Association of Geriatric Nutritional Risk Index With Immune Checkpoint Inhibitor Treatment Duration and Adverse Events in Lung Cancer

Abstract

Background/Aim

Compared to conventional cytotoxic anticancer agent-based therapy, treatment with immune checkpoint inhibitors (ICI) significantly prolongs overall survival. The Geriatric Nutritional Risk Index (GNRI) has been used as a new prognostic indicator in cancer. As nutritional status is associated with prognosis and indicates treatment response, we investigated the effect of the pretreatment GNRI on the (1) occurrence of ICI-induced immune-related adverse events (ir-AE) and (2) association with time to treatment failure (TTF) in ICI monotherapy for lung cancer.

Patients and Methods

In this study, 127 patients with lung cancer who were treated with ICI monotherapy were retrospectively enrolled. Based on a cutoff value of 92 for the GNRI, we investigated intergroup differences in the occurrence of adverse events and their association with TTF in the High-GNRI (≥92) and Low-GNRI (<92) groups. For intergroup comparisons, we used the Student’s t-test, Welch’s t-test, Fisher’s direct probability test, and Mann–Whitney’s U-test, and factors with p<0.05 in the intergroup comparison were extracted as explanatory variables.

Results

Based on the pretreatment GNRI, the median TTF was 5.1 months (95%CI=2.4-7.9 months) in the High-GNRI group and 2.3 months (95%CI=1.6-3.1 months) in the Low-GNRI group, with the High-GNRI group having a significantly longer TTF (p<0.01). The incidence of skin rash (p=0.0129) and pruritus (p<0.01) was significantly higher in the High-GNRI group.

Conclusion

Pretreatment GNRI influences the continuation of ICI monotherapy. The High-GNRI group demonstrated a significantly higher frequency of skin lesions, which may have influenced the prolongation of TTF.

In Japan, 126,548 new cases of lung cancer were reported in 2019 and 75,585 deaths in 2020; thus, lung cancer has the second highest incidence rate by cancer site and the highest cancer mortality rate (1). Lung cancer may be treated surgically if detected early; however, in case of metastasis or recurrence, drug therapy is the mainstay of treatment, wherein cytotoxic anticancer agents, molecular-targeted drugs, and immune-checkpoint inhibitors (ICI) are used. Thus, in the treatment of non–small cell lung cancer, molecular-targeted agents and ICI are the mainstay of treatment. In small cell lung cancer, although cytotoxic anticancer agents have been the mainstay of treatment for a long time, ICI can also be used (2). Nivolumab, pembrolizumab, atezolizumab, and durvalumab are used in the treatment of lung cancer. Treatment with ICI significantly prolongs overall survival compared to conventional treatment with cytotoxic anticancer agents (3-6). However, ICI treatment discontinuation or withdrawal occurs when the disease progresses or serious adverse events occur, which results in the difference in treatment duration. In addition, many cancers are more prevalent in the 60-70 age group and in older adults (1), who may, regardless of disease status, have reduced nutritional status. According to the "Summary of the Results of the National Health and Nutrition Survey in Japan, 40" published by the Ministry of Health, Labour and Welfare, the percentage of men and women of all ages with a body mass index (BMI) of less than 18.5 kg/m² is 12.4% for men and 20.7% for women aged 65 years or more who are undernourished (BMI ≤20 kg/m²). This indicates a trend toward an aging-related decline in nutritional status. The proportion of people aged 65 years or older who are undernourished (BMI ≤20 kg/m²) is 12.4% for men and 20.7% for women, indicating a trend of declining nutritional status with advancing age (7). Furthermore, cancer is a typical pathology that induces undernutrition, and many reports indicate that the nutritional status before and during treatment is associated with resistance to treatment, incidence of adverse events, and oncologic prognosis in various cancer types (8,9). The Geriatric Nutritional Risk Index (GNRI) is a nutritional index that was reported by Bouillanne et al. (10) and, compared to other nutritional indices such as the Nutritional Risk Index (NRI), the GNRI is calculated from the serum albumin level and ideal-body weight ratio. This makes the GNRI a more useful nutritional assessment index for older adults (11) and a prognostic indicator of patient outcomes (12). Furthermore, immune-related adverse events (ir-AE) that are caused by ICIs are associated with the efficacy of ICI therapy (13). Therefore, the GNRI may be related to the efficacy of ICI therapy and the incidence of ir-AEs. However, little is known about the impact of GNRIs on the duration of treatment of ICI and the development of side effects. In this study, we investigated the effect of the pretreatment GNRI on the occurrence of adverse events and its association with TTF in ICI monotherapy for lung cancer.

Patients and Methods

Patients. From January 2016 to November 2022, 127 patients with lung cancer treated with nivolumab or pembrolizumab monotherapy at JR Sapporo Hospital and Sapporo Minami Sanjo Hospital were included. Patients who had received cytotoxic chemotherapeutic agents followed by monotherapy were excluded. Based on the records of physicians, nurses, and pharmacists in the electronic medical records, the background and occurrence of adverse events of the patients in the study were investigated retrospectively.

Data collection. The survey items include the patient’s sex, age, weight, height, BMI, performance status (PS), treatment history, most recent white blood cell count, hemoglobin level, platelet count, neutrophil count, aspartate aminotransferase (AST), alanine aminotransferase (ALT), lactate dehydrogenase (LDH), total bilirubin level, serum creatinine level, serum albumin level, serum sodium level, and serum potassium level. The PS was evaluated by using the Eastern Cooperative Oncology Group (ECOG) rating scale (14). Adverse events included hematologic toxicity, increased LDH, hypoalbuminemia, hyponatremia, hypokalemia, anorexia, skin disorders, diarrhea, constipation, fatigue, interstitial pneumonia, thyroid dysfunction, increased AST, increased ALT , increased blood bilirubin, and increased creatinine, as evaluated by Common Terminology The severity of each case was evaluated using the Common Terminology Criteria for Adverse Events (CTCAE) ver5.0. The severity of each case was defined as the highest grade among all courses. The participants were classified into the High-GNRI group (GNRI ≥92) and the Low-GNRI group (GNRI <92) based on the classification by Bouillanne et al. (10) using GNRI evaluation, which is one of the nutritional evaluation indices, and TTF, patient background, and occurrence of side effects were compared between the two groups.

GNRI was calculated using the following formula:

GNRI=14.89×Alb value (g/dl)+41.7×BMI/22

Statistical analysis. For intergroup comparisons, we used the Student’s t-test, Welch’s t-test, Fisher’s direct probability test, and Mann–Whitney’s U-test. Factors with p<0.05 in the intergroup comparison of laboratory values at treatment initiation were identified as candidate risk factors and extracted as explanatory variables. When explanatory variables were continuous variables, we used the calculated cutoff values in ROC analysis to convert the continuous variables to categorical variables. Univariate and multivariate Cox proportional hazards analyses were performed to analyze factors affecting treatment duration. BellCurve for Excel (Social Survey Research Information Co., Ltd., Tokyo, Japan) was used for statistical analysis.

Ethical considerations. This study was conducted with the approval of the Ethics Committee of JR Sapporo Hospital (Approval No. 2022-27) and the Ethics Committee of Sapporo Minamisanjo Hospital (Approval No. R4-4). Owing to its retrospective nature, no written or oral consent was obtained from the research participants. Opt-out information about the study was made available to the participants (posted in the hospital or on the hospital website), and the participants were guaranteed the opportunity to refuse to have the study conducted. We ensured that all patient confidential information were protected. The data were anonymized prior to handling.

Results

Patient characteristics. The total study population comprised 127 participants; there were 74 in the High-GNRI group and 53 in the Low-GNRI group. There were 97 men and 30 women, with a median age of 71 years. The median body weight of the High-GNRI and Low-GNRI groups was significantly lower (p<0.01): 60.0 kg in the High-GNRI group and 50.6 kg in the Low-GNRI group. The median BMI was significantly lower in the Low-GNRI group; BMI was 23.7 kg/m² in the High-GNRI group and 18.9 kg/m² in the Low-GNRI group (p<0.01). The median serum albumin level was 3.7 and 3.0 g/dl in the High- and Low-GNRI groups, respectively, and was significantly lower in the Low-GNRI group (p<0.01). The median serum albumin level was 3.7 and 3.0 g/dl in the High- and Low-GNRI groups, respectively, and was significantly lower in the Low-GNRI group (p<0.01). The median height was 1.61 and 1.64 m in the High-and Low-GNRI groups, respectively, and was significantly higher in the Low-GNRI group (p=0.027), whereas the PS was significantly higher in the Low-GNRI group (p<0.01). There was no significant difference in age, type of ICI, or previous treatment (Table I).

Table I. Characteristics of the study population.

^aMedian (1^st quartile - 3^rd quartile); ^bFisher’s exact probability test; ^cMann-Whitney’s U-test. GNRI: Geriatric Nutritional Risk Index; ECOG PS: Eastern Cooperative Oncology Group performance status.

Time to treatment failure. The median TTF was 5.1 months (95%CI=2.4-7.9 months) in the High-GNRI group and 2.3 months (95%CI=1.6-3.1 months) in the Low-GNRI group, with the High-GNRI group having a significantly longer TTF (p<0.01) (Figure 1).

Figure 1. Comparison of time-to-treatment failure (TTF) between the High and Low Geriatric Nutritional Risk Index (GNRI) groups. The medians for the High-GNRI and Low-GNRI groups were 5.1 (95%CI=2.4-7.9) and 2.3 (95%CI=1.6-3.1) months, respectively, with a significantly longer TTF in the High-GNRI group (p<0.01).

Univariate and multivariate COX regression proportional hazards analysis of TTF risk factors. Univariate and multivariate Cox regression proportional hazards analysis were used to analyze the factors influencing treatment duration. ROC curves we used to determine cutoff values and Cox proportional hazards regression analysis to determine hazard ratios and 95% confidence intervals. The area under the ROC curve for age was 0.526, and the cutoff was 72 years. The cutoff value for GNRI was 92. In the univariate analysis with the Cox proportional hazards model, TTF was longer for women compared to men (HR=0.728, 95%CI=0.460-1.151, p=0.174). Patients aged 72 or above had a significantly longer TTF compared to those under 72 years (HR=0.677, 95%CI=0.462-0.933, p<0.05). Moreover, PS was longer in the 0-1 group than in the 2-3 group (HR=0.602, 95%CI=0.313-1.157, p=0.128), and patients with a GNRI of 92 and above had a significantly longer TTF than those below 92 (HR=0.603, 95%CI=0.411-0.885, p<0.01). Multivariate analysis showed that TTF was significantly longer for those aged ≥72 compared to those aged <72 (HR=0.657, 95%CI=0.446-0.967, p=0.033), and GNRI was significantly longer for those ≥92 than for those <92 (HR=0.616, 95%CI=0.415-0.914, p=0.016) (Table II).

Table II. Univariate and multivariate analyses using Cox proportional hazards models of time to treatment completion in lung cancer patients treated with immune checkpoint inhibitors.

ECOG PS: Eastern Cooperative Oncology Group performance status; GNRI: Geriatric Nutritional Risk Index.

Incidence of adverse events. Among the adverse events, hemoglobinemia Grade ≥3 (p=0.046), albumin depletion of all grades and Grade ≥3 (p=0.019, 0.011), hyponatremia of all grades (p<0.01), constipation of all grades (p<0.01) were significantly higher in the Low-GNRI group than in the High-GNRI group. The Low-GNRI group had a significantly higher proportion than the High-GNRI group for hyponatremia (p<0.01), constipation (p<0.01), and fatigue (p<0.01) of all grades. In contrast, the incidence of skin rash (p<0.01) and pruritus (p<0.01) of all grades was significantly higher in the High-GNRI group than in the Low-GNRI group (Table III).

Table III. Adverse events of the patients.

^dChi-square for independence test.

Discussion

In the present study, we investigated the relationship between the TTF and the occurrence of adverse events in patients treated with ICI monotherapy for lung cancer, by using the GNRI as an index. In the patient background, the Low-GNRI group had significantly lower weight, BMI, and serum albumin levels than the High-GNRI group (p<0.01), and significantly higher height and PS. Thus, the GNRI is a useful tool to assess nutritional status based on BMI and serum albumin levels. BMI is defined as "low nutritional tendency" below 20 kg/m² (15). The median BMI of the Low-GNRI group was 18.9 kg/m², and since many patients were underweight, BMI was low, which would have affected GNRI. Low BMI and malnutrition are a common problem for patients with cancer, and cancer-related malnutrition is associated with increased morbidity and mortality (16,17). Furthermore, in patients with cancer, malnutrition and cachexia are reflected in hypoalbuminemia, and hypoalbuminemia correlates with a decreased immune response (18). Therefore, we believe that serum albumin levels are a useful indicator of nutritional status. In addition, high serum albumin levels have been shown to contribute to improved survival in patients with cancer (19). As anorexia due to metabolic abnormalities and systemic inflammation may be associated with malnutrition in patients with cancer (13,20), it is advisable to collect information in advance regarding food intake and diet. In terms of patient background, the Low-GNRI group had significantly higher PS than the High-GNRI group; GNRI is an important prognostic factor in relation to PS (21). However, in the present study, the GNRI was not a prognostic factor because no significant difference was found in the multivariate analysis.

The pretreatment GNRI and age were factors that influenced treatment duration in multivariate analysis results; although there was no significant difference in progression-free survival or overall survival between older and younger patients treated with ICI (22,23), the progression-free survival and overall survival were significantly longer in the High-GNRI group (21). This suggests that it is possible to achieve the same efficacy of treatment in older patients as in younger patients. The GNRI is a factor that affects progression-free survival and overall survival in platinum-based chemotherapy for non-small cell lung cancer (24). Furthermore, it has been reported that the progression-free survival and overall survival in the High-GNRI group (≥89.5) were significantly prolonged compared to those in the Low-GNRI group (<89.5) in the treatment of non-small cell lung cancer with ICI (21). In the present study, a comparison was made with TTF, and similar results were obtained. Therefore, we consider GNRI to be a factor that influences TTF and is predictive of treatment efficacy.

The Low-GNRI group had a significantly higher incidence of Grade ≥3 hemoglobin loss than the High-GNRI group (p<0.05). This may have been caused by the low nutritional status, resulting in inadequate supply of protein and iron and decreased synthesis of hemoglobin. Furthermore, the Low-GNRI group had a significantly higher incidence of malaise (p<0.01), and we believe that the decreased hemoglobin level may have been one of the reasons for the higher incidence of malaise. Anemia is an important prognostic factor that triggers malignant tumor progression and treatment resistance, influences treatment efficacy and survival in patients with cancer, and correlates with decreased quality of life in patients (25-27). Therefore, patients in the Low-GNRI group should be considered at risk for anemia.

In the present study, the incidence of skin rash and pruritus was significantly higher in the High-GNRI group. The irAEs include skin disorders that occur following ICI administration, wherein T cells recognize the antigens that are common to lung tumors and the skin and, therefore, target both components simultaneously (28). ICI administration in non-small cell lung cancer significantly prolongs progression-free survival in patients who develop skin lesions (29). In the High-GNRI group, skin lesions occurred significantly more frequently, which may have prolonged the TTF and have had a positive impact on the response rate. The incidence of skin rash and pruritus were all grade 1 and 2. Zhang et al. reported that low severity of irAE expression is associated with treatment efficacy, although no association is seen when the severity is high (30). On the other hand, JIRI et al. reported improved progression-free survival even in a group of patients with severe irAEs (31). These differences require further investigation, as the number of patients with severe irAEs is not large enough and the type of irAE may also play a role in the impact. Therefore, a high GNRI seems to have an impact on treatment efficacy. However, it is important to carefully monitor the skin condition during the treatment period, as even mild skin disorders may affect the patient’s quality of life.

A limitation of this study is its retrospective design and small sample size, in which participants were recruited from only two centers. Future research involving larger cohorts is necessary to more accurately investigate the association between the GNRI and the occurrence of adverse events and TTF.

Conclusion

Pretreatment GNRI influences the continuation of ICI monotherapy. Moreover, the High-GNRI group demonstrated a significantly higher frequency of skin lesions, which may have influenced the prolongation of TTF.

Conflicts of Interest

The Authors have no conflicts of interest to declare in relation to this study.

Authors’ Contributions

Atsuya Shimizu: Conceptualization; Data curation; Miyu Fukasawa, Koharu Mitani, Keisuke Goto, Azusa Wakamoto , Tae Hatsuyama, Takanobu Hoshi, and Isao Hasegawa: Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Hideki Sato: Supervision; Validation; Visualization; Atsuya Shimizu: Writing – original draft; Miyu Fukasawa, Koharu Mitani, Keisuke Goto, Azusa Wakamoto, Tae Hatsuyama, Takanobu Hoshi, Isao Hasegawa, and Hideki Sato: Writing – review & editing. All Authors have approved the submitted version of the manuscript and agree to be accountable for all parts of the study.