Our study showed that the percentage of patients with normal G8 score of more than 14 was only 6.8% in older patients with GI cancer. The rate was much lower in patients registered in our study compared with 17–32% in previous reports involving older patients with solid tumors. 5,6,9−11 In addition, the cut off value of G8 score did not work to predict either OS or SAEs, which was not consistent with previous reports. 6,7 These results indicated that the value of 14 which is defined as cut off value of G8 score may not be clinically useful in older patients with GI cancer, most of whom had abnormal G8.
One reason for the low G8 score in patients with GI cancer was digestive symptoms, such as appetite loss, which causes malnutrition and low BMI. The nutritional status is well reflected in the score, since the G8 score consists of an MNA questionnaire, which primarily focuses on nutrition. A previous study showed that the proportion of abnormal G8 score was significantly different among cancer types. 10 GI cancers, which are prone to low nutrition, are predicted to have lower G8 score compared with other cancer types. In the present study, the score for items related to digestive symptoms was lower compared to patients with other cancer types, whereas the score for other items was similar.11 Based on these results, the lower score in items related to digestive symptoms would contribute to the lower G8 score in older patients with GI cancer. The mean value for BMI of Japanese older patients was about 22 which was significantly lower compared with that of Western patients, which may contribute to the low G8 score. 15,16 A few Japanese reports recommended a revised G8 cut-off score of 9.5–11, however, it was controversial because of the heterogeneity of cancer types in small single institutional study.11–14 The BMI varies among ethnic groups or cancer types; therefore, use of the same cut-off value of 14 may be inappropriate for all cancer patients in the world. We should consider G8 scores of each item unrelated to digestive symptoms when evaluating older patients with GI cancer.
In our study, the patients receiving CT had significantly higher total G8 score as well as better score in terms of mobility, neuropsychological problems, and age compared to those without CT, even with abnormal G8 score. We may put higher value on the three items than others as a reference to decide the tolerability for CT: better mobility, normal neuropsychological function, and younger age. These items would be clinically useful determinants when considering the treatment plan for older or vulnerable patients with GI cancer. In addition, the rate of independent IADL was significantly higher in patients with CT compared to those without CT, even with abnormal G8 score. The IADL is one of the important GA tools that is directly linked to independence of daily living. The IADL consists of question of ability to care for oneself including responsibility for own medications, which can affect the eligibility of CT. Moreover, there was significant difference in OS according to IADL even in G8 abnormal patients. The IADL was scarcely affected by malnutrition in contrast to G8 score that showed no association with OS. Most older patients with GI cancer suffered from malnutrition that led G8 abnormal score; therefore, IADL may be clinically useful to predict prognosis in older GI cancer patients. These findings suggest that an assessment using the GA tools which are little affected by malnutrition may potentially have the clinical utility to determine the optimal treatment plan more accurately in older patients with GI cancer.
In this study, about half of older patients started CT with dose reduction and they had acceptable toxicities even with abnormal G8 score. Recently, the favorable efficacy has been reported in older cancer patients treated with reduced doses.17,18 In our study, 83% of the patients underwent dose reduction during entire course and most continued CT safely and the rate of SAEs and discontinuation was similar to that of previous reports in spite of most patients with abnormal G8 score.19–21 Furthermore, the OS was longer in patients with palliative CT compared to those without CT, even with dose reduction. These results indicate that older GI cancer patients with abnormal G8 score would have a chance of receiving CT safely and effectively through adjusting the dose of the drugs.
There are several limitations in this study. First, this was a retrospective study with gastric, pancreatic, and colorectal cancer patients in a single institution. Therefore, there were several biases including patient selection and the various treatment regimens that could affect OS and AE frequency. Second, treatment choice was affected by multiple factors regardless of the G8 score in clinical practice; therefore, it would be necessary to verify the efficacy of G8 scoring for judging the tolerability of CT by randomized controlled study. Third, as the sample size of the patients with normal G8 score was small, we could not compare OS between patients with normal and abnormal G8 score. Therefore, we compared OS according to G8 score sub-group based on the quartile using the Kaplan-Meier method to evaluate the association between G8 and survival time. Finally, we could not obtain detailed information about intervention such as nutritional guidance and rehabilitation. However, we obtained data regarding dose reduction, which was one of the interventions for patients with abnormal G8 score. 17,18