Search
Close
Search
Advanced Search
Search Menu
AI Discovery Assistant

Abstract

Context

Many studies have reported conflicting evidence on the association between weight change and mortality.

Objective

We investigated the association between weight change and subsequent all-cause mortality, using a large-scale, population-based cohort from the National Health Insurance System health checkup data between 2005 and 2015.

Methods

A total of 11,524,763 subjects older than age 20 years were included. Weight was measured every 2 years and weight change over 4 years was divided into eight categories, from weight loss ≥15% to weight gain ≥20%, for every 5% of weight change. The hazard ratios (HRs) and 95% confidence intervals (CIs) for all-cause mortality were analyzed using multivariable Cox proportional hazard models compared with the stable weight group (weight change <5%) after adjusting for age, sex, smoking, drinking, exercise, diabetes mellitus, hypertension, dyslipidemia, cancer, and income.

Results

Weight loss was associated with increased mortality rates compared with weight gain; the group with weight loss ≥15% had the highest HR for all-cause mortality (HR, 2.598; 95% CI, 2.537 to 2.659). The HR for all-cause mortality in the ≥20% weight gain group was 1.784 (95% CI, 1.695 to 1.877). Across all body mass index (BMI) categories, weight loss ≥15% was associated with increased mortality rates and the highest mortality rates were found in the BMI ≥30 kg/m2 group (HR, 3.469; 95% CI, 2.236 to 5.381).

Conclusions

Weight change over 4 years showed a reverse J-shaped all-cause mortality curve, independent of BMI status. Weight loss was associated with a greater risk of mortality than was weight gain.

Obesity is a risk factor for cardiometabolic diseases, including diabetes mellitus (DM), hypertension (HTN), cardiovascular diseases, and some cancer types (1, 2). High body mass index (BMI) is associated with increased mortality rates, according to J-shaped (3, 4) or U-shaped curves (5, 6) as BMI increases. Both underweight and obesity are associated with increased mortality, and obesity in mid adulthood or childhood is associated with increased all-cause mortality (7, 8). However, weight status changes continuously throughout a lifetime; therefore, changes in weight or BMI are also considered in the association between weight status and mortality.

Weight loss in overweight or obese people and weight gain in underweight people had beneficial effects on health, with lower mortality rates (9, 10). However, many studies have shown conflicting results regarding the association between weight change and mortality. Some studies showed that weight loss was associated with increased mortality, independent of BMI status (11–20), but some studies found no association (21, 22). The association between weight gain and mortality is also controversial, with some studies showing increased mortality (10, 12, 23, 24) and others showing no association (11–17, 21, 22).

Many studies have examined the association between weight change and mortality in the elderly population (9, 11, 13, 16, 17, 20). Some studies examined this association among white people (9, 13–15, 17, 18, 20, 22–24), Mexican Americans (11), and Asians (19, 21, 25). Many studies had relatively few participants (11, 13, 15, 18, 21, 23). However, to our knowledge, there are currently no large-scale studies involving the Korean population. Therefore, we investigated the association between weight change and subsequent all-cause mortality rates using a large-scale, population-based cohort derived from the National Health Insurance System (NHIS) health checkup data.

Materials and Methods

The NHIS database and NHIS health checkup program

The NHIS is a single insurer that manages the National Health Insurance program, with medical information on ∼50 million Koreans. The NHIS includes patients’ demographic data, such as age, sex, area of residence, medical use and transaction information, insurer payment coverage, and claims and deductions data. The source population of the NHIS is derived from the Health Insurance Review and Assessment service. The Health Insurance Review and Assessment database includes data on ~97.0% of the Korean population’s health insurance claims; details of the NHIS database have been described elsewhere (26).

The National Health Insurance program provides a biannual health checkup program for all insured Koreans, except for employee subscribers who undergo yearly health checkups. The NHIS health checkup program consists of four areas: general health checkup, cancer checkup, lifetime transition-period health checkup, and baby/infant health checkup. Anthropometric measurements, such as weight, height, waist circumference, and systolic and diastolic blood pressures are measured by trained examiners. The NHIS health checkup also includes tests for visual and hearing acuity; laboratory tests, such as fasting blood glucose, total cholesterol, triglyceride, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and serum creatinine levels; and urine analysis. Self-reported questionnaires for general health behaviors such as current smoking status, alcohol intake, and regular exercise were reviewed; The Korean Association of Laboratory Quality Control warranted the quality of the laboratory tests; the hospitals that performed the health checkups were certified by the NHIS.

Subjects

We used the NHIS health checkup database between 2005 and 2015 to select subjects older than 20 years who had health checkups over 4 years. A total of 11,676,522 subjects were followed up to 31 December 2015. The following were excluded: subjects with missing data (n = 71,148), those with a history of abrupt height loss ≥5 cm (n = 80,596), and subjects with unmatched mortality information (n = 15) over 4 consecutive years. Finally, 11,524,763 subjects were included in this study (n = 6,327,210 men; n = 5,197,553 women) and the mean observational time [± standard deviation (SD)] was 5.2 ± 1.2 years. This study was approved by the Institutional Review Board of the Korea National Institute for Bioethics Policy (No. P01–201603–21–005) and permission for the use of Health checkup data was granted by the NHIS (NHIS-2016-4-003).

Weight change and BMI

Weight change was calculated as the difference in body weight over 4 years, corresponding to the period between the first and third health checkup. For example, weight change over 4 years was checked during 2005 through 2009, 2006 through 2010, 2007 through 2011, and 2008 through 2012. Based on the study performed by Corrada et al. (20), we defined the weight stable group as weight change within 5%, and we categorized weight change into eight weight change groups by 5% increase or decrease as follows: weight loss ≥15%; loss ≥10% to <15%; loss ≥5% to <10%; weight change <5%; gain ≥5% to <10%; gain ≥10% to <15%; gain ≥15% to <20%; and gain ≥ 20%. The BMI was calculated as weight in kilograms divided by the square of the height in meters; BMI was also checked over 4 years, between the first and third health checkups. We defined obesity as a BMI ≥25 kg/m2 and normal as 18.5 ≤ BMI < 23 kg/m2, according to the World Health Organization recommendations for Asians (27).

General health behavior and sociodemographic variables

Smoking history was categorized as nonsmoker, ex-smoker, or current smoker. Alcohol drinking was categorized into imbibing zero, one to two, or three or more times per week. Regular exercise was defined as vigorous physical activity for at least 20 min/d, and categorized by frequency as follows: none, one to four, and five or more times per week. Income was divided by quartile (Q): Q1 (low income), Q2, Q3, and Q4.

Definition of chronic diseases

DM was defined as fasting plasma glucose levels ≥126 mg/dL (from the NHIS health checkup) or at least one claim per year for an antidiabetic medication prescription under the International Classification of Diseases, 10th Revision (ICD-10) codes E11–14. HTN was defined by blood pressure ≥140/90 mm Hg or at least one claim per year for an antihypertensive medication prescription under ICD-10 codes I10–I15. Dyslipidemia was defined by total cholesterol level ≥240 mg/dL or at least one claim per year for an antihyperlipidemic medication prescription under ICD-10 code E78. Cardiovascular disease was identified when the subject gave an affirmative answer to the following question: “Do you have a history of stroke or acute myocardial infarction?” Cancer was defined by the presence of at least one claim per rare intractable disease meeting the specific diagnostic criteria for cancer, and diagnosed and certified by a physician.

All-cause mortality

The primary outcome was all-cause mortality between 1 January 2009 and 31 December 2015 in each participant, and the number of person-years of follow-up was counted. All-cause mortality was assessed as death during 5.2 years (±SD, 1.2 years) after the last recorded weight.

Statistical analysis

The general characteristics of subjects are expressed as mean ± SD for continuous variables and percentage (SD) for categorical variables, according to the eight weight-change categories. The hazard ratios (HRs) and 95% confidence intervals (CIs) for all-cause mortality according to the eight weight-change categories were analyzed by multivariable Cox proportional hazard models, using the stable weight change (within ±5%) category as a reference group, after adjusting for age and sex in model 1; and age, sex, smoking, alcohol drinking, regular exercise, DM, HTN, dyslipidemia, cancer, and income (Q1) in model 2.The HRs and 95% CIs for mortality according to eight weight change categories were also obtained by constructing a multivariable Cox model for the following different subgroups: five BMI categories, age, sex, smoking status, drinking status, exercise, cancer, and the presence of at least one of the following cardiometabolic diseases: HTN, DM, and dyslipidemia. We also identified four categories of BMI changes over 4 years, as follows: group 1: normal to normal; group 2: obese to obese; group 3: normal to obese; and group 4: obese to normal. The HRs and 95% CIs for body mass change categories were obtained by a multivariable Cox proportional hazard models. Linear trend analysis was performed to determine the P for trend. All statistical analyses were performed using SAS version 9.3 (SAS Institute, Cary, NC), and P < 0.05 for two-tailed t tests was considered statistically significant.

Results

Baseline characteristics

Table 1 shows baseline characteristics of participants. We identified 212,704 deaths attributed to all-cause mortality. Compared with the weight loss groups, weight gain groups had greater proportions of men, smoking, and alcohol drinking. However, the following decreased as weight increased: age and the proportion of participants ≥60 years of age, low income (Q1), and chronic diseases (i.e., DM, HTN, and cancer). A higher proportion of overweight participants (i.e., 23 ≤ BMI<25 kg/m2) was found in the weight stable groups. The highest death rate (8.67%) was found in the group with weight loss ≥15%, and the lowest death rate (1.3%) was observed in the group with weight gain between 5% and 10%. Among the three weight change groups (i.e., weight loss, weight stable, and weight gain), the weight loss group had a higher proportion of patients with DM, HTN, dyslipidemia, and cancer (Supplemental Table 1). Because the prevalence of chronic diseases such as DM, HTN, and dyslipidemia decreased as weight increased, as shown in Table 1, we analyzed the odds ratio for the prevalence of chronic diseases after adjusting for age in the eight groups according to weight change (Supplemental Table 2). In contrast to the prevalence of DM and cancer, the prevalence of HTN and dyslipidemia increased as weight increased. When we calculated the prevalence of groups having more than one of the chronic diseases (among DM, HTN, and dyslipidemia), as weight increased, the risk of having more than one chronic disease increased.

Table 1.
Open in new tab

General Characteristics of Subjects

Weight Change, %
≥ −15<−15 to ≥−10<−10 to ≥−5<−5 to <5≥5 to <10≥10 to <15≥15 to <20≥20
No.86,709269,2711,270,0317,614,3861,623,429459,09412704674,797
Male sex, %34,797 (40·1)118,343 (44.0)632,444 (49.8)4,258,515 (55.9)913,860 (56.3)262,930 (57.3)71,007 (55.9)35,314 (47.21)
Age, y55.6 ± 17.154.4 ± 15.653.6 ± 14.150.8 ± 13.146.5 ± 13.343.1 ± 13.741.1 ± 14.041.7 ± 15.2
Age ≥60 y, %39,008 (45·0)108,976 (40.5)459,456 (36.2)2,049,544 (26.9)303,514 (18.7)67,979 (14.8)17,225 (13.6)12,306 (16.45)
Height, cm159.8 ± 9.6160·8 ± 9.4161.9 ± 9.2163.6 ± 9.1164.6 ± 9.3165.4 ± 9.5165.7 ± 9.5164.3 ± 9.3
Weight, kg53.9 ± 10.457.2 ± 10.560.0 ± 10.564.0 ± 11.166.7 ± 11.868.5 ± 12.470.5 ± 12.872.7 ± 13.3
Mean BMI, kg/m221.0 ± 3.022.1 ± 2.922.8 ± 2.923.8 ± 3.024.5 ± 3.224.9 ± 3.325.6 ± 3.426.8 ± 3.9
BMI, kg/m2
 <18.516,711 (19·3)27,021 (10.0)75,712 (6.0)232,064 (3.1)21,546 (1.3)3762 (0.8)604 (0.5)183 (0.24)
 ≥18.5–<2349,902 (57·6)149,168 (55.4)620,424 (48.9)2,846,777 (37.4)521,763 (32.1)132,389 (28.8)28,809 (22.7)10,717 (14.33)
 ≥23–<2512,094 (14·0)53,371 (19.8)309,699 (24.4)2,002,102 (26.3)419,031 (25.8)115,924 (25.3)31,502 (24.8)15,597 (20.85)
 ≥25–<307488 (8·6)36,799 (13.7)245,378 (19.3)2,312,095 (30.4)575,857 (35.5)172,853 (37.7)53,032 (41.7)33,834 (45.23)
 ≥30514 (0.6)2912 (1.1)18,818 (1.5)221,348 (2.9)85,232 (5.3)34,166 (7.4)13,099 (10.3)14,466 (19.34)
Smoker, %15,208 (17.5)54,053 (20.1)276,354 (21.8)1,747,419 (23.0)402,143 (24.8)122,487 (26.7)34,580 (27.2)17,314 (23.15)
Alcohol drinker, %4099 (4.7)15,182 (5.6)81,599 (6.4)565,655 (7.4)127,110 (7.8)37,090 (8.1)10,403 (8.2)5340 (7.14)
Regularly exercises, %16,849 (19.4)56,929 (21.1)277,096 (21.8)1,548,686 (20.3)275,833 (17.0)69,027 (15.0)17,793 (14.0)9906 (13.24)
Income, Q1, %16,908 (19.5)52,897 (19.6)232,778 (18.3)1,211,958 (15.9)248,853 (15.3)69,413 (15.1)19,298 (15.2)11,919 (15.94)
DM, %15,335 (17.7)44,798 (16.6)185,615 (14.6)733,027 (9.6)107,690 (6.6)27,905 (6.1)7795 (6.1)5510 (7.37)
HTN, %31,164 (35.9)90,899 (33.8)410,114 (32.3)2,226,627 (29.2)403,444 (24.9)100,258 (21.8)26,379 (20.8)17,145 (22.92)
Dyslipidemia, %17,803 (20.5)56,833 (21.1)279,449 (22.0)1,652,296 (21.7)328,947 (20.3)87,664 (19.1)25,026 (19.7)17,721 (23.69)
Cancer, %7009 (8.1)13,304 (4.9)37,351 (2.9)144,896 (1.9)25,534 (1.6)6487 (1.4)1681 (1.3)1043 (1.39)
Died, %7521 (8.7)14,035 (5.2)38,159 (3.0)122,236 (1.6)21,045 (1.3)6317 (1.4)1885 (1.5)1506 (2.01)
Weight Change, %
≥ −15<−15 to ≥−10<−10 to ≥−5<−5 to <5≥5 to <10≥10 to <15≥15 to <20≥20
No.86,709269,2711,270,0317,614,3861,623,429459,09412704674,797
Male sex, %34,797 (40·1)118,343 (44.0)632,444 (49.8)4,258,515 (55.9)913,860 (56.3)262,930 (57.3)71,007 (55.9)35,314 (47.21)
Age, y55.6 ± 17.154.4 ± 15.653.6 ± 14.150.8 ± 13.146.5 ± 13.343.1 ± 13.741.1 ± 14.041.7 ± 15.2
Age ≥60 y, %39,008 (45·0)108,976 (40.5)459,456 (36.2)2,049,544 (26.9)303,514 (18.7)67,979 (14.8)17,225 (13.6)12,306 (16.45)
Height, cm159.8 ± 9.6160·8 ± 9.4161.9 ± 9.2163.6 ± 9.1164.6 ± 9.3165.4 ± 9.5165.7 ± 9.5164.3 ± 9.3
Weight, kg53.9 ± 10.457.2 ± 10.560.0 ± 10.564.0 ± 11.166.7 ± 11.868.5 ± 12.470.5 ± 12.872.7 ± 13.3
Mean BMI, kg/m221.0 ± 3.022.1 ± 2.922.8 ± 2.923.8 ± 3.024.5 ± 3.224.9 ± 3.325.6 ± 3.426.8 ± 3.9
BMI, kg/m2
 <18.516,711 (19·3)27,021 (10.0)75,712 (6.0)232,064 (3.1)21,546 (1.3)3762 (0.8)604 (0.5)183 (0.24)
 ≥18.5–<2349,902 (57·6)149,168 (55.4)620,424 (48.9)2,846,777 (37.4)521,763 (32.1)132,389 (28.8)28,809 (22.7)10,717 (14.33)
 ≥23–<2512,094 (14·0)53,371 (19.8)309,699 (24.4)2,002,102 (26.3)419,031 (25.8)115,924 (25.3)31,502 (24.8)15,597 (20.85)
 ≥25–<307488 (8·6)36,799 (13.7)245,378 (19.3)2,312,095 (30.4)575,857 (35.5)172,853 (37.7)53,032 (41.7)33,834 (45.23)
 ≥30514 (0.6)2912 (1.1)18,818 (1.5)221,348 (2.9)85,232 (5.3)34,166 (7.4)13,099 (10.3)14,466 (19.34)
Smoker, %15,208 (17.5)54,053 (20.1)276,354 (21.8)1,747,419 (23.0)402,143 (24.8)122,487 (26.7)34,580 (27.2)17,314 (23.15)
Alcohol drinker, %4099 (4.7)15,182 (5.6)81,599 (6.4)565,655 (7.4)127,110 (7.8)37,090 (8.1)10,403 (8.2)5340 (7.14)
Regularly exercises, %16,849 (19.4)56,929 (21.1)277,096 (21.8)1,548,686 (20.3)275,833 (17.0)69,027 (15.0)17,793 (14.0)9906 (13.24)
Income, Q1, %16,908 (19.5)52,897 (19.6)232,778 (18.3)1,211,958 (15.9)248,853 (15.3)69,413 (15.1)19,298 (15.2)11,919 (15.94)
DM, %15,335 (17.7)44,798 (16.6)185,615 (14.6)733,027 (9.6)107,690 (6.6)27,905 (6.1)7795 (6.1)5510 (7.37)
HTN, %31,164 (35.9)90,899 (33.8)410,114 (32.3)2,226,627 (29.2)403,444 (24.9)100,258 (21.8)26,379 (20.8)17,145 (22.92)
Dyslipidemia, %17,803 (20.5)56,833 (21.1)279,449 (22.0)1,652,296 (21.7)328,947 (20.3)87,664 (19.1)25,026 (19.7)17,721 (23.69)
Cancer, %7009 (8.1)13,304 (4.9)37,351 (2.9)144,896 (1.9)25,534 (1.6)6487 (1.4)1681 (1.3)1043 (1.39)
Died, %7521 (8.7)14,035 (5.2)38,159 (3.0)122,236 (1.6)21,045 (1.3)6317 (1.4)1885 (1.5)1506 (2.01)
Table 1.
Open in new tab

General Characteristics of Subjects

Weight Change, %
≥ −15<−15 to ≥−10<−10 to ≥−5<−5 to <5≥5 to <10≥10 to <15≥15 to <20≥20
No.86,709269,2711,270,0317,614,3861,623,429459,09412704674,797
Male sex, %34,797 (40·1)118,343 (44.0)632,444 (49.8)4,258,515 (55.9)913,860 (56.3)262,930 (57.3)71,007 (55.9)35,314 (47.21)
Age, y55.6 ± 17.154.4 ± 15.653.6 ± 14.150.8 ± 13.146.5 ± 13.343.1 ± 13.741.1 ± 14.041.7 ± 15.2
Age ≥60 y, %39,008 (45·0)108,976 (40.5)459,456 (36.2)2,049,544 (26.9)303,514 (18.7)67,979 (14.8)17,225 (13.6)12,306 (16.45)
Height, cm159.8 ± 9.6160·8 ± 9.4161.9 ± 9.2163.6 ± 9.1164.6 ± 9.3165.4 ± 9.5165.7 ± 9.5164.3 ± 9.3
Weight, kg53.9 ± 10.457.2 ± 10.560.0 ± 10.564.0 ± 11.166.7 ± 11.868.5 ± 12.470.5 ± 12.872.7 ± 13.3
Mean BMI, kg/m221.0 ± 3.022.1 ± 2.922.8 ± 2.923.8 ± 3.024.5 ± 3.224.9 ± 3.325.6 ± 3.426.8 ± 3.9
BMI, kg/m2
 <18.516,711 (19·3)27,021 (10.0)75,712 (6.0)232,064 (3.1)21,546 (1.3)3762 (0.8)604 (0.5)183 (0.24)
 ≥18.5–<2349,902 (57·6)149,168 (55.4)620,424 (48.9)2,846,777 (37.4)521,763 (32.1)132,389 (28.8)28,809 (22.7)10,717 (14.33)
 ≥23–<2512,094 (14·0)53,371 (19.8)309,699 (24.4)2,002,102 (26.3)419,031 (25.8)115,924 (25.3)31,502 (24.8)15,597 (20.85)
 ≥25–<307488 (8·6)36,799 (13.7)245,378 (19.3)2,312,095 (30.4)575,857 (35.5)172,853 (37.7)53,032 (41.7)33,834 (45.23)
 ≥30514 (0.6)2912 (1.1)18,818 (1.5)221,348 (2.9)85,232 (5.3)34,166 (7.4)13,099 (10.3)14,466 (19.34)
Smoker, %15,208 (17.5)54,053 (20.1)276,354 (21.8)1,747,419 (23.0)402,143 (24.8)122,487 (26.7)34,580 (27.2)17,314 (23.15)
Alcohol drinker, %4099 (4.7)15,182 (5.6)81,599 (6.4)565,655 (7.4)127,110 (7.8)37,090 (8.1)10,403 (8.2)5340 (7.14)
Regularly exercises, %16,849 (19.4)56,929 (21.1)277,096 (21.8)1,548,686 (20.3)275,833 (17.0)69,027 (15.0)17,793 (14.0)9906 (13.24)
Income, Q1, %16,908 (19.5)52,897 (19.6)232,778 (18.3)1,211,958 (15.9)248,853 (15.3)69,413 (15.1)19,298 (15.2)11,919 (15.94)
DM, %15,335 (17.7)44,798 (16.6)185,615 (14.6)733,027 (9.6)107,690 (6.6)27,905 (6.1)7795 (6.1)5510 (7.37)
HTN, %31,164 (35.9)90,899 (33.8)410,114 (32.3)2,226,627 (29.2)403,444 (24.9)100,258 (21.8)26,379 (20.8)17,145 (22.92)
Dyslipidemia, %17,803 (20.5)56,833 (21.1)279,449 (22.0)1,652,296 (21.7)328,947 (20.3)87,664 (19.1)25,026 (19.7)17,721 (23.69)
Cancer, %7009 (8.1)13,304 (4.9)37,351 (2.9)144,896 (1.9)25,534 (1.6)6487 (1.4)1681 (1.3)1043 (1.39)
Died, %7521 (8.7)14,035 (5.2)38,159 (3.0)122,236 (1.6)21,045 (1.3)6317 (1.4)1885 (1.5)1506 (2.01)
Weight Change, %
≥ −15<−15 to ≥−10<−10 to ≥−5<−5 to <5≥5 to <10≥10 to <15≥15 to <20≥20
No.86,709269,2711,270,0317,614,3861,623,429459,09412704674,797
Male sex, %34,797 (40·1)118,343 (44.0)632,444 (49.8)4,258,515 (55.9)913,860 (56.3)262,930 (57.3)71,007 (55.9)35,314 (47.21)
Age, y55.6 ± 17.154.4 ± 15.653.6 ± 14.150.8 ± 13.146.5 ± 13.343.1 ± 13.741.1 ± 14.041.7 ± 15.2
Age ≥60 y, %39,008 (45·0)108,976 (40.5)459,456 (36.2)2,049,544 (26.9)303,514 (18.7)67,979 (14.8)17,225 (13.6)12,306 (16.45)
Height, cm159.8 ± 9.6160·8 ± 9.4161.9 ± 9.2163.6 ± 9.1164.6 ± 9.3165.4 ± 9.5165.7 ± 9.5164.3 ± 9.3
Weight, kg53.9 ± 10.457.2 ± 10.560.0 ± 10.564.0 ± 11.166.7 ± 11.868.5 ± 12.470.5 ± 12.872.7 ± 13.3
Mean BMI, kg/m221.0 ± 3.022.1 ± 2.922.8 ± 2.923.8 ± 3.024.5 ± 3.224.9 ± 3.325.6 ± 3.426.8 ± 3.9
BMI, kg/m2
 <18.516,711 (19·3)27,021 (10.0)75,712 (6.0)232,064 (3.1)21,546 (1.3)3762 (0.8)604 (0.5)183 (0.24)
 ≥18.5–<2349,902 (57·6)149,168 (55.4)620,424 (48.9)2,846,777 (37.4)521,763 (32.1)132,389 (28.8)28,809 (22.7)10,717 (14.33)
 ≥23–<2512,094 (14·0)53,371 (19.8)309,699 (24.4)2,002,102 (26.3)419,031 (25.8)115,924 (25.3)31,502 (24.8)15,597 (20.85)
 ≥25–<307488 (8·6)36,799 (13.7)245,378 (19.3)2,312,095 (30.4)575,857 (35.5)172,853 (37.7)53,032 (41.7)33,834 (45.23)
 ≥30514 (0.6)2912 (1.1)18,818 (1.5)221,348 (2.9)85,232 (5.3)34,166 (7.4)13,099 (10.3)14,466 (19.34)
Smoker, %15,208 (17.5)54,053 (20.1)276,354 (21.8)1,747,419 (23.0)402,143 (24.8)122,487 (26.7)34,580 (27.2)17,314 (23.15)
Alcohol drinker, %4099 (4.7)15,182 (5.6)81,599 (6.4)565,655 (7.4)127,110 (7.8)37,090 (8.1)10,403 (8.2)5340 (7.14)
Regularly exercises, %16,849 (19.4)56,929 (21.1)277,096 (21.8)1,548,686 (20.3)275,833 (17.0)69,027 (15.0)17,793 (14.0)9906 (13.24)
Income, Q1, %16,908 (19.5)52,897 (19.6)232,778 (18.3)1,211,958 (15.9)248,853 (15.3)69,413 (15.1)19,298 (15.2)11,919 (15.94)
DM, %15,335 (17.7)44,798 (16.6)185,615 (14.6)733,027 (9.6)107,690 (6.6)27,905 (6.1)7795 (6.1)5510 (7.37)
HTN, %31,164 (35.9)90,899 (33.8)410,114 (32.3)2,226,627 (29.2)403,444 (24.9)100,258 (21.8)26,379 (20.8)17,145 (22.92)
Dyslipidemia, %17,803 (20.5)56,833 (21.1)279,449 (22.0)1,652,296 (21.7)328,947 (20.3)87,664 (19.1)25,026 (19.7)17,721 (23.69)
Cancer, %7009 (8.1)13,304 (4.9)37,351 (2.9)144,896 (1.9)25,534 (1.6)6487 (1.4)1681 (1.3)1043 (1.39)
Died, %7521 (8.7)14,035 (5.2)38,159 (3.0)122,236 (1.6)21,045 (1.3)6317 (1.4)1885 (1.5)1506 (2.01)

All-cause mortality by weight change

The HR of all-cause mortality by weight change is shown in Table 2. All-cause mortality showed a reverse J-shaped mortality curve according to weight change. Compared with the weight stable group, after adjusting for all covariates, weight loss was associated with a greater mortality than weight gain per the same changes in weight. The group with weight loss ≥15% had the highest HR for all-cause mortality (HR, 2.598; 95% CI, 2.537 to 2.659) and the HRs for weight loss between 10% and 15% and weight gain ≥20% were 1.923 and 1.784, respectively.

Table 2.
Open in new tab

Weight Change and All-Cause Mortality

Weight Change, %DiedDurationIncidenceHR (95% CI)
Model 1aModel 2b
≥ −157521426,389.8317.63883.089 (3.017–3.163)2.598 (2.537–2.659)
<−15 to ≥ −1014,0351,354,671.4210.36042.160 (2.122–2.198)1.923 (1.890–1.957)
<−10 to ≥ −538,1596,527,906.65.84551.425 (1.408–1.441)1.356 (1.365–1.371)
<−5 to <5122,23639,797,312.593.07151 (Ref)1 (Ref)
≥5–<1021,0458,431,358.432.4961.100 (1.084–1.116)1.099 (1.083–1.115)
≥10–<1563172,345,570.422.69321.407 (1.372–1.443)1.388 (1.353–1.423)
≥15–<201885641,186.722.93991.640 (1.567–1.716)1.602 (1.531–1.677)
≥201506375,348.24.01231.829 (1.739–1.925)1.784 (1.695–1.877)
Weight Change, %DiedDurationIncidenceHR (95% CI)
Model 1aModel 2b
≥ −157521426,389.8317.63883.089 (3.017–3.163)2.598 (2.537–2.659)
<−15 to ≥ −1014,0351,354,671.4210.36042.160 (2.122–2.198)1.923 (1.890–1.957)
<−10 to ≥ −538,1596,527,906.65.84551.425 (1.408–1.441)1.356 (1.365–1.371)
<−5 to <5122,23639,797,312.593.07151 (Ref)1 (Ref)
≥5–<1021,0458,431,358.432.4961.100 (1.084–1.116)1.099 (1.083–1.115)
≥10–<1563172,345,570.422.69321.407 (1.372–1.443)1.388 (1.353–1.423)
≥15–<201885641,186.722.93991.640 (1.567–1.716)1.602 (1.531–1.677)
≥201506375,348.24.01231.829 (1.739–1.925)1.784 (1.695–1.877)
a

Model 1 was adjusted for age and sex.

b

Model 2 was adjusted for age, sex, smoking, drinking, exercise, DM, HTN, dyslipidemia, cancer, and income.

Table 2.
Open in new tab

Weight Change and All-Cause Mortality

Weight Change, %DiedDurationIncidenceHR (95% CI)
Model 1aModel 2b
≥ −157521426,389.8317.63883.089 (3.017–3.163)2.598 (2.537–2.659)
<−15 to ≥ −1014,0351,354,671.4210.36042.160 (2.122–2.198)1.923 (1.890–1.957)
<−10 to ≥ −538,1596,527,906.65.84551.425 (1.408–1.441)1.356 (1.365–1.371)
<−5 to <5122,23639,797,312.593.07151 (Ref)1 (Ref)
≥5–<1021,0458,431,358.432.4961.100 (1.084–1.116)1.099 (1.083–1.115)
≥10–<1563172,345,570.422.69321.407 (1.372–1.443)1.388 (1.353–1.423)
≥15–<201885641,186.722.93991.640 (1.567–1.716)1.602 (1.531–1.677)
≥201506375,348.24.01231.829 (1.739–1.925)1.784 (1.695–1.877)
Weight Change, %DiedDurationIncidenceHR (95% CI)
Model 1aModel 2b
≥ −157521426,389.8317.63883.089 (3.017–3.163)2.598 (2.537–2.659)
<−15 to ≥ −1014,0351,354,671.4210.36042.160 (2.122–2.198)1.923 (1.890–1.957)
<−10 to ≥ −538,1596,527,906.65.84551.425 (1.408–1.441)1.356 (1.365–1.371)
<−5 to <5122,23639,797,312.593.07151 (Ref)1 (Ref)
≥5–<1021,0458,431,358.432.4961.100 (1.084–1.116)1.099 (1.083–1.115)
≥10–<1563172,345,570.422.69321.407 (1.372–1.443)1.388 (1.353–1.423)
≥15–<201885641,186.722.93991.640 (1.567–1.716)1.602 (1.531–1.677)
≥201506375,348.24.01231.829 (1.739–1.925)1.784 (1.695–1.877)
a

Model 1 was adjusted for age and sex.

b

Model 2 was adjusted for age, sex, smoking, drinking, exercise, DM, HTN, dyslipidemia, cancer, and income.

All-cause mortality by weight change according to subgroups

We categorized participants by sex, age, smoking status, presence of more than one concomitant chronic disease (i.e., DM, HTN, and dyslipidemia), and cancer. We also analyzed the HR for mortality in the eight weight change groups, setting the stable weight group as the reference group after adjusting for all covariates in Table 3. The HR for mortality was higher among men than among women, except for the groups with extreme weight change (i.e., weight loss ≥15% and weight gain ≥20%). In the group younger than 60 years old, there was a higher HR for mortality among subjects with weight loss and a lower HR for mortality among subjects with weight gain, compared with the group of subjects ≥60 years old. Smoking and cancer groups had lower mortality HRs than those of the nonsmoking and noncancer groups, respectively, in all weight change criteria. The group of subjects with one or more chronic diseases had lower mortality HRs in weight loss, but higher mortality HRs in weight gain compared with the group with no chronic disease.

Table 3.
Open in new tab

Weight Change and All-Cause Mortality by Subgroup

SubgroupWeight Change, %No. of SubjectsDiedHR (95% CI)
Model 1aModel 2b
Men≥−1534,79741142.985 (2.892–3·080)2.472 (2.395–2.551)
<−15 to ≥−10118,34389232.247 (2.198–2.297)1.984 (1.940–2.028)
<−10 to ≥−5632,44425,6071.436 (1.416–1.456)1.364 (1.345–1.383)
<−5 to <54,258,51586,9511 (Ref)1 (Ref)
≥5–<10913,86014,8321.109 (1.090–1.129)1.108 (1.089–1.127)
≥10–<15262,93043691.409 (1.367–1.453)1.391 (1.349–1.434)
≥15–<2071,00712541.624 (1.536–1.717)1.591 (1.504–1.682)
≥2035,3149341.767 (1.657–1.885)1.725 (1.618–1.840)
Women≥−1551,91234073.155 (3.045–3.270)2.725 (2.630–2.824)
<−15 to ≥−10150,92851121.997 (1.939–2.057)1.810 (1.757–1.864)
<−10 to ≥−5637,58712,5521.396 (1.367–1.424)1.336 (1.309–1.363)
<−5 to <53,355,87135,2851 (Ref)1 (Ref)
≥5–<10709,56962131.079 (1.05–1.109)1.081 (1.052–1.111)
≥10–<15196,16419481.398 (1.336–1.463)1.382 (1.321–1.447)
≥15–<2056,0396311.657 (1.531–1.792)1.616 (1.494–1.749)
≥2039,4835721.918 (1.766–2.083)1.880 (1.731–2.042)
Age <60 y≥−1547,7017523.909 (3.636–4.202)2.582 (2.400–2.777)
<−15 to ≥−10160,29519022.625 (2.506–2.750)2.056 (1.962–2.154)
<−10 to ≥−5810,57564861.542 (1.501–1.584)1.391 (1.354–1.430)
<−5 to <55,564,84228,9021 (Ref)1 (Ref)
≥5–<101,319,91556341.065 (1.035–1.096)1.068 (1.038–1.099)
≥10–<15391,11516051.317 (1.252–1.385)1.291 (1.227–1.358)
≥15–<20109,8214511.590 (1.448–1.745)1.518 (1.382–1.666)
≥2062,4913082.021 (1.806–2.261)1.876 (1.676–2.099)
Age ≥60 y≥−1539,00867692.915 (2.843–2.989)2.508 (2.446–2.571)
<−15 to ≥−10108,97612,1332.055 (2.016–2.095)1.856 (1.821–1.892)
<−10 to ≥−5459,45631,6731.390 (1.373–1.408)1.332 (1.315–1.349)
<−5 to <52,049,54493,3341 (Ref)1 (Ref)
≥5–<10303,51415,4111.098 (1.080–1.117)1.094 (1.076–1.113)
≥10–<1567,97947121.398 (1.358–1.439)1.373 (1.333–1.413)
≥15–<2017,22514341.593 (1.512–1.678)1.552 (1.473–1.635)
≥2012,30611981.722 (1.626–1.823)1.680 (1.587–1.778)
Nonsmoking≥−1571,50163133.180 (3.099–3.263)2.692 (2.624–2.763)
<−15 to ≥−10215,21810,8232.149 (2.106–2.192)1.935 (1.897–1.975)
<−10 to ≥−5993,67728,6021.419 (1.400–1.438)1.367 (1.349–1.386)
<−5 to <55,866,96790,7131 (Ref)1 (Ref)
≥5–<101,221,28615,9631.126 (1.107–1.145)1.123 (1.104–1.142)
≥10–<15336,60748151.433 (1.392–1.475)1.405 (1.365–1.447)
≥15–<2092,46614511.669 (1.585–1.758)1.627 (1.545–1.714)
≥2057,48312131.890 (1.786–2.000)1.847 (1.745–1.955)
Smoking≥−1515,20812082.381 (2.248–2.523)2.143 (2.022–2.270)
<−15 to ≥−1054,05332121.971 (1.900–2.044)1.866 (1.799–1.935)
<−10 to ≥−5276,35495571.353 (1.322–1.385)1.317 (1.287–1.348)
<−5 to <51,747,41931,5231 (Ref)1 (Ref)
≥5–<10402,14350821.027 (0.997–1.058)1.028 (0·998–1.059)
≥10–<15122,48715021.321 (1.254–1.391)1.320 (1.253–1.390)
≥15–<2034,5804341.508 (1.372–1.658)1.503 (1.367–1.652)
≥2017,3142931.552 (1.384–1.742)1.535 (1.368–1.722)
No chronic disease≥−1544,12221493.301 (3.159–3.449)2.509 (2.401–2.622)
<−15 to ≥−10141,03140772.178 (2.108–2.250)1.877 (1.817–1.939)
<−10 to ≥−5678,74511,4101.410 (1.380–1.439)1.339 (1.311–1.368)
<−5 to <54,356,10838,7171 (Ref)1 (Ref)
≥5–<101,011,03867421.090 (1.063–1.119)1.077 (1.049–1.105)
≥10–<15299,79419761.378 (1.317–1.442)1.337 (1.278–1.399)
≥15–<2082,8305581.602 (1.473–1.741)1.555 (1.430–1.691)
≥2044,6164091.776 (1.611–1.957)1.738 (1.577–1.916)
Chronic disease≥−1542,58753723.038 (2.955–3.124)2.608 (2.536–2.682)
<−15 to ≥−10128,24099582.169 (2.124–2.215)1.934 (1.894–1.975)
<−10 to ≥−5591,28626,7491.438 (1.418–1.458)1.361 (1.342–1.380)
<−5 to <53,258,27883,5191 (Ref)1 (Ref)
≥5–<10612,39114,3031.103 (1.084–1.123)1.110 (1.090–1.130)
≥10–<15159,30043411.419 (1.377–1.463)1.410 (1.368–1.454)
≥15–<2044,21613271.653 (1.565–1.745)1.621 (1.535–1.711)
≥2030,18110971.848 (1.741–1.961)1.800 (1.696–1.911)
No cancerc≥−1579,70060372.983 (2.906–3.062)2.793 (2.721–2.868)
<−15 to ≥−10255,96711,7252.104 (2.064–2.145)1.989 (1.951–2.028)
<−10 to ≥−51,232,68032,9181.398 (1.381–1.416)1.352 (1.336–1.369)
<−5 to <57,469,490108,2021 (Ref)1 (Ref)
≥5–<101,597,89518,8101.113 (1.096–1.130)1.117 (1.100–1.134)
≥10–<15452,60756101.418 (1.381–1.457)1.412 (1.374–1.450)
≥15–<20125,36516751.651 (1.574–1.733)1.625 (1.548–1.705)
≥2073,75413751.883 (1.786–1.986)1.836 (1.741–1.936)
Cancerc≥−15700914841.938 (1.837–2.045)1.901 (1.802–2.006)
<−15 to ≥−1013,30423101.602 (1.533–1.674)1.566 (1.498–1.636)
<−10 to ≥−537,35152411.353 (1.310–1.396)1.320 (1.278–1.362)
<−5 to <5144,89614,0341 (Ref)1 (Ref)
≥5–<1025,53422350.986 (0.943–1.031)0.989 (0.946–1.034)
≥10–<1564877071.262 (1.170–1.361)1.257 (1.166–1.356)
≥15–<2016812101.444 (1.260–1.654)1.426 (1.244–1.634)
≥2010431311.377 (1.159–1.635)1.362 (1.147–1.618)
SubgroupWeight Change, %No. of SubjectsDiedHR (95% CI)
Model 1aModel 2b
Men≥−1534,79741142.985 (2.892–3·080)2.472 (2.395–2.551)
<−15 to ≥−10118,34389232.247 (2.198–2.297)1.984 (1.940–2.028)
<−10 to ≥−5632,44425,6071.436 (1.416–1.456)1.364 (1.345–1.383)
<−5 to <54,258,51586,9511 (Ref)1 (Ref)
≥5–<10913,86014,8321.109 (1.090–1.129)1.108 (1.089–1.127)
≥10–<15262,93043691.409 (1.367–1.453)1.391 (1.349–1.434)
≥15–<2071,00712541.624 (1.536–1.717)1.591 (1.504–1.682)
≥2035,3149341.767 (1.657–1.885)1.725 (1.618–1.840)
Women≥−1551,91234073.155 (3.045–3.270)2.725 (2.630–2.824)
<−15 to ≥−10150,92851121.997 (1.939–2.057)1.810 (1.757–1.864)
<−10 to ≥−5637,58712,5521.396 (1.367–1.424)1.336 (1.309–1.363)
<−5 to <53,355,87135,2851 (Ref)1 (Ref)
≥5–<10709,56962131.079 (1.05–1.109)1.081 (1.052–1.111)
≥10–<15196,16419481.398 (1.336–1.463)1.382 (1.321–1.447)
≥15–<2056,0396311.657 (1.531–1.792)1.616 (1.494–1.749)
≥2039,4835721.918 (1.766–2.083)1.880 (1.731–2.042)
Age <60 y≥−1547,7017523.909 (3.636–4.202)2.582 (2.400–2.777)
<−15 to ≥−10160,29519022.625 (2.506–2.750)2.056 (1.962–2.154)
<−10 to ≥−5810,57564861.542 (1.501–1.584)1.391 (1.354–1.430)
<−5 to <55,564,84228,9021 (Ref)1 (Ref)
≥5–<101,319,91556341.065 (1.035–1.096)1.068 (1.038–1.099)
≥10–<15391,11516051.317 (1.252–1.385)1.291 (1.227–1.358)
≥15–<20109,8214511.590 (1.448–1.745)1.518 (1.382–1.666)
≥2062,4913082.021 (1.806–2.261)1.876 (1.676–2.099)
Age ≥60 y≥−1539,00867692.915 (2.843–2.989)2.508 (2.446–2.571)
<−15 to ≥−10108,97612,1332.055 (2.016–2.095)1.856 (1.821–1.892)
<−10 to ≥−5459,45631,6731.390 (1.373–1.408)1.332 (1.315–1.349)
<−5 to <52,049,54493,3341 (Ref)1 (Ref)
≥5–<10303,51415,4111.098 (1.080–1.117)1.094 (1.076–1.113)
≥10–<1567,97947121.398 (1.358–1.439)1.373 (1.333–1.413)
≥15–<2017,22514341.593 (1.512–1.678)1.552 (1.473–1.635)
≥2012,30611981.722 (1.626–1.823)1.680 (1.587–1.778)
Nonsmoking≥−1571,50163133.180 (3.099–3.263)2.692 (2.624–2.763)
<−15 to ≥−10215,21810,8232.149 (2.106–2.192)1.935 (1.897–1.975)
<−10 to ≥−5993,67728,6021.419 (1.400–1.438)1.367 (1.349–1.386)
<−5 to <55,866,96790,7131 (Ref)1 (Ref)
≥5–<101,221,28615,9631.126 (1.107–1.145)1.123 (1.104–1.142)
≥10–<15336,60748151.433 (1.392–1.475)1.405 (1.365–1.447)
≥15–<2092,46614511.669 (1.585–1.758)1.627 (1.545–1.714)
≥2057,48312131.890 (1.786–2.000)1.847 (1.745–1.955)
Smoking≥−1515,20812082.381 (2.248–2.523)2.143 (2.022–2.270)
<−15 to ≥−1054,05332121.971 (1.900–2.044)1.866 (1.799–1.935)
<−10 to ≥−5276,35495571.353 (1.322–1.385)1.317 (1.287–1.348)
<−5 to <51,747,41931,5231 (Ref)1 (Ref)
≥5–<10402,14350821.027 (0.997–1.058)1.028 (0·998–1.059)
≥10–<15122,48715021.321 (1.254–1.391)1.320 (1.253–1.390)
≥15–<2034,5804341.508 (1.372–1.658)1.503 (1.367–1.652)
≥2017,3142931.552 (1.384–1.742)1.535 (1.368–1.722)
No chronic disease≥−1544,12221493.301 (3.159–3.449)2.509 (2.401–2.622)
<−15 to ≥−10141,03140772.178 (2.108–2.250)1.877 (1.817–1.939)
<−10 to ≥−5678,74511,4101.410 (1.380–1.439)1.339 (1.311–1.368)
<−5 to <54,356,10838,7171 (Ref)1 (Ref)
≥5–<101,011,03867421.090 (1.063–1.119)1.077 (1.049–1.105)
≥10–<15299,79419761.378 (1.317–1.442)1.337 (1.278–1.399)
≥15–<2082,8305581.602 (1.473–1.741)1.555 (1.430–1.691)
≥2044,6164091.776 (1.611–1.957)1.738 (1.577–1.916)
Chronic disease≥−1542,58753723.038 (2.955–3.124)2.608 (2.536–2.682)
<−15 to ≥−10128,24099582.169 (2.124–2.215)1.934 (1.894–1.975)
<−10 to ≥−5591,28626,7491.438 (1.418–1.458)1.361 (1.342–1.380)
<−5 to <53,258,27883,5191 (Ref)1 (Ref)
≥5–<10612,39114,3031.103 (1.084–1.123)1.110 (1.090–1.130)
≥10–<15159,30043411.419 (1.377–1.463)1.410 (1.368–1.454)
≥15–<2044,21613271.653 (1.565–1.745)1.621 (1.535–1.711)
≥2030,18110971.848 (1.741–1.961)1.800 (1.696–1.911)
No cancerc≥−1579,70060372.983 (2.906–3.062)2.793 (2.721–2.868)
<−15 to ≥−10255,96711,7252.104 (2.064–2.145)1.989 (1.951–2.028)
<−10 to ≥−51,232,68032,9181.398 (1.381–1.416)1.352 (1.336–1.369)
<−5 to <57,469,490108,2021 (Ref)1 (Ref)
≥5–<101,597,89518,8101.113 (1.096–1.130)1.117 (1.100–1.134)
≥10–<15452,60756101.418 (1.381–1.457)1.412 (1.374–1.450)
≥15–<20125,36516751.651 (1.574–1.733)1.625 (1.548–1.705)
≥2073,75413751.883 (1.786–1.986)1.836 (1.741–1.936)
Cancerc≥−15700914841.938 (1.837–2.045)1.901 (1.802–2.006)
<−15 to ≥−1013,30423101.602 (1.533–1.674)1.566 (1.498–1.636)
<−10 to ≥−537,35152411.353 (1.310–1.396)1.320 (1.278–1.362)
<−5 to <5144,89614,0341 (Ref)1 (Ref)
≥5–<1025,53422350.986 (0.943–1.031)0.989 (0.946–1.034)
≥10–<1564877071.262 (1.170–1.361)1.257 (1.166–1.356)
≥15–<2016812101.444 (1.260–1.654)1.426 (1.244–1.634)
≥2010431311.377 (1.159–1.635)1.362 (1.147–1.618)

Abbreviation: Ref, reference.

a

Model 1 was adjusted for age and sex.

b

Model 2 was adjusted for age, sex, smoking, drinking, exercise, DM, HTN, dyslipidemia, cancer, and income.

c

Cancer was not adjusted in model 2.

Table 3.
Open in new tab

Weight Change and All-Cause Mortality by Subgroup

SubgroupWeight Change, %No. of SubjectsDiedHR (95% CI)
Model 1aModel 2b
Men≥−1534,79741142.985 (2.892–3·080)2.472 (2.395–2.551)
<−15 to ≥−10118,34389232.247 (2.198–2.297)1.984 (1.940–2.028)
<−10 to ≥−5632,44425,6071.436 (1.416–1.456)1.364 (1.345–1.383)
<−5 to <54,258,51586,9511 (Ref)1 (Ref)
≥5–<10913,86014,8321.109 (1.090–1.129)1.108 (1.089–1.127)
≥10–<15262,93043691.409 (1.367–1.453)1.391 (1.349–1.434)
≥15–<2071,00712541.624 (1.536–1.717)1.591 (1.504–1.682)
≥2035,3149341.767 (1.657–1.885)1.725 (1.618–1.840)
Women≥−1551,91234073.155 (3.045–3.270)2.725 (2.630–2.824)
<−15 to ≥−10150,92851121.997 (1.939–2.057)1.810 (1.757–1.864)
<−10 to ≥−5637,58712,5521.396 (1.367–1.424)1.336 (1.309–1.363)
<−5 to <53,355,87135,2851 (Ref)1 (Ref)
≥5–<10709,56962131.079 (1.05–1.109)1.081 (1.052–1.111)
≥10–<15196,16419481.398 (1.336–1.463)1.382 (1.321–1.447)
≥15–<2056,0396311.657 (1.531–1.792)1.616 (1.494–1.749)
≥2039,4835721.918 (1.766–2.083)1.880 (1.731–2.042)
Age <60 y≥−1547,7017523.909 (3.636–4.202)2.582 (2.400–2.777)
<−15 to ≥−10160,29519022.625 (2.506–2.750)2.056 (1.962–2.154)
<−10 to ≥−5810,57564861.542 (1.501–1.584)1.391 (1.354–1.430)
<−5 to <55,564,84228,9021 (Ref)1 (Ref)
≥5–<101,319,91556341.065 (1.035–1.096)1.068 (1.038–1.099)
≥10–<15391,11516051.317 (1.252–1.385)1.291 (1.227–1.358)
≥15–<20109,8214511.590 (1.448–1.745)1.518 (1.382–1.666)
≥2062,4913082.021 (1.806–2.261)1.876 (1.676–2.099)
Age ≥60 y≥−1539,00867692.915 (2.843–2.989)2.508 (2.446–2.571)
<−15 to ≥−10108,97612,1332.055 (2.016–2.095)1.856 (1.821–1.892)
<−10 to ≥−5459,45631,6731.390 (1.373–1.408)1.332 (1.315–1.349)
<−5 to <52,049,54493,3341 (Ref)1 (Ref)
≥5–<10303,51415,4111.098 (1.080–1.117)1.094 (1.076–1.113)
≥10–<1567,97947121.398 (1.358–1.439)1.373 (1.333–1.413)
≥15–<2017,22514341.593 (1.512–1.678)1.552 (1.473–1.635)
≥2012,30611981.722 (1.626–1.823)1.680 (1.587–1.778)
Nonsmoking≥−1571,50163133.180 (3.099–3.263)2.692 (2.624–2.763)
<−15 to ≥−10215,21810,8232.149 (2.106–2.192)1.935 (1.897–1.975)
<−10 to ≥−5993,67728,6021.419 (1.400–1.438)1.367 (1.349–1.386)
<−5 to <55,866,96790,7131 (Ref)1 (Ref)
≥5–<101,221,28615,9631.126 (1.107–1.145)1.123 (1.104–1.142)
≥10–<15336,60748151.433 (1.392–1.475)1.405 (1.365–1.447)
≥15–<2092,46614511.669 (1.585–1.758)1.627 (1.545–1.714)
≥2057,48312131.890 (1.786–2.000)1.847 (1.745–1.955)
Smoking≥−1515,20812082.381 (2.248–2.523)2.143 (2.022–2.270)
<−15 to ≥−1054,05332121.971 (1.900–2.044)1.866 (1.799–1.935)
<−10 to ≥−5276,35495571.353 (1.322–1.385)1.317 (1.287–1.348)
<−5 to <51,747,41931,5231 (Ref)1 (Ref)
≥5–<10402,14350821.027 (0.997–1.058)1.028 (0·998–1.059)
≥10–<15122,48715021.321 (1.254–1.391)1.320 (1.253–1.390)
≥15–<2034,5804341.508 (1.372–1.658)1.503 (1.367–1.652)
≥2017,3142931.552 (1.384–1.742)1.535 (1.368–1.722)
No chronic disease≥−1544,12221493.301 (3.159–3.449)2.509 (2.401–2.622)
<−15 to ≥−10141,03140772.178 (2.108–2.250)1.877 (1.817–1.939)
<−10 to ≥−5678,74511,4101.410 (1.380–1.439)1.339 (1.311–1.368)
<−5 to <54,356,10838,7171 (Ref)1 (Ref)
≥5–<101,011,03867421.090 (1.063–1.119)1.077 (1.049–1.105)
≥10–<15299,79419761.378 (1.317–1.442)1.337 (1.278–1.399)
≥15–<2082,8305581.602 (1.473–1.741)1.555 (1.430–1.691)
≥2044,6164091.776 (1.611–1.957)1.738 (1.577–1.916)
Chronic disease≥−1542,58753723.038 (2.955–3.124)2.608 (2.536–2.682)
<−15 to ≥−10128,24099582.169 (2.124–2.215)1.934 (1.894–1.975)
<−10 to ≥−5591,28626,7491.438 (1.418–1.458)1.361 (1.342–1.380)
<−5 to <53,258,27883,5191 (Ref)1 (Ref)
≥5–<10612,39114,3031.103 (1.084–1.123)1.110 (1.090–1.130)
≥10–<15159,30043411.419 (1.377–1.463)1.410 (1.368–1.454)
≥15–<2044,21613271.653 (1.565–1.745)1.621 (1.535–1.711)
≥2030,18110971.848 (1.741–1.961)1.800 (1.696–1.911)
No cancerc≥−1579,70060372.983 (2.906–3.062)2.793 (2.721–2.868)
<−15 to ≥−10255,96711,7252.104 (2.064–2.145)1.989 (1.951–2.028)
<−10 to ≥−51,232,68032,9181.398 (1.381–1.416)1.352 (1.336–1.369)
<−5 to <57,469,490108,2021 (Ref)1 (Ref)
≥5–<101,597,89518,8101.113 (1.096–1.130)1.117 (1.100–1.134)
≥10–<15452,60756101.418 (1.381–1.457)1.412 (1.374–1.450)
≥15–<20125,36516751.651 (1.574–1.733)1.625 (1.548–1.705)
≥2073,75413751.883 (1.786–1.986)1.836 (1.741–1.936)
Cancerc≥−15700914841.938 (1.837–2.045)1.901 (1.802–2.006)
<−15 to ≥−1013,30423101.602 (1.533–1.674)1.566 (1.498–1.636)
<−10 to ≥−537,35152411.353 (1.310–1.396)1.320 (1.278–1.362)
<−5 to <5144,89614,0341 (Ref)1 (Ref)
≥5–<1025,53422350.986 (0.943–1.031)0.989 (0.946–1.034)
≥10–<1564877071.262 (1.170–1.361)1.257 (1.166–1.356)
≥15–<2016812101.444 (1.260–1.654)1.426 (1.244–1.634)
≥2010431311.377 (1.159–1.635)1.362 (1.147–1.618)
SubgroupWeight Change, %No. of SubjectsDiedHR (95% CI)
Model 1aModel 2b
Men≥−1534,79741142.985 (2.892–3·080)2.472 (2.395–2.551)
<−15 to ≥−10118,34389232.247 (2.198–2.297)1.984 (1.940–2.028)
<−10 to ≥−5632,44425,6071.436 (1.416–1.456)1.364 (1.345–1.383)
<−5 to <54,258,51586,9511 (Ref)1 (Ref)
≥5–<10913,86014,8321.109 (1.090–1.129)1.108 (1.089–1.127)
≥10–<15262,93043691.409 (1.367–1.453)1.391 (1.349–1.434)
≥15–<2071,00712541.624 (1.536–1.717)1.591 (1.504–1.682)
≥2035,3149341.767 (1.657–1.885)1.725 (1.618–1.840)
Women≥−1551,91234073.155 (3.045–3.270)2.725 (2.630–2.824)
<−15 to ≥−10150,92851121.997 (1.939–2.057)1.810 (1.757–1.864)
<−10 to ≥−5637,58712,5521.396 (1.367–1.424)1.336 (1.309–1.363)
<−5 to <53,355,87135,2851 (Ref)1 (Ref)
≥5–<10709,56962131.079 (1.05–1.109)1.081 (1.052–1.111)
≥10–<15196,16419481.398 (1.336–1.463)1.382 (1.321–1.447)
≥15–<2056,0396311.657 (1.531–1.792)1.616 (1.494–1.749)
≥2039,4835721.918 (1.766–2.083)1.880 (1.731–2.042)
Age <60 y≥−1547,7017523.909 (3.636–4.202)2.582 (2.400–2.777)
<−15 to ≥−10160,29519022.625 (2.506–2.750)2.056 (1.962–2.154)
<−10 to ≥−5810,57564861.542 (1.501–1.584)1.391 (1.354–1.430)
<−5 to <55,564,84228,9021 (Ref)1 (Ref)
≥5–<101,319,91556341.065 (1.035–1.096)1.068 (1.038–1.099)
≥10–<15391,11516051.317 (1.252–1.385)1.291 (1.227–1.358)
≥15–<20109,8214511.590 (1.448–1.745)1.518 (1.382–1.666)
≥2062,4913082.021 (1.806–2.261)1.876 (1.676–2.099)
Age ≥60 y≥−1539,00867692.915 (2.843–2.989)2.508 (2.446–2.571)
<−15 to ≥−10108,97612,1332.055 (2.016–2.095)1.856 (1.821–1.892)
<−10 to ≥−5459,45631,6731.390 (1.373–1.408)1.332 (1.315–1.349)
<−5 to <52,049,54493,3341 (Ref)1 (Ref)
≥5–<10303,51415,4111.098 (1.080–1.117)1.094 (1.076–1.113)
≥10–<1567,97947121.398 (1.358–1.439)1.373 (1.333–1.413)
≥15–<2017,22514341.593 (1.512–1.678)1.552 (1.473–1.635)
≥2012,30611981.722 (1.626–1.823)1.680 (1.587–1.778)
Nonsmoking≥−1571,50163133.180 (3.099–3.263)2.692 (2.624–2.763)
<−15 to ≥−10215,21810,8232.149 (2.106–2.192)1.935 (1.897–1.975)
<−10 to ≥−5993,67728,6021.419 (1.400–1.438)1.367 (1.349–1.386)
<−5 to <55,866,96790,7131 (Ref)1 (Ref)
≥5–<101,221,28615,9631.126 (1.107–1.145)1.123 (1.104–1.142)
≥10–<15336,60748151.433 (1.392–1.475)1.405 (1.365–1.447)
≥15–<2092,46614511.669 (1.585–1.758)1.627 (1.545–1.714)
≥2057,48312131.890 (1.786–2.000)1.847 (1.745–1.955)
Smoking≥−1515,20812082.381 (2.248–2.523)2.143 (2.022–2.270)
<−15 to ≥−1054,05332121.971 (1.900–2.044)1.866 (1.799–1.935)
<−10 to ≥−5276,35495571.353 (1.322–1.385)1.317 (1.287–1.348)
<−5 to <51,747,41931,5231 (Ref)1 (Ref)
≥5–<10402,14350821.027 (0.997–1.058)1.028 (0·998–1.059)
≥10–<15122,48715021.321 (1.254–1.391)1.320 (1.253–1.390)
≥15–<2034,5804341.508 (1.372–1.658)1.503 (1.367–1.652)
≥2017,3142931.552 (1.384–1.742)1.535 (1.368–1.722)
No chronic disease≥−1544,12221493.301 (3.159–3.449)2.509 (2.401–2.622)
<−15 to ≥−10141,03140772.178 (2.108–2.250)1.877 (1.817–1.939)
<−10 to ≥−5678,74511,4101.410 (1.380–1.439)1.339 (1.311–1.368)
<−5 to <54,356,10838,7171 (Ref)1 (Ref)
≥5–<101,011,03867421.090 (1.063–1.119)1.077 (1.049–1.105)
≥10–<15299,79419761.378 (1.317–1.442)1.337 (1.278–1.399)
≥15–<2082,8305581.602 (1.473–1.741)1.555 (1.430–1.691)
≥2044,6164091.776 (1.611–1.957)1.738 (1.577–1.916)
Chronic disease≥−1542,58753723.038 (2.955–3.124)2.608 (2.536–2.682)
<−15 to ≥−10128,24099582.169 (2.124–2.215)1.934 (1.894–1.975)
<−10 to ≥−5591,28626,7491.438 (1.418–1.458)1.361 (1.342–1.380)
<−5 to <53,258,27883,5191 (Ref)1 (Ref)
≥5–<10612,39114,3031.103 (1.084–1.123)1.110 (1.090–1.130)
≥10–<15159,30043411.419 (1.377–1.463)1.410 (1.368–1.454)
≥15–<2044,21613271.653 (1.565–1.745)1.621 (1.535–1.711)
≥2030,18110971.848 (1.741–1.961)1.800 (1.696–1.911)
No cancerc≥−1579,70060372.983 (2.906–3.062)2.793 (2.721–2.868)
<−15 to ≥−10255,96711,7252.104 (2.064–2.145)1.989 (1.951–2.028)
<−10 to ≥−51,232,68032,9181.398 (1.381–1.416)1.352 (1.336–1.369)
<−5 to <57,469,490108,2021 (Ref)1 (Ref)
≥5–<101,597,89518,8101.113 (1.096–1.130)1.117 (1.100–1.134)
≥10–<15452,60756101.418 (1.381–1.457)1.412 (1.374–1.450)
≥15–<20125,36516751.651 (1.574–1.733)1.625 (1.548–1.705)
≥2073,75413751.883 (1.786–1.986)1.836 (1.741–1.936)
Cancerc≥−15700914841.938 (1.837–2.045)1.901 (1.802–2.006)
<−15 to ≥−1013,30423101.602 (1.533–1.674)1.566 (1.498–1.636)
<−10 to ≥−537,35152411.353 (1.310–1.396)1.320 (1.278–1.362)
<−5 to <5144,89614,0341 (Ref)1 (Ref)
≥5–<1025,53422350.986 (0.943–1.031)0.989 (0.946–1.034)
≥10–<1564877071.262 (1.170–1.361)1.257 (1.166–1.356)
≥15–<2016812101.444 (1.260–1.654)1.426 (1.244–1.634)
≥2010431311.377 (1.159–1.635)1.362 (1.147–1.618)

Abbreviation: Ref, reference.

a

Model 1 was adjusted for age and sex.

b

Model 2 was adjusted for age, sex, smoking, drinking, exercise, DM, HTN, dyslipidemia, cancer, and income.

c

Cancer was not adjusted in model 2.

All-cause mortality by weight change in five BMI categories

We also analyzed the HRs for all-cause mortality of weight change in the five BMI categories, after adjusting all covariates (Fig. 1). Across all BMI categories, weight loss ≥15% was associated with increased mortality rates. Among the weight loss ≥15% groups, the highest mortality rates were observed in subjects with BMI ≥30 kg/m2 (HR, 3.469; 95% CI, 2.236 to 5.381) and the lowest mortality rates were observed in the subjects with BMI <18.5 kg/m2 (HR, 1.830; 95% CI, 1.744 to 1.921). In underweight (BMI <18.5 kg/m2), weight loss was associated with increased mortality rates, as were 5% to 10% and 10% to 15% weight gains, but weight gain of ≥15% was not associated with increased mortality rates. Compared with the normal BMI group, the overweight and obese groups had higher HRs for all-cause mortality for each weight change, and the weight loss group had higher HRs than weight gain group. In the BMI ≥30 kg/m2 group, each weight loss had higher HRs, but each weight gain had lower HRs compared with the normal BMI group.

Weight change and all-cause mortality in the five BMI groups, adjusted for age, sex, smoking, drinking, exercise, DM, HTN, dyslipidemia, cancer, and income. *Not statistically significant.
Figure 1.

Weight change and all-cause mortality in the five BMI groups, adjusted for age, sex, smoking, drinking, exercise, DM, HTN, dyslipidemia, cancer, and income. *Not statistically significant.

Open in new tabDownload slide

BMI category change and all-cause mortality

We also analyzed HRs for mortality by category of BMI change according to weight change (Fig. 2). Both the BMI stable group [i.e., group 1 (normal to normal BMI) and group 2 (obesity to obesity BMI) together] showed a reverse J-shaped mortality curve. Compared with group 1, weight loss was associated with lower HR for mortality in group 2. In group 3 (obesity to normal BMI), increases in weight loss were associated with increased mortality rates. In group 4 (normal to obesity BMI), increases in weight gain were also associated with increased mortality rates, but the HRs for mortality in group 4 were lower than those in group 3.

Graphs indicating results of change in BMI by category, and all-cause mortality according to weight change. (a) Normal to normal BMI change group. (b) Obese to obese BMI change group. (c) Obese to normal BMI change group. (d) Normal to obese BMI change group. Data were adjusted for age, sex, smoking, drinking, exercise, DM, HTN, dyslipidemia, cancer, and income. *Not statistically significant.
Figure 2.

Graphs indicating results of change in BMI by category, and all-cause mortality according to weight change. (a) Normal to normal BMI change group. (b) Obese to obese BMI change group. (c) Obese to normal BMI change group. (d) Normal to obese BMI change group. Data were adjusted for age, sex, smoking, drinking, exercise, DM, HTN, dyslipidemia, cancer, and income. *Not statistically significant.

Open in new tabDownload slide

Discussion

In this study, weight change over 4 years was associated with increased all-cause mortality independent of BMI status, and weight loss was associated with higher mortality rates than was weight gain, overall, and in subgroup analysis. The HR for mortality with weight loss ≥15% was 2.598 (95% CI, 2.537 to 2.659), and with weight gain ≥20% was 1.784 (95% CI 1.695 to 1.877). Among the groups with weight loss ≥15%, BMI ≥30 kg/m2 was associated with the highest mortality rates (HR, 3.469; 95% CI, 2.236 to 5.381).

Weight loss was associated with increased mortality rates (11–20) even after excluding subjects with coexisting morbidity (13, 14, 17–19, 23). In this study, weight loss caused more deaths than did weight gain and this association was still maintained in subgroup analysis in the group without chronic disease and no-cancer group. In the first National Health and Nutrition Examination Survey, with a 5-year follow-up, weight loss ≥15% increased mortality by twofold compared with weight loss <5% in men and women (28). In a large-scale Japanese cohort study over 8.7 years, a reverse J-shaped association between weight change during a 5-year period and all-cause mortality was observed, and weight loss was associated with a greater HR for all-cause mortality than was weight gain (25). However, other studies showed no association between weight loss and mortality in healthy Japanese-American men (21) and Dutch women older than age 50 years (22).

The mechanism underlying the association between weight loss and mortality is unclear, although changes in body composition, such as muscle mass decline, particularly among the elderly, has been considered (29). Weight loss, particularly among the elderly, is associated with frailty, which correlates with morbidity, falls, disability, hospitalization, and mortality (30). Weight loss is associated with oxidative DNA damage, which is hypothesized to exert a carcinogenic effect (31) and may also weaken the immune system (32), increasing the incidence of cancer and cancer-related mortality. In underweight individuals, weight loss ≥5% and a 5% to 15% weight gain were associated with increased mortality. In a Swiss national cohort study, the underweight group had increased mortality risk (33). Moreover, in a US population study, the underweight participants who lost >5% of their weight had a 1.87 times increased risk for all-cause mortality (20). However, the effect of weight gain in the underweight subjects’ mortality is controversial. Weight gain of 5% to 15% was not associated with increased mortality rate in the US population study, but >15% weight gain was associated with decreased mortality in underweight elderly people (20). Moreover, one study found a beneficial effect of weight gain on mortality (34). In this study, weight gain ≥15% in underweight individuals was not associated with decreased mortality (34). This result may be due to the small number of underweight individuals gaining ≥15% of their weight. Moreover, the Japanese study (25) reported the highest mortality rates among patients with weight loss ≥5 kg, particularly in the BMI <22 kg/m2 group (HR, 2.85 in men and 3.62 in women). However, our study shows conflicting results, with the highest HR for mortality in the BMI ≥30 kg/m2 group with weight loss ≥15% compared with the group with BMI <18.5 kg/m2 and weight loss ≥15%. This difference may be due to the effects of the different exclusion criteria. The Japanese study excluded subjects with BMI <14 kg/m2 and ≥40 kg/m2, and chronic diseases, such as cancer, cerebrovascular disease, chronic liver disease, and myocardial disease. Moreover, different effects of weight and BMI on mortality or morbidity in Korean and Japanese subjects may also have different effects (35). Further research, including genomic studies, is required to evaluate the differences among ethnicities (36).

Many studies have reported the association between weight gain and increased mortality (12, 23, 24). However, some studies found no association between weight gain and mortality (11–17, 21, 22, 37). The conflicting findings may result from study design, subject age, duration of weight change or observation period, and cause of death. Weight gain also increases the risk of DM, HTN, and cardiovascular disease risk factors (38–40), with a resulting increase in inflammation, which may increase mortality (41). Obesity-related cancers, including colorectal, prostate, and breast cancer, are more common in the obese population and weight gain is associated with increased cancer-related mortality (42, 43). Nevertheless, weight gain increases mortality to a lesser extent than weight loss, and this may be explained by weight tolerance associated with aging. People tend to gain weight over the life span (44, 45), so people have some tolerance to weight gain and, therefore, there may be a decreased effect of weight gain, compared with weight loss, on mortality (24). Moreover, weight loss with concomitant chronic diseases is largely due to loss of lean body mass, which is replaced by fat mass, resulting in greater increases in mortality in the weight loss group than in the weight gain group (24). However, in two studies including the reanalysis of Framingham Heart Study and the Tecumseh Community Health Study data, authors suggested that weight loss via fat mass may reduce mortality (46). Therefore, the lower mortality rates in the subjects in the group who had one or more chronic diseases than in subjects without chronic diseases may be explained by intentional weight loss, although we could not know intention of weight change in this study. More precisely designed studies are needed to reveal this result.

We also examined the categorical BMI change and all-cause mortality according to weight change. In the Ohsaki study, change from normal BMI (i.e., 18.5 to 24.9 kg/m2) to overweight BMI (i.e., 25 to 29.9 kg/m2) was associated with decreased all-cause mortality (HR, 0.87), but change from normal BMI to obese BMI was associated with increased all-cause mortality (HR, 1.38), and stable obese BMI was also associated with increased all-cause mortality (HR, 2.21) (47). In our study, increased mortality was observed in the normal to obese BMI group, consistent with the Ohsaki study, although, the same weight change was associated with a greater increase in mortality in the obese to normal BMI group than in the normal to obese BMI group. Moreover, stable normal weight subjects had an increased mortality rate compared with stable obese subjects with the same weight loss. These results suggest that normal-weight people may be more sensitive to weight loss than obese subjects, and weight loss may be more harmful than weight gain, contrary to popular belief, even in obese subjects. A Swedish study found that overweight or obese people had increased rates of noncancer mortality after weight loss; therefore, we postulated that increased mortality in obese subjects may be caused by noncancer mortality in this weight loss group (48), although we did not know the cause of death in this study. Another potential explanation is that in persons who are obese and maintain obese status for a long time, the sustained effect of excess weight and fat mass or weight change become less harmful (49).

Although we could not examine the cause of death in this study, we analyzed the association between weight change and mortality in subgroup analysis. Some studies have found that BMI values at younger ages have a greater predictive power with respect to mortality than BMI values at older ages (50, 51). Moreover, weight loss in the elderly mainly results from preexisting or new onset of disease, and weight gain may cause few adverse effects on mortality in the elderly (49). This is explained by the fact that cumulative body mass has already exerted an effect on mortality from early adulthood; therefore, weight gain in later life has little effect on mortality (24). Weight change resulted in greater increases in mortality in the noncancer group than in the cancer group. This implies that substantial weight change in individuals without cancer reflects development of a new condition not captured or reflected in their data; for example, the effect of educational interventions about weight change in patients with cancer. The effects of weight change on mortality were also greater among nonsmokers than among smokers in this study. A similar association was also reported in a Japanese study (25) and in Western studies (13–15, 17–19, 23); therefore, we postulated that smoking may modify the weight change effect on mortality such that the cumulative effect of smoking supersedes the effect of weight change.

This study has several limitations. First, we did not know whether weight change was intentional or unintentional. Intentional weight loss to improve health is associated with lower mortality (9, 10). The exclusion of participants losing weight intentionally may affect these results, particularly for obese participants. Second, we did not assess body composition. Weight loss involving muscle mass will increase mortality to a greater extent than weight loss involving fat mass (29); however, the measurement of body composition in large cohort studies is challenging. Third, we were not able to design this study to assess time and duration of comorbidities, which also affect weight status. Fourth, the duration of weight change and the mean observational time were too short to fully assess the association between weight change and mortality. Fifth, we did not know the cause of death; studies on causes of death according to weight change are required. Sixth, this study only included Korean participants and results cannot be generalized to other ethnicities. The effect of weight may be different by cancer type and ethnicity. Stomach and esophageal cancer are common in Korea and Japan but not in Western countries, and are more prevalent in low-BMI populations (52).

Nevertheless, this study has important strengths. To the best of our knowledge, this is the first study to examine weight change and all-cause mortality in a large population of Korean adults. Second, we conducted a nationwide study involving a homogeneous participant group. Furthermore, we analyzed various subgroups, adjusting for covariates potentially affecting mortality.

In conclusion, weight change over 4 years showed a reverse J-shaped all-cause mortality curve independent of BMI status. Both weight loss and weight gain were associated with increased mortality and weight loss was associated with a greater mortality risk, particularly in the BMI ≥30 kg/m2 group. Further research is required to establish the association between weight change and mortality by taking into account the intentional elements of weight loss and body composition. Moreover, studies on cause-specific mortality according to weight change are also required to determine the underlying mechanisms.

Abbreviations

     
  • BMI

    body mass index

    •  
  • CI

    confidence interval

    •  
  • DM

    diabetes mellitus

    •  
  • HR

    hazard ratio

    •  
  • HTN

    hypertension

    •  
  • ICD-10

    International Classification of Diseases, 10th Revision

    •  
  • NHIS

    National Health Insurance System

    •  
  • Q

    quartile

    •  
  • SD

    standard deviation.

Acknowledgments

We thank the Korean National Health Insurance Corporation and all the participants of the study and health checkup.

Financial Support: This study was supported by Grant HC16C2285 from the Korean Health Technology and Research and Development project, Ministry of Health and Welfare, Republic of Korea.

Disclosure Summary: The authors have nothing to disclose.

References

1.

Mokdad
AH
,
Ford
ES
,
Bowman
BA
,
Dietz
WH
,
Vinicor
F
,
Bales
VS
,
Marks
JS
.
Prevalence of obesity, diabetes, and obesity-related health risk factors, 2001
.
JAMA
.
2003
;
289
(
1
):
76
–
79
.

Google Scholar

Crossref
Search ADS
PubMed

 
2.

Yoon
KH
,
Lee
JH
,
Kim
JW
,
Cho
JH
,
Choi
YH
,
Ko
SH
,
Zimmet
P
,
Son
HY
.
Epidemic obesity and type 2 diabetes in Asia
.
Lancet
.
2006
;
368
(
9548
):
1681
–
1688
.

Google Scholar

Crossref
Search ADS
PubMed

 
3.

Jee
SH
,
Sull
JW
,
Park
J
,
Lee
SY
,
Ohrr
H
,
Guallar
E
,
Samet
JM
.
Body-mass index and mortality in Korean men and women
.
N Engl J Med
.
2006
;
355
(
8
):
779
–
787
.

Google Scholar

Crossref
Search ADS
PubMed

 
4.

Calle
EE
,
Thun
MJ
,
Petrelli
JM
,
Rodriguez
C
,
Heath
CW
Jr .
Body-mass index and mortality in a prospective cohort of U.S. adults
.
N Engl J Med
.
1999
;
341
(
15
):
1097
–
1105
.

Google Scholar

Crossref
Search ADS
PubMed

 
5.

Adams
KF
,
Schatzkin
A
,
Harris
TB
,
Kipnis
V
,
Mouw
T
,
Ballard-Barbash
R
,
Hollenbeck
A
,
Leitzmann
MF
.
Overweight, obesity, and mortality in a large prospective cohort of persons 50 to 71 years old
.
N Engl J Med
.
2006
;
355
(
8
):
763
–
778
.

Google Scholar

Crossref
Search ADS
PubMed

 
6.

Tsugane
S
,
Sasaki
S
,
Tsubono
Y
.
Under- and overweight impact on mortality among middle-aged Japanese men and women: a 10-y follow-up of JPHC study cohort I
.
Int J Obes Relat Metab Disord
.
2002
;
26
(
4
):
529
–
537
.

Google Scholar

Crossref
Search ADS
PubMed

 
7.

Gunnell
DJ
,
Frankel
SJ
,
Nanchahal
K
,
Peters
TJ
,
Davey Smith
G
.
Childhood obesity and adult cardiovascular mortality: a 57-y follow-up study based on the Boyd Orr cohort
.
Am J Clin Nutr
.
1998
;
67
(
6
):
1111
–
1118
.

Google Scholar

PubMed
OpenURL Placeholder Text

 
8.

Must
A
,
Strauss
RS
.
Risks and consequences of childhood and adolescent obesity
.
Int J Obes Relat Metab Disord
.
1999
;
23
(
Suppl 2
):
S2
–
S11
.

Google Scholar

Crossref
Search ADS
PubMed

 
9.

French
SA
,
Folsom
AR
,
Jeffery
RW
,
Williamson
DF
.
Prospective study of intentionality of weight loss and mortality in older women: the Iowa Women’s Health Study
.
Am J Epidemiol
.
1999
;
149
(
6
):
504
–
514
.

Google Scholar

Crossref
Search ADS
PubMed

 
10.

Gregg
EW
,
Gerzoff
RB
,
Thompson
TJ
,
Williamson
DF
.
Intentional weight loss and death in overweight and obese U.S. adults 35 years of age and older
.
Ann Intern Med
.
2003
;
138
(
5
):
383
–
389
.

Google Scholar

Crossref
Search ADS
PubMed

 
11.

Amador
LF
,
Al Snih
S
,
Markides
KS
,
Goodwin
JS
.
Weight change and mortality among older Mexican Americans
.
Aging Clin Exp Res
.
2006
;
18
(
3
):
196
–
204
.

Google Scholar

Crossref
Search ADS
PubMed

 
12.

Andres
R
,
Muller
DC
,
Sorkin
JD
.
Long-term effects of change in body weight on all-cause mortality. A review
.
Ann Intern Med
.
1993
;
119
(
7 Pt 2
):
737
–
743
.

Google Scholar

Crossref
Search ADS
PubMed

 
13.

Dey
DK
,
Rothenberg
E
,
Sundh
V
,
Bosaeus
I
,
Steen
B
.
Body mass index, weight change and mortality in the elderly. A 15 y longitudinal population study of 70 y olds
.
Eur J Clin Nutr
.
2001
;
55
(
6
):
482
–
492
.

Google Scholar

Crossref
Search ADS
PubMed

 
14.

Drøyvold
WB
,
Lund Nilsen
TI
,
Lydersen
S
,
Midthjell
K
,
Nilsson
PM
,
Nilsson
JA
,
Holmen
J
;
Nord-Trøndelag Health Study
.
Weight change and mortality: the Nord-Trøndelag Health Study
.
J Intern Med
.
2005
;
257
(
4
):
338
–
345
.

Google Scholar

Crossref
Search ADS
PubMed

 
15.

Higgins
M
,
D’Agostino
R
,
Kannel
W
,
Cobb
J
,
Pinsky
J
.
Benefits and adverse effects of weight loss. Observations from the Framingham Study
.
Ann Intern Med
.
1993
;
119
(
7 Pt 2
):
758
–
763
.

Google Scholar

Crossref
Search ADS
PubMed

 
16.

Ho
SC
,
Woo
J
,
Sham
A
.
Risk factor change in older persons, a perspective from Hong Kong: weight change and mortality
.
J Gerontol
.
1994
;
49
(
6
):
M269
–
M272
.

Google Scholar

Crossref
Search ADS
PubMed

 
17.

Newman
AB
,
Yanez
D
,
Harris
T
,
Duxbury
A
,
Enright
PL
,
Fried
LP
;
Cardiovascular Study Research Group
.
Weight change in old age and its association with mortality
.
J Am Geriatr Soc
.
2001
;
49
(
10
):
1309
–
1318
.

Google Scholar

Crossref
Search ADS
PubMed

 
18.

Pamuk
ER
,
Williamson
DF
,
Serdula
MK
,
Madans
J
,
Byers
TE
.
Weight loss and subsequent death in a cohort of U.S. adults
.
Ann Intern Med
.
1993
;
119
(
7 Pt 2
):
744
–
748
.

Google Scholar

Crossref
Search ADS
PubMed

 
19.

Sauvaget
C
,
Ramadas
K
,
Thomas
G
,
Vinoda
J
,
Thara
S
,
Sankaranarayanan
R
.
Body mass index, weight change and mortality risk in a prospective study in India
.
Int J Epidemiol
.
2008
;
37
(
5
):
990
–
1004
.

Google Scholar

Crossref
Search ADS
PubMed

 
20.

Corrada
MM
,
Kawas
CH
,
Mozaffar
F
,
Paganini-Hill
A
.
Association of body mass index and weight change with all-cause mortality in the elderly
.
Am J Epidemiol
.
2006
;
163
(
10
):
938
–
949
.

Google Scholar

Crossref
Search ADS
PubMed

 
21.

Iribarren
C
,
Sharp
DS
,
Burchfiel
CM
,
Petrovitch
H
.
Association of weight loss and weight fluctuation with mortality among Japanese American men
.
N Engl J Med
.
1995
;
333
(
11
):
686
–
692
.

Google Scholar

Crossref
Search ADS
PubMed

 
22.

Maru
S
,
van der Schouw
YT
,
Gimbrère
CH
,
Grobbee
DE
,
Peeters
PH
.
Body mass index and short-term weight change in relation to mortality in Dutch women after age 50 y
.
Am J Clin Nutr
.
2004
;
80
(
1
):
231
–
236
.

Google Scholar

PubMed
OpenURL Placeholder Text

 
23.

Lee
IM
,
Paffenbarger
RS
Jr .
Change in body weight and longevity
.
JAMA
.
1992
;
268
(
15
):
2045
–
2049
.

Google Scholar

Crossref
Search ADS
PubMed

 
24.

Adams
KF
,
Leitzmann
MF
,
Ballard-Barbash
R
,
Albanes
D
,
Harris
TB
,
Hollenbeck
A
,
Kipnis
V
.
Body mass and weight change in adults in relation to mortality risk
.
Am J Epidemiol
.
2014
;
179
(
2
):
135
–
144
.

Google Scholar

Crossref
Search ADS
PubMed

 
25.

Nanri
A
,
Mizoue
T
,
Takahashi
Y
,
Noda
M
,
Inoue
M
,
Tsugane
S
.
Japan Public Health Center- based Prospective Study Group. Weight change and all-cause, cancer and cardiovascular disease mortality in Japanese men and women: the Japan Public Health Center-Based Prospective Study
.
Int J Obes (Lond)
.
2010
;
34
(
2
):
348
–
356
.

Google Scholar

OpenURL Placeholder Text

 
26.

Song
SO
,
Jung
CH
,
Song
YD
,
Park
CY
,
Kwon
HS
,
Cha
BS
,
Park
JY
,
Lee
KU
,
Ko
KS
,
Lee
BW
.
Background and data configuration process of a nationwide population-based study using the Korean national health insurance system
.
Diabetes Metab J
.
2014
;
38
(
5
):
395
–
403
.

Google Scholar

Crossref
Search ADS
PubMed

 
27.

World Health Organization
.
The Asia-Pacific perspective: redefining obesity and its treatment
.
2000
. Available at: http://www.wpro.who.int/nutrition/documents/docs/Redefiningobesity.pdf. Accessed 25 August 2017.

28.

Pamuk
ER
,
Williamson
DF
,
Madans
J
,
Serdula
MK
,
Kleinman
JC
,
Byers
T
.
Weight loss and mortality in a national cohort of adults, 1971-1987
.
Am J Epidemiol
.
1992
;
136
(
6
):
686
–
697
.

Google Scholar

Crossref
Search ADS
PubMed

 
29.

Miller
SL
,
Wolfe
RR
.
The danger of weight loss in the elderly
.
J Nutr Health Aging
.
2008
;
12
(
7
):
487
–
491
.

Google Scholar

Crossref
Search ADS
PubMed

 
30.

Bortz
II
WM
.
A conceptual framework of frailty: a review
.
J Gerontol A Biol Sci Med Sci
.
2002
;
57
(
5
):
M283
–
M288
.

Google Scholar

Crossref
Search ADS
PubMed

 
31.

Ames
BN
.
Endogenous oxidative DNA damage, aging, and cancer
.
Free Radic Res Commun
.
1989
;
7
(
3-6
):
121
–
128
.

Google Scholar

Crossref
Search ADS
PubMed

 
32.

Shade
ED
,
Ulrich
CM
,
Wener
MH
,
Wood
B
,
Yasui
Y
,
Lacroix
K
,
Potter
JD
,
McTiernan
A
.
Frequent intentional weight loss is associated with lower natural killer cell cytotoxicity in postmenopausal women: possible long-term immune effects
.
J Am Diet Assoc
.
2004
;
104
(
6
):
903
–
912
.

Google Scholar

Crossref
Search ADS
PubMed

 
33.

Roh
L
,
Braun
J
,
Chiolero
A
,
Bopp
M
,
Rohrmann
S
,
Faeh
D
;
Swiss National Cohort Study Group
.
Mortality risk associated with underweight: a census-linked cohort of 31,578 individuals with up to 32 years of follow-up
.
BMC Public Health
.
2014
;
14
:
371
.

Google Scholar

Crossref
Search ADS
PubMed

 
34.

Andres
R
,
Elahi
D
,
Tobin
JD
,
Muller
DC
,
Brant
L
.
Impact of age on weight goals
.
Ann Intern Med
.
1985
;
103
(
6 ( Pt 2)
):
1030
–
1033
.

Google Scholar

Crossref
Search ADS
PubMed

 
35.

Ma
RC
,
Chan
JC
.
Type 2 diabetes in East Asians: similarities and differences with populations in Europe and the United States
.
Ann N Y Acad Sci
.
2013
;
1281
:
64
–
91
.

Google Scholar

Crossref
Search ADS
PubMed

 
36.

Park
JW
,
Park
J
,
Jee
SH
.
ADIPOQ gene variants associated with susceptibility to obesity and low serum adiponectin levels in healthy Koreans
.
Epidemiol Health
.
2011
;
33
:
e2011003
.

Google Scholar

Crossref
Search ADS
PubMed

 
37.

Wannamethee
SG
,
Shaper
AG
,
Walker
M
.
Weight change, weight fluctuation, and mortality
.
Arch Intern Med
.
2002
;
162
(
22
):
2575
–
2580
.

Google Scholar

Crossref
Search ADS
PubMed

 
38.

Ford
ES
,
Williamson
DF
,
Liu
S
.
Weight change and diabetes incidence: findings from a national cohort of US adults
.
Am J Epidemiol
.
1997
;
146
(
3
):
214
–
222
.

Google Scholar

Crossref
Search ADS
PubMed

 
39.

Huang
Z
,
Willett
WC
,
Manson
JE
,
Rosner
B
,
Stampfer
MJ
,
Speizer
FE
,
Colditz
GA
.
Body weight, weight change, and risk for hypertension in women
.
Ann Intern Med
.
1998
;
128
(
2
):
81
–
88
.

Google Scholar

Crossref
Search ADS
PubMed

 
40.

Norman
JE
,
Bild
D
,
Lewis
CE
,
Liu
K
,
West
DS
;
CARDIA Study
.
The impact of weight change on cardiovascular disease risk factors in young black and white adults: the CARDIA study
.
Int J Obes Relat Metab Disord
.
2003
;
27
(
3
):
369
–
376
.

Google Scholar

Crossref
Search ADS
PubMed

 
41.

Fogarty
AW
,
Glancy
C
,
Jones
S
,
Lewis
SA
,
McKeever
TM
,
Britton
JR
.
A prospective study of weight change and systemic inflammation over 9 y
.
Am J Clin Nutr
.
2008
;
87
(
1
):
30
–
35
.

Google Scholar

PubMed
OpenURL Placeholder Text

 
42.

Playdon
MC
,
Bracken
MB
,
Sanft
TB
,
Ligibel
JA
,
Harrigan
M
,
Irwin
ML
.
Weight gain after breast cancer diagnosis and all-cause mortality: systematic review and meta-analysis
.
J Natl Cancer Inst
.
2015
;
107
(
12
):
djv275
.

Google Scholar

Crossref
Search ADS
PubMed

 
43.

Bonn
SE
,
Wiklund
F
,
Sjölander
A
,
Szulkin
R
,
Stattin
P
,
Holmberg
E
,
Grönberg
H
,
Bälter
K
.
Body mass index and weight change in men with prostate cancer: progression and mortality
.
Cancer Causes Control
.
2014
;
25
(
8
):
933
–
943
.

Google Scholar

Crossref
Search ADS
PubMed

 
44.

Jacobsen
BK
,
Njølstad
I
,
Thune
I
,
Wilsgaard
T
,
Løchen
ML
,
Schirmer
H
.
Increase in weight in all birth cohorts in a general population: the Tromsø Study, 1974-1994
.
Arch Intern Med
.
2001
;
161
(
3
):
466
–
472
.

Google Scholar

Crossref
Search ADS
PubMed

 
45.

Drøyvold
WB
,
Nilsen
TI
,
Krüger
O
,
Holmen
TL
,
Krokstad
S
,
Midthjell
K
,
Holmen
J
.
Change in height, weight and body mass index: longitudinal data from the HUNT study in Norway
.
Int J Obes
.
2006
;
30
(
6
):
935
–
939
.

Google Scholar

Crossref
Search ADS

 
46.

Allison
DB
,
Zannolli
R
,
Faith
MS
,
Heo
M
,
Pietrobelli
A
,
VanItallie
TB
,
Pi-Sunyer
FX
,
Heymsfield
SB
.
Weight loss increases and fat loss decreases all-cause mortality rate: results from two independent cohort studies
.
Int J Obes Relat Metab Disord
.
1999
;
23
(
6
):
603
–
611
.

Google Scholar

Crossref
Search ADS
PubMed

 
47.

Shimazu
T
,
Kuriyama
S
,
Ohmori-Matsuda
K
,
Kikuchi
N
,
Nakaya
N
,
Tsuji
I
.
Increase in body mass index category since age 20 years and all-cause mortality: a prospective cohort study (the Ohsaki Study)
.
Int J Obes
.
2009
;
33
(
4
):
490
–
496
.

Google Scholar

Crossref
Search ADS

 
48.

Nilsson
PM
,
Nilsson
JA
,
Hedblad
B
,
Berglund
G
,
Lindgärde
F
.
The enigma of increased non-cancer mortality after weight loss in healthy men who are overweight or obese
.
J Intern Med
.
2002
;
252
(
1
):
70
–
78
.

Google Scholar

Crossref
Search ADS
PubMed

 
49.

Ferrucci
L
,
Alley
D
.
Obesity, disability, and mortality: a puzzling link
.
Arch Intern Med
.
2007
;
167
(
8
):
750
–
751
.

Google Scholar

Crossref
Search ADS
PubMed

 
50.

Jeffreys
M
,
McCarron
P
,
Gunnell
D
,
McEwen
J
,
Smith
GD
.
Body mass index in early and mid-adulthood, and subsequent mortality: a historical cohort study
.
Int J Obes Relat Metab Disord
.
2003
;
27
(
11
):
1391
–
1397
.

Google Scholar

Crossref
Search ADS
PubMed

 
51.

Stevens
J
,
Cai
J
,
Pamuk
ER
,
Williamson
DF
,
Thun
MJ
,
Wood
JL
.
The effect of age on the association between body-mass index and mortality
.
N Engl J Med
.
1998
;
338
(
1
):
1
–
7
.

Google Scholar

Crossref
Search ADS
PubMed

 
52.

Reeves
GK
,
Pirie
K
,
Beral
V
,
Green
J
,
Spencer
E
,
Bull
D
;
Million Women Study Collaboration
.
Cancer incidence and mortality in relation to body mass index in the Million Women Study: cohort study
.
BMJ
.
2007
;
335
(
7630
):
1134
.

Google Scholar

Crossref
Search ADS
PubMed

 

Author notes

*

These authors contributed equally to this study.

Copyright © 2017 Endocrine Society
Issue Section:
Clinical Research Articles > Obesity and Adipocyte Biology
Download all slides
  • Supplementary data

    • Supplementary data

    Supplemental Table 1 - docx file