Abstract
Introduction Infectious complications are a major cause of morbidity and mortality in acute myeloid leukemia (AML). While cytarabine and an anthracycline (7+3) is standard induction, lower intensity regimens such as hypomethylating agents (HMA) alone or with venetoclax (HMA/Ven) are used in older or unfit patients. Comparative data on infectious risk and mortality between induction regimens remain limited.
Methods We retrospectively analyzed 251 newly diagnosed adult AML patients who received induction with 7+3, HMA, or HMA/Ven at the University of Maryland Medical Center between April 2013 to January 2022. For HMA/Ven, Ven was administered for 14, 21, or 28 days. Data was collected from electronic records. Neutropenia was defined as moderate (ANC <1000/μL) or severe (ANC <500/μL). Infections were documented using established criteria based on clinical, laboratory, microbiologic, and radiologic evidence.
Results Subjects in 7+3 were younger (mean 54.4) than HMA and HMA/Ven (69.0 and 68.0; p<0.001). ECOG ≥2 was more common in HMA and HMA/Ven (32.8% and 31.0%) versus 7+3 (15.5%; p=0.011).
Baseline ANC did not differ across groups (median 0.79 HMA, 0.67 HMA/Ven, 0.55 7+3; p=0.45). ANC nadir was lowest in 7+3 (0), followed by HMA/Ven (0.02), and HMA (0.1; p<0.001), Time to ANC nadir was shortest in 7+3 (median 9 days), followed by HMA/Ven and HMA (13.5 and 15 days; p<0.001).
Moderate neutropenia occurred in 89.2% of subjects, with highest rates in 7+3 and HMA/Ven (100% and 95.2%) and lowest in HMA (78.8%; p<0.001). Severe neutropenia occurred in 85.3% of subjects, with highest rates in 7+3 and HMA/Ven (100% and 90.5%) and lowest in HMA (72.9%; p<0.001). Median time spent in moderate neutropenia (43 days HMA, 43 HMA/Ven, 27 7+3; p=0.096) and severe neutropenia (37.5 HMA, 36 HMA/Ven, 27 7+3; p=0.25) was longer in lower-intensity groups, but did not reach statistical significance.
Incidence of infection within 60 days occurred in 45.6% of subjects, highest in 7+3 (58.3%), compared to HMA and HMA/Ven (37.9% and 42.9%, respectively, p=0.014). Median time to first infection was 10 days, similar across groups. Most common infections were PNA (n=37; 13 in 7+3, 23 HMA, 1 HMA/Ven), bacteremia (n=28, 11, 8, and 9 respectively), and fungal (n=25, 14, 8, and 3). Overall, 23.9% of subjects required ICU care within 60 days of induction, with similar rates across groups. Infection was associated with increased 60-day mortality (HR 2.71, p=0.002) and reduced overall survival (HR 1.36, p=0.034). Sixty-day mortality was highest in HMA (24.6%) and HMA/VEN (23.8%) versus 7+3 (5.1%; p=0.001). Kaplan-Meier analysis showed trend toward differential overall survival by regimen (log-rank p=0.077), with highest survival in 7+3. In univariate Cox regression, HMA was associated with inferior overall survival compared to 7+3 (HR 1.47, p=0.026), while HMA/Ven had a non-significant trend (HR 1.38, p=0.14).
Discussion Induction regimen selection was associated with distinct patterns of cytopenia, infection, and mortality. Patients receiving 7+3 experienced the deepest neutropenia and highest incidence of infection, but lowest 60-day mortality, likely reflecting selection bias.
Infection itself was associated with increased 60-day mortality and worse survival outcomes, highlighting its clinical impact. Both 7+3 and HMA/Ven were associated with profound cytopenias, with HMA/Ven demonstrating more pronounced myelosuppression than HMA, evidenced by lower ANC nadirs and higher rates of moderate/severe neutropenia. However, this did not translate differential infection incidence or mortality within 60 days.
Our findings highlight the complex interplay between induction regimen intensity, cytopenia, infectious complications, and early mortality. While HMA/Ven is associated with greater myelosuppression than HMA alone, this did not translate into significantly higher infection rates or 60-day mortality. Additionally, while 7+3 was associated with greater myelosuppression and incidence of infection, this regimen had the lowest 60-day mortality, likely reflecting baseline differences in patient fitness that limit direct comparison. These results underscore the importance of individualized induction strategies that incorporate regimen associated complications.
