Skip to main content Accessibility help
×
Home
Hostname: page-component-55597f9d44-mm7gn Total loading time: 0.839 Render date: 2022-08-14T18:44:02.589Z Has data issue: true Feature Flags: { "shouldUseShareProductTool": true, "shouldUseHypothesis": true, "isUnsiloEnabled": true, "useRatesEcommerce": false, "useNewApi": true } hasContentIssue true

Upper arm length and knee height are associated with diabetes in the middle-aged and elderly: evidence from the China Health and Retirement Longitudinal Study

Published online by Cambridge University Press:  18 May 2022

Bingjie He
Affiliation:
Department of Epidemiology and Biostatistics, School of Public Health, Peking University, 38 Xueyuan Road, Haidian District, Beijing100191, People’s Republic of China
Zhengyang Li
Affiliation:
Department of Endocrinology, Shandong Provincial Hospital Affiliated to Shandong First Medical University, Huaiyin District, Jinan, People’s Republic of China
Lu Xu
Affiliation:
Department of Epidemiology and Biostatistics, School of Public Health, Peking University, 38 Xueyuan Road, Haidian District, Beijing100191, People’s Republic of China
Lili Liu
Affiliation:
Department of Epidemiology and Biostatistics, School of Public Health, Peking University, 38 Xueyuan Road, Haidian District, Beijing100191, People’s Republic of China
Shengfeng Wang*
Affiliation:
Department of Epidemiology and Biostatistics, School of Public Health, Peking University, 38 Xueyuan Road, Haidian District, Beijing100191, People’s Republic of China
Siyan Zhan*
Affiliation:
Department of Epidemiology and Biostatistics, School of Public Health, Peking University, 38 Xueyuan Road, Haidian District, Beijing100191, People’s Republic of China Research Center of Clinical Epidemiology, Peking University Third Hospital, Haidian District, Beijing, People’s Republic of China Center for Intelligent Public Health, Institute for Artificial Intelligence, Peking University, Beijing, People’s Republic of China
Yongfeng Song*
Affiliation:
Department of Endocrinology, Shandong Provincial Hospital Affiliated to Shandong First Medical University, Huaiyin District, Jinan, People’s Republic of China Shandong Institute of Endocrine & Metabolic Diseases, Shandong First Medical University, Jinan, People’s Republic of China
Rights & Permissions[Opens in a new window]

Abstract

Objective:

To determine if limb lengths, as markers of early life environment, are associated with the risk of diabetes in China.

Design:

We performed a cohort analysis using data from the China Health and Retirement Longitudinal Study (CHARLS), and multivariable-adjusted Cox proportional hazard regression models were used to examine the associations between baseline limb lengths and subsequent risk of diabetes.

Setting:

The CHARLS, 2011–2018.

Participants:

The study confined the eligible subject to 10 711 adults aged over 45 years from the CHARLS.

Results:

During a mean follow-up period of 6·13 years, 1358 cases of incident diabetes were detected. When controlling for potential covariates, upper arm length was inversely related to diabetes (hazard ratio (HR) 0·95, 95 % CI (0·91, 0·99), P = 0·028), and for every 1-cm difference in knee height, the risk of diabetes decreased by about 4 % (HR 0·96, 95 % CI (0·93, 0·99), P = 0·023). The association between upper arm length and diabetes was only significant among females while the association between knee height and diabetes was only significant among males. In analyses stratified by BMI, significant associations between upper arm length/knee height and diabetes only existed among those who were underweight (HR 0·91, 95 % CI (0·83, 1·00), P = 0·049, HR 0·92, 95 % CI (0·86, 0·99), P = 0·031).

Conclusions:

Inverse associations were observed between upper arm length, knee height and the risk for diabetes development in a large Asian population, suggesting early life environment, especially infant nutritional status, may play an important role in the determination of future diabetes risk.

Type
Research Paper
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright
© The Author(s), 2022. Published by Cambridge University Press on behalf of The Nutrition Society

Diabetes mellitus is one of the largest epidemics around the world and its prevalence is increasing in both developed and developing countries, which poses a significant burden to individuals and society, especially type 2 diabetes mellitus in the middle-aged and elderly(Reference Zimmet, Alberti and Magliano1Reference Jia, Weng and Zhu3). Type 2 diabetes mellitus is a complex disease and the underlying causes have not been fully elucidated. Several studies have suggested that early life environment plays a role in susceptibility to chronic disease such as CVD, hypertension, dementia and type 2 diabetes mellitus in later life(Reference Huang, Carlson and Fitzpatrick4,Reference Paajanen, Oksala and Kuukasjärvi5) . Limb lengths are considered as indicators of children’s socio-economic conditions, environmental status and psychological stress, which means that differences in these anthropometric measures may reflect nutritional or other deficits throughout childhood(Reference Gunnell, Smith and Frankel6,Reference Gunnell, Okasha and Smith7) . Recently, several studies have discovered a significant association between tall stature and the lower risk of glucose intolerance and type 2 diabetes(Reference Brown, Byrne and Clark8Reference Wittenbecher, Kuxhaus and Boeing11).

A number of epidemiological studies have investigated the association between anthropometric parameters and risk of type 2 diabetes mellitus, while the results are inconsistent(Reference Lawlor, Ebrahim and Davey Smith9Reference Johnston, Harris and Retnakaran14). In cross-sectional analyses, some studies suggested that leg length instead of height had the strong association with diabetes(Reference Lawlor, Ebrahim and Davey Smith9,Reference Liu, Tan and Jeynes10) , and one research indicated that femur length might be the key component in this relevance(Reference Liu, Tan and Jeynes10). Besides, another cross-sectional analysis found that arm lengths (including total arm length and upper arm length) were inversely related to diabetes, and forearm length, height, leg length and lower leg length were not associated with diabetes in Japanese Americans(Reference Smits, Boyko and Utzschneider12). In prospective researches, a couple of studies also confirmed that leg length was inversely associated with the risk of developing diabetes(Reference Wittenbecher, Kuxhaus and Boeing11,Reference Weitzman, Wang and Pankow13,Reference Johnston, Harris and Retnakaran14) , but other part of limb lengths including knee height or arm length, which are also thought to be particularly sensitive to nutritional status during childhood(Reference McEnry and Palloni15), has not been investigated. In addition, recent studies have found that the inverse association between limb length and diabetes was only significant in the individuals with normal fasting blood glucose or people who are lean(Reference Rhee, Cho and Kwon16), and the inverse association has been observed significantly in whites, but not in African Americans(Reference Weitzman, Wang and Pankow13).

This association remains controversial and these studies were heterogeneous with respect to insufficient sample size (462 subjects(Reference Johnston, Harris and Retnakaran14) or 658 participants(Reference Smits, Boyko and Utzschneider12)), study design (cross-sectional analysis), differences in outcome measurement, differences in anthropometric measures across different races and other characteristics that may contribute to inconsistencies in the literature. Previous studies mainly focused on Caucasians, Japanese Americans and African Americans, and the difference in anthropometric measures across different races suggested the need for new research in other population(Reference Nakanishi and Nethery17,Reference Wulan, Westerterp and Plasqui18) . No prospective studies have been conducted on this topic among Asian population. Therefore, using the data from the China Health and Retirement Longitudinal Study (CHARLS) which offers a large national sample of both men and women with a relatively long period of follow-up, the present study was designed to investigate whether upper arm length and knee height are related to the risk of developing diabetes.

Materials and methods

Database and study population

The data used in this study came from the survey of the CHARLS. The CHARLS, a nationally representative longitudinal survey, was conducted biannually among the Chinese population aged 45 years and over, including information regarding demographics, health status and functioning, blood sample data and physical examinations(Reference Zhao, Hu and Smith19).

To date, a total of four surveys have been conducted in 2011 (visit 1), 2013 (visit 2), 2015 (visit 3) and 2018 (visit 4). Details of the design and conduct of CHARLS were described elsewhere(Reference Zhao, Hu and Smith19). The baseline survey recruited 17 424 individuals, and 13 978 (78·9 %) of them provided anthropometric and physical performance measures. For the present study, only participants who had physical examination data, did not diagnose with diabetes at baseline and had at least one follow-up were included. Individuals with missing data on upper arm length and knee height (322), with missing data on covariates (368) and with unknown diabetic status at baseline or during follow-up (661) were excluded (please note that some participants were excluded for meeting more than one criteria), accounting for 9·60 % of the population that should be included and leaving 10 711 participants included in this study.

Measurements

Trained researchers interviewed participants face to face in their homes using computer-assisted personal interviewing technology to collect variables including socio-demographic information, health behaviour, medical history and medication usage. Physical parameters were measured by the trained investigators with standardised equipment at every follow up, and venous blood sample collection was done in 2011 (visit 1) and 2015 (visit 3)(Reference Zhao, Hu and Smith19,Reference Chen, Crimmins and Hu20) . These blood samples were transported from all study sites to the study headquarters in Beijing where they were assayed for biochemical parameters including glucose, glycated HbA1c and lipid profiles by trained research staff(Reference Chen, Crimmins and Hu20).

Exposure measurement

Anthropometric measures followed the survey protocol(Reference Zhao, Strauss and Yang21). Upper arm length and knee height were measured by trained investigators using DongfangTM XTCL-I Martin rule. Upper arm length was defined as the distance between the olecranon process and the superior border of the acromion process, and knee height was defined as the distance from the sole of the foot to the anterior surface of the thigh, with the ankle and knee each flexed to a 90° angle. The quartiles of the limb length were divided according to gender-specific cut-off values(Reference Wittenbecher, Kuxhaus and Boeing11,Reference Rhee, Cho and Kwon16) .

Evaluation of diabetes

At baseline and any follow-up examination, subjects with one or more of the followings were determined as having diabetes: (1) fasting plasma glucose ≥ 126 mg/dl; (2) non-fasting plasma glucose ≥ 200 mg/dl; (3) HbA1c ≥ 6·5 %; (4) self-reported prior diagnosis of diabetes by a doctor; and (5) self-reported use of antidiabetic medications(22,Reference Park, Xun and Li23) .

Covariate assessment

Based on biologic plausibility or previous studies(Reference Liu, Tan and Jeynes10,Reference Smits, Boyko and Utzschneider12,Reference Weitzman, Wang and Pankow13,22) , the following candidate covariates that were collected by trained researchers through questionnaires and physical measurements were examined as potential confounding factors: age (years), sex (male, female), marriage (single, married/divorced), smoking (never, current/former), drinking (never, current/former), weight (kg), BMI (kg/m2), waist circumference (cm), hypertension and dyslipidaemia. BMI was calculated by dividing weight (kg) by the square of height (m), and BMI could be classified into four levels (underweight: <18·5 kg/m2, normal: 18·5–24 kg/m2, overweight: 24–28 kg/m2, obesity: ≥28 kg/m2)(Reference Chen and Lu24). Hypertension was defined by self-reported physician diagnosis of hypertension or blood pressure measurements ≥ 140/90 mmHg or use of antihypertensive agents(25). Dyslipidaemia was defined by self-reported physician diagnosis of dyslipidaemia, or total cholesterol ≥ 240 mg/dl and LDL ≥ 160 mg/dl, or TAG ≥ 200 mg/dl, or HDL cholesterol < 40 mg/dl, or use of anti-dyslipidaemia agents(26).

Statistical analysis

Baseline characteristics of the participants in our study were presented as means with standard deviations for continuous variables, with differences between groups by diabetes status (with/without) calculated using t test, or percentages for categorical variables, with differences between groups by diabetes status calculated using χ 2 test. The Shapiro–Wilk test was used to assess for data normality. Diabetes incidence rates and 95 % CI were calculated as events per 1000 person-years using Poisson regression models. The relationship between upper arm length/knee height and onset of diabetes was conducted by survival analysis. The survival time was defined as the period from the date of first interview to the date of onset of diabetes, loss to follow-up or last interview in 2018, whichever came first. The Cox proportional hazards regression was used to obtain adjusted hazard ratios (HR) and 95 % CI. The proportional hazards assumption was tested by Schoenfeld residuals. Four models were considered to adjust potential covariates defined a priori. Model 1 was adjusted for age and sex. Model 2 was additionally adjusted for weight. In Model 3, adjustment further included marriage, smoking and drinking. Model 4 was additionally adjusted for hypertension and dyslipidaemia. To make our results nationally representative, a sample weight, which was calculated using an inverse probability method, was used under the correction for household and individual non-response as well as non-participation in the anthropometric measure. All the multivariable analyses were weighted in our study.

Subgroup analyses by sex (male, female), age (<60, ≥60), BMI (underweight, normal, overweight, obesity), hypertension status (without hypertension, with hypertension) or dyslipidaemia status (without dyslipidaemia, with dyslipidaemia) were performed using model 4. The interaction effect between limb length and a covariate was tested by including a two-way interaction term in the final model. We also performed sensitivity analysis to test the robustness of our results: we refit the Cox model to adjust the hazard of diabetes for the competing risk of death using the Fine–Gray model which estimated subdistribution hazard.

Stata 15.0 was used for all statistical analyses, and two-sided P value less than 0·05 was considered as statistically significant.

Results

Baseline characteristics

A total of 10 711 participants were included in this study, with 5114 (47·75 %) males and 5597 (52·25 %) females. The mean baseline age of all participants was 59·20 (sd 9·41) years. The participants were followed up for an average of 6·13 (sd 1·59) years, which resulted in 1358 cases of incident diabetes. Baseline characteristics by diabetes status are shown in Table 1. Compared with those without diabetes, participants with diabetes were older, female, more likely to be overweight and obese, and had shorter upper arm length at baseline. They were also less likely to be regular smokers and regular alcohol drinkers and more likely to had hypertension and dyslipidaemia at baseline. We excluded 1137 subjects because of unknown diabetic status or missing exposure and covariate information, and they were older, more likely to be female, divorced and underweight, less likely to be regular smokers and regular alcohol drinkers, and more likely to had hypertension and less likely to had dyslipidaemia at baseline (see online supplementary material, Supplemental Table S1).

Table 1 Baseline characteristics of study subjects subdivided by diabetes outcome

* 182 participants had missing values for waist circumference.

Upper arm length/knee height and risks of diabetes

During the study period (65 648·11 person-years, mean follow-up time 6·13 years), a total of 1358 cases of incident diabetes were detected and the incidence rate was 20·69 per 1000 person-years (95 % CI 19·61, 21·80). According to the different group of limb length, the incidence rates of diabetes during the follow-up were 22·25 (95 % CI 20·00, 24·68) per 1000 person-years in the first quartile (short limb length) of upper arm length, 21·30 (95 % CI 19·11, 23·66) in the second quartile, 19·62 (95 % CI 17·58, 21·83)in the third quartile and 19·72 (95 % CI 17·70, 21·92) in the fourth quartile (long limb length), with respective incidence rates of 21·07 (95 % CI 18·88, 23,43), 20·04 (95 % CI 17·97, 22·29), 19·13 (95 % CI 17·05, 21·40) and 22·35 (95 % CI 20·21, 24·65) per 1000 person-years in the first, second, third and fourth quartile of knee height (Table 2). With the increase of upper arm length, the incidence of diabetes had a downward trend (P = 0·062).

Table 2 Incidence of diabetes during the follow-up

Quartiles of upper arm length: female: Q1: < 31·5 cm, Q2: 31·5–32·5 cm, Q3: 32·5–33·9 cm, Q4: ≥ 33·9 cm; male: Q1: < 34·0 cm, Q2: 34·0–35·3 cm, Q3: 35·3–36·5 cm, Q4: ≥ 36·5 cm.

Quartiles of knee height: female: Q1: < 44·7 cm, Q2: 44·7–46·4 cm, Q3: 46·4–48·0 cm, Q4: ≥ 48·0 cm; male: Q1: < 48·0 cm, Q2: 48·0–49·9 cm, Q3: 49·9–51·5 cm, Q4: ≥ 51·5 cm.

The average upper arm length of participants with diabetes was 33·61 cm (sd: 2·54), significantly shorter than those without diabetes (33·87 cm, sd 2·50; P < 0·001). When controlling for explanatory and confounding variables using Cox proportional hazards regression models, model 1 adjusted for age and sex, mode 2 adjusted for age, sex and weight, model 3 adjusted for age, sex, weight, marriage, smoking and drinking and model 4 adjusted for age, sex, weight, marriage, smoking, drinking, hypertension and dyslipidaemia all showed that upper arm length was inversely related to diabetes (model 4: adjusted HR per 1-cm difference 0·95, 95 % CI (0·91, 0·99); P = 0·028) (Table 3).

Table 3 HR (95 % CI) of incident diabetes for upper arm length and knee height, adjusted for possible explanatory and confounding factors

HR, hazard ratio.

Model 1: adjusted for age and sex; model 2: adjusted for age, sex and weight; model 3: adjusted for age, sex, weight, marriage, smoking and drinking; model 4: adjusted for age, sex, weight, marriage, smoking, drinking, hypertension and dyslipidaemia.

The average knee height of participants with diabetes was 47·79 cm (sd 3·34), shorter than those without diabetes (47·93cm, sd 3·36), but the difference was marginal significant (P = 0·147). However, adjusted for all potential covariates, for every 1-cm difference in knee height, the risk of diabetes decreased by about 4 % (model 4: adjusted HR 0·96, 95 % CI (0·93, 0·99); P = 0·023) (Table 3).

Subgroup analyses

The results of subgroup analyses are presented in Table 4 and Fig. 1. The association between upper arm length and diabetes was only significant among females (adjusted HR 0·94, 95 % CI (0·90, 0·98)), while the association between knee height and diabetes was only significant among males (adjusted HR 0·94, 95 % CI (0·90, 0·98)), and the interaction between knee height and gender was significant (P = 0·042). In different age group, the associations between upper arm length/knee height and diabetes were only significant in the younger group. In analyses stratified by BMI, significant associations only existed among those who were underweight (adjusted HR 0·91, 95 % CI (0·83, 1·00); P = 0·049/adjusted HR 0·92, 95 % CI (0·86, 0·99); P = 0·031). Subgroup analyses by hypertension status and dyslipidaemia status showed that significant associations only existed among those who had no hypertension and no dyslipidaemia, and the interaction term between knee height and dyslipidaemia in the fully adjusted model was significant (P = 0·032) (Table 4).

Table 4 Association of upper arm length/knee height with development of diabetes according to subgroups of various variables

HR, hazard ratio.

* P < 0·05.

** P < 0·01.

***P < 0·001.

Adjusted for age, sex, weight, marriage, smoking, drinking, hypertension and dyslipidaemia.

Fig. 1 Hazard ratios (HR) and 95 % CI for the risk of diabetes, by sex, age, BMI, hypertension status or diabetes status

Sensitivity analyses

The competing risk analysis did not materially alter the main results described above (upper arm length: adjusted HR per 1-cm difference 0·95, 95 % CI (0·91, 0·99), P = 0·029; knee height: adjusted HR per 1-cm difference 0·96, 95 % CI (0·93, 0·99), P = 0·017) (see online supplementary material, Supplemental Table S2).

Discussion

In this large prospective cohort study of adults in China, inverse associations were observed between upper arm length, knee height and the risk for diabetes development, and these results were significant after rigorous adjustment. To the best of our knowledge, this is the first study to evaluate the associations of different limb lengths with the risk of diabetes development in Asia population.

Leg length was considered to be a marker for poor environmental conditions in early life(Reference Johnston, Harris and Retnakaran14,Reference Wadsworth, Hardy and Paul27) . Shorter leg length was found to be independently related to the risk of diabetes in white Americans(Reference Weitzman, Wang and Pankow13), British Women(Reference Lawlor, Ebrahim and Davey Smith9), Canada population(Reference Johnston, Harris and Retnakaran14) and Germany population(Reference Wittenbecher, Kuxhaus and Boeing11), but the association was not found in African Americans(Reference Weitzman, Wang and Pankow13), Japanese Americans(Reference Smits, Boyko and Utzschneider12) or Chinese(Reference Schooling, Jiang and Lam28,Reference Conway, Shu and Zhang29) . Another study using data from the National Health and Nutrition Examination Survey in the USA indicated that femur length might be the key height component in diabetes risk association(Reference Liu, Tan and Jeynes10). Our study showed knee height as another component of leg length was significantly associated with incidence of diabetes in Chinese, but the result was faint. First, the inconsistent results of these studies on the relationship between leg length and diabetes may be due to differences in ethnicity reflected in different susceptibility to diabetes or different body composition(Reference Golden, Yajnik and Phatak30,Reference Kagawa, Binns and Hills31) . There are racial/ethnic differences in incidence of type 2 diabetes mellitus, and these differences might be associated with many other risk factors including being obese and diet for type 2 diabetes mellitus(Reference Golden, Yajnik and Phatak30). Also, many studies have shown that compared with whites, Asians have different body composition(Reference Wulan, Westerterp and Plasqui18,Reference Deurenberg, Deurenberg-Yap and Guricci32) , and there are differences in BMI and relative leg length among Asians(Reference Gurrici, Hartriyanti and Hautvast33). These ethnic differences might contribute to the inconsistency of the association between leg length and diabetes(Reference Smits, Boyko and Utzschneider12,Reference Weitzman, Wang and Pankow13) . Second, the small sample size in the African Americans study and the Japanese Americans study could be an additional explanation for the absence of association, and homogeneity of the population and inappropriate adjustment for adult BMI in the two Chinese studies, which only recruited residents in Shanghai or Guangzhou, might explain the different results from our study. The two Chinese studies recruited permanent Guangzhou residents aged 50 years and older(Reference Schooling, Jiang and Lam28) or population living in urban communities of Shanghai(Reference Conway, Shu and Zhang29) respectively, which could suggest that environmental factors in early life might be similar in two study population, and therefore, the difference in leg length or knee height would be much smaller than in a population with large variation in socio-economic status, and researchers could not found the association between leg length and diabetes. Also, the population differences in our study may not be large enough to find a greater effect. In addition, we only have data on knee height and no data on leg length, which prevents us from fully explaining the association between leg length and its components with diabetes in the Chinese population. Besides, greater knee height was associated with lower risks of dementia and Alzheimer disease in US women(Reference Huang, Carlson and Fitzpatrick4), and this association was related to nutrition in early life, as is the case with type 2 diabetes as hypothesised by the thrifty phenotype hypothesis(Reference Kim, Stewart and Shin34), which could also provide evidence for knee height as a good indicator for susceptibility to chronic disease during adulthood.

We have shown that upper arm length was inversely associated with diabetes and the result was also faint after adjustment for confounding factors. Related research for arm length is relatively limited, but arm length has also been shown to be related to early life environment(Reference Huang, Carlson and Fitzpatrick4,Reference Kim, Stewart and Shin34) . In a study conducted in the USA, a cross-sectional analysis of more than 600 Japanese Americans found that total arm length and upper arm length were negatively correlated with diabetes(Reference Smits, Boyko and Utzschneider12), which is consistent with our research. Because the average upper arm length and knee height variation were faint, whether it has clinical significance remains to be studied.

The underlying mechanism for the association between limb lengths and diabetes has not been fully understood. Adult limb lengths can be affected by genetic and environmental factors, but previous studies have shown that limb lengths are associated with early childhood nutrition, especially infant nutrition(Reference Wadsworth, Hardy and Paul27,Reference Kim, Stewart and Shin34,Reference Mitchell35) . Among the relatively undernourished populations, such as the Chinese population in our study, limb lengths may be more determined by nutritional conditions in the early life(Reference Mueller and Pereira36Reference Wang, Zou and Yang38). The possible explanation is the thrifty phenotype hypothesis, which believes that poor nutrition in fetal and early infant life are detrimental to the development and function of the Beta cells of the islets of Langerhans and such defects predispose to the later development of diabetes(Reference Hales and Barker39Reference Hales and Barker41). A recent experiment in rats also confirmed this view. The study found that postpartum nutritional deficiencies and subsequent catch-up growth reduced insulin sensitivity(Reference Alexeev, Lönnerdal and Griffin42). Our epidemiological study found that upper arm length and knee height, as indicators of environmental conditions in early childhood, were negatively correlated with diabetes, which could provide indirect evidence that poor nutritional status in early life may lead to long-term metabolic disorders and is related to the development of diabetes. As for different parts of the limb length, human beings follow a cephalo-caudal gradient of growth, which is characterised by the faster growth rate of the lower limbs, especially the knee height, than other body segments during the period from birth to puberty(Reference Bogin and Varela-Silva43). Some studies have found that different anatomical regions of the human body (humerus, femur, tibia, hand or foot) have different sensitivity to environmental pressure during growth(Reference Holliday and Ruff44), and the ‘distal blood flow’ hypothesis assumes that because it is the last area to receive oxygenated blood and blood nutrients, the tibia is more susceptible to adverse factors such as hypoxia than other limbs(Reference Lampl, Kuzawa and Jeanty45,Reference Pomeroy, Stock and Stanojevic46) . Therefore, knee height might be more sensitive to nutrient shortage, infection and socio-economic environment in early life. This assumption still needs further verification.

In terms of subgroup analyses, the associations did have differences between genders. The association between upper arm length and diabetes was only significant among females, while the association between knee height and diabetes was only significant among males. The gender difference observed in the effect of upper arm length and knee height on diabetes risk possibly strongly associated with menarche and the menstrual cycle. In girls, increases in body fat can affect the initiation of menarche, ensuing early puberty, which influence final height in females(Reference Janghorbani, Momeni and Dehghani47). When the association was separately analysed in groups divided by BMI, only those who were underweight showed significant inverse association between limb lengths and diabetes, which might indicate that the effects of obesity overwhelmed the effects of short limb lengths when individual was overweight or obese(Reference Rhee, Cho and Kwon16,Reference Costacou and Mayer-Davis48,Reference Chang, Song and Wang49) . Several studies have revealed that early exposure to poor nutrition may also increase the risk for obesity in adulthood(Reference Chang, Song and Wang49Reference Bhargava, Sachdev and Fall52) and obesity is an important risk factor for type 2 diabetes mellitus(22), which indicates that obesity and limb lengths all reflect the nutritional status of early life to some extent and the effects of limb lengths on diabetes may be greatly affected in overweight and obese people. The competing risk analysis considering the effect of death also did not materially alter the main results, which reflected the robustness of the results of this study.

Strengths and limitations

To our knowledge, this is the first study to investigate the association of different limb lengths with diabetes in Asian population, using nationally representative longitudinal data with relatively large sample size and an average follow-up time of 6·13 years. Nevertheless, it has some limitations. First, we only have upper arm length and knee height data and have no other limb length data, such as total arm length, arm span, leg length and femur length. Therefore, we cannot fully elucidate the association of different components of limb lengths with diabetes, but we still added the evidence of the association of limb lengths with diabetes considering the correlation between limb lengths. A second limitation is that we cannot rule out the fact that some of the subjects in our study might have type 1 diabetes. However, considering that type 2 diabetes mellitus accounts for the vast majority of diabetes patients and that the onset age of diabetes in our study population was over 45 years old, it is likely that all patients in this analysis had type 2 diabetes(Reference Sun, Zhang and Duan37,Reference Peng, Zhao and Yang53) .

Conclusion

In conclusion, inverse associations were observed between upper arm length, knee height and the risk for diabetes development in a large Asian population. Since upper arm length and knee height may be more determined by early life nutritional conditions, our results are consistent with the hypothesis that poor nutrition in early life may lead to long-term metabolic disorders and be associated with the development of diabetes. The results of this study suggest that attention should be paid to prenatal, postpartum and early childhood nutrition to prevent adulthood diabetes, and limb length might be used as a predictor for diabetes prediction. More studies with complete limb lengths data are still needed to confirm our findings and clarify the potential underlying mechanisms.

Acknowledgements

Acknowledgements: None. Financial support: This work has been supported by grants from National Natural Science Foundation of China (81922016, 81870607), the Natural Science Foundation of Shandong Province (ZR2020ZD14, ZR2019JQ25), the National Key Research and Development Program of China (2018YFC1311704) and 2021 Chronic Disease Control and Health Education Research Project (BJMB0012021025014). The funders/sponsors had no role in the design and conduct of the study; collection management, analysis or interpretation of the data; preparation, review or approval of the manuscript; and decision to submit the manuscript for publication. Authorship: Conception: S.Y.Z., S.F.W., Y.F.S. and B.J.H.; design: S.F.W. and B.J.H.; administrative support: S.F.W., S.Y.Z. and Y.F.S.; provision of study materials: B.J.H.; collection and assembly of data: B.J.H., L.X. and L.L.L.; data analysis and interpretation: B.J.H., L.X., L.L.L. and Z.Y.L.; manuscript writing: B.J.H. and Z.Y.L.; final approval of manuscript: all authors. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving research study participants were approved by the Ethical Review Committee of Peking University (IRB00001052-11015), and all participants signed informed consent.

Conflict of interest:

There are no conflicts of interest.

Supplementary material

For supplementary material accompanying this paper visit https://doi.org/10.1017/S1368980022001215

Footnotes

Bingjie He and Zhengyang Li contributed equally to this manuscript

References

Zimmet, P, Alberti, KG, Magliano, DJ et al. (2016) Diabetes mellitus statistics on prevalence and mortality: facts and fallacies. Nat Rev Endocrinol 12, 616622.CrossRefGoogle ScholarPubMed
American Diabetes Association (2019) 1. Improving care and promoting health in populations: standards of medical care in diabetes-2019. Diabetes Care 42, S7S12.CrossRefGoogle Scholar
Jia, W, Weng, J, Zhu, D et al. (2019) Standards of medical care for type 2 diabetes in China 2019. Diabetes Metab Res Rev 35, e3158.CrossRefGoogle ScholarPubMed
Huang, TL, Carlson, MC, Fitzpatrick, AL et al. (2008) Knee height and arm span: a reflection of early life environment and risk of dementia. Neurology 70, 18181826.CrossRefGoogle ScholarPubMed
Paajanen, TA, Oksala, NK, Kuukasjärvi, P et al. (2010) Short stature is associated with coronary heart disease: a systematic review of the literature and a meta-analysis. Eur Heart J 31, 18021809.CrossRefGoogle ScholarPubMed
Gunnell, DJ, Smith, GD, Frankel, SJ et al. (1998) Socio-economic and dietary influences on leg length and trunk length in childhood: a reanalysis of the Carnegie (Boyd Orr) survey of diet and health in prewar Britain (1937–39). Paediatr Perinat Epidemiol 12, Suppl. 1, 96113.CrossRefGoogle Scholar
Gunnell, D, Okasha, M, Smith, GD et al. (2001) Height, leg length, and cancer risk: a systematic review. Epidemiol Rev 23, 313342.CrossRefGoogle ScholarPubMed
Brown, DC, Byrne, CD, Clark, PM et al. (1991) Height and glucose tolerance in adult subjects. Diabetologia 34, 531533.CrossRefGoogle ScholarPubMed
Lawlor, DA, Ebrahim, S & Davey Smith, G (2002) The association between components of adult height and Type II diabetes and insulin resistance: British women’s heart and health study. Diabetologia 45, 10971106.Google ScholarPubMed
Liu, J, Tan, H & Jeynes, B (2009) Is femur length the key height component in risk prediction of type 2 diabetes among adults? Diabetes Care 32, 739740.CrossRefGoogle ScholarPubMed
Wittenbecher, C, Kuxhaus, O, Boeing, H et al. (2019) Associations of short stature and components of height with incidence of type 2 diabetes: mediating effects of cardiometabolic risk factors. Diabetologia 62, 22112221.CrossRefGoogle ScholarPubMed
Smits, MM, Boyko, EJ, Utzschneider, KM et al. (2012) Arm length is associated with type 2 diabetes mellitus in Japanese-Americans. Diabetologia 55, 16791684.CrossRefGoogle ScholarPubMed
Weitzman, S, Wang, CH, Pankow, JS et al. (2010) Are measures of height and leg length related to incident diabetes mellitus? The ARIC (atherosclerosis risk in communities) study. Acta Diabetol 47, 237242.CrossRefGoogle ScholarPubMed
Johnston, LW, Harris, SB, Retnakaran, R et al. (2013) Short leg length, a marker of early childhood deprivation, is associated with metabolic disorders underlying type 2 diabetes: the PROMISE cohort study. Diabetes Care 36, 35993606.CrossRefGoogle ScholarPubMed
McEnry, M & Palloni, A (2010) Early life exposures and the occurrence and timing of heart disease among the older adult Puerto Rican population. Demography 47, 2343.CrossRefGoogle ScholarPubMed
Rhee, EJ, Cho, JH, Kwon, H et al. (2019) Relation between baseline height and new diabetes development: a nationwide population-based study. Diabetes Metab J 43, 794803.CrossRefGoogle ScholarPubMed
Nakanishi, Y & Nethery, V (1999) Anthropometric comparison between Japanese and Caucasian American male university students. Appl Hum Sci 18, 911.CrossRefGoogle ScholarPubMed
Wulan, SN, Westerterp, KR & Plasqui, G (2010) Ethnic differences in body composition and the associated metabolic profile: a comparative study between Asians and Caucasians. Maturitas 65, 315319.CrossRefGoogle ScholarPubMed
Zhao, Y, Hu, Y, Smith, JP et al. (2014) Cohort profile: the China health and retirement longitudinal study (CHARLS). Int J Epidemiol 43, 6168.CrossRefGoogle Scholar
Chen, X, Crimmins, E, Hu, PP et al. (2019) Venous blood-based biomarkers in the China health and retirement longitudinal study: rationale, design, and results from the 2015 wave. Am J Epidemiol 188, 18711877.CrossRefGoogle ScholarPubMed
Zhao, Y, Strauss, J, Yang, G et al. (2013) China Health and Retirement Longitudinal Study: 2011–2012 National Baseline User’s Guide. China: National School of Development, Peking University.Google Scholar
Community Development Society (2018) Chinese guidelines for prevention and treatment of type 2 Diabetes (2017 revision). Chin J Diabetes Mellit 10, 467.Google Scholar
Park, JS, Xun, P, Li, J et al. (2016) Longitudinal association between toenail zinc levels and the incidence of diabetes among American young adults: the CARDIA trace element study. Sci Rep 6, 23155.CrossRefGoogle ScholarPubMed
Chen, C & Lu, FC (2004) The guidelines for prevention and control of overweight and obesity in Chinese adults. Biomed Environ Sci 17, 136.Google ScholarPubMed
Joint Committee for Guideline Revision (2019) 2018 Chinese guidelines for prevention and treatment of hypertension-a report of the revision committee of Chinese guidelines for prevention and treatment of hypertension. J Geriatr Cardiol 16, 182241.Google Scholar
Joint Committee for Developing Chinese Guidelines on Prevention and Treatment of Dyslipidemia in Adults (2016) Chinese guidelines on prevention and treatment of dyslipidemia in adults (2016 revision). Chin Circ J 31, 937950.Google Scholar
Wadsworth, ME, Hardy, RJ, Paul, AA et al. (2002) Leg and trunk length at 43 years in relation to childhood health, diet and family circumstances; evidence from the 1946 national birth cohort. Int J Epidemiol 31, 383390.CrossRefGoogle ScholarPubMed
Schooling, CM, Jiang, C, Lam, TH et al. (2007) Height, its components, and cardiovascular risk among older Chinese: a cross-sectional analysis of the Guangzhou Biobank cohort study. Am J Public Health 97, 18341841.CrossRefGoogle ScholarPubMed
Conway, BN, Shu, XO, Zhang, X et al. (2012) Age at menarche, the leg length to sitting height ratio, and risk of diabetes in middle-aged and elderly Chinese men and women. PLoS ONE 7, e30625.CrossRefGoogle ScholarPubMed
Golden, SH, Yajnik, C, Phatak, S et al. (2019) Racial/ethnic differences in the burden of type 2 diabetes over the life course: a focus on the USA and India. Diabetologia 62, 17511760.CrossRefGoogle ScholarPubMed
Kagawa, M, Binns, CB & Hills, AP (2007) Body composition and anthropometry in Japanese and Australian Caucasian males and Japanese females. Asia Pac J Clin Nutr 16, Suppl. 1, 3136.Google ScholarPubMed
Deurenberg, P, Deurenberg-Yap, M & Guricci, S (2002) Asians are different from Caucasians and from each other in their body mass index/body fat per cent relationship. Obes Rev 3, 141146.CrossRefGoogle ScholarPubMed
Gurrici, S, Hartriyanti, Y, Hautvast, JG et al. (1999) Differences in the relationship between body fat and body mass index between two different Indonesian ethnic groups: the effect of body build. Eur J Clin Nutr 53, 468472.CrossRefGoogle ScholarPubMed
Kim, JM, Stewart, R, Shin, IS et al. (2003) Limb length and dementia in an older Korean population. J Neurol Neurosurg Psychiatry 74, 427432.CrossRefGoogle Scholar
Mitchell, HS (1962) Nutrition in relation to stature. J Am Diet Assoc 40, 521524.CrossRefGoogle Scholar
Mueller, NT & Pereira, MA (2015) Leg length and type 2 diabetes: what’s the link? Curr Opin Clin Nutr Metab Care 18, 452456.CrossRefGoogle ScholarPubMed
Sun, Y, Zhang, L, Duan, W et al. (2018) Association between famine exposure in early life and type 2 diabetes mellitus and hyperglycemia in adulthood: results from the China health and retirement longitudinal study (CHARLS). J Diabetes 10, 724733.CrossRefGoogle Scholar
Wang, Z, Zou, Z, Yang, Z et al. (2018) The association between fetal-stage exposure to the China famine and risk of diabetes mellitus in adulthood: results from the China health and retirement longitudinal study. BMC Public Health 18, 1205.CrossRefGoogle Scholar
Hales, CN & Barker, DJ (2001) The thrifty phenotype hypothesis. Br Med Bull 60, 520.CrossRefGoogle ScholarPubMed
Vaag, AA, Grunnet, LG, Arora, GP et al. (2012) The thrifty phenotype hypothesis revisited. Diabetologia 55, 20852088.CrossRefGoogle ScholarPubMed
Hales, CN & Barker, DJ (2013) Type 2 (non-insulin-dependent) diabetes mellitus: the thrifty phenotype hypothesis. Int J Epidemiol 42, 12151222.CrossRefGoogle ScholarPubMed
Alexeev, EE, Lönnerdal, B & Griffin, IJ (2015) Effects of postnatal growth restriction and subsequent catch-up growth on neurodevelopment and glucose homeostasis in rats. BMC Physiol 15, 3.CrossRefGoogle ScholarPubMed
Bogin, B & Varela-Silva, MI (2010) Leg length, body proportion, and health: a review with a note on beauty. Int J Environ Res Public Health 7, 10471075.CrossRefGoogle ScholarPubMed
Holliday, TW & Ruff, CB (2001) Relative variation in human proximal and distal limb segment lengths. Am J Phys Anthropol 116, 2633.CrossRefGoogle Scholar
Lampl, M, Kuzawa, CW & Jeanty, P (2003) Prenatal smoke exposure alters growth in limb proportions and head shape in the midgestation human fetus. Am J Hum Biol 15, 533546.CrossRefGoogle ScholarPubMed
Pomeroy, E, Stock, JT, Stanojevic, S et al. (2012) Trade-offs in relative limb length among Peruvian children: extending the thrifty phenotype hypothesis to limb proportions. PLoS ONE 7, e51795.CrossRefGoogle ScholarPubMed
Janghorbani, M, Momeni, F & Dehghani, M (2012) Hip circumference, height and risk of type 2 diabetes: systematic review and meta-analysis. Obes Rev 13, 11721181.CrossRefGoogle ScholarPubMed
Costacou, T & Mayer-Davis, EJ (2003) Nutrition and prevention of type 2 diabetes. Annu Rev Nutr 23, 147170.CrossRefGoogle Scholar
Chang, X, Song, P, Wang, M et al. (2018) The risks of overweight, obesity and abdominal obesity in middle age after exposure to famine in early life: evidence from the China’s 1959–1961 famine. J Nutr Health Aging 22, 11981204.CrossRefGoogle ScholarPubMed
Martins, PA, Hoffman, DJ, Fernandes, MT et al. (2004) Stunted children gain less lean body mass and more fat mass than their non-stunted counterparts: a prospective study. Br J Nutr 92, 819825.CrossRefGoogle ScholarPubMed
Bruce, KD & Hanson, MA (2010) The developmental origins, mechanisms, and implications of metabolic syndrome. J Nutr 140, 648652.CrossRefGoogle ScholarPubMed
Bhargava, SK, Sachdev, HS, Fall, CH et al. (2004) Relation of serial changes in childhood body-mass index to impaired glucose tolerance in young adulthood. N Engl J Med 350, 865875.CrossRefGoogle ScholarPubMed
Peng, J, Zhao, F, Yang, X et al. (2021) Association between dyslipidemia and risk of type 2 diabetes mellitus in middle-aged and older Chinese adults: a secondary analysis of a nationwide cohort. BMJ Open 11, e042821.CrossRefGoogle ScholarPubMed
Figure 0

Table 1 Baseline characteristics of study subjects subdivided by diabetes outcome

Figure 1

Table 2 Incidence of diabetes during the follow-up

Figure 2

Table 3 HR (95 % CI) of incident diabetes for upper arm length and knee height, adjusted for possible explanatory and confounding factors

Figure 3

Table 4 Association of upper arm length/knee height with development of diabetes according to subgroups of various variables

Figure 4

Fig. 1 Hazard ratios (HR) and 95 % CI for the risk of diabetes, by sex, age, BMI, hypertension status or diabetes status

You have Access Open access

Save article to Kindle

To save this article to your Kindle, first ensure coreplatform@cambridge.org is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about saving to your Kindle.

Note you can select to save to either the @free.kindle.com or @kindle.com variations. ‘@free.kindle.com’ emails are free but can only be saved to your device when it is connected to wi-fi. ‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

Find out more about the Kindle Personal Document Service.

Upper arm length and knee height are associated with diabetes in the middle-aged and elderly: evidence from the China Health and Retirement Longitudinal Study
Available formats
×

Save article to Dropbox

To save this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you used this feature, you will be asked to authorise Cambridge Core to connect with your Dropbox account. Find out more about saving content to Dropbox.

Upper arm length and knee height are associated with diabetes in the middle-aged and elderly: evidence from the China Health and Retirement Longitudinal Study
Available formats
×

Save article to Google Drive

To save this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you used this feature, you will be asked to authorise Cambridge Core to connect with your Google Drive account. Find out more about saving content to Google Drive.

Upper arm length and knee height are associated with diabetes in the middle-aged and elderly: evidence from the China Health and Retirement Longitudinal Study
Available formats
×
×

Reply to: Submit a response

Please enter your response.

Your details

Please enter a valid email address.

Conflicting interests

Do you have any conflicting interests? *