OBESITY Body mass index and chronic obstructive pulmonary disease-related mortality: a nationally representative prospective study of 220 000 men in China

of 10

Please download to get full document.

View again

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
10 pages
0 downs
OBESITY Body mass index and chronic obstructive pulmonary disease-related mortality: a nationally representative prospective study of 220 000 men in China
  OBESITY Body mass index and chronic obstructivepulmonary disease-related mortality: anationally representative prospectivestudy of 220000 men in China Ling Yang, 1 * Maigeng Zhou, 2 Margaret Smith, 1 Gonghuan Yang, 2 Richard Peto, 1 Jun Wang, 2 Jillian Boreham, 1  Yisong Hu 2 and Zhengming Chen 1 1 Clinical Trial Service Unit & Epidemiological Studies Unit (CTSU), University of Oxford, Oxford, UK and  2 Center of Public HealthSurveillance and Information Service, Chinese Center for Disease Control and Prevention, Beijing, P.R. China*Corresponding author: CTSU, Richard Doll Building, Old Road Campus, University of Oxford, Oxford OX3 7LF, UK.E-mail: ling.yang@ctsu.ox.ac.uk Accepted  15 February 2010 Background  Low body mass index (BMI) is associated with chronic obstructivepulmonary disease (COPD) in populations where many are over- weight. Substantial uncertainty remains about the relationship inpopulations with lower mean BMI levels, and about the relevanceto it of the effects of smoking or of reverse causality. Methods  A nationally representative prospective cohort study included221194 Chinese men aged 40–79 years in 1990–91, who were fol-lowed up for 15 years or to the age of 80 years. Hazard ratios forCOPD-related mortality vs baseline BMI were adjusted for age,smoking, drinking and other factors. To reduce reverse causality,main analyses excluded all men with prior history of any respira-tory diseases or abnormal lung function at baseline, leaving 2960COPD-related deaths (16% of all deaths). Results  The mean baseline BMI was 21.7 kg/m 2 . There was a highly sig-nificant inverse association between BMI and COPD-related mortal-ity among men without any apparent impairment of lung function. Approximately 90% of men had a baseline BMI <25 kg/m 2 , andamong them, 5 kg/m 2 lower BMI was associated with 31% (95%confidence interval 18–45%) higher COPD-related mortality. Theexcess risk persisted after restricting the analysis to never-smokersor excluding the first 5 years of follow-up. Conclusions  Low BMI is associated with increased COPD mortality in a relativelylean adult male population in China where COPD is one of the mostcommon causes of death. Keywords  Body mass index, smoking, prospective study, chronic obstructivepulmonary disease, mortality Published by Oxford University Press on behalf of the International Epidemiological Association   The Author 2010; all rights reserved. Advance Access publication 16 April 2010  International Journal of Epidemiology  2010; 39 :1027–1036doi:10.1093/ije/dyq051 1027   a  t  F  un d  a Ã  § Ã  £  o C  o or  d  e n a Ã  § Ã  £  o d  e A p e r f   e i  Ã  §  o a m e n t   o d  e P  e  s  s  o a l   d  e  Nà- v e l   S  u p e r i   or  on N o v e m b  e r 1  0  ,2  0 1 2 h  t   t   p :  /   /  i   j   e  . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om   Introduction Chronic obstructive pulmonary disease (COPD) is aleading cause of death and disability worldwide, 1  with more than 200 million affected individuals andapproximately 3 million annual deaths. 2 In China,COPD causes about 1.3 million deaths every year 3 and the prevalence of spirometric-detected COPD inpeople aged  5 40 years has been reported to be 8%, with three-quarters of those affected having clinicallymanifest disease. 4 Despite this high disease burden,epidemiological evidence about the aetiology of COPD in China is limited. 5 In Western populations, where cigarette smokinghas persisted for many decades, smoking is respon-sible for most COPD deaths. 6,7 In China, the situationis different, with only about half of the male COPDdeaths in the 1980s being attributable to smoking, 8 and with high rates even among individuals whohave never smoked, especially in certain inland ruralregions. 4,9,10 It has been suggested that, in addition tosmoking, other factors such as particular types of indoor air pollution, chronic infection and malnutri-tion may also contribute significantly to risk of thedisease. 1,11–14 Several hospital-based studies have shown that lowbody mass index (BMI; weight in kilograms dividedby the square of the height in metres) is an independ-ent prognostic factor for both long-term survival fol-lowing diagnosis of COPD 15–20 and disease severity. 21 In cross-sectional studies, COPD is more prevalentamong lean people. 6,22–25 Recent data from large pro-spective cohort studies have also shown that low BMIis associated with COPD. However, most such studieshave been in European-origin populations wherethe prevalence of COPD is relatively low, especiallyamong non-smokers, and few have low BMI persist-ently throughout childhood and adult life. 26–29 Furthermore, most such studies have included onlysmall number of deaths and did not control appropri-ately for smoking or for the effects of pre-existingdisease on BMI (reverse causality). 29 We report a15-year prospective study of the association betweenBMI and COPD mortality in a nationally representa-tive cohort of 220000 men in China recruited during1990–91, including 148000 current smokers and59000 never-smokers. Materials and methods Baseline survey  Details of the study design, participants and baselinesurvey methods have been reported previously. 30,31 Briefly, the study involves 45 areas (23 urban and22 rural) randomly selected from China’s 145Disease Surveillance Points (DSPs). The DSP system was established in the mid-1980s to provide a nation-ally representative sample of mortality statistics forthe entire country. 32,33  A typical surveillance pointcovers a defined population of approximately50000–100000 residents in four to eight geographic-ally defined administrative units (i.e. catchment areaof urban street committees or clusters of rural villages). During 1990–91, all men aged  5 40 yearsfrom two or three randomly selected units withineach of the 45 areas were invited to participate inthe survey, and   80% of invitees attended. Insurvey clinics set up for the study, trained health workers administered a standardized questionnaire, which included education, occupation, tobacco use,alcohol consumption, dietary patterns, exposure toindoor air pollution, self-reported medical historyand health status. Physical measurements includedblood pressure (BP), height, weight and peak expira-tory flow rate. Waist and hip circumference andsubsequent weight changes were not measured. Follow-up for cause-specific mortality  The vital status of each study participant was moni-tored regularly by DSP staff through their death regis-tries previously established in these areas. In addition,active confirmation of vital status was done annuallythrough local residential committees. Causes of death were sought chiefly from official death certificates,supplemented, if necessary, by review of medical re-cords. 32 The underlying cause of each death wascoded centrally by DSP staff in Beijing, withoutknowledge of the baseline information, using theninth revision of the International Classification of Disease (ICD-9). In the few deaths without anyrecent medical attention, standard procedures wereused by local DSP staff to determine the probablecause from symptoms or signs described by familymembers. 32 COPD-related death was defined asdeath from chronic bronchitis, emphysema, asthma,bronchiectasis, other COPD, or pulmonary heart dis-ease (ICD-9 codes 490–496 and 415–417).Since it is often difficult to assign an underlyingcause reliably for death in old age, all analyses wererestricted to deaths occurring between age 40 and79 years, with censoring when men reached 80 yearsof age, died from other causes or moved away from theoriginal study area during follow-up. This report isbased on follow-up data to 1 January 2006 amongthe 221194 men who were aged 40–79 years andhad baseline BMI in the range of 10–50 kg/m 2 . Statistical analysis Cox proportional hazards models were used to calcu-late hazard ratios (HRs), with BMI as the exposure variable and COPD-related death as the outcome. All analyses were stratified by area (45 strata) andby 5-year age group at risk, and were adjusted simul-taneously for education, tobacco use (never, former,current), alcohol consumption (at least weekly, not),regular exposure to indoor domestic or occupationalair pollution from coal or biomass (yes, no), and 1028  INTERNATIONAL JOURNAL OF EPIDEMIOLOGY   a  t  F  un d  a Ã  § Ã  £  o C  o or  d  e n a Ã  § Ã  £  o d  e A p e r f   e i  Ã  §  o a m e n t   o d  e P  e  s  s  o a l   d  e  Nà- v e l   S  u p e r i   or  on N o v e m b  e r 1  0  ,2  0 1 2 h  t   t   p :  /   /  i   j   e  . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om   consumption of certain food stuffs (times/week), suchas meat, vegetables and fruit. Baseline BMI wasdivided into five categories, with cut-off points at18.5, 20, 22.5 and 25 kg/m 2 . The HR forCOPD-related mortality was calculated for each BMIcategory, with the BMI category of 20.0–22.5 kg/m 2 asthe reference group. The 95% confidence interval (CI)for each log HR was estimated using the ‘floatingabsolute risk’ method, 34  which facilitates many differ-ent comparisons and tests for trend between differentBMI categories, rather than just pair-wise compari-sons between one arbitrarily chosen reference groupand each of the other categories. Trend tests wereconducted treating BMI as a continuous variablein the Cox model. The analyses were done inSAS version 9.1 (SAS Institute Inc., Cary, NC, USA)under Windows XP. Results For the 221194 participants (mean baseline age54.3 years), overall mean BMI was 21.7 (standarddeviation 2.7) kg/m 2 , with 9.2% being underweight(BMI <18.5 kg/m 2 ), and 11.1% being either over- weight (BMI 25 to <30 kg/m 2 , 10.3%) or obese(BMI 5 30 kg/m 2 , 0.8%). The mean age-adjusted BMI was higher among urban than among rural men(23.1 vs 21.2 kg/m 2 ). After adjusting for area andage, it was strongly positively related to BP (Table 1), with similar gradients for smokers and never-smokers(data not shown). Men with high BMI were slightlybetter educated, and somewhat more likely to drinkalcohol. The prevalence of smoking was stronglyinversely associated with BMI, but among currentsmokers, the mean number of cigarette smoked per Table 1  Baseline characteristics of participants and standardized mortality rate from COPD-related disease by BMIcategory, all men aged 40–79 years at baseline Baseline characteristics a AllBMI categories (kg/m 2 ) <18.5 18.5–19.9 20–22.4 22.5–24.9  5 25Total population 221194 20338 37748 89079 49418 24611Mean BMI (kg/m 2 ) 21.7 17.6 19.4 21.3 23.5 26.9Mean age (years) 54.3 59.0 55.4 53.4 53.1 54.4Living in urban area (%) 27.2 17.9 15.5 19.3 35.6 64.4Mean weight (kg) 59.0 48.4 53.0 57.8 63.4 72.3Mean height (cm) 164.5 165.5 165.1 164.7 163.9 163.5Mean SBP (mm Hg) 124.1 119.4 121.3 123.4 125.8 131.2Mean peak flow (l/min) 395.2 360.8 381.9 397.7 406.8 412.1Education 5 6 years (%) 33.1 32.8 31.4 32.3 34.3 36.6Ex-smokers (%) 6.3 6.1 5.2 5.5 6.8 9.8Current smokers (%) 67.1 72.3 71.6 68.7 63.7 56.8Number of cigarettes/day 11.9 11.6 11.6 11.9 12.1 12.3 Age started smoking (years) 22.2 22.0 22.1 22.2 22.4 22.7Regular alcohol drinking (%) 33.5 30.6 31.9 33.7 34.6 35.1Exposure to indoor air pollution (%) 10.9 11.1 10.9 10.7 11.0 11.1Intake of meat (times/week) 3.5 3.3 3.3 3.4 3.6 3.8Intake of vegetable (times/week) 13.8 13.7 13.7 13.7 13.8 13.9Intake of fruit (times/week) 1.3 1.2 1.2 1.3 1.4 1.6History of major chronic disease (%) 19.4 27.7 21.0 17.7 17.2 20.2Prior respiratory disease (%) 15.9 24.6 17.9 14.6 13.6 14.5Chronic respiratory symptoms (%) 28.4 36.6 30.8 27.0 25.4 29.0Self-reported poor health status (%) 7.3 13.5 8.4 6.6 5.5 6.6Mortality from COPD b Number of deaths 8769 1860 1980 3244 1264 421Mortality rate /100000 255.6 490.9 301.1 235.1 194.5 189.2 a Except for locality and age, all other characteristics were adjusted for individual area and age by 5-year age group, by directstandardization to the whole study population. b Mortality rates were standardized to the geographic area and the 5-year age group structure in the study population aged40–79 years.SBP, systolic BP. BMI AND COPD-RELATED MORTALITY  1029   a  t  F  un d  a Ã  § Ã  £  o C  o or  d  e n a Ã  § Ã  £  o d  e A p e r f   e i  Ã  §  o a m e n t   o d  e P  e  s  s  o a l   d  e  Nà- v e l   S  u p e r i   or  on N o v e m b  e r 1  0  ,2  0 1 2 h  t   t   p :  /   /  i   j   e  . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om   day was not associated with BMI (which implies that,among smokers, the dose of tobacco per kilogram of body weight was inversely related to BMI). Men withlow BMI were more likely to report a prior historyof disease at baseline (Table 1).During the 15-year follow-up, 40907 died before80 years of age, 8769 (21.4%) from COPD-related con-ditions. Of the remainder, 34165 (1% per annum) were lost to follow-up, mainly because of demolitionof whole neighbourhoods for redevelopment. Thoselost to follow-up did not differ significantly fromthose not in the main baseline characteristics suchas age, SBP, BMI or prevalence of ever-smoking. 31 The standardized COPD mortality rate was morethan twice as high among men with than without ahistory of respiratory disease at baseline (540 vs 203per 100000, respectively). It was also about twice ashigh among rural than urban men (290 vs 156 respec-tively), and higher among ever-smokers than amongnever-smokers (265 vs 223 respectively). Overall,there was a highly significant inverse associationbetween BMI and COPD-related mortality, withthe death rate in the lowest BMI group(<18.5kg/m 2 ) being 2.5-fold that in the highestgroup ( 5 25kg/m 2 ) (Table 1).Peak expiratory flow rate was strongly inverselyassociated with COPD-related mortality (Figure 1).This inverse association was particularly steepamong men with impaired lung function at baseline[i.e. a diagnosis of prior tuberculosis (TB), chronicbronchitis, asthma, emphysema or pulmonary heartdisease; self-reported chronic respiratory symptoms(i.e. cough or shortness of breath); or in the lowestquintile of age- and height-adjusted peak expiratoryflow rate] (Figure 2). Among these participants, each5 kg/m 2 lower baseline BMI was associated with 118%(95% CI 104–134%) higher risk of COPD-related deathfor BMI <25 kg/m 2 , whereas in men without im-paired lung function, the increase in risk was shal-lower [30% (17–44%)]. Exclusion of the first 5 years Figure 1  HRs for death from COPD-related disease vsbaseline age- and height-adjusted peak flow (l/min) amongmen aged 40–79 years. Analyses were stratified by area andage, and adjusted simultaneously for education, tobacco,alcohol, exposure to indoor domestic or occupational airpollution and dietary patterns. The HRs are plotted on afloating absolute scale. Each square has an area inverselyproportional to the standard error (SE) of the log risk.Vertical lines indicate the corresponding 95% CIs (HR*exp   1.96SE). Numbers above CIs are of deaths and thosebelow CIs are the HRs Figure 2  HRs for death from COPD-related disease vsbaseline BMI among men aged 40–79 years (i) withimpaired lung function at baseline (white squares) and(ii) without impaired lung function at baseline (blacksquares). Conventions as in Figure 1; trends were fittedby linear regression through mean BMI for each categoryamong men with baseline BMI <25 kg/m 2 1030  INTERNATIONAL JOURNAL OF EPIDEMIOLOGY   a  t  F  un d  a Ã  § Ã  £  o C  o or  d  e n a Ã  § Ã  £  o d  e A p e r f   e i  Ã  §  o a m e n t   o d  e P  e  s  s  o a l   d  e  Nà- v e l   S  u p e r i   or  on N o v e m b  e r 1  0  ,2  0 1 2 h  t   t   p :  /   /  i   j   e  . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om   of follow-up did not materially alter these relation-ships [94% (76–113%) higher risk per 5 kg/m 2 lowerBMI for men with impaired lung function vs 34%(17–53%) for men without it]. To limit the effectsof reverse causality, those with impaired lung func-tion at baseline were excluded from subsequentanalyses. Among men without any apparent impairment of lung function, the relationship of BMI with theCOPD HR was similar during different periods of follow-up (  P ¼ 0.4 for heterogeneity among the threefollow-up periods, Figure 3). Moreover, further exclu-sion of those who reported poor health status at base-line did not materially change the association (data Figure 3  HRs for death from COPD-related disease vs baseline BMI among men aged 40–79 years with no prior respiratorydisease or impaired lung function at baseline in ( a ) the first 5 years, ( b ) the second 5 years and ( c ) the third 5 years of follow-up. Conventions as in Figures 1 and 2 Figure 4  HRs for death from COPD-related disease vs baseline BMI among men aged 40–79 years with no prior respira-tory disease or impaired lung function at baseline in ( a ) ever-smokers and ( b ) never-smokers. Conventions as in Figures 1and 2BMI AND COPD-RELATED MORTALITY  1031   a  t  F  un d  a Ã  § Ã  £  o C  o or  d  e n a Ã  § Ã  £  o d  e A p e r f   e i  Ã  §  o a m e n t   o d  e P  e  s  s  o a l   d  e  Nà- v e l   S  u p e r i   or  on N o v e m b  e r 1  0  ,2  0 1 2 h  t   t   p :  /   /  i   j   e  . oxf   or  d  j   o ur n a l   s  . or  g /  D o wnl   o a  d  e  d f  r  om 
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks