Background: Following HSCT, regardless of whether it is autologous or allogeneic, engraftment fever (EF), which mimics “hyper-acute” GVHD in allogeneic HSCT when associated with skin rash, is often reported to occur at the time of neutrophil recovery. It has been speculated that it may be easier to distinguish EF from infections or “hyper-acute GVHD” in HSCT with a reduced-intensity regimen (RIST) compared to conventional HSCT with a myeloablative regimen, since the former is associated with a lower incidence of regimen-related toxicities including severe mucositis, myelosuppression and organ damage. To clarify this point, we analyzed data from patients who underwent RIST with peripheral blood cells from an HLA-matched sibling donor between April 2000 and August 2003 at National Cancer Center Hospital in Japan.

Patients and Methods: There were 89 patients [55 men, 34 women; median age, 52 years (range, 16–65)] and the diagnosis was AML in 16, CML in 4, MDS in 13, ML in 27, solid tumor in 23 and others in 6. The conditioning regimen consisted of fludarabine 180 mg/me2 or cladribine 0.66 mg/kg plus busulfan 8 mg/kg, with (n=30) or without (n=59) rabbit antithymocyte globulin (ATG) 5 mg/kg. GVHD prophylaxis was cyclosporine alone (n=86), or cyclosporine + methotrexate (n=3). All of the patients but one received IV granulocyte-colony stimulating factor daily from day 6 to engraftment. Engraftment was defined as the first day the absolute neutrophil count exceeded 500 μl/l for two consecutive days. Engraftment fever was defined as non-infectious fever ( ≥38.3°C without clinical or laboratory signs of infection ) within 4 days before or after engraftment.

Results: Fever was documented in 21 patients in the early engraftment phase and 6 were suspected or proven to have infection; thus, 15 patients (17%) were believed to have EF. In this cohort, EF was associated with skin rash in 5 patients, weight gain in 6, liver dysfunction in 2, and transient hypoxia in 2. In comparing patients with (n=15) and without (n=74) EF, there were no differences in sex, age, use of ATG, median number of infused CD34+ cells (3.39 vs 3.40 x 10e6/kg), or median day to engraftment (day11 vs day11). However, the incidence of EF was significantly different among tumor types; the frequency was 8 of 27 (30%) in malignant lymphoma, compared with 7 of 62 (11%) in patients with other types of tumor (Pearson’s X2 p=0.034 ). EF was not seen in patients with AML (p=0.047 vs other, including ML). There was no significant correlation between EF and the incidence of acute GVHD (grade II-IV) (p=0.199). In a multivariate analysis, ML significantly predicted the development of EF (Table 1).

Conclusion: These findings suggest that there may be a close relationship between tumor type and EF, and in this cohort, EF was significantly associated with ML.

Influence of the clinical factors on the development of EF in a multivariate analysis

valiableHR95% CIP value
ATG: rabbit antithymocyte globulin, ML: malignant lymphma 
age (<40 vs ≥ 40 years ) 0.792 0.226–2.776 0.716 
ATG (with vs without) 1.574 0.405–6.120 0.512 
ML (vs other) 3.823 1.061–13.776 0.040 
valiableHR95% CIP value
ATG: rabbit antithymocyte globulin, ML: malignant lymphma 
age (<40 vs ≥ 40 years ) 0.792 0.226–2.776 0.716 
ATG (with vs without) 1.574 0.405–6.120 0.512 
ML (vs other) 3.823 1.061–13.776 0.040 

Author notes

Corresponding author

Sign in via your Institution