant exposure may improve outcomes.

The authors connect this to model-based dosing, the same nomogram used in the PARACHUTE and PRAISE IR trials, which sets rATG dose and start date from weight, absolute lymphocyte count, stem cell source and transplant type. They describe it as an easy, low-cost tool usable even by centers without the InsightRx platform, and note that an upcoming BMT CTN 2502 IMPACT trial will assess model-based dosing of several conditioning agents, including this rATG nomogram, in patients with inborn errors of immunity and sickle cell disease.

For a reader weighing this evidence, the actionable threshold is the 30 AUxd/mL boundary and the contrast with zero exposure, both specific to this mainly adult, T-replete, registry-based cohort. The relapse hazard ratios for groups B and C are the strongest adjusted signals in the paper; the survival advantage of group A over no ATG is suggestive but not statistically separated in the adjusted model.

Sources used in this answer