Abstract
Objectives
This study aimed to examine the association between the Geriatric Nutritional Risk Index (GNRI) and mortality outcomes, encompassing all-cause, cardiovascular, and diabetes-associated deaths, among adults living with obesity.
Methods
This prospective cohort analysis employed nationally representative survey information from 8834 adults with obesity (BMI ≥30 kg/m2) enrolled in ten consecutive cycles of the National Health and Nutrition Examination Survey (NHANES) spanning 1999 to 2018. Vital status was determined via linkage to the National Death Index, with follow-up extending through December 31, 2019. Multivariate Cox proportional hazards models, restricted cubic spline analyses, and stratified subgroup evaluations were performed.
Results
Throughout the observation period, 2110 deaths from all causes, 715 cardiovascular-related deaths, and 111 diabetes-attributed deaths were recorded. Following comprehensive covariate adjustment, low GNRI (<98) exhibited markedly higher risks of all-cause death (hazard ratio [HR]: 1.67; 95% confidence interval [CI]: 1.43, 1.96), cardiovascular death (HR: 1.61; 95% CI: 1.20, 2.13), and diabetes-associated death (HR: 2.44; 95% CI: 1.41, 4.17) relative to those with high GNRI (≥98). Non-linear relationships were detected for all-cause and cardiovascular mortality, with threshold values identified at GNRI levels of 100.5 and 98.7, respectively. The observed associations persisted consistently across the majority of demographic and clinical strata.
Conclusions
Lower GNRI scores demonstrated independent associations with heightened mortality among adults with obesity. These results offer supportive evidence for incorporating nutritional risk screening into routine clinical management approaches for this patient population.
Keywords: Obesity, GNRI, Mortality, Cardiovascular death, Nutritional assessment
Highlights
-
•
Lower GNRI is independently associated with mortality in adults with obesity.
-
•
Diabetes-related mortality shows strongest association with low GNRI.
-
•
Nonlinear dose-response relationships with clinically actionable thresholds
1. Introduction
Obesity is among the most widespread metabolic conditions globally [1], with contemporary estimates suggesting that more than 650 million adults meet criteria for obesity [2]. Large-scale dose-response meta-analytic data have established non-linear relationships between obesity measures and both cardiovascular and all-cause mortality risks [3], along with heightened susceptibility to cardiovascular disease and type 2 diabetes mellitus [2]. Nevertheless, BMI alone fails to capture the metabolic diversity within the obese population. Sarcopenic obesity, characterized by concurrent excessive fat mass and reduced skeletal muscle [4], affects approximately 9% of adults with obesity and confers a 51% increased mortality risk [5], underscoring both the prognostic importance of nutritional status and the need for validated screening tools to identify high-risk individuals within this population.
Originally developed to evaluate nutritional risk in older hospitalized patients, the Geriatric Nutritional Risk Index (GNRI) is derived from serum albumin levels and the ratio of actual to ideal body weight [6]. Its prognostic utility has since been validated in various clinical contexts, including patients undergoing hemodialysis [7], those with acute myocardial infarction [8], [9], and other disease settings. More recently, GNRI has demonstrated predictive value for cardiovascular outcomes in middle-aged, community-dwelling adults [10], extending its applicability beyond traditional geriatric populations. Despite this evidence, data specifically examining the association between GNRI and mortality among adults with obesity remain limited. This knowledge gap is particularly relevant because the GNRI formula constrains the weight ratio to 1.0 when actual weight exceeds ideal weight [6]. Consequently, in individuals with obesity, GNRI is driven predominantly by serum albumin and may serve as a sensitive marker of subclinical protein depletion in this metabolically complex population. In this context, GNRI may provide information beyond body size itself by capturing a nutritional-inflammatory dimension of risk that is not adequately reflected by BMI alone.
Accordingly, the current investigation leveraged nationally representative survey data from the U.S. NHANES database covering 1999 through 2018 to examine the associations of GNRI with all-cause, cardiovascular, and diabetes-related mortality in adults with obesity. GNRI was analyzed both as a categorical variable using the clinically established threshold of 98 and as a continuous exposure. Restricted cubic spline regression and threshold effect analyses were performed to detect potential non-linear associations, and stratified analyses were conducted to evaluate whether these relationships were consistent across demographic and clinical subgroups. The findings from this study may provide epidemiological evidence supporting the incorporation of nutritional risk assessment into comprehensive obesity care.
2. Methods
2.1. Data source and participants
The present investigation followed the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) checklist to maintain methodological rigor and reporting transparency [11]. Study data originated from NHANES, an ongoing cross-sectional survey program that collects health information from a nationally representative sample of non-institutionalized U.S. residents through structured interviews, standardized physical assessments, and laboratory testing [12].
A total of 101,316 adult participants (age ≥ 18 years) from ten consecutive NHANES cycles (1999–2018) were initially considered for inclusion. Participants were sequentially excluded based on the following criteria: BMI below 30 kg/m2 (n = 66,299), absence of BMI measurements (n = 13,517), incomplete data required for GNRI computation (n = 12,658), and lacking mortality follow-up records (n = 8). The resultant analytical cohort consisted of 8834 adults with obesity with available GNRI calculations and mortality outcome data (Supplementary Fig. 1). Ethical clearance for the NHANES protocol was granted by the National Center for Health Statistics Research Ethics Review Board, with documented informed consent acquired from each participant. Given that this analysis exclusively utilized publicly available, anonymized datasets, supplementary institutional ethics approval was deemed unnecessary.
2.2. Definition of obesity and GNRI
Anthropometric data were collected by trained personnel in accordance with standardized protocols. Body mass index was computed as weight in kilograms divided by height in meters squared. Obesity was defined as BMI ≥30 kg/m2, consistent with World Health Organization criteria. Nutritional risk was quantified using GNRI, an index that integrates serum albumin concentration with the proportion of actual body weight relative to ideal body weight [6]. The GNRI score was calculated according to the established formula: GNRI = [1.489 × serum albumin (g/L)] + [41.7 × (actual body weight / ideal body weight)] [13]. The Lorentz formula was applied to estimate ideal body weight, with sex-specific calculations: for males, ideal body weight (kg) = height (cm) − 100 − [(height − 150) / 4]; for females, ideal body weight (kg) = height (cm) − 100 − [(height − 150) / 2.5] [6]. Per established convention, actual-to-ideal body weight ratios exceeding 1.0 were capped at 1 [6]. Serum albumin was measured using standardized NHANES laboratory protocols. For categorical analysis, GNRI was dichotomized using a threshold of 98, in which GNRI ≥98 indicates no nutrition-related risk [6], [13]. This literature-based threshold was retained for the primary analysis to preserve clinical interpretability and to avoid the optimism bias associated with deriving and validating a cutoff within the same dataset. To explore dose-response associations, GNRI was also modeled as a continuous variable using restricted cubic splines. In sensitivity analyses, GNRI was additionally examined across quartiles and according to the original four-category classification (<82, 82–<92, 92–<98, and ≥98) proposed by Bouillanne et al [6].
2.3. Definition of mortality
Participant follow-up commenced from their NHANES examination date and continued until death occurrence or the censoring date of December 31, 2019. Mortality status and underlying cause of death were ascertained via probabilistic record linkage with the CDC National Death Index database.
The primary outcome measure was death from any cause. Secondary outcome measures included cardiovascular mortality, defined by ICD-10 codes I00-I09, I11, I13, I20-I51, and I60-I69 [14], as well as diabetes-related mortality, identified using codes E10-E14 [15].
2.4. Covariates
Selection of covariates was guided by their clinical relevance, consistency with published guidelines, and evidence from previous epidemiological research. These variables encompassed demographic characteristics, lifestyle-related factors, and comorbid health conditions.
Demographic variables comprised age (treated continuously), sex (male versus female), and race/ethnicity (grouped as Mexican American, non-Hispanic White, non-Hispanic Black, or Other) [16]. Educational level was stratified into three tiers: less than high school education, high school graduate or equivalent credential, and college-level education or beyond. Marital status categories included currently married or living with partner, divorced/separated/widowed, and never married. The poverty-to-income ratio (PIR), reflecting household socioeconomic standing, was categorized into three levels: <1.3, 1.3 to <3.5, and ≥3.5 [17].
Lifestyle-related covariates included tobacco use, alcohol intake, and physical activity patterns. Smoking history was classified according to lifetime consumption of at least 100 cigarettes into never, former, or current smoker categories. Alcohol use was defined as lifetime consumption of 12 or more alcoholic drinks [17]. Physical activity was assessed via the Global Physical Activity Questionnaire (GPAQ) and categorized as low or high based on weekly MET-minutes [18].
Hypertension was defined as any of the following: physician-diagnosed hypertension, systolic blood pressure ≥ 140 mmHg, or diastolic blood pressure ≥ 90 mmHg. Diabetes mellitus was identified when participants met any criterion including: physician diagnosis, use of insulin or oral glucose-lowering medications, fasting glucose ≥126 mg/dL, or HbA1c ≥6.5%. Dyslipidemia was ascertained based on abnormal lipid profiles (elevated total cholesterol, triglycerides, or LDL-cholesterol; reduced HDL-cholesterol) or use of lipid-modifying therapy [19]. History of cardiovascular disease (CVD) was determined by self-reported doctor-diagnosed conditions including: heart failure, ischemic heart disease, angina, prior myocardial infarction, or cerebrovascular accident [20].
2.5. Statistical analysis
Participant characteristics at baseline were tabulated according to GNRI status, with continuous measures reported as means with standard errors and categorical measures as counts with weighted percentages. Between-group comparisons utilized survey-weighted t-tests for continuous outcomes and chi-square tests for categorical outcomes. The association of GNRI with mortality endpoints was evaluated using Cox proportional hazards modeling, yielding hazard ratios (HR) with accompanying 95% confidence intervals (CI). GNRI was modeled both categorically (<98 versus ≥98, with ≥98 serving as reference) and continuously. For continuous analyses, HR denotes the proportional change in risk per one-unit GNRI decrement. In piecewise Cox regression for threshold analysis, HR similarly represents risk change per one-unit GNRI reduction to demonstrate the protective influence of higher GNRI values. Sequential covariate adjustment proceeded through three models: Model 1 was unadjusted; Model 2 adjusted for age, sex, race/ethnicity, educational level, marital status, and PIR; Model 3 further incorporated smoking status, physical activity, alcohol use, BMI, hypertension, diabetes, dyslipidemia, and prevalent CVD.
Verification of the proportional hazards assumption employed scaled Schoenfeld residual testing, which revealed no substantive violations. Covariate multicollinearity was assessed through variance inflation factors (VIFs), with all values remaining below the threshold of 5. Kaplan-Meier methodology generated survival curves stratified by GNRI category, with inter-group differences assessed via log-rank statistics. Restricted cubic spline (RCS) analysis with knots positioned at the 5th, 35th, 65th, and 95th percentiles of GNRI distribution explored non-linear patterns, referencing the median GNRI value. Upon detecting significant non-linearity (P < 0.05), two-segment piecewise Cox models were fitted to characterize threshold effects.
Table 1 displays baseline characteristics from observed data. Missing covariates were addressed through single imputation: predictive mean matching for continuous variables and logistic regression for categorical variables. Imputed data informed all subsequent analyses, encompassing Cox regression and stratified subgroup evaluations. Pre-planned subgroup analyses were stratified by demographic, lifestyle, and clinical factors, with interaction terms testing effect modification. Statistical analyses employed R software (version 4.4.0), applying a two-sided P < 0.05 threshold for significance. Survey weights, clustering variables, and stratification factors were incorporated throughout to account for the complex multistage sampling design and produce nationally representative estimates. Weight selection adhered to NHANES analytic guidelines based on variables included in each analysis.
Table 1.
Baseline characteristics of study participants by GNRI status.
| Characteristics | Total (n = 8834) |
GNRI < 98 (n = 999) |
GNRI ≥ 98 (n = 7835) |
P value |
|---|---|---|---|---|
| Age, years, Mean (SE) | 62.41 (0.15) | 63.93 (0.42) | 62.26 (0.15) | <0.001 |
| BMI, kg/m2, Mean (SE) | 35.48 (0.08) | 37.79 (0.30) | 35.25 (0.08) | <0.001 |
| GNRI, Mean (SE) | 103.44 (0.10) | 94.70 (0.12) | 104.32 (0.08) | <0.001 |
| Sex, n (%) | <0.001 | |||
| Male | 3911 (44.72) | 319 (28.88) | 3592 (46.31) | |
| Female | 4923 (55.28) | 680 (71.12) | 4243 (53.69) | |
| Race/ethnicity, n (%) | <0.001 | |||
| Mexican American | 1496 (5.54) | 145 (6.38) | 1351 (5.45) | |
| Non-Hispanic White | 4093 (75.26) | 356 (62.77) | 3737 (76.52) | |
| Non-Hispanic Black | 2157 (11.60) | 366 (21.61) | 1791 (10.60) | |
| Other | 1088 (7.59) | 132 (9.23) | 956 (7.43) | |
| Education, n (%) | 0.006 | |||
| Less than high school | 2780 (18.82) | 346 (23.58) | 2434 (18.34) | |
| High school or equivalent | 2142 (26.67) | 227 (26.27) | 1915 (26.71) | |
| College or above | 3903 (54.52) | 425 (50.15) | 3478 (54.95) | |
| Marital status, n (%) | <0.001 | |||
| Married or living with a partner | 5185 (64.86) | 499 (53.57) | 4686 (66.00) | |
| Widowed/divorced/separated | 2966 (28.76) | 396 (38.10) | 2570 (27.82) | |
| Never married | 605 (6.38) | 100 (8.33) | 505 (6.18) | |
| PIR, n (%) | <0.001 | |||
| <1.3 | 2334 (18.37) | 295 (23.73) | 2039 (17.84) | |
| 1.3 to <3.5 | 3186 (37.27) | 383 (41.63) | 2803 (36.84) | |
| ≥3.5 | 2451 (44.36) | 209 (34.64) | 2242 (45.32) | |
| Age group, n (%) | 0.001 | |||
| <65 years | 4918 (62.65) | 528 (56.62) | 4390 (63.25) | |
| ≥65 years | 3916 (37.35) | 471 (43.38) | 3445 (36.75) | |
| Smoking status, n (%) | 0.130 | |||
| Never | 4434 (50.19) | 483 (48.31) | 3951 (50.38) | |
| Former | 3260 (37.16) | 345 (36.25) | 2915 (37.26) | |
| Current | 1134 (12.64) | 171 (15.44) | 963 (12.36) | |
| Alcohol consumption, n (%) | <0.001 | |||
| No | 3262 (34.78) | 450 (46.19) | 2812 (33.67) | |
| Yes | 5009 (65.22) | 449 (53.81) | 4560 (66.33) | |
| Physical activity, n (%) | <0.001 | |||
| Low | 3758 (47.55) | 506 (58.27) | 3252 (46.50) | |
| High | 3628 (52.45) | 329 (41.73) | 3299 (53.50) | |
| Hypertension, n (%) | 0.019 | |||
| No | 2401 (30.98) | 222 (25.74) | 2179 (31.50) | |
| Yes | 6421 (69.02) | 773 (74.26) | 5648 (68.50) | |
| Diabetes, n (%) | <0.001 | |||
| No | 5162 (65.65) | 440 (52.03) | 4722 (67.03) | |
| Yes | 3478 (34.35) | 544 (47.97) | 2934 (32.97) | |
| Hyperlipidemia, n (%) | <0.001 | |||
| No | 1249 (12.79) | 182 (18.45) | 1067 (12.22) | |
| Yes | 7585 (87.21) | 817 (81.55) | 6768 (87.78) | |
| Cardiovascular disease, n (%) | <0.001 | |||
| No | 6819 (80.22) | 692 (71.48) | 6127 (81.10) | |
| Yes | 1947 (19.78) | 299 (28.52) | 1648 (18.90) |
Abbreviations: GNRI, Geriatric Nutritional Risk Index; BMI, body mass index; PIR, poverty income ratio; SE, standard error.
3. Results
3.1. Study population
The analytic sample comprised 8834 adults with obesity (BMI ≥30 kg/m2) from NHANES 1999–2018 cycles (Supplementary Fig. 1). Among these, 7835 individuals (88.69%) had GNRI values ≥98, while 999 (11.31%) had GNRI <98. Compared to the high GNRI group, participants with low GNRI were older on average (63.93 vs. 62.26 years; P < 0.001), had higher BMI measurements (37.79 vs. 35.25 kg/m2; P < 0.001), and included a greater proportion of women (71.12% vs. 53.69%; P < 0.001).
The low GNRI group showed lower educational achievement (P = 0.006), higher prevalence of widowed/divorced/separated status (P < 0.001), and lower household income-to-poverty ratio (P < 0.001). Furthermore, this group exhibited higher rates of physical inactivity (58.27% vs. 46.50%; P < 0.001), diabetes (47.97% vs. 32.97%; P < 0.001), hypertension (74.26% vs. 68.50%; P = 0.019), and prevalent CVD (28.52% vs. 18.90%; P < 0.001). In contrast, dyslipidemia prevalence was lower in the low GNRI group (81.55% vs. 87.78%; P < 0.001). No significant between-group difference was observed for smoking status (P = 0.130). Detailed baseline characteristics appear in Table 1.
3.2. GNRI and mortality outcomes
During the observation period, 2110 all-cause deaths occurred, accompanied by 715 cardiovascular-related and 111 diabetes-associated deaths. Survey-weighted mortality rates comparing low versus high GNRI groups were 28.37% versus 18.86% for all-cause, 9.18% versus 6.37% for cardiovascular, and 2.70% versus 0.90% for diabetes-related deaths.
Fig. 1 displays Cox regression findings. After full covariate adjustment, every 1-unit GNRI decrement was associated with elevated mortality hazards: all-cause (HR = 1.05, 95% CI: 1.03–1.06), cardiovascular (HR = 1.04, 95% CI: 1.02–1.06), and diabetes-related (HR = 1.12, 95% CI: 1.08–1.18; all P < 0.001). When analyzed categorically, low GNRI (<98) was associated with 67% greater all-cause mortality risk (HR = 1.67, 95% CI: 1.43–1.96), 61% greater cardiovascular mortality risk (HR = 1.61, 95% CI: 1.20–2.13), and 144% greater diabetes-related mortality risk (HR = 2.44, 95% CI: 1.41–4.17) versus high GNRI (≥98; all P < 0.001). Kaplan-Meier analysis revealed significantly poorer survival in the low GNRI group for all endpoints (log-rank P < 0.001; Fig. 2). Proportional hazards assumptions held for all outcomes (all P > 0.05), with VIFs remaining below 5.
Fig. 1.
Forest plot for the associations between GNRI and mortality.
Hazard ratios and 95% confidence intervals for all-cause, cardiovascular, and diabetes-related mortality. GNRI was analyzed as a continuous variable (per 1-unit decrease) and a categorical variable (<98 vs. ≥98). Model 1: unadjusted. Model 2: adjusted for age, sex, race/ethnicity, education, marital status, and poverty income ratio. Model 3: additionally adjusted for smoking status, physical activity, alcohol consumption, body mass index, hypertension, diabetes, hyperlipidemia, and cardiovascular disease. GNRI, Geriatric Nutritional Risk Index; HR, hazard ratio; CI, confidence interval.
Fig. 2.
Kaplan-Meier survival curves according to GNRI status.
Curves depict cumulative survival for all-cause mortality (A), cardiovascular mortality (B), and diabetes-related mortality (C). P values were calculated using the log-rank test. GNRI, Geriatric Nutritional Risk Index.
3.3. Nonlinear relationship and threshold effect analysis
RCS modeling revealed non-linear GNRI-mortality relationships for all-cause death (overall P < 0.001, non-linearity P = 0.005) and cardiovascular death (overall P < 0.001, non-linearity P = 0.028). Conversely, diabetes-related mortality exhibited a linear pattern (overall P < 0.001, non-linearity P = 0.535). Detailed results appear in Fig. 3.
Fig. 3.
Restricted cubic spline curves for the dose-response relationships between GNRI and mortality.
Curves depict hazard ratios (solid lines) and 95% confidence intervals (shaded areas) for all-cause mortality (A), cardiovascular mortality (B), and diabetes-related mortality (C). The median GNRI value was used as the reference. Knots were placed at the 5th, 35th, 65th, and 95th percentiles. All models were fully adjusted. GNRI, Geriatric Nutritional Risk Index; HR, hazard ratio; CI, confidence interval.
Segmented Cox regression identified threshold values at GNRI of 100.5, 98.7, and 102.7 for all-cause, cardiovascular, and diabetes-related mortality, respectively (Table 2). For all-cause mortality, each 1-unit GNRI reduction below the threshold corresponded to HR 1.10 (95% CI: 1.08–1.11); above this point, HR was 1.02 (95% CI: 1.01–1.04). Cardiovascular mortality showed HRs of 1.10 (95% CI: 1.05–1.15) below and 1.02 (95% CI: 0.99–1.04) above the inflection point. Segmented models provided superior fit compared to linear models for all-cause (P < 0.001) and cardiovascular mortality (P = 0.005), though not for diabetes-related mortality (P = 0.428).
Table 2.
Nonlinear and threshold effect analysis of GNRI on mortality outcomes.
| GNRI |
All-cause mortality |
Cardiovascular mortality |
Diabetes-related mortality |
|---|---|---|---|
| HR (95% CI) P | HR (95% CI) P | HR (95% CI) P | |
| Model 1 | |||
| Straight-line effect | 1.05 (1.04–1.06) <0.001 | 1.04 (1.03–1.06) <0.001 | 1.11 (1.06–1.15) <0.001 |
| Model 2 | |||
| Inflection point (K) | 100.456 | 98.694 | 102.749 |
| <K effect | 1.10 (1.08–1.11) <0.001 | 1.10 (1.05–1.15) <0.001 | 1.12 (1.06–1.18) <0.001 |
| ≥K effect | 1.02 (1.01–1.04) 0.010 | 1.02 (0.99–1.04) 0.185 | 1.03 (0.93–1.15) 0.599 |
| Log Likelihood Ratio Test | <0.001 | 0.005 | 0.428 |
Model 1: Fully adjusted Cox regression with linear exposure assumption. Model 2: Segmented Cox regression with inflection point (K) determined via log-likelihood ratio testing. Hazard ratios reflect risk change per 1-unit GNRI reduction. Covariates in all models include age, sex, race/ethnicity, education, marital status, PIR, smoking, alcohol use, BMI, hypertension, diabetes, dyslipidemia, and CVD history.
Abbreviations: GNRI, Geriatric Nutritional Risk Index; HR, hazard ratio; CI, confidence interval.
3.4. Subgroup analyses
Supplementary Fig. 2 presents stratified all-cause mortality analyses, with cardiovascular mortality results in Supplementary Fig. 3.
Low GNRI (<98) showed consistent associations with heightened all-cause mortality across all examined subgroups, with HRs spanning 1.15 to 2.78. No significant effect modification was detected (all interaction P > 0.05). Associations persisted across age strata (HR = 1.82, 95% CI: 1.33–2.44 for age < 65; HR = 1.75, 95% CI: 1.39–2.17 for age ≥ 65; interaction P = 0.394), sex categories (HR = 2.08, 95% CI: 1.61–2.70 for males; HR = 1.52, 95% CI: 1.19–1.89 for females; interaction P = 0.119), and diabetes status (HR = 1.61, 95% CI: 1.23–2.08 without diabetes; HR = 1.82, 95% CI: 1.45–2.27 with diabetes; interaction P = 0.540). Mexican Americans exhibited the largest point estimate (HR = 2.78, 95% CI: 1.89–4.17), though racial/ethnic interaction was non-significant (P = 0.528). Non-significant findings were observed in the “Other” race/ethnicity category (HR = 1.85, 95% CI: 0.98–3.45, P = 0.057) and among never-married individuals (HR = 1.15, 95% CI: 0.62–2.13, P = 0.670).
Regarding cardiovascular mortality, low GNRI-associated HRs varied from 0.88 to 4.55 across strata. Significant effect modification emerged for smoking (interaction P = 0.016) and hypertension status (interaction P = 0.028). The GNRI-cardiovascular mortality relationship was most pronounced among current smokers (HR = 3.33, 95% CI: 1.67–6.67, P < 0.001), remained significant in never smokers (HR = 1.85, 95% CI: 1.27–2.78, P = 0.002), but was attenuated in former smokers (HR = 1.11, 95% CI: 0.69–1.75, P = 0.676). By hypertension status, stronger associations appeared in normotensive participants (HR = 3.45, 95% CI: 1.82–6.25, P < 0.001) than in hypertensive individuals (HR = 1.52, 95% CI: 1.10–2.08, P = 0.011). Mexican Americans again showed the strongest effect (HR = 4.55, 95% CI: 2.27–9.09; P < 0.001). Non-significant associations were noted among never-married participants (HR = 0.88, 95% CI: 0.30–2.63, P = 0.815), those with less than high school education (HR = 1.33, 95% CI: 0.88–2.00, P = 0.180), and individuals with prevalent CVD (HR = 1.43, 95% CI: 0.93–2.17, P = 0.104).
Subgroup analyses for diabetes-related mortality were precluded by insufficient event counts (n = 111).
3.5. Sensitivity analyses
To evaluate the robustness of the primary findings, GNRI was additionally examined by quartiles and according to the original Bouillanne four-category classification. In the quartile-based analysis, using the lowest quartile (Q1, GNRI 72.97–99.77, n = 2488) as the reference, the multivariable-adjusted HRs for all-cause mortality were 0.716 (95% CI 0.617–0.830) for Q2, 0.761 (95% CI 0.652–0.889) for Q3, and 0.542 (95% CI 0.464–0.633) for Q4 (P for trend <0.001). Similar inverse associations were observed for cardiovascular mortality (P for trend <0.001) and diabetes-related mortality (P for trend <0.001). Complete results are presented in Supplementary Table S1. When the original Bouillanne four-category classification was applied (<82, 82–<92, 92–<98, ≥98), the major risk group (GNRI <82) contained 11 participants (0.12%) and the moderate risk group (82–<92) contained 120 participants (1.36%), resulting in wide confidence intervals for these categories. A decreasing pattern of mortality risk across higher GNRI categories was observed for all three endpoints (Supplementary Table S2).
4. Discussion
Leveraging nationally representative NHANES data spanning 1999–2018, this investigation examined how GNRI relates to mortality endpoints—including all-cause, cardiovascular, and diabetes-associated deaths—in an obese adult cohort. Key findings are as follows: First, lower GNRI was significantly associated with heightened mortality risk across all three outcome categories, with each one-unit GNRI reduction corresponding to 5%, 4%, and 12% increases in all-cause, cardiovascular, and diabetes-related mortality, respectively. Second, categorical analyses revealed that low GNRI (<98) was independently associated with elevated mortality; relative to the high GNRI group (≥98), hazard ratios were 1.67 for all-cause, 1.61 for cardiovascular, and 2.44 for diabetes-related mortality.
Our results corroborate and build upon the expanding literature establishing GNRI as a reliable prognostic indicator across various patient populations. In elderly heart failure patients, meta-analytic evidence has shown that the lowest GNRI stratum confers approximately twofold increased risk of all-cause death and major adverse cardiovascular events [21]. This finding has been consistently replicated in severe heart failure cohorts (adjusted HR = 2.03) [22] and in patients undergoing transcatheter tricuspid valve repair (adjusted HR = 1.53) [23]. Furthermore, GNRI has shown predictive utility in hemodialysis patients, where serial improvements in GNRI values were associated with diminished mortality [24], as well as in patients with chronic kidney disease [25], type B aortic dissection [26], non-cardiac surgery [27], osteoporotic fractures [28], renal cell carcinoma [29], and critical illness [30]. The current results corroborate these findings and confirm that low GNRI is independently associated with elevated mortality risk even within the metabolically heterogeneous population of adults with obesity.
However, several aspects of the present study extend beyond prior research. First, whereas most existing evidence derives from hospitalized, elderly, or disease-specific cohorts, this analysis provides a comprehensive evaluation of GNRI and mortality in a large, nationally representative sample of community-dwelling adults with obesity. This potential relevance is particularly important in obesity, because GNRI may capture a broader state of metabolic vulnerability that is not adequately reflected by BMI alone. In this context, GNRI may complement BMI-based risk assessment by identifying high-risk individuals whose excess body weight might otherwise obscure underlying nutritional compromise. Second, this study systematically examined diabetes-related mortality and identified a notably strong association (HR = 2.44), consistent with recent evidence that malnutrition assessed by nutritional indices is associated with increased mortality in heart failure patients with and without diabetes [31]. Third, through RCS and threshold effect analyses, nonlinear dose–response relationships were characterized and specific inflection points were identified at 100.5 for all-cause mortality and 98.7 for cardiovascular mortality. Below these thresholds, the mortality risk gradient steepens substantially, patterns consistent with those observed in aortic dissection [26] and sepsis risk among critically ill patients [30]. These threshold values may serve as clinically actionable targets for identifying individuals most likely to benefit from nutritional assessment and intervention.
Several interconnected pathophysiological mechanisms may explain the observed associations between low GNRI and elevated mortality risk among adults with obesity. A low GNRI may reflect protein-energy malnutrition and chronic low-grade inflammation. Such chronic inflammation has been implicated in the pathogenesis of sarcopenia and metabolic dysfunction, and the coexistence of malnutrition, inflammation, and atherosclerosis, known as the MIA (malnutrition-inflammation-atherosclerosis) syndrome, has been shown to confer substantially elevated mortality risk (HR = 3.46) in vascular disease populations [32]. Hypoalbuminemia functions as both a negative acute-phase reactant and a marker of hepatic synthetic capacity, reflecting convergent inflammatory processes associated with insulin resistance, endothelial dysfunction, and atherogenesis [33]. Beyond inflammation, lower GNRI in individuals with obesity may signal sarcopenic obesity, characterized by excess adiposity with diminished muscle mass [4], [34]. Because the GNRI formula constrains the weight ratio at 1.0 when actual weight exceeds ideal weight, GNRI variation in obesity is driven predominantly by serum albumin, potentially serving as a surrogate marker for muscle protein depletion. Indeed, the construct captured by GNRI in obesity likely extends beyond nutritional status to encompass sarcopenia and frailty. A meta-analysis of 11 observational studies demonstrated that each 1-unit increase in GNRI was associated with a 6.9% reduction in sarcopenia risk [35], and GNRI has been shown to discriminate both sarcopenia (AUC = 0.683) and frailty (AUC = 0.741) in hospitalized older adults, with the three conditions coexisting in over one quarter of patients [36]. In elderly patients with type 2 diabetes, GNRI correlated positively with skeletal muscle index and handgrip strength [37]. These observations suggest that a low GNRI in individuals with obesity may signal a catabolic phenotype driven by chronic inflammation and muscle protein depletion, consistent with the concept that the obesity paradox is more closely related to nutritional-metabolic vulnerability than to body weight per se [38]. At the mechanistic level, sarcopenic obesity engenders adverse metabolic consequences including adipokine dysregulation and heightened inflammation, with inflammation and insulin resistance mediating 6–21% of the association between sarcopenic obesity and hypertension [39], and meta-analytic evidence indicating approximately 51% increased mortality risk [5]. The nonlinear associations observed herein, with inflection points near 100.5 and 98.7, may reflect thresholds below which subclinical sarcopenia becomes clinically consequential. Additionally, nutritional deficiencies, including vitamin D insufficiency commonly associated with malnutrition, compromise immune function and increase susceptibility to infections. The particularly strong association with diabetes-related mortality (HR = 2.44) likely reflects convergent immunometabolic perturbations amplified by nutritional deficits [31]. Collectively, these mechanisms support a “malnutrition within obesity” paradigm wherein low GNRI identifies metabolically vulnerable individuals despite excess body weight.
The present findings have several clinical and public health implications. GNRI, derived from readily available clinical parameters, may serve as a practical and cost-effective tool for identifying metabolically vulnerable individuals among patients with obesity in primary care and specialist settings [1], [6]. The inflection points identified in this study (100.5 for all-cause and 98.7 for cardiovascular mortality) may serve as clinically actionable thresholds to guide nutritional evaluation and intervention. Furthermore, these findings suggest that comprehensive obesity management should extend beyond weight reduction alone to encompass concurrent assessment and optimization of nutritional status, with particular attention to dietary protein quality and serum albumin levels. At the public health level, strategies addressing the obesity epidemic may benefit from a paradigm shift, moving from “weight loss” toward “healthy weight with adequate nutrition.” Nonetheless, given the observational design, whether targeted nutritional interventions can improve clinical outcomes in adults with obesity with low GNRI warrants confirmation through randomized controlled trials.
Strengths of this investigation include the use of large-scale nationally representative NHANES survey data, a prospective cohort framework with extended follow-up, and implementation of rigorous analytical approaches encompassing complex survey weighting, multiple imputation, and RCS modeling. Nevertheless, certain limitations warrant acknowledgment. The observational nature of this study precludes establishment of causality, and unmeasured confounding remains possible. Furthermore, reverse causality cannot be excluded. Subclinical or undiagnosed conditions, such as occult malignancy or chronic organ dysfunction, may simultaneously reduce serum albumin concentrations and increase mortality risk, potentially confounding the observed GNRI-mortality association. Although the design with mortality follow-up through the National Death Index establishes temporal precedence of GNRI measurement, it does not eliminate the possibility that preclinical disease processes were already present at baseline and contributed to both lower GNRI and subsequent mortality. GNRI determination occurred solely at baseline, preventing assessment of temporal changes—a notable consideration given prior reports linking GNRI improvement trajectories with survival advantage [24]. In addition, direct comparison between GNRI and other nutritional screening instruments, such as the Mini Nutritional Assessment [40] or the Malnutrition Universal Screening Tool [41], was not performed. Therefore, the present study does not allow conclusions regarding the superiority of GNRI over other nutritional screening tools. Limited diabetes-related deaths (n = 111) constrained statistical power for this outcome, and findings may not generalize beyond U.S. adults. Future research should prioritize intervention trials evaluating nutritional supplementation in individuals with obesity with low GNRI, mechanistic studies exploring underlying biological pathways, and validation studies confirming the identified thresholds across diverse populations.
5. Conclusion
In this large, nationally representative cohort of adults with obesity, lower GNRI demonstrated independent associations with increased all-cause, cardiovascular, and diabetes-related mortality, whereas higher GNRI values were associated with lower mortality risk—particularly pronounced above threshold values of 100.5 for all-cause and 98.7 for cardiovascular mortality. Stratified analyses confirmed consistency of these relationships across diverse demographic and clinical subsets. These observations underscore the importance of incorporating nutritional risk assessment into obesity care paradigms, especially for high-risk individuals. Routine GNRI assessment may facilitate identification of metabolically vulnerable patients who could benefit from targeted nutritional interventions.
CRediT authorship contribution statement
Xueqing Yang: Writing – original draft, Formal analysis, Data curation, Conceptualization. Kun Na: Writing – review & editing, Methodology, Formal analysis. Chenghui Yan: Writing – review & editing, Supervision, Funding acquisition. Yaling Han: Writing – review & editing, Supervision, Funding acquisition.
Ethical statement
Ethical clearance for the NHANES protocol was granted by the National Center for Health Statistics Research Ethics Review Board, with documented informed consent acquired from each participant. Given that this analysis exclusively utilized publicly available, anonymized datasets, supplementary institutional ethics approval was deemed unnecessary.
Declaration of Generative AI and AI-assisted technologies in the writing process
No AI-assisted tools were used during the preparation of this manuscript.
Funding
This study was supported by the National Natural Science Foundation of China (82470303).
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
The authors would like to thank all the staff and participants involved in the NHANES data collection.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.ahjo.2026.100796.
Appendix A. Supplementary data
Supplementary tables
Supplementary Fig. 1.
Study flow chart.
Stratified analyses for all-cause mortality according to GNRI status.
Forest plot showing hazard ratios and 95% confidence intervals for low GNRI (<98) versus high GNRI (≥98, reference) across subgroups. P for interaction was calculated using the likelihood ratio test. All models were adjusted for covariates except for the stratification variable. GNRI, Geriatric Nutritional Risk Index; HR, hazard ratio; CI, confidence interval; PIR, poverty income ratio.
Stratified analyses for cardiovascular mortality according to GNRI status.
Forest plot showing hazard ratios and 95% confidence intervals for low GNRI (<98) versus high GNRI (≥98, reference) across subgroups. P for interaction was calculated using the likelihood ratio test. All models were adjusted for covariates except for the stratification variable. GNRI, Geriatric Nutritional Risk Index; HR, hazard ratio; CI, confidence interval; PIR, poverty income ratio.
Data availability
Publicly available datasets were analyzed in this study. The data can be found at the National Health and Nutrition Examination Survey (NHANES) website (https://www.cdc.gov/nchs/nhanes/).
References
- 1.Ahmed S.K., Mohammed R.A. Obesity: prevalence, causes, consequences, management, preventive strategies and future research directions. Metab. Open. 2025;27 doi: 10.1016/j.metop.2025.100375. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Sorensen T.I.A., Martinez A.R., Jorgensen T.S.H. Epidemiology of obesity. Handb. Exp. Pharmacol. 2022;274:3–27. doi: 10.1007/164_2022_581. [DOI] [PubMed] [Google Scholar]
- 3.Wang L., Ding H., Deng Y., Huang J., Lao X., Wong M.C.S. Associations of obesity indices change with cardiovascular outcomes: a dose-response meta-analysis. Int. J. Obes. (Lond) 2024;48(5):635–645. doi: 10.1038/s41366-024-01485-8. [DOI] [PubMed] [Google Scholar]
- 4.Donini L.M., Busetto L., Bischoff S.C., Cederholm T., Ballesteros-Pomar M.D., Batsis J.A., et al. Definition and diagnostic criteria for Sarcopenic obesity: ESPEN and EASO consensus statement. Obes. Facts. 2022;15(3):321–335. doi: 10.1159/000521241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Liu C., Wong P.Y., Chung Y.L., Chow S.K., Cheung W.H., Law S.W., et al. Deciphering the “obesity paradox” in the elderly: a systematic review and meta-analysis of sarcopenic obesity. Obes. Rev. 2023;24(2) doi: 10.1111/obr.13534. [DOI] [PubMed] [Google Scholar]
- 6.Bouillanne O., Morineau G., Dupont C., Coulombel I., Vincent J.P., Nicolis I., et al. Geriatric nutritional risk index: a new index for evaluating at-risk elderly medical patients. Am. J. Clin. Nutr. 2005;82(4):777–783. doi: 10.1093/ajcn/82.4.777. [DOI] [PubMed] [Google Scholar]
- 7.Xiong J., Wang M., Zhang Y., Nie L., He T., Wang Y., et al. Association of Geriatric Nutritional Risk Index with mortality in hemodialysis patients: a meta-analysis of cohort studies. Kidney Blood Press. Res. 2018;43(6):1878–1889. doi: 10.1159/000495999. [DOI] [PubMed] [Google Scholar]
- 8.Nakamura T., Haraguchi Y., Matsumoto M., Ishida T., Momomura S.I. Prognostic impact of malnutrition in elderly patients with acute myocardial infarction. Heart Vessels. 2022;37(3):385–391. doi: 10.1007/s00380-021-01922-y. [DOI] [PubMed] [Google Scholar]
- 9.Tang S., Wu C., Su Y., Li Y. Prognostic value of the geriatric nutritional risk index in mortality prediction among critically ill acute myocardial infarction patients. BMC Cardiovasc. Disord. 2025;25(1):152. doi: 10.1186/s12872-025-04546-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Park J.Y., Bu S.Y. The ability of the geriatric nutritional risk index to predict the risk of heart diseases in Korean adults: a Korean genome and epidemiology study cohort. Front. Nutr. 2023;10 doi: 10.3389/fnut.2023.1276073. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.von Elm E., Altman D.G., Egger M., Pocock S.J., Gøtzsche P.C., Vandenbroucke J.P. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. J. Clin. Epidemiol. 2008;61(4):344–349. doi: 10.1016/j.jclinepi.2007.11.008. [DOI] [PubMed] [Google Scholar]
- 12.Ahluwalia N., Dwyer J., Terry A., Moshfegh A., Johnson C. Update on NHANES dietary data: focus on collection, release, analytical considerations, and uses to inform public policy. Adv. Nutr. 2016;7(1):121–134. doi: 10.3945/an.115.009258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Zheng W., Zhou C., Miao J., Yang Y., Hu X., Wang H., et al. Association between geriatric nutritional risk index and overactive bladder in the elderly population: a cross-sectional study. Front. Nutr. 2025;12 doi: 10.3389/fnut.2025.1537549. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Di D., Zhang R., Zhou H., Wei M., Cui Y., Zhang J., et al. Exposure to phenols, chlorophenol pesticides, phthalate and PAHs and mortality risk: a prospective study based on 6 rounds of NHANES. Chemosphere. 2023;329 doi: 10.1016/j.chemosphere.2023.138650. [DOI] [PubMed] [Google Scholar]
- 15.Laclaustra M., Moreno-Franco B., Lou-Bonafonte J.M., Mateo-Gallego R., Casasnovas J.A., Guallar-Castillon P., et al. Impaired sensitivity to thyroid hormones is associated with diabetes and metabolic syndrome. Diabetes Care. 2018;42(2):303–310. doi: 10.2337/dc18-1410. [DOI] [PubMed] [Google Scholar]
- 16.Dong G., Gan M., Xu S., Xie Y., Zhou M., Wu L. The neutrophil-lymphocyte ratio as a risk factor for all-cause and cardiovascular mortality among individuals with diabetes: evidence from the NHANES 2003-2016. Cardiovasc. Diabetol. 2023;22(1):267. doi: 10.1186/s12933-023-01998-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Xiao Q., Cai B., Yin A., Huo H., Lan K., Zhou G., et al. L-shaped association of serum 25-hydroxyvitamin D concentrations with cardiovascular and all-cause mortality in individuals with osteoarthritis: results from the NHANES database prospective cohort study. BMC Med. 2022;20(1):308. doi: 10.1186/s12916-022-02510-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.MacGregor K.A., Gallagher I.J., Moran C.N. Relationship between insulin sensitivity and menstrual cycle is modified by BMI, fitness, and physical activity in NHANES. J. Clin. Endocrinol. Metab. 2021;106(10):2979–2990. doi: 10.1210/clinem/dgab415. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wang G., Fang L., Chen Y., Ma Y., Zhao H., Wu Y., et al. Association between exposure to mixture of heavy metals and hyperlipidemia risk among U.S. adults: a cross-sectional study. Chemosphere. 2023;344 doi: 10.1016/j.chemosphere.2023.140334. [DOI] [PubMed] [Google Scholar]
- 20.Zhang Q., Xiao S., Jiao X., Shen Y. The triglyceride-glucose index is a predictor for cardiovascular and all-cause mortality in CVD patients with diabetes or pre-diabetes: evidence from NHANES 2001-2018. Cardiovasc. Diabetol. 2023;22(1):279. doi: 10.1186/s12933-023-02030-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Li H., Cen K., Sun W., Feng B. Prognostic value of geriatric nutritional risk index in elderly patients with heart failure: a meta-analysis. Aging Clin. Exp. Res. 2021;33(6):1477–1486. doi: 10.1007/s40520-020-01656-3. [DOI] [PubMed] [Google Scholar]
- 22.Pagnesi M., Serafini L., Chiarito M., Stolfo D., Baldetti L., Inciardi R.M., et al. Impact of malnutrition in patients with severe heart failure. Eur. J. Heart Fail. 2024;26(7):1585–1593. doi: 10.1002/ejhf.3285. [DOI] [PubMed] [Google Scholar]
- 23.Pagnesi M., Adamo M., Stolz L., Pancaldi E., Kresoja K.P., von Stein J., et al. Malnutrition and outcomes in patients with tricuspid regurgitation undergoing transcatheter tricuspid valve repair. Eur. J. Heart Fail. 2025;27(7):1304–1314. doi: 10.1002/ejhf.3623. [DOI] [PubMed] [Google Scholar]
- 24.Wang Q., Ye Z., Rao J., Zhao W., Liu J., Chen Y., et al. Assessing the predictive power of nutritional indices on all-cause and cardiovascular mortality in hemodialysis patients: a longitudinal study. Ren. Fail. 2024;46(2) doi: 10.1080/0886022X.2024.2431140. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Chen L., Yan M., Li J., Zhao X., Zeng L., Gao Z., et al. Association of geriatric nutritional risk index with renal prognosis and all-cause mortality among older patients with chronic kidney disease: a secondary analysis of CKD-ROUTE study. Ren. Fail. 2025;47(1) doi: 10.1080/0886022X.2025.2449720. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Zhao K., Niu J., He Y., Kong L., Zhao W., Lu Q., et al. The role of geriatric nutritional risk index in predicting survival of type B aortic dissection patients after thoracic endovascular aortic repair. J. Nutr. Health Aging. 2025;29(7) doi: 10.1016/j.jnha.2025.100572. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Zhao W., Liu L., Zhang J., Shi L., Zhang C., Ma Y., et al. Preoperative geriatric nutritional risk index predicts prognosis and postoperative complications in elderly patients undergoing non-cardiac surgery: a retrospective cohort study. BMC Geriatr. 2025;25(1):101. doi: 10.1186/s12877-024-05667-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Chen M.X., Feng L.L., Lu K., Li C., Wei Y.L., Jin J., et al. Association between geriatric nutritional risk index and all-cause mortality in individuals with osteoporotic fractures: a retrospective cohort study. Aging Clin. Exp. Res. 2025;37(1):77. doi: 10.1007/s40520-025-02978-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Zhou H., Lv D., Cui F., Qian G., Li J., Wen J., et al. Prognostic value of the geriatric nutritional risk index in patients with non-metastatic clear cell renal cell carcinoma: a propensity score matching analysis. Nutr. J. 2024;23(1) doi: 10.1186/s12937-024-01010-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Jin Y., Zhou T., Hou C., Zhang H., Xu B. Relationship between the geriatric nutritional risk index and sepsis in elderly critically ill patients: a retrospective cohort study. Eur. J. Med. Res. 2025;30(1):130. doi: 10.1186/s40001-025-02389-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Prokopidis K., Chen Y., Liu Y., Zhong Z., Morwani-Mangnani J., Cuthbertson D.J., et al. Assessing the link of malnutrition with diabetes and mortality risk in heart failure patients. ESC Heart Fail. 2025;12(4):3103–3113. doi: 10.1002/ehf2.15263. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Ohmure K., Kanda D., Ikeda Y., Tokushige A., Sonoda T., Arikawa R., et al. Impact of co-presence of malnutrition-inflammation-atherosclerosis factors on prognosis in lower extremity artery disease patients after endovascular therapy. Cardiovasc. Interv. Ther. 2025;40(1):102–111. doi: 10.1007/s12928-024-01058-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Argano C., Mirarchi L., Amodeo S., Orlando V., Torres A., Corrao S. The role of vitamin D and its molecular bases in insulin resistance, diabetes, metabolic syndrome, and cardiovascular disease: state of the art. Int. J. Mol. Sci. 2023;24(20) doi: 10.3390/ijms242015485. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Nishikawa H., Enomoto H., Nishiguchi S., Iijima H. Sarcopenic obesity in liver cirrhosis: possible mechanism and clinical impact. Int. J. Mol. Sci. 2021;22(4) doi: 10.3390/ijms22041917. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Xie D., Huang T., Yang C., Li C., Chen F. Predictive performance of the geriatric nutritional risk index for sarcopenia: a meta-analysis. Arch. Gerontol. Geriatr. 2025;138 doi: 10.1016/j.archger.2025.105985. [DOI] [PubMed] [Google Scholar]
- 36.Rasheedy D., El-Kawaly W.H. The accuracy of the geriatric nutritional risk index in detecting frailty and sarcopenia in hospitalized older adults. Aging Clin. Exp. Res. 2020;32(12):2469–2477. doi: 10.1007/s40520-020-01492-5. [DOI] [PubMed] [Google Scholar]
- 37.Matsuura S., Shibazaki K., Uchida R., Imai Y., Mukoyama T., Shibata S., et al. Sarcopenia is associated with the Geriatric Nutritional Risk Index in elderly patients with poorly controlled type 2 diabetes mellitus. J. Diabetes Investig. 2022;13(8):1366–1373. doi: 10.1111/jdi.13792. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Shirahama Y., Tabata N., Ishi M., Kusaka H., Hanatani S., Usuku H., et al. Nutritional and inflammatory status in the obesity paradox of coronary artery disease. Clin. Nutr. 2026;57 doi: 10.1016/j.clnu.2025.11.024. [DOI] [PubMed] [Google Scholar]
- 39.Zou Y., Ye H., Xu Z., Yang Q., Zhu J., Li T., et al. Obesity, sarcopenia, Sarcopenic obesity, and hypertension: mediating role of inflammation and insulin resistance. J. Gerontol. A Biol. Sci. Med. Sci. 2025;80(3) doi: 10.1093/gerona/glae284. [DOI] [PubMed] [Google Scholar]
- 40.Vellas B., Guigoz Y., Garry P.J., Nourhashemi F., Bennahum D., Lauque S., et al. The Mini nutritional assessment (MNA) and its use in grading the nutritional state of elderly patients. Nutrition. 1999;15(2):116–122. doi: 10.1016/s0899-9007(98)00171-3. [DOI] [PubMed] [Google Scholar]
- 41.Stratton R.J., Hackston A., Longmore D., Dixon R., Price S., Stroud M., et al. Malnutrition in hospital outpatients and inpatients: prevalence, concurrent validity and ease of use of the ‘malnutrition universal screening tool’ (‘MUST’) for adults. Br. J. Nutr. 2004;92(5):799–808. doi: 10.1079/bjn20041258. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary tables
Stratified analyses for all-cause mortality according to GNRI status.
Forest plot showing hazard ratios and 95% confidence intervals for low GNRI (<98) versus high GNRI (≥98, reference) across subgroups. P for interaction was calculated using the likelihood ratio test. All models were adjusted for covariates except for the stratification variable. GNRI, Geriatric Nutritional Risk Index; HR, hazard ratio; CI, confidence interval; PIR, poverty income ratio.
Stratified analyses for cardiovascular mortality according to GNRI status.
Forest plot showing hazard ratios and 95% confidence intervals for low GNRI (<98) versus high GNRI (≥98, reference) across subgroups. P for interaction was calculated using the likelihood ratio test. All models were adjusted for covariates except for the stratification variable. GNRI, Geriatric Nutritional Risk Index; HR, hazard ratio; CI, confidence interval; PIR, poverty income ratio.
Data Availability Statement
Publicly available datasets were analyzed in this study. The data can be found at the National Health and Nutrition Examination Survey (NHANES) website (https://www.cdc.gov/nchs/nhanes/).