Lower c-D-Index in Acute Myeloid Leukemia with Febrile Neutropenia during Chemotherapy

Background: The D-index is a clinical tool used to assess the total severity of a patient's neutropenia. The D-index was originally proposed by Portugal et al. as a quantitative evaluation of the severity of a patient's neutropenia. The utility of the D-index has been validated in various clinical settings. The D-index has recently been investigated as a potential predictor of febrile neutropenia (FN) and infections resulting from neutropenia. The D-index varies according to the ending period for the calculation. If the period until the onset of FN is included, this is called the cumulative D-index (c-D-index) and reflects an accumulation of neutropenia until FN. The c-D-index is thought to represent more accurate and ongoing severity of patient's neutropenia.

Purpose: To examine whether the D-index/c-D-index are a useful predictor of the onset of FN or various infections in acute myeloid leukemia (AML) patients treated with chemotherapy.

Methods: This study included consecutive AML patients treated in our hospital between November 1998 and December 2018, retrospectively. The inclusion criteria were age 18-74 years at the onset, Eastern Cooperative Oncology Group (ECOG) Performance Status (PS) 0-2, hematologic diagnosis of AML, and treatment with Japan Adult Leukemia Study Group (JALSG) 97 or JALSG 201 protocol regimens. We used the JALSG disease risk as the risk factors at the onset of AML. We extracted the cases that fulfilled all the above criteria. We collected all FN events from the start of a patient's induction chemotherapy to the end of the final consolidation therapy. We collected complete FN events during chemotherapy and evaluated associations with the severity of FN and infectious events. FN was defined as a fever ≥38°C with a neutrophil count <500 cells/μL. If a fever ≥38°C was observed again after the body temperature had decreased to <37.5°C for over 48 h, it was defined as a new febrile episode.

Results: A total of 51 patients (25 women, 26 men; median age 53 years, range 18-74 years) met the eligibility criteria. These patients experienced 171 FN events (68 induction and 103 consolidation chemotherapy episodes). The JALSG prognosis score indicated good risk in 16 subjects, intermediate risk in 19, poor risk in 14, and unknown in 2. The response rate to induction chemotherapy was 56.8% (29/51), and 80.4% (41/51) of patients were alive at the end of follow-up. In the univariate analysis of the risk factors for the onset of FN, the D-index, the c-D-index, PS, remission status, diabetes mellitus, and neutrophil count at the start of chemotherapy were identified. Multivariate analysis showed the c-D-index, but not the D-index, as a risk factor for the onset of FN (P = 0.00903). The statistically significant explanatory factors for microbiologically proven infections were high-dose Ara-C (P = 0.00103) and diabetes mellitus (P = 0.0251). The most prolonged period of FN was observed in subjects in whom the site of infection was respiratory comparing to other sites; blood stream infection (P = 0.0018), oral cavity (P = 0.0026), and gastro-intestine (P = 0.047), respectively. The D-index was significantly higher in patients with FN (P = 0.0028). Conversely, the c-D-index was significantly lower in patients with FN (P = 0.0016). The mean c-D-index in patients with FN was 3842 ± 1811 (the 75% tile range is 5517), and the mean c-D-index in patients without FN was 5597 ± 4289. The median day of the onset of FN was day 14 ± 5. Among the patients with FN, c-D-index were lower in patients with documented infection compared to those whom without it (2681 vs 4295, P = 0.0006).

Conclusion: The D-index did not predict the onset of infection. However, the c-D-index can predict the onset of FN. The c-D-index depict the onset of FN within the mean value of 3842 at the median day 14. Among FN events, the onset of microbiologically proven infections were significantly associated to high-dose Ara-C and diabetes mellitus. Among FN patients, the documented infection occurred in significantly lower c-D-index patients. This may suggest documented infection will be diagnosed in the earlier stage of febrile episode among patients with AML during chemotherapy. Respiratory symptoms on the onset of FN are a risk factor for the prolonged duration of FN. Future research should focus on whether comorbidity at diagnosis and the value of the D-index at the onset of FN are associated with poor treatment outcomes in AML.