In a prospective cohort trial including patients with solid tumors (44.5 %) and hematological neoplasia (55.5 %), Wedding et al. identified comorbidity as an independent prognostic factor of survival (HR = 1.424; 95 % CI, 1.012–2.003) in addition to age (HR = 1.019; 95 % CI, 1.007–1.032), tumor type (HR = 1.832; 95 % CI, 1.314–2.554), and performance status (HR = 1.455; 95 % CI, 1.059–2.000) [20].

A study of patients admitted to the ICU for hematologic and solid tumors (mostly lymphomas and leukemias) identified Charlson score as an independent predictor of ICU and hospital mortality [32].

Comorbidity was also associated with hospitalizations for neutropenia in older NHL patients. In the Iowa SEER/Medicare database, 21.7 % of patients aged 66 and older treated with chemotherapy had a hospitalization for febrile neutropenia; 41 % were hospitalized during the first cycle and 22 % during the second cycle. A positive Charlson score was an independent predictor with an associated adjusted hazard ratio of 1.50 (95 % CI, 1.03–2.17) [33]. A study in 1,355 community-treated patients of all ages treated with CHOP identified liver comorbity as a risk factor for hospitalization for febrile neutropenia, along with age >65, albumin <3.5, baseline absolute neutrophil count <1,500, planned dose intensity ≥80 % standard, and no early use of G-CSF [34]. A review of the 1992–2002 SEER data for the impact of prophylactic growth factors in 13,283 patients aged 65 and older (median 74.9) receiving chemotherapy for non-Hodgkin’s lymphoma recorded comorbidity according to the Klabunde index [35]. On multivariate analysis, comorbidity was associated with an increased risk of febrile neutropenia and documented infection, and a worse OS. The odds ratio (OR) was 1.23 (1.11–1.31) for a score of 1 and 1.15 (1.01–1.31) for a score of 2 or more for febrile neutropenia; 1.27 (1.16–1.39) for a score of 1 and 1.67 (1.50–1.86) for a score of 2 or more for infection. The OR of death was 1.28 (1.21–1.34) for a score of 1, and 1.75 (1.65–1.86) for a score of 2 or more. A retrospective study analyzed NHL and prostate patients treated at Moffitt for the impact of hyperglycemia on toxicity from chemotherapy [36]. The subgroup of 162 NHL patients were all treated with CHOP-rituximab (CHOP-R). Comorbidity, as measured with the CIRS-G, did not impact the occurrence of severe toxicity, whereas in prostate cancer patients the total CIRS-G score was associated with grade 3–4 nonhematologic toxicity. In NHL patients, grade 3–4 nonhematologic toxicity was associated with hyperglycemia, both at baseline and during treatment.

An international study assessed the impact of comorbidity on the quality of life (QOL) of CLL patients: 1,482 patients answered a web questionnaire that included an evaluation of their Charlson score [37]. Overall, CLL patients had a similar QOL to that of the general population, but their emotional wellbeing scores were dramatically lower. Factors associated with lower overall QOL on multivariate analysis were older age, greater fatigue, severity of comorbid illnesses, and undergoing current treatment for CLL. Wedding et al. identified that elderly cancer patients with severe comorbidities (CIRS-G grade 3 or 4) had a worse quality of life, independently from functional status [38]. Forty-six percent of these patients had hematologic malignancies.

An epidemiologic study by the Eindhoven cancer registry was conducted in Hodgkin’s and NHL patients aged 60 and older. It identified a prevalence of comorbidity by Charlson score of 58 % for NHL and 62 % for Hodgkin’s patients. This was associated with a decline in administration of chemotherapy and a 10–20 % decline in 5-year survival [39]..

Although transplant strategies are mostly offered to younger patients, transplantation is increasingly used in the fit elderly. Autologous transplant for multiple myeloma is a case in point. Fit patients in their lower seventies with high-risk disease are sometimes offered autologous or reduced intensity conditioning (RIC) allogeneic transplant. Wildes et al. compared patients aged 60 and over with younger patients for outcome of autologous stem cell transplant for relapsed NHL. They identified that comorbidity rated by Charlson score, rather than age, predicted transplant-related mortality and OS [40]. Labonté et al. compared the performance of the Charlson score, the HCT-CI [23], and the modified pretransplantation assessment of mortality (mPAM) in patients receiving autologous transplant for multiple myeloma [41]. All indexes performed similarly in being associated with serious organ toxicity and length of stay. Artz et al. compared the performance of Charlson and the Kaplan-Feinstein score (KFS) in patients receiving RIC for NHL (median age 52, range 17–70) [42]. The KFS was more sensitive and more strongly associated with transplant-related mortality. When combining the KFS and ECOG performance status (PS), a high-risk category could be identified. The patients with a KFS >2, and ECOG PS >1, or an alteration on both scores, had 4.1 times the risk of transplant-related mortality compared to the low risk group (50 % at 6 months vs 15 %). OS was also significantly reduced in older patients. Pollack et al. reviewed the impact of comorbidity measured by HCT-CI in NHL patients receiving RIC [43]. The median age was 53 years (range 32–74). Patients with 3+ comorbidities had more transplant-related mortality (26.3 % vs 4.5 % at 100 days, and 36.8 vs 13.6 % at 1 year). There was no association with disease-related mortality. Farina et al. similarly reviewed the association of HCT-CI with outcome in patients receiving RIC for lymphoma or myeloma [44]. HCT-CI was an independent determinant of OS (P < .001), PFS (P = .002), and non-relapse mortality (P = .03). Karnofsky PS was the other independent predictor for OS and non-relapse mortality. The effect was similar for lymphomas and myeloma.