Glucose intolerance and cardiovascular risk factors in Hong Kong: Data from two occupation-based cross-sectional surveys

Results Mean (SD) 2-h plasma glucose values were 5.6 mmol/l (2.1) in 1990 and 6.5 mmol/l (2.5) in 2001–2003, an apparent increase of 0.5 mmol/l (95% CI 0.3 to 0.7, p < 0.001) after age and sex adjustment. However, there was no significant difference in the age-standardised prevalence of glucose intolerance, overweight or obesity. There were significantly smaller proportions of women with hypertension and hyperlipidaemia and male smokers in the second compared to the first survey. Conclusions We observed a relatively adverse glycaemia profile, which may have worsened over time, in two healthy populations of survey respondents, with comparatively low rates of most CVD risk factors. This has implications for the future burden of disease associated with hyperglycaemia in this population. Keywords Glucose Diabetes Cardiovascular disease Hong Kong 1 Introduction The prevalence of diabetes is increasing worldwide in parallel with an upward shift in the distribution of blood glucose. Data from China and other Asian countries suggests that this shift is accelerating due to rapid social and cultural change, alongside ageing and growing populations [1–3] . The rising prevalence of diabetes may not accurately reflect the true current and future burden of disease associated with abnormalities of glucose tolerance given the linear continuous nature of the association between plasma glucose (PG) and cardiovascular (CVD) risk that extends below the diabetic threshold [4] . As such, it is important to examine the overall distribution of PG and related CVD risk factors in populations, and how they might change over time. A high proportion of glucose intolerance, overweight, obesity, hypertension, and hyperlipidaemia was observed in cross-sectional population surveys in Hong Kong in the 1990s [5,6] . The “ Better Health for Better Hong Kong ” (BHBHK) public health campaign was introduced in 2001 and aimed to promote healthy lifestyle choices to prevent obesity, hypertension, and hyperlipidaemia, as well as diabetes [7] . Using data from two similar cross-sectional surveys in Hong Kong, we examined the distribution of glucose intolerance and related CVD risk factors in an adult, working Chinese ethnic population in 1990 and in 2001–2003. 2 Subjects, materials and methods 2.1 Data source Both cross-sectional surveys examined glucose intolerance and CVD risk factors in individuals recruited from occupational settings. In 1990, all employees from a public utility company and a regional hospital were invited by letter to participate and complete data were available for 1496 participants (∼75% response rate) (Survey A) [8] . A similar cross-sectional survey was conducted in 2002 under the auspices of the BHBHK campaign (Survey B) [7] . A stratified sample of participants was randomly selected using computer-generated codes in accordance with the distribution of occupational groups in Hong Kong and invited for a health assessment (41% response rate). Complete data on 534 participants were available for this analysis, collected using identical measures and standard operating procedures as employed in the first survey. Both studies were approved by the Chinese University of Hong Kong Clinical Research Ethics Committee. 2.2 Assessment The methods have been described in detail previously [7,8] . In brief, all participants attended the study sites (their worksites for the first survey, and the Prince of Wales Hospital Clinical Trial Unit for the second survey) after an overnight fast. Demographic data, past medical history, smoking and drinking habits were recorded. Height and weight (measured to the nearest 0.1 kg) were measured with the participant wearing light clothing without shoes using the same weighing scales for each survey. BMI was calculated as weight (in kg) divided by the square of the height (in m). Overweight was defined as BMI ≥ 23 kg/m 2 and obesity as BMI ≥ 27.5 kg/m 2 [9,10] . Waist circumference (WC, in cm) was taken as the minimum circumference between the umbilicus and xiphoid process and measured to the nearest 0.5 cm. After sitting for at least 5 min, blood pressure (BP) was measured in the right arm using a standard mercury sphygmomanometer (Survey A) or the Dinamapp machine (Survey B). Korotkoff sound V was taken as the diastolic BP for sphygmomanometer measurements. The mean value of two readings measured 1 min apart was used. Hypertension was defined as BP ≥ 140/90 mm Hg and/or known history of hypertension [11] . Blood was taken after a 12 h fast for measurement of PG, HbA 1c , total cholesterol (TC), triglyceride (TG), and HDL-cholesterol (HDL-C). LDL-cholesterol (LDL-C) was calculated using the Friedewald's formula [12] . Hyperlipidaemia was defined as TC ≥ 5.2 mmol/l and/or LDL-C ≥ 2.6 mmol/l [13] . All participants underwent a 75 g OGTT. Fasting and 2-h PG and insulin were measured during the OGTT. The laboratory assays have been described previously [8] . In brief, TC and TG were assayed enzymatically with commercial reagents (Baker Instruments Corporation, Allentown, PA, USA) on a Cobas Mira analyzer (Hoffmann-La Roche and Co, Basle, Switzerland). Inter-assay CVs were: TC, 1.9% at 6.4 mmol/l; TG, 2.6% at 1.9 mmol/l; and HDL, 5.4% at 0.86 mmol/l. PG was measured by a glucose oxidase method (Diagnostic Chemicals Reagent kit). The intra-assay CV of glucose was 2% at 6.6 mmol/l. HbA 1c was measured using an automated ion-exchange chromatography system (Bio-Rad Laboratory, Hercules, United States). The intra- and inter-assay coefficients were 3.1% or less at values at or below 8.5% and up to 5% at values above 8.5%. Insulin assay was performed using a radioimmunoassay kit (Pharmacia Insulin RIA 100, Pharmacia diagnostics AB, Uppsala, Sweden). The lower limit of detection was <2 μU/ml. The inter-assay CV was 5%. Insulin resistance (IR) was calculated using a computer-based HOMA method: IR = fasting serum insulin/(22.5e −ln fasting plasma glucose ) [14] . Participants were assigned into categories according to the 1999 WHO classification [15] : IFG (impaired fasting glucose) = fasting PG ≥ 6.1 and <7.0 mmol/l and (if measured) 2-h PG < 7.8 mmol/l; IGT (impaired glucose tolerance) = fasting PG < 7.0 mmol/l and 2-h PG ≥ 7.8 and <11.1 mmol/l; diabetes = fasting PG ≥ 7.0 mmol/l and/or 2-h PG ≥ 11.1 mmol/l. Our definition of diabetes included both known (clinically diagnosed) and unknown (undiagnosed prevalent) cases of the disease. 2.3 Statistical analysis Statistical analysis was performed using Stata (version 10.1) software. Type I error was set at 0.05 for all tests. All results are expressed as mean ± SD or number (%) where appropriate. To examine glucose intolerance and related CVD risk factor prevalence in the two surveys, data were stratified by sex, and age-standardized by the direct method using the Hong Kong 2007 population (Hong Kong Hospital Authority Statistical Report [16] ). Age-specific rates were calculated by summing the number of individuals with diabetes in each 5-year age group and dividing by the total population for that age group. Each age specific rate was then multiplied by the standard population weighting for that age group and summed to produce the age-standardised rate. Linear regression modelling was used to examine age- and sex-adjusted differences in the means of continuous CVD risk factors in the two surveys, where CVD risk factor was entered as the dependent variable and the survey time (A or B) as the independent variable. 3 Results Table 1 shows the characteristics of men and women taking part in Survey A (1990) and Survey B (2002). The mean age of men was 37 years in Survey A and 45 years in Survey B. The proportion of male current or ex-smokers was significantly different between the two surveys (Survey A vs. B: 36% vs. 26%, p = 0.004). The mean BMI was in the overweight category for both surveys, while mean TC and LDL-C were in the hyperlipidaemic category. Only small proportions of the male population were prescribed anti-hypertensive or diabetes drugs in both 1990 and 2001–2003. For women, the mean age was 39 years in Survey A and 43 years in Survey B. There were no differences in the proportion of current or ex-smokers between surveys. The mean BMI was in the overweight category for both surveys, while mean LDL-C was in the hyperlipidaemic category. Again, only small proportions of women were prescribed anti-hypertensive or diabetic drugs in both time periods. The crude prevalence of known diabetes (clinically diagnosed) was 1.8% in Survey A and 1.9% in Survey B. Table 2 shows the age-standardised prevalence of diabetes (including both clinically diagnosed and undiagnosed prevalent diabetes), IFG, IGT and related CVD risk factors in men and women in Survey A and Survey B. In men, the age-standardised prevalence of diabetes and IFG was similar in the two surveys. While increases in the frequency of IGT, glucose intolerance and dysglycaemia were observed, these did not reach statistical significance. There were no significant differences in the age-standardised prevalence of overweight and obesity. Declines in hypertension and hyperlipidaemia also did not reach statistical significance. In women, the age-standardised prevalence of IGT, diabetes, glucose intolerance and dysglycaemia was higher in 2001–2003 compared to 1990, but none of these differences reached statistical significance. There were no significant differences in the prevalence of IFG. While the prevalence of overweight and obesity declined between the two surveys, these differences did not reach statistical significance. The age-standardised prevalence of hypertension was significantly lower in the second survey ( p < 0.001). Similarly, the prevalence of hyperlipidaemia was lower in Survey B than in Survey A ( p < 0.001). Fig. 1 shows the distribution of 2-h PG. Table 3 shows the age- and sex-adjusted difference in continuous CVD risk factors between the two surveys mean 2-h PG was 0.5 mmol/l higher in 2001–2003 than in 1990 (95% CI 0.3 to 0.7, p < 0.001), while mean HbA 1c was 1.0% higher in Survey B (95% CI 0.7 to 1.3, p < 0.001). There were no significant differences in mean BMI and WC. Mean systolic BP was 3 mm Hg lower in Survey B, while there was no significant difference in mean diastolic BP. Mean HDL-C and TG values were higher in 2001–2003 compared to 1990, while LDL-C values were 0.26 mmol/l lower on average. There was no difference in HOMA values between the two surveys. 4 Discussion In two healthy populations of working Hong Kong Chinese survey respondents with comparatively low rates of most CVD risk factors, individuals exhibited a relatively adverse glycaemic profile. Mean 2-h PG and HbA 1c values were significantly higher in 2001–2003 compared to 1990. Corresponding increases in the prevalence of glucose intolerance and dysglycaemia did not reach statistical significance. There were significantly smaller proportions of women with hypertension and hyperlipidaemia and male smokers in Survey B. This suggests that Hong Kong faces a continued public health challenge in relation to hyperglycaemia, diabetes and CVD. The significant proportion of the population with glucose intolerance, overweight, obesity, hypertension, and hyperlipidaemia has been observed in other surveys in Hong Kong [5,6] , and more recently in mainland China [1,3,11,17] . The decrease in LDL-C levels, alongside improvements in HDL-C and systolic BP between 1990 and 2001–2003, may be due to increased awareness of CVD risk factors and better case finding and treatment. While we do not have data on statin use in the two surveys, there was a significantly higher proportion of men prescribed anti-hypertensive medication in Survey B compared to Survey A. The “ Better Health for Better Hong Kong ” public health campaign took place between 2001 and 2004 and aimed to promote healthy lifestyle choices to prevent obesity, hypertension, hyperlipidaemia, and diabetes [7] . This improvement in certain CVD risk factors has been observed in other developed populations [18,19] . The decline in the prevalence of smoking in men also mirrors trends seen in developed countries [18,19] . To the best of our knowledge, this is the first study which documents the possibility of increasing values of 2-h PG and HbA 1c over time in Chinese living in an affluent city such as Hong Kong. Given the rapid socioeconomic development in mainland China, these secular changes provide a useful indication of the potential disease burden in the 1.2 billion people living in this region. A similar transition in glucose intolerance has been observed in populations of West African-origin, with increasing levels of age-standardised diabetes from Cameroon to Jamaica and Britain alongside increasing levels of BMI [20] . Similarly, there was an increase in crude diabetes prevalence in Qingdao between 2001–2002 and 2006, in parallel with an increase in BMI in rural areas [3] . However, the changes in glycaemia observed in Hong Kong appear to have occurred without a significant increase in weight. Our finding that rates of overweight and obesity have been relatively stable in Hong Kong over 10 years is supported by other published data in this region [21,22] . Interestingly, these trends have been observed alongside a doubling of central obesity in men [21] , as well as the apparent increases in glycaemia seen in the two surveys in this analysis. Dietary and physical activity patterns, which may have been expected to contribute to secular changes in CVD risk factors, have remained largely static in Hong Kong [23] . Other factors, such as psychosocial stress and increased sedentary behaviour [24] , which may contribute to the increasing prevalence of dysglycaemia, have yet to be explored. 4.1 Strengths and limitations Both cross-sectional surveys included relatively large numbers of participants drawn from working populations in Hong Kong. However, the small number of participants in some CVD risk factor categories when undertaking analyses by sex means we may have been under-powered to detect possible differences. Participants were measured in different locations in each survey, however standard operating procedures were identical. The same laboratory was used to analyse blood samples and similar test procedures were undertaken. Participants in the second cross-sectional survey were randomly selected using computer-generated codes in proportions corresponding to the distributions of different occupational groups and socio-economic classes in Hong Kong, which suggests that results from Survey B may be generalisable to working populations. Further, given that the data from working populations were age standardised and adjusted for sex, we believe the two groups were demographically similar and comparable. However, we do not have any information on non-responders, and the response rate was different between surveys, which limits the wider generalisability of our findings. The self-selected nature of the participants means that the samples are likely to be subject to healthy volunteer bias, which has been observed in a number of epidemiological studies [25–27] . Analysis of attendance in a long-term prospective cohort, the British Regional Heart study, showed that non-attendees were older than attendees, had higher BP, were more likely to smoke, were more likely to work in manual occupations, and had significantly increased mortality rates [28] . As such, the prevalence of adverse CVD risk factors in Hong Kong Chinese workers is likely to be higher than that found in this study. If participants in the second survey were healthier than non-respondents then the apparent adverse shift in the population distribution of glycaemia over time may be underestimated. We cannot exclude the possibility that the differences we observed between the two surveys were due to in-out migration. This may be particularly pertinent given that the transfer of sovereignty from the UK to the People's Republic of China took place in 1997, mid-way between the two surveys. 4.2 Public health challenge Our data demonstrate that the difference in the prevalence of diabetes and IGT using arbitrary cut points along the glucose distribution occur through small shifts in frequency distribution along this scale over time. As such, while the proportion of the population with glucose intolerance did not differ significantly between the two surveys, this finding overlooks the large population shift we observed in 2 h PG and HbA 1c . This shift in distribution warns of a future rapid rise in the number of people with diabetes and IGT as the exponential part of the glucose distribution shifts towards the diagnostic thresholds. Furthermore, it highlights a future increase in the risk of CVD events associated with hyperglycemia in the population. While better detection and treatment may have contributed to the decline in hypertension and hyperlipiaemia seen in women, the shift in the population curve of glucose demonstrates that increasing attention to earlier and more intensive treatment of CVD risk factors can only be part of the solution. Population-based strategies should be considered a priority alongside identification and treatment of individuals at high risk. This approach aims to shift the whole distribution of exposure in a favourable direction to alter some of the society's norms of behaviour [29] regarding diet and physical activity. This paper also highlights the need to report population distributions of glucose rather than simply prevalence rates of diabetes, and to describe appropriately standardised shifts in distributions to accurately reflect the increasing public health burden of hyperglycaemia. 5 Conclusion In two healthy populations of Hong Kong working survey respondents with comparatively low rates of most CVD risk factors, individuals exhibited a relatively adverse glycaemic profile. The shift in the distribution of PG from 1990 to 2001–2003, a continuous linear CVD risk factor, has important implications for the future prevalence of diabetes and IGT in this population, as well as the burden of CVD associated with hyperglycaemia. Hong Kong is often considered as the epitome of future China and our findings may serve as useful references as the mainland continues to modernize. Population-based strategies aiming to shift the distribution of CVD risk factors, including hyperglycaemia, alongside individual-based approaches among those at high risk, will be needed for disease prevention. Conflict of interest There are no conflicts of interest. Acknowledgement With many thanks to Stephen Sharp (MRC Epidemiology Unit) for his statistical expertise and help with the analysis. References [1] J.C. Chan V. Malik W. Jia T. Kadowaki C.S. Yajnik K.H. Yoon Diabetes in Asia: epidemiology, risk factors, and pathophysiology JAMA 301 May (20) 2009 2129 2140 [2] K.H. Yoon J.H. Lee J.W. Kim J.H. Cho Y.H. Choi S.H. Ko Epidemic obesity and type 2 diabetes in Asia Lancet 368 November (9548) 2006 1681 1688 [3] F. Ning Z.C. Pang Y.H. Dong W.G. Gao H.R. Nan S.J. Wang Risk factors associated with the dramatic increase in the prevalence of diabetes in the adult Chinese population in Qingdao, China Diabet Med 26 September (9) 2009 855 863 [4] E.B. Levitan Y. Song E.S. Ford S. Liu Is nondiabetic hyperglycemia a risk factor for cardiovascular disease? A meta-analysis of prospective studies Arch Intern Med 164 October (19) 2004 2147 2155 [5] G.N. Thomas S.Y. Ho K.S. Lam E.D. Janus A.J. Hedley T.H. Lam Impact of obesity and body fat distribution on cardiovascular risk factors in Hong Kong Chinese Obes Res 12 November (11) 2004 1805 1813 [6] G.T. Ko J.S. Tang Prevalence of obesity, overweight and underweight in a Hong Kong community: the United Christian Nethersole Community Health Service (UCNCHS) primary health care program 1996–1997 Asia Pac J Clin Nutr 15 2 2006 236 241 [7] G.T. Ko J.C. Chan A.W. Chan P.T. Wong S.S. Hui S.D. Tong Low levels of awareness of suboptimal health conditions in a high-risk working population: the “better health for better Hong Kong” health promotion campaign Int J Behav Med 14 2 2007 63 69 [8] C.S. Cockram J. Woo E. Lau J.C. Chan A.Y. Chan J. Lau The prevalence of diabetes mellitus and impaired glucose tolerance among Hong Kong Chinese adults of working age Diabetes Res Clin Pract 21 July (1) 1993 67 73 [9] WHO/IASO/IOTF Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies Lancet 363 January (9403) 2004 157 163 [10] WHO/IASO/IOTF The Asia-Pacific perspective: redefining obesity and its treatment 2000 WHO/IASO/IOTF Melbourne, Australia [11] European Society of Hypertension-European Society of Cardiology European Society of Hypertension-European Society of Cardiology guidelines for the management of arterial hypertension J Hypertens 21 June (6) 2003 1011 1053 [12] W.T. Friedewald R.I. Levy D.S. Fredrickson Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge Clin Chem 18 June (6) 1972 499 502 [13] NCEP Executive summary of the third report of The National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III) JAMA 285 May (19) 2001 2486 2497 [14] D.R. Matthews J.P. Hosker A.S. Rudenski B.A. Naylor D.F. Treacher R.C. Turner Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man Diabetologia 28 July (7) 1985 412 419 [15] WHO Definition, diagnosis, and classification of diabetes mellitus and its complications: report of a WHO consultation 1999 WHO Geneva [16] Hong Kong Hospital Authority. http://www.ha.org.hk/visitor/ha_visitor_index.asp?Parent_ID=105&Content_ID=339&Dimension=100 . [17] W.P. Jia C. Pang L. Chen Y.Q. Bao J.X. Lu H.J. Lu Epidemiological characteristics of diabetes mellitus and impaired glucose regulation in a Chinese adult population: the Shanghai Diabetes Studies, a cross-sectional 3-year follow-up study in Shanghai urban communities Diabetologia 50 February (2) 2007 286 292 [18] A. Rosengren H. Eriksson B. Larsson K. Svardsudd G. Tibblin L. Welin Secular changes in cardiovascular risk factors over 30 years in Swedish men aged 50: the study of men born in 1913, 1923, 1933 and 1943 J Intern Med 247 January (1) 2000 111 118 [19] B. Unal J.A. Critchley S. Capewell Modelling the decline in coronary heart disease deaths in England and Wales, 1981–2000: comparing contributions from primary prevention and secondary prevention BMJ 331 September (7517) 2005 614 [20] J.C. Mbanya J.K. Cruickshank T. Forrester B. Balkau J.Y. Ngogang L. Riste Standardized comparison of glucose intolerance in west African-origin populations of rural and urban Cameroon, Jamaica, and Caribbean migrants to Britain Diabetes Care 22 March (3) 1999 434 440 [21] T.C. Ko J. Chan A. Chan P. Wong S. Hui F. Chow Doubling over ten years of central obesity in Hong Kong Chinese working men Chin Med J 120 July (13) 2007 1151 1154 [22] G.T. Ko J.S. Tang J.C. Chan Worsening trend of central obesity despite stable or declining body mass index in Hong Kong Chinese between 1996 and 2005 Eur J Clin Nutr 64 May (5) 2010 549 552 [23] K.L. Chou The prevalence and clustering of four major lifestyle risk factors in Hong Kong Chinese older adults J Aging Health 20 October (7) 2008 788 803 [24] D.W. Dunstan J. Salmon N. Owen T. Armstrong P.Z. Zimmet T.A. Welborn Associations of TV viewing and physical activity with the metabolic syndrome in Australian adults Diabetologia 48 November (11) 2005 2254 2261 [25] R. Benfante D. Reed C. MacLean A. Kagan Response bias in the Honolulu Heart Program Am J Epidemiol 130 December (6) 1989 1088 1100 [26] K.D. Lindsted G.E. Fraser M. Steinkohl W.L. Beeson Healthy volunteer effect in a cohort study: temporal resolution in the Adventist Health Study J Clin Epidemiol 49 July (7) 1996 783 790 [27] P. Froom S. Melamed E. Kristal-Boneh J. Benbassat J. Ribak Healthy volunteer effect in industrial workers J Clin Epidemiol 52 August (8) 1999 731 735 [28] M.C. Thomas M. Walker L.T. Lennon A.G. Thomson F.C. Lampe A.G. Shaper Non-attendance at re-examination 20 years after screening in the British Regional Heart Study J Public Health Med 24 December (4) 2002 285 291 [29] G. Rose Sick individuals and sick populations Int J Epidemiol 14 1 1985 32 38