Handgrip Strength as a Predictor of Nutritional Status in Chinese Elderly Inpatients at Hospital Admission

The study protocol was reviewed and approved by the Institutional Review Board of Chinese PLA General Hospital. All procedures in this study were conducted in accordance with the Helsinki Declaration, and subjects gave written informed consent.

In this cross-sectional study, a total number of 1, 343 patients (691 men and 652 women) between the ages of 65 and 90 years, with a mean age of 73.8 years (SD = 5.9), receiving medical care at internal medicine of Chinese PLA General Hospital were included. Data were obtained by accredited practicing dietitians via interviews and clinical screening between May 2014 and December 2015. Briefly, the inclusion criteria were participants who were between 65 and 90 years of age, were admitted for 24 h to the internal medicine wards, and could be measured for handgrip strength. Those who had injury, malformation, or severe rheumatoid arthritis in both hands; pregnant women; and critically ill or unconscious patients were excluded.

Body weight and height were recorded to the nearest 0.1 kg and 0.1 cm. Body mass index (BMI) was computed as weight (kg) divided by height (m) squared. Waist circumference (WC) was measured to the nearest 0.1 cm at the umbilical level horizontally in a standing position, and calf circumference (CC) was measured to the nearest 0.1 cm at the thickest point in a relaxed and standing position.

Clinical data including age, gender, hospital department, and diagnosis were collected by nurses at hospital admission. Following admission, fasting blood samples were taken from each patient for laboratory assays including serum levels of albumin (ALB), total protein (TP), creatinine (Cr), triglycerides (TG), total cholesterol (TC), hemoglobin (Hb), blood urea nitrogen, and C-reactive protein (CRP).

Patientsa' nutrition risk was assessed by using the NRS 2002^[7], including initial screening, which consists of four questions as a pre-screening, and final screening, which consists of three aspects: impaired nutritional status, severity of disease, and age. Nutritional assessments were categorized as at nutritional risk (NRS 2002 score ≥ 3) or no nutritional risk (NRS 2002 score < 3).

Nutritional status was determined by using SGA, which was carried out as proposed by Detsky et al.^[21]. The following parameters were assessed: body weight change in the previous 6 months and in the last 15 days; nutritional history including appetite, dietary intake, and gastrointestinal symptoms; gastrointestinal derangements including diarrhea, vomiting, and nausea; functional physical capacity; and physical assessment of fat loss, edema, muscle wasting, and ascites. Patients were classified as a'well nourisheda' (SGA A) or a'malnourished or at risk of malnutritiona' (SGA B or C).

HGS was measured using a Jamar dynamometer (Sammons Preston, USA) according to a previously described method^[22]. The Jamar dynamometer has a range of 1-90 kg and an accuracy of 0.1 kg. The arm was positioned at the side of the body and the dynanometer held with the elbow flexed at 90°. The patient was asked to squeeze the device as hard as possible for 3 s. The measurement was repeated twice more at intervals of at least 30 s. Three measurements were taken from each of the patienta's hands. For the present analysis, the maximal value of three consecutive measurements was registered in kilogram units. Then, the mean HGS result was calculated, and the HGS of the dominant hand was recorded. The patienta's dominant HGS was used for the assessment.

The data were tabulated using Excel 2007, and statistical analyses for all phases of the study were performed using SPSS 18.0 (SPSS Inc., Chicago IL, USA). Quantitative variables of symmetric and asymmetric distributions were described as the mean ± SD or medians and 25^th-75^th percentile values, while categorical variables were presented as numbers (proportions).

Frequencies were compared using the χ² test. Means and SD values were compared with the Studenta's t-test. The Mann-Whitney U test was used to compare asymmetrically distributed quantitative variables. A multivariate logistic regression analysis was used to evaluate significant associations between dominant HGS and other variables while adjusting for potential confounding factors, such as age and disease status. As there are large differences in HGS between genders, the analysis was stratified by gender. Finally, if there was a statistically significant independent association between dominant HGS and nutritional status according to NRS 2002 and SGA, ROC curves were generated to determine the optimal cut-off points for dominant HGS by maximizing Youdena's index (sensitivity + specificity -1)^[23]. With the identified cut-off points, the sensitivity, specificity, area under the curve (AUC), and 95% confidence intervals (Cls) were calculated. The significance level was set at P < 0.05.

The characteristics of the 1, 343 patients are shown in Table 1. There was no difference in age between men and women. Respiratory (14.59%), endocrinological (15.26%), neurological (22.93%), gastroenterological (11.39%), and cardiological (25.76%) diseases were common at hospital admission. BMI, WC, CC, and HGS; serum levels of TP, Cr, TG, TC, Hb, and CRP; and the number of internal medicine patients were significantly different between women and men.

The classification of nutritional status by NRS 2002 showed that 63.81% of the population was at nutritional risk (men 69.75%, women 57.52%), while the classification of nutritional status by SGA showed that 28.22% were malnourished (men 25.57%, women 31.13%). Women were at a lower nutritional risk, but at a higher risk of being malnourished (Figure 1).

Figure 1.  The percentage of nutritional risk or malnutrition in elderly inpatients. Data are presented as proportions (%). Statistically significant differences from the man are indicated as ^*P < 0.05.

The left HGS, right HGS, mean HGS, optimal HGS, BMI, WC, and CC and the serum levels of ALB, TG, and Hb of inpatients with malnutrition or nutritional risk were significantly lower than those of well-nourished inpatients, both male and female. However, in women, age and CRP were higher, and the serum level of TP was significantly lower in malnutrition or nutritional risk inpatients than in well-nourished inpatients (Table 2).

The multivariate logistic regression model showed that, in male subjects, according to NRS 2002, the odds ratio (OR) of optimal HGS was 0.93 (95% CI 0.90-0.97, P < 0.001), the age OR was 1.05 (95% CI 1.01-1.10, P = 0.018), and the BMI was 0.83 (95% CI 0.74-0.92, P = 0.001). According to SGA, the optimal HGS was 0.93 (95% CI 0.89-0.96, P < 0.001), the BMI was 0.87 (95% CI 0.77-0.99, P = 0.032), the ALB was 0.85 (95% CI 0.78-0.94, P = 0.001), and the TP was 1.07 (95% CI 1.01-1.13, P = 0.017). For female subjects, according to NRS 2002, the OR of optimal HGS was 0.93 (95% CI 0.89-0.98, P = 0.002), the age was 0.93 (95% CI 0.89-0.98, P = 0.002), and the BMI was 0.80 (95% CI 0.72-0.88, P < 0.001). According to SGA, the OR of optimal HGS was 0.93 (95% CI 0.88-0.98, P = 0.003), the BMI was 0.84 (95% CI 0.76-0.94, P = 0.001), and the CC OR was 0.90 (95% CI 0.82-0.99, P = 0.031). There was no significant difference in any other variable (Table 3).

To determine the optimal HGS value to detect nutritional risk and malnutrition in men and women, we calculated the optimal HGS cut-off values by maximizing the Youden Index by sex and age (Table 4 and Figure 2). Based on the ROC curve of NRS 2002, we identified the optimal cut-off points as 27.5 kg (65-74 years) and 21.0 kg (75-90 years) for men and 17.0 kg (65-74 years) and 14.6 kg (75-90 years) for women. Likewise, for SGA, we determined the optimal cut-off points as 24.9 kg (65-74 years) and 20.8 kg (75-90 years) for men and 15.2 kg (65-74 years) and 13.5 kg (75-90 years) for women.

Figure 2.  Gender-and age-specific ROC curves for identifying malnutrition or nutrition risk according to different cut-points for optimal HGS. AUC is indicated in the figure. The AUC is significantly different from 0.5, P < 0.05 for all.

Our study found a strong correlation between HGS and nutritional status in elderly inpatients at hospital admission, in agreement with previous literature^{[18-19, 24-26]}. Our findings also demonstrated that HGS cut-off values can identify a high proportion of elderly inpatients with malnutrition or at nutritional risk and can be used for assessing the nutritional status of elderly inpatients at hospital admission in China.

The prevalence of malnutrition or nutritional risk was greater than 40% among elderly inpatients^{[1-2, 27]}. This study showed a 64% prevalence of nutritional risk and a 28% prevalence of malnourishment in elderly inpatients at hospital admission, and the prevalence of nutritional risk and HGS was lower and that of malnourishment was higher in women than in men. Age and sex were the two most important independent determining factors of HGS^[28-29]. HGS decreases with increasing age^[30-31], and the rate of loss of HGS was shown to be twice as fast in older groups compared with that in younger groups. Moreover, older women had 43% lower HGS than older men^{[30, 32]}. This study involved elderly inpatients who were assessed by SGA or NRS 2002 to identify malnutrition or nutritional risk. We found that HGS, BMI, WC, and CC and the serum level of ALB, TG, and Hb of inpatients with malnutrition or nutritional risk were significantly lower, and the age was higher than that of well-nourished inpatients, both male and female. However, in women serum levels of TP were significantly lower, and those of CRP were significantly higher. This showed that HGS, like other indicators, was a stronger predictor of nutritional status. We further analyzed the adjusted multivariate logistic regression model and showed that lower HGS was associated with malnutrition or nutritional risk status in patients of both sexes. Similar to other studies, HGS and PG-SGA scores were significantly correlated^[19], as was impaired handgrip strength with poor nutritional status (BMI < 18.5 kg/m²)^[26]. A systematic review highlighted HGS as a good indicator of nutrition status^[18], and HGS is an excellent tool to assess nutrition status at the bedside in patients with cirrhosis. Furthermore, HGS has higher diagnostic accuracy than other anthropometric tests such as mid-arm muscle circumference and triceps skin fold thickness^[33].

HGS reflects overall body muscle strength and can indicate nutritional deprivation before a change occurs in body composition^{[18, 34]}. HGS can be a reliable first screening tool for nutritional risk in hospitals^[35] and can independently predict nutritional status and changes in nutritional status defined by PG-SGA score and category^[19]. A study reported that the overall HGS cut-points for the likelihood of mobility limitation were 37 kg for men and 21 kg for women^[36]. Another study reported that only men with low handgrip strength had an increased risk of functional decline at discharge, with a specificity of 91.3% and a cut-off point of 20.65 kg for handgrip strength, whereas women had no significant association between handgrip strength and functional decline^[37]. This study reported that the HGS cut-points for the likelihood of nutritional risk were 27.5 kg (65-74 years) and 21.0 kg (75-90 years) for men and 17.0 kg (65-74 years) and 14.6 kg (75-90 years) for women, and the HGS cut-points for the likelihood of malnutrition were 24.9 kg (65-74 years) and 20.8 kg (75-90 years) for men and 15.2 kg (65-74 years) and 13.5 kg (75-90 years) for women. Guerra et al.^[23] reported that the cut-off value for men older than 65 years was 30.2 kg when screening undernourished patients classified by PG-SGA. The risk of poor nutrition might be a predictor of lower physical performance. The HGS cut-off value for physical performance was 16.2 kg in elderly female cardiac inpatients^[38] and 22.7 kg in elderly post-cardiac surgery patients complicated by diabetes mellitus^[39].

There was a high prevalence of malnutrition in cirrhotic outpatients, especially when assessed by HGS, which was superior to SGA and the prognostic nutritional index (PNI)^[40]. As stated earlier, elderly inpatients also had a high prevalence of malnutrition or nutritional risk; according to the HGS cut-points, the prevalence of nutritional risk and malnutrition was, respectively, 51.1% and 27.5% for men and 51.2% and 41.4% for women. HGS identified a high proportion of inpatients at nutritional risk, indicating that it was a reliable first screening tool for nutritional risk in elderly inpatients.

Our analysis was based on a large sample size (n = 1, 343) and included male and female inpatients aged more than 65 years from a variety of wards. There are also some limitations to this study. First, there is no consensus on the assessment protocol for HGS^[18]. This limits the comparison between our study and others that used different methodologies. Second, there is no gold standard for evaluating elderly patientsa' nutritional status. Our study was based on the SGA and NRS 2002, which can assess malnutrition and screen for nutritional risk, respectively. Third, the lack of HGS data on elderly patients in China also limited comparisons of our study with others. Fourth, the HGS is a simple parameter, which provides valuable information in addition to the assessment of nutritional status. Finally, the study could not clarify the cause-effect association between HGS and malnutrition or nutritional risk, as this was a cross-sectional analysis. Reliable HGS cut-off values remain to be proposed and validated.

In conclusion, elderly inpatients were associated with a greater incidence of malnutrition or nutritional risk. Our research demonstrates that HGS can independently predict nutrition status, as assessed by SGA and screened by NRS 2002. HGS cut-off values for nutritional risk screening at hospital admission were 27.5 kg (65-74 years) and 21.0 kg (75-90 years) for men and 17.0 kg (65-74 years) and 14.6 kg (75-90 years) for women, and the HGS cut-off values for malnutrition assessment were 24.9 kg (65-74 years) and 20.8 kg (75-90 years) for men and 15.2 kg (65-74 years) and 13.5 kg (75-90 years) for women. These HGS cut-off values should be tested in other samples and validated in future studies.

No conflict of interest to declare.

The authors thank all the workers for their contribution and the participants for their cooperation.