Primary Prophylaxis with Pegfilgrastim in Patients with Newly Diagnosed Diffuse Large B-Cell Lymphoma: A Nation-Wide, Propensity Score Based Analysis

Background: In the daily practice, clinical effectiveness of pegfilgrastim compared to that of daily filgrastim in patients with malignant lymphoma is still unclear. This study aimed to clarify the effectiveness of pegfilgrastim versus daily filgrastim, using a national inpatient database in Japan.

Study Design and Methods: We retrospectively reviewed 18095 patients with newly diagnosed diffuse large B-cell lymphoma (DLBCL) who received the first R-CHOP treatment between July 2007 and March 2017. Patients who received primary G-CSF prophylaxis were divided into pegfilgrastim group and filgrastim group. Outcomes included incidence of febrile neutropenia, all-cause in-hospital death, length of hospital stay, and total costs. To account for measured confounding, patients' characteristics were adjusted using two different propensity score utilizing methods: propensity score matching (PSM) and stabilized inverse probability of treatment weighting (IPTW). Instrumental variable (IV) analysis was also performed to account for unmeasured confounding.

Results: We identified 1299 patients in the pegfilgrastim group and 2203 patients in the filgrastim group. FN occurred in 380 of the 3502 (11%) total patients with a median of 9 days (interquartile range [IQR] 7-10) after the initiation of the chemotherapy. All-cause in-hospital death occurred in 94 (2.7%) of 3502 patients. Median of hospital stay was 17 days (IQR 13-22), and total medical costs during hospitalization were 7300 (IQR 5700-9500) in USD.

By one-to-one PSM, 1294 patients for both the pegfilgrastim and filgrastim groups were selected. Pegfilgrastim was administered 2 days (median, IQR 1-3) after the chemotherapy. Daily filgrastim administration was initiated 5 days (median, IQR 3-6) after the chemotherapy, and continued for 6 days (median, IQR 2-7). Compared with the filgrastim group, the pegfilgrastim group showed significant lower risk of the FN incidence (risk difference 6.1%, 95% confidence interval [CI] 4.1-8.1) and all-cause in-hospital mortality (risk difference 0.9%, 95% CI 0.05-1.8). In similar, stabilized IPTW showed that pegfilgrastim group was significantly associated with the lower risk of FN (risk difference 7.0%, 95% CI 5.1-8.9) and in-hospital mortality (risk difference 1.8%, 95% CI 0.9-2.7). In PSM, the length of hospital stay and total costs during hospitalization were also significantly decreased in the pegfilgrastim group compared to the filgrastim group (percent reduction 34% [95% CI 31-37], percent reduction 12% [95% CI 9-15], respectively). Stabilized IPTW also showed the significant percent reduction of the pegfilgrastim group to the filgrastim group for length of hospital stay and total costs (31% [95% CI 25-35], 19% [95% CI 13-24], respectively).

In the sensitivity analyses, IV methods showed significantly lower FN susceptibility (odds ratio 0.14, 95% CI 0.07-0.26). However, pegfilgrastim was not associated with the reduction in susceptibility of in-hospital mortality (odds ratio 0.93, 95% CI 0.23-3.70). The ratio of the pegfilgrastim group to the filgrastim group for length of hospital stay and total costs were 55% (95% CI 51-60) and 78% (95% CI 73-84), respectively.

Conclusion: Pegfilgrastim contributed to lower susceptibility of FN. Total length of hospital stay and medical costs were also decreased in the pegfilgrastim group.