Association of Glasgow Prognostic Score with frailty, mortality and adverse health outcomes in older patients with cancer: A prospective cohort study

Introduction

To balance benefits and risks of cancer treatment in older patients, prognostic information is needed. The Glasgow Prognostic Score (GPS), composed of albumin and C-reactive protein (CRP), might provide such information. This study first aims to investigate the association between GPS and frailty, functional decline, and health-related quality of life (HRQoL) decline as indicators of health problems in older patients with cancer. The second aim is to study the predictive value of GPS for mortality, in addition to clinical predictors.

Materials and Methods

This prospective cohort study included patients aged ≥70 years with a solid malignant tumor who underwent a geriatric assessment and blood sampling before treatment initiation. GPS was calculated using serum albumin and CRP measured in batch, categorized into normal (0) and abnormal GPS (1–2). Outcomes were all-cause mortality and a composite outcome of decline in daily functioning and/or HRQoL, or mortality at one year follow-up. Daily functioning was assessed by Activities of Daily Living and Instrumental Activities of Daily Living questionnaires and HRQoL by the EQ-5D-3L and EQ-VAS questionnaires.

Results

In total, 192 patients with a median age of 77 years (interquartile range 72.3–81.0) were included. Patients with abnormal GPS were more often frail compared to those with normal GPS (79 % vs. 63 %, p = 0.03). Patients with abnormal GPS had higher mortality rates after one year compared to those with normal GPS (48 % vs. 23 %, p < 0.01) in unadjusted analysis. Abnormal GPS was associated with increased mortality risk (hazard ratio 2.8, 95 % CI 1.7–4.8). The area under the receiver operating characteristics curve of age, distant metastasis, tumor site, comorbidity, and malnutrition combined was 0.73 (0.68–0.83) for mortality prediction, and changed to 0.78 (0.73–0.86) with GPS (p = 0.10). The composite outcome occurred in 88 % of patients with abnormal GPS versus 83 % with normal GPS (p = 0.44).

Discussion

Abnormal GPS was associated with frailty and mortality. The addition of GPS to clinical predictors showed a numerically superior mortality prediction in this cohort of older patients with cancer, although not statistically significant. While GPS may improve the stratification of future older patients with cancer, larger studies including older patients with similar tumor types are necessary to evaluate its clinical usefulness.

Trial Registration

The TENT study is retrospectively registered at the Netherlands Trial Register (NTR), trial number NL8107. Date of registration: 22-10-2019.