Introduction: Previous studies have demonstrated that frailty is associated with mortality among patients with non-Hodgkin lymphoma including diffuse large B-cell lymphoma (DLBCL). However, no studies have examined frailty in an unselected population-based sample of patients with DLBCL, nor have data on health care utilization been considered as a potential mediator of this relationship.

Objective: To determine whether frailty is associated with one-year survival in an unselected population of patients with DLBCL, and examine whether its impact is mediated by health care utilization during chemotherapy treatment.

Methods: A retrospective cohort study was conducted using population-based health care data in Ontario, Canada. Patients >65 years diagnosed with DLBCL or transformed follicular lymphoma between January 2006 and December 2017 and receiving first-line chemo-immunotherapy were included. Frailty was defined by modifying a previously validated score developed for use with population-based data in Ontario, comprising 30 multidimensional variables (McIsaac, Ann Surg. 2019;270(1):102-108). Patients were categorized as "frail" (score >0.21) vs. "non-frail" (score ≤0.21). Covariates included age, number of comorbidities based on the Johns Hopkins Aggregated Diagnosis Groups (ADGs), and health care utilization during chemotherapy (defined as emergency department (ED) visit or inpatient hospitalization not resulting in death during treatment, and analysed as a time-varying covariate). Cox regression was performed to examine the association between frailty and one-year mortality (primary outcome). Secondary outcomes included health care utilization, chemo-immunotherapy exposure, and cause of death.

Results: 5,527 patients were included in the study. 5,216 patients (94%) had de novo DLBCL, and 311 (6%) of patients had transformed follicular lymphoma. The median age was 75 years (IQR 70-80), and 48% (N=2672) were female (Table 1). 2,699 (49%) of patients were classified as frail (Table 2). Frail patients tended to be older (median age 76 (IQR 71-81) vs. 74 years (IQR 70-79)).

The difference in mortality between frail and non-frail patients was most pronounced in the initial year following start of treatment (Figure 1a). Within 90 days of first-line rituximab, 14% (N=370) of frail vs. 7% (N=185) of non-frail patients had died (p<0.0001). Within one-year of first-line treatment, 32% (N=868) of frail patients had died compared to 20% (N=553) of non-frail patients (unadjusted HR 1.8, 95% CI 1.6-2.0, p<0.0001, Figure 1b). Among frail patients who died within 1 year (N=868), 34% (N=298) had only received 1 cycle of chemotherapy. In multivariable modelling controlling for age, number of ADG comorbidities, and health care utilization during chemotherapy, frailty (binary exposure) remained independently associated with one-year mortality (adjusted HR 1.6, 95% CI 1.5-1.8, p<0.0001).

The relationship between frailty and survival remained consistent when measured in quartiles (HR 1.6 (95% CI 1.3-1.9) for Q2, 2.0 (95% CI 1.7-2.4) for Q3, 2.7 (95% CI 2.3-3.2) for Q4, p<0.0001, Figure 1b). Frail patients were significantly more likely to receive only 1 cycle of chemotherapy than non-frail patients (14% vs. 7%, p<0.0001). Frail patients also had higher health care utilization during chemotherapy (mean ED visits 0.75 + 1.47 vs. 0.59 + 1.17, p<0.001; mean hospitalizations 0.9 + 1.12 vs. 0.72 + 1.05, p<0.001). Frail patients were also more likely to die of DLBCL (38.3 vs. 29%, p<0.0001).

Conclusion: Frailty is significantly associated with one-year mortality in patients with newly diagnosed DLBCL, even after adjusting for age, comorbidities, and health care utilization. Frailty appears to be associated with poor tolerability of chemotherapy and a higher likelihood for requiring acute hospital-based care, and future analyses will explore whether this is related to patients suffering increased treatment-related toxicity. Future analyses of these data will also address whether frail patients who die within one-year of first-line treatment have different clinical characteristics compared to frail patients who survive beyond one year. Future prospective studies may help clinicians understand whether any frailty-related variables are modifiable and the role of alternative treatment strategies for vulnerable patients.


Prica:astra zeneca: Honoraria; seattle genetics: Honoraria; Gilead: Honoraria.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution