Risk Factors for Early Death after Immunochemotherapy in Older Patients with Diffuse Large B-Cell Lymphoma (DLBCL)

Background: DLBCL is a curable malignancy for a majority of patients >65 years old, but they are at higher risk of toxicities and potential early death after chemotherapy. We evaluated risk factors for death and other severe adverse events during the first month of rituximab-based immunochemotherapy among Medicare beneficiaries with DLBCL, using Medicare claims linked to the Surveillance, Epidemiology and End Results (SEER-Medicare) database.

Methods: We selected patients aged ≥65 years, diagnosed with DLBCL between 2003 and 2012, excluding cases diagnosed by autopsy, with incomplete Medicare claims, primary central nervous system lymphoma, or first chemotherapy in the inpatient setting. Eligible patients initiated an outpatient regimen containing rituximab, cyclophosphamide, vincristine, with doxorubicin (R-CHOP-like), or a version without anthracycline. The primary outcome was defined as death within 30 days from the start of chemotherapy. Secondary outcomes included: admission to a hospital, critical care unit or a nursing home, and a cardiac event within this 30-day window. Risk factors were studied in multivariable logistic regression models adjusting for age, sex, race, Medicaid (poverty) status, DLBCL stage, B symptoms, and receipt of anthracycline. We stepwise-selected validated indicators of performance status and comorbidities (based on Medicare claims from the preceding year) which showed strongest associations with the outcomes, using Bonferroni correction. Results are expressed as odds ratios (OR) with 95% confidence intervals (CI).

Results: We identified 6,425 patients with median age of 76 years, of whom 49% were women, 88% white non-Hispanic, 44% had stage III/IV lymphoma, and 84% received R-CHOP-like regimen. Cumulative incidence of death was 2.1% at day 30, and 12.3% at day 180 after chemotherapy. Prophylactic granulocyte growth factor was administered to 63% of patients during the first treatment cycle.

In a multivariable model, the risk of early death was significantly higher for patients older than 75 years (OR vs. 65-70 years, 2.07, CI 1.04-4.14) or ≥80 years (OR, 3.22, CI, 1.66-6.24), and those with B-symptoms (OR, 1.90, CI, 1.13-3.20), but there was no significant difference by sex, race, stage, poverty status or anthracycline use. The risk was also associated with chronic kidney disease (OR, 3.37, CI, 2.07-5.46), poor performance status (OR, 2.08, CI, 1.22-3.54), prior use of walking aids (OR, 2.26, CI, 1.38-3.71), prior hospitalization (OR, 1.68, CI, 1.14-2.47), or a history of upper endoscopy (OR, 1.73, CI, 1.19-2.51). The risk of early death was only 1.1% for patients with <2 of those factors (79% of cases) while it was 7.2% for those with ≥4 factors (6.8% of cases).

The risk of other outcomes within 30 days of chemotherapy was: 24% for hospitalization (8% with a diagnosis of febrile neutropenia), 11% for a cardiac event, 8% for critical care, and 3% for nursing home admission. The median time to hospitalization was 9 days (interquartile range, 7 to 14). The risk of early hospitalization was associated with increasing age, female sex (OR, 1.14, CI, 1.01-1.28), B symptoms (OR, 1.49, CI, 1.27-1.74), stage IV lymphoma (OR, 1.20, CI, 1.02-1.40), prior myocardial infarction (OR, 2.56, CI, 1.57-4.18), renal disease (OR, 1.66, CI, 1.31-2.11), prior hospitalization (OR, 1.50, CI, 1.32-1.69) or a history of upper endoscopy (OR, 1.20, CI, 1.05-1.38).

Conclusions: Among older DLBCL patients who receive contemporary rituximab-based chemotherapy, 1 in 50 die during the first month of treatment, and 1 in 4 are hospitalized. Easily identifiable factors can distinguish groups at highest risk of early death, who may benefit from preventive strategies such as the prephase treatment (Pfreundschuh, Blood 2010), or from novel, personalized therapeutic approaches. Despite national guidelines, prophylactic granulocytic growth factors are not administered to over 1/3 of patients, indicating an opportunity to lower the risk of adverse events. Withholding doxorubicin was not associated with a lower risk of early death or hospitalization.