Abstract
Little is known about the optimal measure of handgrip strength for predicting all-cause mortality and whether this association is modified by age or sex. We used data from the 2011–2014 National Health and Nutrition Examination Survey (NHANES), 9,583 adults aged ≥ 20 years were included. Equal-length grip strength was measured using a digital handheld Takei dynamometer. We defined five measurements of grip strength, i.e., the average of handgrip strength (HGS), maximum of grip strength (MGS), HGS/body mass index (BMI), HGS/height (HT)2, and MGS/weight, and three indicators of low grip strength, namely, low reference grip strength, lowest 20% grip strength, and low grip strength in sarcopenia. Information on deaths were obtained through linkage to National Death Index (NDI). Cox regression was used to assess the association of grip strength with mortality risk. HGS, MGS, HGS/BMI, HGS/HT2, and MGS/weight were all inversely associated with all-cause mortality, with HGS or HGS/HT2 (the area under the curve (AUC) = 0.714) being the optimal predictor of mortality, followed by MGS (AUC = 0.712). Participants with low grip strength showed increased risk of mortality regardless of which indicator was used, and the highest effect size was seen for lowest 20% grip strength group (hazard ratio (HR) = 2.20 for men, 2.52 for women). The above-mentioned correlations were consistently found in people of different ages and sexes. This study suggests the simplest measure of absolute grip strength (HGS, MGS) was the optimal index for predicting all-cause mortality, followed by HGS/HT2. Keeping an adequate level of handgrip strength may be beneficial to reduce the risk of mortality.
Supplementary Information
The online version contains supplementary material available at 10.1038/s41598-024-80487-y.
Keywords: Grip strength, All-cause mortality, Optimal measurement index, Interactions
Subject terms: Epidemiology, Preventive medicine, Risk factors
Introduction
Hand grip strength is a simple indicator of upper limb muscle strength, and a reflection of overall skeletal muscle strength1. As a measurement index, it has the advantages of easy to obtain, repeated application and cost effectiveness. It can quickly and quantitatively evaluate muscle function, which is one of the basic measurement indexes of physical examination2,3.
Currently, the measurements of grip strength are not uniform, limiting the comparability of research results, and the optimal grip strength index is unclear4. Common grip strength indicators include the average of handgrip strength (HGS)5 and maximum of grip strength (MGS)6, which are often called absolute grip strength. However, in general, grip strength is easily affected by body measurement indicators such as height, weight, and body mass index (BMI). Absolute grip strength as a simple indicator of grip strength cannot take into account the differences in body size, so some researchers combine absolute grip strength with body size indicators as a relative grip strength indicator7–9, e.g., HGS/BMI, MGS/body weight, HGS/height (HT)2, HGS/muscle tissue content, and HGS/muscle mass. Of these, HGS/BMI is a widely used measure in academic research, while MGS/weight was officially used in National Standards for Students’ Physical Fitness and Health in China to evaluate the muscle strength of adolescents10. Meanwhile, a study based on the National Health and Nutrition Examination Survey (NHANES)11, found height to be the optimal body size associated with HGS, suggesting that the square of height be used to normalize HGS. Some studies showed that relative grip strength measurements may be a better indicator for muscle weakness and more predictive of adverse outcomes8,12, whereas others not13. The application conditions of absolute and relative grip strength indexes need to be further studied.
Low grip strength represents a decline in muscle function, and increases the risk of frailty, disease, and death14–17. However, there is no consensus on the definition of low grip strength18. One of the definitions was “low reference grip strength”, defined as the HGS lower than the population reference value of grip strength calculated by the equation according to the gender, age, height, and weight of the sample population19. Another definition was “lowest 20% grip strength”, defined as the lowest 20% of grip strength in the population sampled adjusted for gender and BMI according to the definition of frailty20,21. “Sarcopenia low grip strength” was defined as MGS less than 26 kg in men or less than 16 kg in women22.
All-cause mortality, which encompasses deaths from all causes, reflects the overall health status of a population and provides the most intuitive demonstration of the impact of exposures on the population from a holistic perspective23. Previous evidence suggested an inverse association between hand grip strength and all-cause mortality24–26. However, the comparison and predictive ability of mortality for different grip strength measurements have not reached a consensus27. Meanwhile, whether the association between grip strength and death differed by age or sex was inconsistent28,29.
Therefore, the aim of this study was to determine the association of the indicators of grip strength level and low grip strength with the risk of all-cause mortality in adults registered in NHANES, and to assess the optimal measure of grip strength for predicting all-cause mortality.
Methods
The NHANES is a national, periodic, cross-sectional survey which aims to obtain the number, distribution and influence factors of disease and disability in the United States. The data were acquired from a complex, multistage probability sampling design in noninstitutionalized civilian resident, and information about questionnaire, physical examination, and laboratory examination were collected. The NHANES study was approved by the Research Ethics Review Board of the National Center for Health Statistics, and all study participants provided written informed consent. More details about NHANES have been described elsewhere30. This study was a secondary data analysis of publicly available federal data without personal identifiers, therefore, did not require Institutional Review Board review.
We used data from the 2011–2014 cycle of NHANES. A total of 19,931 people were recruited and 19,151 people attended the mobile examination center (MEC) physical examination. We included adults aged ≥ 20 years and excluded those pregnant and with incomplete or unqualified death information recall or missing data on grip strength and BMI, finally 9,583 subjects were included in this study (Figure S1).
Assessment of grip strength
Equal-length grip strength was measured using a digital handheld Takei dynamometer (Model T.K.K.5401). After the inspector’s demonstration adjustment and practice test, participants took a standing position and squeezed the dynamometer as hard as they could with one hand, then repeated the test with the other hand. Grip strength was measured 3 times in each hand, alternating hands between trials with at least 60 s between each test.
We defined five measurements of handgrip strength. HGS was calculated as the average of the maximum values of the readings in both hands, expressed in kg. HGS/BMI was obtained by the ratio of HGS to BMI. HGS/HT2 was calculated as HGS/body height2 (in m). MGS was the maximum value of six measurements of both hands, and MGS/weight was MGS/body weight (in kg) *100.
We also defined three measurements of low grip strength. Low reference grip strength referred to the HGS lower than the calculated reference value in men and women respectively, according to the formula raised by Ying-Chih Wang19. Lowest 20% grip strength was defined as the lowest 20% of the MGS of the sample population stratified by gender and BMI20, specific cutoff value was shown in Table S1. Low grip strength in sarcopenia was defined as MGS less than 26 kg for men or less than 16 kg for women22.
Assessment of covariates
Information on covariates was obtained through questionnaires, physical examinations, and laboratory tests. Covariates considered in this study mainly included gender (men, women), age (< 65, ≥ 65)/(in years), race (Mexican American, Other Hispanic, Non-Hispanic White, Non-Hispanic Black, Non-Hispanic Asian and other non-Hispanic groups), education level (less than 9th grade, 9-11th grade, high school graduate, some college or AA degree and college graduate or above), marital status (live with someone, live alone), poverty-to-income ratio (PIR) grouped into quartiles (≤ 1.06, 1.07–2.15, 2.16–4.19, > 4.19), BMI (kg/m2: <18.5, 18.5–24.9, 25-29.9, ≥ 30), smoking status (never, former, current), alcohol consumption (never, current), regular physical activity (MET-min/week), composite cognitive function score (CFS, points), and underlying chronic diseases (no, yes) included hypertension, hyperlipidemia, diabetes, cardiovascular diseases (CVD), and cancer. Detailed definition of covariates was displayed in Supplementary Text S1.
Assessment of mortality
Information on deaths were obtained through linkage to National Death Index (NDI) to 31 December 2019 via participants’ serial numbers (https://www.cdc.gov/nchs/data-linkage/mortality-public.htm).
Statistical analysis
All statistical analyses were conducted using Stata 15.0. A two-sided P < 0.05 was considered statistically significant. We adjusted the data using appropriate survey weights, strata, and primary sampling units in accordance with the complex sample design of the NHANES and oversampling of subgroups of the sample.
Baseline characteristics of participants were presented by sex-specific quartiles of HGS. Analysis of variance was used for continuous variables and the chi-square test was used for categorical variables, with correction for sample weights. The follow-up time of the study was calculated from the NHANES 2011–2014 examination date until the last known date alive or censored, or 31 December 2019, whichever occurred first. Cox proportional hazards models was used to analyze the association of grip strength measurements which included HGS, HGS/BMI, HGS/HT2, MGS, and MGS/weight and low grip strength with all-cause mortality, indicators of low grip strength included low reference grip strength, lowest 20% grip strength, and low grip strength in sarcopenia. To achieve comparability of grip strength measures, we standardized both absolute and relative grip strength measurements to investigate their association with the risk of all-cause mortality. Considering the difference in grip strength values between men and women, we analyzed the data by gender.
In Cox proportional hazards regression analyses, model 1 was adjusted for age and race, fully adjusted model was further adjusted for education level, marital status, PIR, smoking status, alcohol consumption, regular exercise, baseline chronic diseases (hypertension, hyperlipidemia, diabetes, CVD, and cancer).
To assess the possible non-linear association of handgrip measurements with mortality risk, we performed restricted cubic spline regression, with Z-scores of handgrip strength measurements modelled as natural cubic splines with four knots (the 5th, 35th, 65th, and 95th percentiles). The receiver operating characteristic (ROC) curve was drawn, and the area under the curve (AUC) was calculated to compare the ability of different grip strength indexes in predicting the risk of all-cause mortality. To further validate the predictive ability of the different grip strength indices, we computed the Harrell C-index (which estimates the probability of agreement between observed and predicted responses) for a model that includes fully-adjusted models of five different grip strength indices, and compared their ability to predict all-cause mortality by dividing the samples completely randomly into a training set and a testing set and reporting the corresponding C-indexes. The differences were finally verified by two-by-two comparisons.
Based on gender stratification, we further conducted age stratification analysis (< 65 years, ≥ 65 years). To examine the robustness of our findings, we did the following three sensitivity analyses: First, we repeated the main analysis in individuals aged ≥ 65 years while further adjusting for CFS. Second, subjects who died in the first year after MEC follow-up were excluded to reduce the influence of reverse causation. Third, subjects who already had cancer or CVD at baseline were excluded since these diseases may affect grip strength.
Results
The mean age of all participants was 47.4 ± 16.8 years, and 49.3% were men. In general, with the decline of grip strength, participants were more likely to be older, being non-Hispanic Asian, and have lower level of education and lower income (i.e., lower PIR) (Table 1, Figure S2). They also tended to be former smokers, current drinkers, have less regular physical activity and lower values of BMI, and have higher prevalence of chronic diseases (hypertension, hyperlipidemia, diabetes, CVD, and cancer).
Table 1.
Baseline characteristics of participants by sex-specific quartiles of hand grip strength (HGS).
| All (n = 9,583) |
Q4 (highest) (n = 2,390) |
Q3 (n = 2,382) |
Q2 (n = 2,409) |
Q1 (lowest) (n = 2,402) |
P value | |
|---|---|---|---|---|---|---|
| Age (years) | 47.4 (16.8) | 39.2 (11.7) | 44.5 (13.7) | 49.7 (16.6) | 59.6 (20.2) | < 0.001 |
| Age range | 20–80 | 20–80 | 20–80 | 20–80 | 20–76 | |
| Men (%) | 49.3 | 51.8 | 51.5 | 47.0 | 45.6 | 0.029 |
| Race (%) | < 0.001 | |||||
| Mexican American | 8.0 | 7.2 | 8.6 | 8.1 | 8.3 | |
| Other Hispanic | 5.8 | 4.7 | 5.1 | 7.1 | 6.7 | |
| Non-Hispanic White | 67.3 | 65.5 | 69.9 | 66.7 | 66.6 | |
| Non-Hispanic Black | 11.2 | 17.7 | 10.0 | 8.2 | 7.9 | |
| Non-Hispanic Asian | 5.0 | 2.1 | 3.6 | 7.3 | 7.9 | |
| Others | 2.7 | 2.8 | 2.9 | 2.5 | 2.4 | |
| Educational level (%) | < 0.001 | |||||
| Less than 9th grade | 4.5 | 2.2 | 3.2 | 5.0 | 8.9 | |
| 9-11th grade | 10.7 | 9.0 | 9.6 | 10.9 | 14.1 | |
| High school graduate | 21.0 | 19.9 | 21.1 | 20.8 | 22.6 | |
| Some college or AA degree | 32.8 | 36.9 | 33.9 | 30.3 | 28.9 | |
| College graduate or above | 30.9 | 32.0 | 32.2 | 32.9 | 25.4 | |
| Marital status (%)* | 0.001 | |||||
| Live with someone | 61.8 | 64.7 | 63.2 | 62.3 | 55.4 | |
| Live alone | 38.2 | 35.3 | 36.8 | 37.7 | 44.6 | |
| Poverty-to-income ratio (%) | < 0.001 | |||||
| ≤ 1.06 | 16.4 | 16.1 | 14.7 | 15.8 | 19.9 | |
| 1.07–2.15 | 19.2 | 19.4 | 16.9 | 18.6 | 23.9 | |
| 2.16–4.19 | 26.2 | 25.3 | 27.6 | 27.1 | 24.0 | |
| >4.19 | 31.9 | 34.0 | 35.0 | 32.0 | 24.6 | |
| Smoke status (%) | < 0.001 | |||||
| Never | 56.1 | 58.3 | 55.7 | 53.8 | 56.2 | |
| Former | 24.0 | 20.3 | 23.1 | 26.2 | 27.7 | |
| Current | 19.9 | 21.3 | 21.1 | 20.0 | 16.0 | |
| Alcohol status (%) | < 0.001 | |||||
| Never | 74.7 | 77.4 | 76.2 | 75.2 | 65.5 | |
| Current | 19.4 | 17.4 | 18.1 | 20.3 | 28.7 | |
| MET-h/w | 56.2 (91.8) | 75.8 (104.4) | 59.1 (86.3) | 51.9 (87.0) | 31.3 (69.8) | < 0.001 |
| BMI (kg/m2) | 29.0 (6.9) | 30.5 (6.9) | 28.8 (6.3) | 28.2 (6.6) | 28.1 (7.5) | < 0.001 |
| Prevalence of chronic diseases (%) | ||||||
| Hypertension | 39.7 | 31.3 | 35.6 | 39.8 | 56.7 | < 0.001 |
| Hyperlipemia | 54.0 | 48.0 | 51.3 | 55.2 | 64.4 | < 0.001 |
| Diabetes | 13.4 | 6.9 | 10.9 | 14.4 | 24.4 | < 0.001 |
| CVD | 8.3 | 2.9 | 5.0 | 8.6 | 19.9 | < 0.001 |
| Cancer | 10.3 | 5.3 | 8.9 | 11.0 | 18.1 | < 0.001 |
HGS: the average of hand grip strength; BMI, body mass index; CVD, cardiovascular disease; MET, metabolic equivalent. Values are expressed as mean (SD) or percentage. Analysis of variance was used for continuous variables and the chi-square test was used for categorical variables, with correction for sample weights. *"Live with someone" includes married and living together, "Live alone" includes widowed, divorced, separated and unmarried.
Over a median of 6.75 years of follow-up, 805 deaths occurred. As shown in Table 2, in men, when grip strength was used as a continuous variable, the fully adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) for per 5 kg lower of HGS and MGS were 1.36 (1.26, 1.48) and 1.34 (1.24, 1.46); and the HRs (95% CIs) for one unit decrease of HGS/BMI, MGS/weight and HGS/HT2 were 3.65 (2.32, 5.75), 1.04 (1.02, 1.05) and 1.21 (1.15, 1.27), respectively. When we treated grip strength as a quartile variable, there was an increased mortality risk with the decline of grip strength (Ptrend <0.05). Compared with the highest quartile (Q4), the lowest quartile (Q1) had about 2-fold higher hazard for all-cause mortality, irrespective of which grip strength measures was applied.
Table 2.
Association between different hand grip strength measurements and risk of all-cause mortality in men.
| Q4 (highest) HR (95%CI) |
Q3 HR (95%CI) |
Q2 HR (95%CI) |
Q1 (lowest) HR (95%CI) |
Continuous* | P trend | |
|---|---|---|---|---|---|---|
| HGS | ||||||
| Number of deaths | 26 | 45 | 86 | 296 | 453 | |
| Mortality rate (/1,000 PYs) | 3.2 | 5.5 | 10.6 | 40.1 | 14.2 | |
| Model 1 | Reference | 1.54 (0.63, 3.75) | 1.98 (0.94, 4.16) | 4.55 (2.44, 8.49) | 1.46 (1.36, 1.57) | < 0.001 |
| Fully adjusted model | Reference | 1.27 (0.53, 3.03) | 1.54 (0.76, 3.11) | 3.00 (1.72, 5.22) | 1.36 (1.26, 1.48) | < 0.001 |
| HGS/BMI | ||||||
| Number of deaths | 36 | 48 | 112 | 257 | 453 | |
| Mortality rate (/1,000 PYs) | 4.4 | 5.9 | 14.0 | 34.4 | 14.2 | |
| Model 1 | Reference | 1.20 (0.57, 2.51) | 1.79 (1.04, 3.07) | 3.16 (1.85, 5.39) | 4.51 (2.95, 6.90) | < 0.001 |
| Fully adjusted model | Reference | 1.29 (0.61, 2.72) | 1.99 (1.19, 3.35) | 2.88 (1.66, 5.00) | 3.65 (2.32, 5.75) | < 0.001 |
| HGS/HT2 | ||||||
| Number of deaths | 29 | 40 | 82 | 302 | 453 | |
| Mortality rate (/1,000 PYs) | 3.5 | 4.9 | 10.2 | 40.4 | 14.2 | |
| Model 1 | Reference | 1.10 (0.48, 2.55) | 1.47 (0.75, 2.90) | 3.23 (1.82, 5.72) | 1.26 (1.20, 1.32) | < 0.001 |
| Fully adjusted model | Reference | 1.00 (0.45, 2.25) | 1.31 (0.69, 2.48) | 2.39 (1.43, 3.99) | 1.21 (1.15, 1.27) | < 0.001 |
| MGS | ||||||
| Number of deaths | 31 | 43 | 76 | 303 | 453 | |
| Mortality rate (/1,000 PYs) | 3.8 | 5.3 | 9.5 | 40.5 | 14.2 | |
| Model 1 | Reference | 1.66 (0.72, 3.81) | 1.68 (0.82, 3.47) | 4.66 (2.57, 8.45) | 1.44 (1.34, 1.55) | < 0.001 |
| Fully adjusted model | Reference | 1.47 (0.66, 3.29) | 1.33 (0.68, 2.59) | 3.16 (1.88, 5.29) | 1.34 (1.24, 1.46) | < 0.001 |
| MGS/weight | ||||||
| Number of deaths | 32 | 67 | 96 | 258 | 453 | |
| Mortality rate (/1,000 PYs) | 3.9 | 8.2 | 12.0 | 34.5 | 14.2 | |
| Model 1 | Reference | 1.60 (0.78, 3.30) | 1.36 (0.75, 2.49) | 3.00 (1.62, 5.56) | 1.04 (1.02, 1.06) | < 0.001 |
| Fully adjusted model | Reference | 1.89 (0.86, 4.16) | 1.72 (0.89, 3.32) | 3.09 (1.55, 6.15) | 1.04 (1.02, 1.05) | 0.001 |
HGS: the average of hand grip strength; MGS: Maximum grip strength; MGS/weight: MGS/body weight *100; HGS/HT2: HGS/height2; PYs: person-years. *Continuous was calculated as per 5 kg decrease in HGS and MGS, and per one unit decrease in HGS/BMI, HGS/HT2 and MGS/weight. Model 1: adjusted for age and race; Fully adjusted model: Model 1 + adjusted for education level, marital status, poverty-to-income ratio, smoking status, alcohol drinking, regular exercise, and baseline chronic diseases (cancer, cardiovascular disease, hypertension, hyperlipidemia, diabetes); Cox regression was corrected for sample weights. HR (95%CI) in bold indicates P ≤ 0.05.
In Table 3 for women, the associations of continuous grip strength were similar to men, the HRs (95% CIs) for per 5 kg lower of HGS and MGS were 1.49 (1.30, 1.69) and 1.45 (1.27, 1.66); and the HRs (95% CIs) for one unit decrease of HGS/BMI, MGS/weight and HGS/HT2 were 2.22 (1.28, 3.83), 1.02 (1.00, 1.03) and 1.23 (1.16, 1.32), respectively. When we treated grip strength as a quartile variable, except for HGS/BMI, the lowest quartile (Q1) in grip strength measurements were associated with a higher hazard for all-cause mortality compared with the highest quartile (Q4). After standardization of grip strength, the highest risk of all-cause death per 1-SD decrease in grip strength was found in HGS (men; adjusted HR = 1.81, 95% CI = 1.55–2.11, women; 1.62, 1.38–1.90) and HGS/HT2 (men; adjusted HR = 1.81, 95% CI = 1.55–2.10, women; 1.62, 1.38–1.90), followed by MGS (men; 1.79, 1.53–2.09, women; 1.60, 1.36–1.90) (Table S2).
Table 3.
Association between different hand grip strength measurements and risk of all-cause mortality in women.
| Q4 (highest) HR (95%CI) |
Q3 HR (95%CI) |
Q2 HR (95%CI) |
Q1 (lowest) HR (95%CI) |
Continuous* | P trend | |
|---|---|---|---|---|---|---|
| HGS | ||||||
| Number of deaths | 22 | 50 | 49 | 231 | 352 | |
| Mortality rate (/1,000 PYs) | 2.7 | 6.2 | 5.9 | 30.3 | 10.9 | |
| Model 1 | Reference | 1.55 (0.86, 2.82) | 1.43 (0.75, 2.73) | 3.47 (1.81, 6.66) | 1.67 (1.45, 1.93) | < 0.001 |
| Fully adjusted model | Reference | 1.50 (0.84, 2.67) | 1.38 (0.72, 2.64) | 2.81 (1.49, 5.27) | 1.49 (1.30, 1.69) | 0.001 |
| HGS/BMI | ||||||
| Number of deaths | 32 | 48 | 60 | 212 | 352 | |
| Mortality rate (/1,000 PYs) | 3.9 | 5.8 | 7.4 | 27.6 | 10.9 | |
| Model 1 | Reference | 0.87 (0.46, 1.65) | 0.65 (0.37, 1.14) | 1.85 (1.00, 3.41) | 4.44 (2.45, 8.07) | 0.001 |
| Fully adjusted model | Reference | 0.80 (0.46, 1.38) | 0.58 (0.36, 0.92) | 1.32 (0.81, 2.16) | 2.22 (1.28, 3.83) | 0.022 |
| HGS/HT2 | ||||||
| Number of deaths | 22 | 39 | 58 | 233 | 352 | |
| Mortality rate (/1,000 PYs) | 2.7 | 4.8 | 7.1 | 30.5 | 10.9 | |
| Model 1 | Reference | 1.67 (0.83, 3.38) | 1.53 (0.79, 2.98) | 3.75 (2.03, 6.92) | 1.32 (1.23, 1.41) | 0.001 |
| Fully adjusted model | Reference | 1.86 (0.93, 3.71) | 1.58 (0.80, 3.14) | 3.28 (1.82, 5.93) | 1.23 (1.16, 1.32) | 0.022 |
| MGS | ||||||
| Number of deaths | 21 | 50 | 47 | 234 | 352 | |
| Mortality rate (/1,000 PYs) | 2.6 | 6.1 | 5.7 | 30.4 | 10.9 | |
| Model 1 | Reference | 1.38 (0.63, 3.03) | 1.13 (0.50, 2.55) | 2.92 (1.30, 6.54) | 1.63 (1.41, 1.88) | 0.001 |
| Fully adjusted model | Reference | 1.34 (0.62, 2.91) | 1.12 (0.50, 2.47) | 2.44 (1.13, 5.27) | 1.45 (1.27, 1.66) | 0.004 |
| MGS/weight | ||||||
| Number of deaths | 23 | 59 | 72 | 198 | 352 | |
| Mortality rate (/1,000 PYs) | 2.8 | 7.2 | 8.9 | 25.7 | 10.9 | |
| Model 1 | Reference | 1.28 (0.69, 2.39) | 1.07 (0.56, 2.06) | 2.38 (1.25, 4.51) | 1.03 (1.02, 1.05) | < 0.001 |
| Fully adjusted model | Reference | 1.19 (0.67, 2.10) | 0.98 (0.56, 1.72) | 1.72 (1.01, 2.91) | 1.02 (1.00, 1.03) | 0.009 |
HGS: the average of hand grip strength; MGS: Maximum grip strength; MGS/weight: MGS/body weight *100; HGS/HT2: HGS/height2; PYs: person-years. *Continuous was calculated as per 5 kg decrease in HGS and MGS, and per one unit decrease in HGS/BMI, HGS/HT2 and MGS/weight. Model 1: adjusted for age and race; Fully adjusted model: Model 1 + adjusted for education level, marital status, poverty-to-income ratio, smoking status, alcohol drinking, regular exercise, and baseline chronic diseases (cancer, cardiovascular disease, hypertension, hyperlipidemia, diabetes); Cox regression was corrected for sample weights. HR (95%CI) in bold indicates P ≤ 0.05.
For low grip strength, although different definitions were applied, the higher mortality risk was consistently found in the low grip strength group compared with the normal grip strength group (Table 4). We observed an interaction of low reference grip strength and sex on the risk of all-cause mortality (Pinteraction =0.013), the HR (95% CI) was 1.91 (1.54, 2.35) in men, but lost significance in fully adjusted model for women. Subgroup analyses by age showed no difference in association among those < 65 years and ≥ 65 years, no matter expressed in continuous grip strength or low grip strength (Table S3-S4, Pinteraction >0.05).
Table 4.
Association between low grip strength of different definitions and the risk of all-cause mortality.
| Men | Women | P interaction | |||
|---|---|---|---|---|---|
| Normal | Low | Normal | Low | ||
| Low reference grip strength | 0.013 | ||||
| Number of deaths | 211 | 242 | 189 | 163 | |
| Mortality rate (/1,000 PYs) | 10.6 | 20.3 | 9.0 | 14.5 | |
| Model 1 | Reference | 2.12 (1.70, 2.64) | Reference | 1.61 (1.22, 2.11) | |
| Fully adjusted model | Reference | 1.91 (1.54, 2.35) | Reference | 1.22 (0.99, 1.51) | |
| Lowest 20% grip strength | 0.174 | ||||
| Number of deaths | 189 | 264 | 128 | 224 | |
| Mortality rate (/1,000 PYs) | 7.2 | 45.5 | 4.9 | 37.1 | |
| Model 1 | Reference | 2.70 (2.12, 3.44) | Reference | 3.03 (2.11, 4.34) | |
| Fully adjusted model | Reference | 2.20 (1.71, 2.83) | Reference | 2.52 (1.83, 3.46) | |
| Men ≤ 26 kg, Women ≤ 16 kg | 0.661 | ||||
| Number of deaths | 395 | 58 | 298 | 54 | |
| Mortality rate (/1,000 PYs) | 12.6 | 106.6 | 9.4 | 101.3 | |
| Model 1 | Reference | 2.94 (1.97, 4.39) | Reference | 3.37 (2.34, 4.86) | |
| Fully adjusted model | Reference | 2.06 (1.33, 3.19) | Reference | 2.29 (1.55, 3.38) | |
PYs: person-years. Model 1: adjusted for age (except for low reference grip strength) and race; Fully adjusted model: Model 1 + adjusted for education level, marital status, poverty-to-income ratio, smoking status, alcohol drinking, regular exercise, and baseline chronic diseases (cancer, cardiovascular disease, hypertension, hyperlipidemia, diabetes); Cox regression was corrected for sample weights. HR (95%CI) in bold indicates P ≤ 0.05.
In the restricted cubic spline regression, we observed a non-linear inverse association of Z-scores of grip strength with mortality risk, which were similar across different hand strength measurements (Figure S3). By comparing the AUC and C-indices of the five grip strength measurements, HGS or HGS/HT2 (AUC = 0.714; C-index = 0.860) was the best predictor of all-cause mortality risk, followed by MGS (AUC = 0.712; C-index = 0.860) (Fig. 1, Table S5). In sensitivity analyses, after adjustment for cognitive function among older adults, the effect sizes for all grip strength indicators decreased, or even disappeared (Table S6). when we excluded deaths during the first year of follow-up, or excluded participants with CVD or cancer at baseline, the result remained almost unchanged (Table S7-S8).
Fig. 1.
Receiver operating characteristic (ROC) curve of all-cause mortality risk after adjusting for age, gender and race predicted with different hand grip strength measurements HGS: the average of hand grip strength; MGS: maximum grip strength; MGS/weight: MGS/body weight *100; HGS/HT2: HGS/height2. There was no statistically significant difference between HGS, MGS and HGS/HT2, no statistically significant difference between HGS/BMI and MGS/weight. The C-indexes of HGS, MGS and HGS/HT2 were higher than that of HGS/BMI and MGS/weight (P < 0.05).
Discussion
The main finding of this study was that decreased grip strength in adults was associated with an increased risk for all-cause mortality. Using five different measures of grip strength, we found similar non-linear inverse correlations. HGS or HGS/HT2 had the best ability to predict all-cause mortality, followed by MGS. Participants with low grip strength were at increased risk of mortality, and the HR was highest in the lowest 20% grip strength group. We found no significant interaction of age, sex, and grip strength, except for low reference grip strength and gender.
Our finding of inverse associations between grip strength and all-cause mortality were in agreement with previous studies. A meta-analysis conducted on 42 studies including 3,002,203 participants showed that grip strength was an independent predictor of all-cause mortality, the HRs (95% CIs) with per-5-kg decrease in grip strength was 1.16 (1.12, 1.20) for all-cause mortality24. These results were comparable to our findings, where we found that per 5-kg lower HGS was associated with an HR of 1.36 in men for all-cause mortality, and an HR of 1.49 in women. Cai, Y. et al.27 similarly found that the HR (95% CI) for all-cause mortality per 5-kg reduction in grip strength was 1.11 (1.06, 1.18) in men and 1.17 (1.08, 1.28) in women.
In our study, we found that after standardizing all grip strength indicators, the association of HGS, MGS, and HGS/HT2 with the risk of all-cause mortality were stronger than HGS/BMI and MGS/weight. Inconsistently, a study, also based on NHANES, found that compared with the sum of the maximum values of both hands (GS), GS/BMI had a stronger correlation with cardiovascular biomarkers as a relative grip strength indicator7. Wonjeong Jeong et al.31 also found that HGS/BMI was more associated with risk of all-cause mortality compared with HGS, the associations persisted after adjustment for chronic diseases. Similarly, Yanan Gao et al.9 also recommended HGS/BMI or HGS/body weight as the best choice for grip strength expression to predict the risk of CVD risk factors. However, by comparing AUC in our study, we recommend HGS or MGS as the optimal predictor of all-cause mortality risk. Among the relative grip strength indicators, we recommend HGS/HT2 as the best predictor of all-cause mortality risk, and its AUC and C-index were basically the same as that of HGS, and the strength of the association with all-cause mortality risk was similar to that of HGS. Alan M Nevill et al. in NHANES11, by fitted centile curves for normalized HGS using the Generalized Additive Model for Location, Scale, and Shape (GAMLSS) found that height was one of the optimal body dimensions for signals associated with HGS. Meanwhile, Ho FKW et al.13 compared different expressions of grip strength in adults from UK Biobank and found no difference in the association between absolute and relative grip strength and all-cause mortality. They believed that the simplest method of handgrip strength measurement, i.e., the absolute unit (kg), was perfectly suitable for predicting health outcomes in clinical practice.
We found participants with low grip strength carried higher mortality risk, irrespective of which definition was used, which was in line with previous studies21,28,32. However, few studies have compared the effect of different measures of low grip strength in the same population. Our study showed that the HRs was highest when using the definition of “lowest 20% grip strength” (HR = 2.20 for men, 2.52 for women), possibly because this definition could identify participants with the worst muscle strength, and better distinguish them from people with normal grip strength.
Overall, we did not find significant modifying effect of gender and age on the association between grip strength and all-cause mortality, except for the interaction between low reference grip strength and gender, where the effect of low grip strength was only found in men. Our results were basicly consistent with the previous conclusion of Strand BH et al. in Tromsø database26. However, this was opposite to the results of a previous KORA-based study33, which found that the association between muscular strength and all-cause mortality tended to be stronger in women. Actually, most previous studies suggested no interaction of grip strength and gender, on the risk of death24,26,28.
Existing evidence stratified by age also suggested inconclusive results. A study based on UK Biobank demonstrated that the hazard ratio of grip strength with all-cause death was higher in the younger age group in both genders28. Similar results were also found in the China Health and Retirement Longitudinal Study (CHARLS) study, where the association between MGS and all-cause mortality was stronger in young men (HR = 0.29, 95% CI: 0.18–0.45) than in older men (HR = 0.49, 95% CI:0.33–0.73)34. Whereas Rachel et al.29 found that the association between low grip strength and all-cause mortality was higher in people over 70 years old (HR = 1.80, 95% CI: 1.48–2.18) than those of under 60 years old (HR = 1.43, 95% CI: 1.07–1.91). Further research is needed to examine the interaction effect of age and sex.
Grip strength, as a simple, safe, reliable and low-cost indicator, can evaluate muscle function1, nutritional status15, and etc., and it is easy to popularize and apply in the clinical settings. Our study showed that HGS, MGS and HGS/HT2 had similar predictive ability of all-cause mortality, which implies that both absolute and relative grip strength measurements were effective predictors of mortality risk. Of course, we highly recommend HGS or MGS, as they are more direct and convenient for clinical application. Researchers can use grip strength as a simple indicator for risk stratification of the population and identify people with low grip strength, so that interventions can be taken to reduce the risk of mortality.
Our study had several advantages. First, we used relatively comprehensive measures of grip strength, including absolute grip strength (HGS, MGS), relative grip strength (HGS/BMI, MGS/weight, HGS/HT2), and low grip strength, to examine and compare their associations with all-cause mortality. Meanwhile, by using the relative grip strength index, the influence of body shape factors (such as height, weight, BMI) can be effectively excluded to some extent, so that more reliable conclusions can be obtained. Second, this study used a large representative sample from the American NHANES. The grip strength was standardized (Z score) to further study the correlation between grip strength and all-cause mortality. The Harrell C-index and AUC were used to obtain the optimal index of grip strength to predict the risk of all-cause death. However, our study had some limitations. First, as the NHANES is a repeated cross-sectional study, it cannot provide data on changes in grip strength, so we could only examine the association of baseline grip strength with all-cause death. Nevertheless, previous studies have found that baseline grip strength measures are the best predictors of cardiovascular mortality35. Second, as an outcome variable in handgrip strength studies, the composition of all-cause mortality tends to be more complex. For example, accidental death may not be intrinsically related to grip strength, so if accidental death accounted for a substantial proportion during the study, it is likely to affect the validity of the study results. Third, we only employed three widely-used relative grip strength indicators (HGS/BMI, MGS/weight, and HGS/HT2), exploring other combinations such as MGS/BMI or HGS/weight may also be of great value in the future. Finally, although we have accounted for as many confounding factors as possible, residual confounding still exists.
Conclusion
The present study has shown that all five measures of grip strength level are significantly and inversely associated with all-cause mortality, and the simple absolute grip strength measurements (HGS, MGS) are the best predictor of all-cause mortality, followed by HGS/HT2. Our results show that low grip strength is associated with an increased risk of all-cause mortality, with the strongest association observed in the lowest 20% grip strength group. The above results remain significant in different ages and genders. We recommend that the simplest measure of absolute grip strength (HGS, MGS) be used as the optimal indicator for predicting all-cause mortality. Keeping an adequate level of handgrip strength may be beneficial to reduce the risk of mortality.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Acknowledgements
We acknowledge the staff at the National Center for Health Statistics at the CDC, who design, collect, administer the NHANES data and release the data available for public use. We are thankful to all study participants for their cooperation.
Author contributions
JF conceived and designed the study. JF and LC analyzed the data. LC drafted the manuscript. JF and LC helped the interpretation of the results. JF contributed to the critical revision of the manuscript for important intellectual content and approved the final version of the manuscript. All authors reviewed and approved the final manuscript. DZ is the guarantor.
Funding
This work was supported by Natural Science Foundation of Shandong Province (ZR2023QH188), China Postdoctoral Science Foundation (2023M731839), Qingdao Postdoctoral Innovation Project (QDBSH20230102012), Qingdao University Scientific Research Startup Fund (DC2200002531), Mount Taishan Scholar Youth Program. The funders had no role in the study design, data collection, data analysis and interpretation, writing of the report, or the decision to submit the article for publication.
Data availability
Publicly available datasets were analyzed in this study. This data can be found here: https://wwwn.cdc.gov/nchs/nhanes/ (accessed on 14 February 2023).
Declarations
Ethics approval and consent to participate
The NHANES study was approved by the Research Ethics Review Board of the National Center for Health Statistics, and all study participants provided written informed consent. This study was a secondary data analysis of publicly available federal data without personal identifiers, therefore, did not require Institutional Review Board review.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Porto, J. M. et al. Relationship between grip strength and global muscle strength in community-dwelling older people. Arch. Gerontol. Geriatr.82, 273–278. 10.1016/j.archger.2019.03.005 (2019). [DOI] [PubMed] [Google Scholar]
- 2.Bohannon, R. W. Muscle strength: clinical and prognostic value of hand-grip dynamometry. Curr. Opin. Clin. Nutr. Metab. Care. 18, 465–470. 10.1097/mco.0000000000000202 (2015). [DOI] [PubMed] [Google Scholar]
- 3.Bohannon, R. W. Grip strength: an indispensable biomarker for older adults. Clin. Interv. Aging. 14, 1681–1691. 10.2147/cia.S194543 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Schaap, L. A. et al. Grip strength measurement: towards a standardized approach in sarcopen ia research and practice. Eur. Geriatr. Med.7, 247–255. 10.1016/j.eurger.2015.11.012 (2016). [Google Scholar]
- 5.Leong, D. P. et al. Prognostic value of grip strength: findings from the prospective Urban Rural Epidemiology (PURE) study. Lancet (London England). 386, 266–273. 10.1016/s0140-6736(14)62000-6 (2015). [DOI] [PubMed] [Google Scholar]
- 6.Yoo, J. I., Choi, H. & Ha, Y. C. Mean Hand grip strength and cut-off value for sarcopenia in Korean adults using KNHANES VI. J. Korean Med. Sci.32, 868–872. 10.3346/jkms.2017.32.5.868 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Lawman, H. G. et al. Associations of relative handgrip strength and cardiovascular disease biomarkers in U.S. adults, 2011–2012. Am. J. Prev. Med.50, 677–683. 10.1016/j.amepre.2015.10.022 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Parra-Soto, S., Pell, J. P., Celis-Morales, C. & Ho, F. K. Absolute and relative grip strength as predictors of cancer: prospective cohort study of 445 552 participants in UK Biobank. J. cachexia Sarcopenia Muscle. 13, 325–332. 10.1002/jcsm.12863 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Gao, Y. et al. Comparison of five expressions of handgrip strength for predicting cardiovascular disease risk factors in Chinese middle-aged community residents. Front. Public. Health. 10, 903036. 10.3389/fpubh.2022.903036 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Yan, X. J. et al. The application of the National standards for Students’ Physical Health (2014 revision) in SPSS. Chin. J. Prev. Med.54, 708–712. 10.3760/cma.j.cn112150-20200306-00270 (2020). [DOI] [PubMed] [Google Scholar]
- 11.Nevill, A. M., Tomkinson, G. R., Lang, J. J., Wutz, W. & Myers, T. D. How should adult handgrip strength be normalized? Allometry reveals new insights and associated reference curves. Med. Sci. Sports. Exerc.54, 162–168. 10.1249/mss.0000000000002771 (2022). [DOI] [PubMed] [Google Scholar]
- 12.Alley, D. E. et al. Grip strength cutpoints for the identification of clinically relevant weakness. J. Gerontol. A. 69, 559–566. 10.1093/gerona/glu011 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ho, F. K. W. et al. The association of grip strength with health outcomes does not differ if grip strength is used in absolute or relative terms: a prospective cohort study. Age Ageing. 48, 684–691. 10.1093/ageing/afz068 (2019). [DOI] [PubMed] [Google Scholar]
- 14.Xie, K. H., Han, X., Zheng, W. J. & Zhuang, S. F. Low grip strength and increased mortality hazard among middle-aged and older Chinese adults with chronic diseases. Biomed. Environ. Sci.: BES. 36, 213–221. 10.3967/bes2023.013 (2023). [DOI] [PubMed] [Google Scholar]
- 15.Norman, K., Stobäus, N., Gonzalez, M. C., Schulzke, J. D. & Pirlich, M. Hand grip strength: outcome predictor and marker of nutritional status. Clin. Nutr.30, 135–142. 10.1016/j.clnu.2010.09.010 (2011). [DOI] [PubMed] [Google Scholar]
- 16.Kim, Y. et al. Independent and joint associations of grip strength and adiposity with all-cause and cardiovascular disease mortality in 403,199 adults: the UK Biobank study. Am. J. Clin. Nutr.106, 773–782. 10.3945/ajcn.117.156851 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Liu, W. et al. The association of grip strength with cardiovascular diseases and all-cause mortality in people with hypertension: findings from the prospective urban rural epidemiology China study. J. Sport Health Sci.10, 629–636. 10.1016/j.jshs.2020.10.005 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Dodds, R. M. et al. Global variation in grip strength: a systematic review and meta-analysis of normative data. Age Ageing. 45, 209–216. 10.1093/ageing/afv192 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wang, Y. C., Bohannon, R. W., Li, X., Sindhu, B. & Kapellusch, J. Hand-grip strength: normative reference values and equations for individuals 18 to 85 years of age residing in the United States. J. Orthop. Sports Phys. Ther.48, 685–693. 10.2519/jospt.2018.7851 (2018). [DOI] [PubMed] [Google Scholar]
- 20.Fried, L. P. et al. Frailty in older adults: evidence for a phenotype. J. Gerontol. Ser. A Biol. Sci. Med. Sci.56, M146-156, 10.1093/gerona/56.3.m146 (2001). [DOI] [PubMed]
- 21.Chainani, V. et al. Objective measures of the frailty syndrome (hand grip strength and gait speed) and cardiovascular mortality: a systematic review. Int. J. Cardiol.215, 487–493. 10.1016/j.ijcard.2016.04.068 (2016). [DOI] [PubMed] [Google Scholar]
- 22.Studenski, S. A. et al. The FNIH sarcopenia project: rationale, study description, conference recommendations, and final estimates. J. Gerontol. A. 69, 547–558. 10.1093/gerona/glu010 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Global burden of 288 causes of death and life expectancy decomposition. In 204 countries and territories and 811 subnational locations, 1990–2021: a systematic analysis for the global burden of disease study 2021. Lancet (London England). 403, 2100–2132. 10.1016/s0140-6736(24)00367-2 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Wu, Y., Wang, W., Liu, T. & Zhang, D. Association of grip strength with risk of all-cause mortality, cardiovascular diseases, and cancer in community-dwelling populations: a meta-analysis of prospective cohort studies. J. Am. Med. Dir. Assoc.18, 551.e517–551.e535. 10.1016/j.jamda.2017.03.011 (2017). [DOI] [PubMed] [Google Scholar]
- 25.Pavasini, R. et al. Grip strength predicts cardiac adverse events in patients with cardiac disorders: an individual patient pooled meta-analysis. Heart (British Cardiac Society). 105, 834–841. 10.1136/heartjnl-2018-313816 (2019). [DOI] [PubMed] [Google Scholar]
- 26.Strand, B. H. et al. The association of grip strength from midlife onwards with all-cause and cause-specific mortality over 17 years of follow-up in the Tromsø study. J. Epidemiol. Commun. Health. 70, 1214–1221. 10.1136/jech-2015-206776 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Cai, Y. et al. Linear association between grip strength and all-cause mortality among the elderly: results from the SHARE study. Aging Clin. Exp. Res.33, 933–941. 10.1007/s40520-020-01614-z (2021). [DOI] [PubMed] [Google Scholar]
- 28.Celis-Morales, C. A. et al. Associations of grip strength with cardiovascular, respiratory, and cancer outcomes and all cause mortality: prospective cohort study of half a million UK Biobank participants. BMJ361, k1651. 10.1136/bmj.k1651 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Cooper, R., Kuh, D. & Hardy, R; Mortality Review Group; FALCon and HALCyon Study Teams. Objectively measured physical capability levels and mortality: systematic review and meta-analysis. BMJ 341, c4467. 10.1136/bmj.c4467 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Zipf, G. et al. National health and nutrition examination survey: plan and operations, 1999-2010. Vital Health Stat.1, 1–37. (2013). [PubMed]
- 31.Jeong, W., Moon, J. Y. & Kim, J. H. Association of absolute and relative hand grip strength with all-cause mortality among middle-aged and old-aged people. BMC Geriatrics.23, 321. 10.1186/s12877-023-04008-8 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Mickute, M. et al. Individual frailty phenotype components and mortality in adults with type 2 diabetes: a UK Biobank study. Diabetes Res. Clin. Pract.195, 110155. 10.1016/j.diabres.2022.110155 (2023). [DOI] [PubMed] [Google Scholar]
- 33.Arvandi, M. et al. Gender differences in the association between grip strength and mortality in older adults: results from the KORA-age study. BMC geriatrics.16, 201. 10.1186/s12877-016-0381-4 (2016). [DOI] [PMC free article] [PubMed]
- 34.Wang, Y. et al. Association of handgrip strength with all-cause mortality: a nationally longitudinal cohort study in China. J. Sci. Med. Sport. 25, 878–883. 10.1016/j.jsams.2022.08.005 (2022). [DOI] [PubMed] [Google Scholar]
- 35.Prasitsiriphon, O. & Pothisiri, W. Associations of grip strength and change in grip strength with all-cause and cardiovascular mortality in a European older population. Clin. Med. Insights Cardiol.12, 1179546818771894. 10.1177/1179546818771894 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Publicly available datasets were analyzed in this study. This data can be found here: https://wwwn.cdc.gov/nchs/nhanes/ (accessed on 14 February 2023).