Contact: Audrey Hu audrey.hu@wecistanche.com

Bodyweight fluctuation is associated with rapid kidney function decline

Young Su Joo, et al

Abstract Objective: This study aimed to evaluate the effects of body weight fluctuations on kidney function deterioration in a prospective cohort of individuals with normal kidney function.

Methods: Data were obtained from the Korean Genome and Epidemiology Study. Bodyweight fluctuations were determined using average successive variability (ASV), which was defined as the average absolute body weight change using repeated measurements for all participants. The decline of the estimated glomerular filtration rate (eGFR) over time was calculated using linear regression analysis of serial eGFR measurements for each patient. Rapid eGFR decline was defined as an average eGFR decline>3 mL/min/1.73 m2 per year.

Results: A total of 6,790 participants were analyzed. During a median follow-up of 11.7 years, rapid eGFR decline was observed in 913 (13.4%) participants. When the participants were categorized into tertiles according to ASV, rapid eGFR decline was more prevalent in the highest ASV tertile group than in the lowest. Analyses using multiple logistic regression models revealed that the risk of rapid eGFR decline was increased in the highest ASV tertile group compared with the lowest (odds ratio: 1.66).

Conclusions: Bodyweight fluctuations were significantly associated with an increased risk of rapid kidney function decline in participants with normal kidney function.

Cistanche is good for kidney function

INTRODUCTION

Chronic kidney disease (CKD) is a major public health concern because of its effects on increasing morbidity and mortality risks (1). Despite blood glucose control and hypertension management, which are key factors affecting kidney function, the worldwide CKD prevalence is rapidly increasing (2). Considering that established CKD is irrevocable, identifying adjustable risk factors and modifying detrimental lifestyles associated with CKD development are critical starting points to alleviate the CKD burden.

Obesity is a significant risk factor for CKD (3,4). In participants with normal kidney function, BMI is an independent risk factor for new-onset kidney disease (5). Furthermore, obesity is associated with a faster progression to end-stage kidney disease in patients with various kidney diseases, including immunoglobin A (IgA) nephropathy and focal segmental glomerulosclerosis (6,7). Based on these findings, weight reduction is commonly recommended as an effective intervention to reduce CKD risk in individuals with obesity (8). However, weight loss is frequently accompanied by weight regain, casually termed weight cycling or the 'yo-yo' effect.

Bodyweight fluctuations have been repeatedly shown to be related to metabolic derangements or adverse cardiovascular outcomes (9,10). An analysis of the Framingham Heart Study follow-up data revealed that fluctuations in body weight were associated with higher mortality and morbidity due to increased coronary heart disease risks (11). Moreover, weight cycling has also been reported as a significant risk factor for the development of diabetes (12). However, the relationship between body weight fluctuations and kidney function remains unknown. Therefore, this study aimed to evaluate the association between body weight variability and rapid kidney function decline in a prospective, community-based cohort composed of participants with normal kidney function.

METHODS

Study population

Data were extracted from the Korean Genome and Epidemiology Study (KoGES), a prospective, community-based cohort study. Detailed cohort profiles and methods concerning the development of KoGES have been described elsewhere (13). The study cohort was composed of participants aged 40 to 69 years who were residents of Ansan (urban area) or Ansung (rural area), Republic of Korea. The participants underwent detailed health examinations and various health-related surveys at baseline and biannually from 2001 to 2014. Participants with recorded body weight measurements from baseline and two or more follow-ups were initially screened. Participants with known kidney disease, those with an estimated glomerular filtration rate (eGFR) < 60 mL/min/1.73 m2 or proteinuria at baseline, those with follow-up intervals > 4 years, or those diagnosed with cancer during the study period were excluded. Ultimately, a total of 6,790 participants were included. In the sensitivity analysis using the calibrated Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) creatinine equation, a total of 6,954 participants were included after excluding participants using the same criteria as was used in the main analysis (Figure 1). All participants voluntarily enrolled and provided written informed consent, which included a general agreement covering data usage for public and academic use. This study was conducted according to the Declaration of Helsinki and was approved by the Institutional Review Board of Yonsei University Health System Clinical Trial Center.

Echinacoside of cistanche can improve kidney function

Data collection

All participants underwent comprehensive health examinations and completed questionnaires regarding their health and lifestyles during enrollment. Demographic and socioeconomic data and medical history were obtained at their baseline visit. Comorbidity surveys, anthropometric measurements, and laboratory evaluations were performed biannually. Bodyweight and height were measured by trained researchers using a calibrated scale or an automated scale with a stadiometer. All anthropometric measurements were performed following specific protocols in which the participants took off their shoes, wore light clothing, and stood on the scale on an even surface while measurements were taken. Participants with blood pressures > 140/90 mm Hg or those taking antihypertensive medications were considered hypertensive. Those with glycated hemoglobin levels > 6.5%, fasting blood glucose concentrations > 126 mg/dL after >8 hours fasting, post-load glucose concentrations > 200 mg/dL in the 75-g oral glucose tolerance test, or those receiving diabetic treatments were considered to have diabetes mellitus. Cardiovascular disease (CVD) was defined as having a history of myocardial infarction, coronary artery disease, congestive heart failure, peripheral artery disease, or cerebrovascular disease. Muscle mass was assessed using multifrequency bioelectrical impedance analyses (InBody 3.0, BioSpace). Single-day dietary data for the total energy intake (in kilocalories) were estimated using a semi-quantitative dietary recall food-frequency questionnaire (FFQ). The FFQ details, as well as the validity and reproducibility of the FFQ, have been reported previously (14).

Blood samples were obtained after >8 hours of fasting and were analyzed in a central laboratory. Urine samples were collected in the morning after the first voiding. Urine tests were performed on fresh urine samples with urine reagent strips and processed using semi-automatic urine analyzers, which were calibrated regularly. Urine protein amounts were determined as absent, trace, 1+, 2+, or 3+, which approximately correlated with urine protein levels of <10, 10 to 20, >30, >100, and >500 mg/dL, respectively. The presence of proteinuria was considered for a urinalysis result higher than trace levels. Serum creatinine levels were measured by the Jaffe assay. eGFR was calculated using the four-variable Modification of Diet in Renal Disease (MDRD) equation (15). For sensitivity analysis, the creatinine levels were reduced by a calibration factor of 5% for standardization to isotope dilution mass spectrometry reference method (16). These creatinine values were used to estimate eGFR with the CKD-EPI creatinine equation (17).

Measures of weight variability

Weight variability, i.e., the variation in each participant’s weight between visits, was calculated from the average absolute difference between successive weights and defined as the average successive variability (ASV). Specifically, ASV was calculated using bodyweight at each follow-up (Bwt_n) using the following formula: ([|Bwt_baseline– Bwt₁| + ·· + |Bwt_n-1 – Bwt_n|]/visit frequency [n]). The directions of body weight changes were also considered in the analyses. These were calculated by subtracting the individual participant's mean weight during the total observation period from the baseline visit weight and defined as either stable (weight change < 5%), increasing (positive weight change ≥ 5%), or decreasing (negative weight change ≥ 5%) (18).

Outcome measurement

The annual eGFR decline for each participant was calculated using linear least-squares regression models, as previously reported (19). Individual eGFR measurements during the entire follow-up duration were included in the models. In order to estimate the annual eGFR decline, the regression line that was closest to the patient's eGFR measurements, in the sense of minimizing the sum of the squared deviations of the individual eGFR measurements from the line, was determined. Annual eGFR decline, described in units of “milliliters per minutes per 1.73 m2 per year,” was defined as the slope of the least-squares regression line of the eGFR versus time data for each patient. Rapid eGFR decline was defined as a calculated annual eGFR decline>3 mL/min/1.73 m2 during the follow-up period (19).

Cistanche can improve kidney function

Statistical analysis

The analyses were performed using ASV as a continuous variable and tertiles of the ASV as categorical variables. Multivariable linear regressions using four models were used to evaluate the linear association between eGFR decline rate and ASV and compare the adjusted annual eGFR decline rates according to ASV groups. Means and 95% confidence intervals were reported. Model 1 adjusted for baseline weight, age, and sex. Model 2 adjusted Model 1 for marital status, education, alcohol consumption, smoking status, and weight change direction. Model 3 adjusted Model 2 for systolic blood pressure (SBP), high-density lipoprotein cholesterol (HDL-C), C-reactive protein (CRP), albumin, homeostatic model assessment for insulin resistance score (HOMA-IR), physical activity, muscle mass, total daily energy intake, and eGFR. Model 4 adjusted for age, sex, study site, education, diabetes, hypertension, smoking, and baseline BMI. Logistic regression analyses were used to evaluate the relationship between body weight variability and rapid eGFR decline. In analyses in which body weight variability was used as a continuous variable, models have been constructed in which the variability measure was entered to calculate the odds ratios (OR) for outcomes per the increase in ASV values. The variables that were used in the multivariable linear regression models were also used in the logistic regression models. For sensitivity analyses, the impact of body weight variability was further analyzed using standard deviation, coefficient of variation, residual standard deviation, percentage of change of absolute body weight changes compared with baseline body weight (ASV percentage), and average successive BMI variability (ASBMIV). ASV percentage was calculated by dividing each participant’s ASV by their baseline body weight. ASBMIVs were calculated as the average of the absolute value change in BMI values between each visit. In addition, the latent class trajectory modeling was used to investigate the association between ASV and the different trajectories of eGFR (20). The optimal number of latent classes was determined using the following criteria: 1) Bayesian information criterion; 2) mean posterior probability of class membership; 3) the number of participants in each trajectory, and 4) interpretability (21). Logistic regression was used to evaluate the relationship between ASV and the trajectory class with the most rapidly declining eGFR. All analyses were performed using STATA software (version 16.1, StataCorp) and R language software (version 3.4.3, R Foundation for Statistical Computing). p < 0.05 were considered statistically significant.

RESULTS

Baseline characteristics

The baseline characteristics of the participants are shown in Table 1. The mean age was 52.0 (8.7) years, and 48.8% were male. The mean eGFR was 91.9 (15.9) mL/min/1.73 m2. The body weight means and the range was 63.2 (10.1) kg and 32.1 to 105 kg, respectively. The range of baseline eGFR was 60 to 155 mL/min/1.73 m2. Among the participants, a total of 5,852 (86.2%) participants visited and measured body weight more than five times during the follow-up duration (Supporting Information Figure S1). In addition, the minimum measurement interval was 2 years, whereas the longest interval between visits was 4 years. Measurements were made at every 2-year interval in 58.0% of the participants (Supporting Information Figure S2). The mean ASV of the participants was 1.9 kg (Supporting Information Figure S3). There was no significant correlation between ASV and baseline eGFR (Pearson correlation, ρ=−0.0115, p=0.34; Supporting Information Figure S4).

Annual eGFR decline association with body weight variability

During a median follow-up of 11.7 years (range: 5.2-12.7), the overall mean annual eGFR decline rate of the participants was −1.57 mL/ min/1.73 m2 per year. The overall annual eGFR decline rate showed a significant negative relationship with ASV (annual eGFR decline rate per 1-kg increase in ASV: −0.10 mL/min/1/73 m2 per year; 95% CI: −0.14 to −0.06; Supporting Information Table S1). When the eGFR decline rates were compared among the ASV tertiles, the rate of eGFR decline was more rapid in the third ASV tertile compared with the first tertile. This difference in annual eGFR decline remained significant even after adjustments were made for confounding variables (Table 2). A significant annual eGFR decline difference was not found between the first and second ASV tertiles.

Rapid eGFR decline

Rapid eGFR decline, defined as an annual eGFR decline > 3 mL/ min/1.72 m2 per year, was observed in 913 (13.4%) participants. Rapid eGFR decline was more prevalent in the tertile with the highest ASV compared with the lowest (Tertile 1: 283 [12.5%]; Tertile 2: 261 [11.5%]; Tertile 3: 369 [16.3%]; p for trend <0.001). The proportion of participants with rapid eGFR decline was comparable between the first and second ASV tertile groups.

Impact of body weight variability on rapid eGFR decline

When the risk of rapid eGFR decline was stratified among the ASV tertiles using multivariable logistic regression models (Table 3), the OR significantly increased in participants in the highest ASV tertile group compared with those in the lowest (OR: 1.48; 95% CI: 1.24- 1.76). This association remained significant after adjusting for confounding variables (OR: 1.66; 95% CI: 1.29-2.14). Subsequently, the associations between rapid eGFR decline and ASV, treated as a continuous variable, were also evaluated (Table 4). Each 1-kg increase in ASV was related to a 24% increase in the risk of rapid eGFR decline (OR: 1.24; 95% CI, 1.16-1.33). This relationship between rapid eGFR decline and ASV was robust even after adjustments were made for confounding factors, including baseline eGFRs and overall weight change directions (OR: 1.29; 95% CI: 1.17-1.42).

Subgroup analysis

The relationship between body weight variability and rapid eGFR decline was further investigated in subgroups stratified by age (<60 or≥60 y), sex (male or female), BMI (<24.5 kg/m2 and ≥24.5 kg/m2 ), smoking habits, body weight directions (decreasing, stable, and increasing direction), diabetes (with or without), and hypertension (with or without). No significant interactions were found in any of the subgroups, suggesting that the relationship between increased body weight variabilities and the risk of rapid eGFR declines were consistently significant across these subgroups (Figure 2).

Sensitivity analysis

Several subsequent evaluations were performed for sensitivity analyses. First, assessments were performed excluding participants with a follow-up duration < 5 years, those with CVD, or those with a BMI > 35 (Supporting Information Tables S2-S4). An evaluation was also done including participants with a follow-up interval greater than 4 years (Supporting Information Table S5). Second, assessments were further performed using other measures of variability, including ASV percentage, ASBMIV, standard deviation, coefficient of variation, and residual standard deviation (Supporting Information Table S6). These evaluations yielded results consistent with the main analyses. Third, when rapid eGFR decline was more stringently defined as an eGFR decline rate > 5 mL/min/1.73 m2 per year, the results were also in concordance with the main findings (Supporting Information Table S7). Fourth, an analysis using CKD-EPI-derived eGFR values also revealed similar findings to the main results (Supporting Information Table S8). Finally, when the trajectory of eGFR was classified into three distinct patterns (Supporting Information Table S9, Supporting Information Figure S6), higher ASV values were found to be significantly related to the most rapidly declining trajectory class (Supporting Information Table S10).

DISCUSSION

In this study, body weight fluctuations were prevalent among the general population with normal kidney function. They were associated with an increased risk of rapid eGFR decline, which was independent of baseline body weight and traditional risk factors. The association was also significant regardless of overall weight gain or loss during the follow-up period.

Obesity is a risk factor for CKD development and progression. Population-based studies have repeatedly documented an independent association between BMI and the risk of incident CKD (22,23). In addition, excessive body weight and adiposity have been found to be associated with rapid disease progression of CKD resulting from various etiologies (24). Recently, weight loss induced by surgical interventions in patients with obesity and altered kidney function has been shown to normalize eGFR (25). However, studies involving nonsurgical interventions for weight loss have reported conflicting results (6,26). In a prospective intervention of patients with diabetes, obesity, and CKD, a hypocaloric diet for 1 year resulted in weight reduction and an improvement in eGFR (27). However, in a study of non-diabetic women with obesity, although lifestyle modifications resulted in BMI decreases, eGFRs remained stable (28). In another study, a 1-year structured weight-loss program resulted in significant weight loss accompanied by significant eGFR reduction (29). One possible explanation for these various weight-loss effects on kidney function may be due to the fact that weight change variability during follow-up was not considered in these previous studies. This possibility is supported by the finding of the current study, which shows that participants allocated to the highest ASV tertile were associated with having an increased risk of rapid eGFR decline, although an overall increase in body weight during follow-up was noted in only 7% of these participants. The result of this study implies that, in addition to the well-known risk of obesity, frequent body weight change could also be a factor accelerating kidney function decline. Therefore, previous trends in body weight should also be taken into account along with the current body weight when stratifying CKD risk. Moreover, when weight loss is recommended to lower the risk of CKD, it will be imperative to take steps to prevent rebound weight gain from occurring. Nonetheless, further prospective clinical investigations would be needed to evaluate whether minimizing weight fluctuations helps avoid rapid kidney function decline.

Cistanche improve kidney function

Body weights were significantly higher in individuals with greater body weight variability during follow-up. This could have affected the increased risk of rapid kidney function decline found in those with increased body weight variability in several ways. First, there is the possibility that the obesity itself, rather than the increased weight variability, could have been the key effector of the increased risk of rapid kidney function decline. However, considering that the risk elevation was still significant even after adjustments were made for baseline body weight, the likelihood of this possibility having a substantial impact is less probable. Another point to consider is that the impact of ASV would differ according to the baseline bodyweight of the individual. The same amount of weight change would be more influential in participants with lower baseline body weights than in those with higher baseline body weights. Therefore, the impact of body weight fluctuations as a percentage of change of baseline weight was evaluated in the sensitivity analyses. These sensitivity analyses yielded similar results to the main analysis, suggesting that weight variability itself, rather than the baseline body weight, played a meaningful role in determining the risk of rapid kidney function decline.

Preexisting comorbidities could have been another cause of the observed body weight fluctuations. Comorbidities, including hypertension, diabetes, and CVDs, were more common in participants with higher ASV values. Furthermore, lifestyle factors that negatively influenced overall health, including positive smoking history and lower daily physical activity, were more prevalent among those with higher ASV values. These findings leave open the possibility that the weight loss observed in these participants was a consequence of illness. However, the fact that overall weight gain, rather than weight loss, was more prevalent in the group with more comorbidities lowers the chance that the weight loss associated with bodyweight fluctuations, was attributable to illness. Additionally, the association between body weight fluctuations and rapid eGFR declines was significant even after adjustments were made for underlying comorbidities and lifestyle factors, suggesting that fluctuations in body weight could have affected kidney function regardless of their overall health condition.

The prevalence, as well as the incidence of hypertension and diabetes, showed a slight but gradual increase with the increase in ASV. This finding is in accordance with previous reports that have shown significant associations between body weight fluctuations and diabetes, hypertension, or metabolic syndromes (30-32). The presence of hypertension and diabetes could have had some influence on kidney function deterioration related to body weight fluctuations. Hypertension and diabetes are well-known factors affecting kidney function (23). Therefore, sequential relationships in which body weight fluctuations increase the risk of kidney function decline through the presence of hypertension and diabetes could be probable. Nonetheless, the significant relationship between ASV and rapid kidney function decline was maintained even after adjusting for SBP and HOMA-IR. Therefore, the possibility of hypertension- and diabetes-independent effect of body weight fluctuations on kidney function should also be considered.

The mechanism responsible for the effect of body weight fluctuations on rapid kidney function decline is unclear, but several possibilities should be considered. Bodyweight fluctuations increase fasting insulin levels and aggravate insulin resistance (30,33). Insulin resistance activates the renin-angiotensin-aldosterone system, increases oxidative stress, and decreases endothelial nitric oxide production, which are well-recognized factors in inducing kidney pathology (34,35). Thus, insulin resistance could be an intermediate between body weight fluctuations and kidney function de-TABLE 2 cline. Another possibility to consider is the association with blood pressure. Animal studies have demonstrated that body weight fluctuations suppress the normal day-night differences in blood pressure, resulting in a “non-dipping” pattern (36,37), which is associated with a faster progression of kidney insufficiency (38). Further investigations with 24-hour ambulatory blood pressure monitoring in individuals experiencing body weight fluctuations are necessary to assess this possibility. Repeated overshoots in glomerular pressure could also play a role. Considering that glomerular filtration rate is modulated by food intake, fluctuations in food intake associated with bodyweight cycling would lead to glomerular filtration rate fluctuations (39). The long-term harm of increased glomerular pressure has been well described (40). Additionally, increased glomerular pressure during the weight-gain period could aggravate the kidneys, allowing greater vulnerability to traditional risk factors. A previous report showing that hyperlipidemia synergistically aggravates glomerular sclerosis when glomerular hypertension is present supports this possibility (41).

This study has several strengths. First, data were obtained from a prospective cohort with up to 12 years of follow-up. Second, various confounding factors that affect kidney function, including lifestyle and socioeconomic variables, were included in the analyses. Third, body weight was measured with a calibrated scale at every visit, and self-reported values were not used. However, this study also has several limitations. First, differentiating intentional weight loss from unintentional weight loss was not possible. Abrupt weight loss could be a result of other illnesses that affected kidney function. However, participants with incident cancers or histories of malignancies, which would most likely result in unintentional weight loss, were excluded from the analysis. Moreover, whereas intentional weight loss may be triggered by obesity-related conditions, participants in the higher ASV tertile groups had higher BMI levels and had diabetes and hypertension more frequently. Second, despite the prospective cohort design, only an association, not causation, could be inferred because of the observational nature of the study. Third, there is a possibility that genetic or environmental factors may have functioned as unmeasured biases and that these factors could not be taken into account. Finally, regardless of the frequent periodic data attainment, which lowers the chance of bias, the interval between the measurements, as well as the number of measurements, would inevitably have some effect on the relationship between ASV and rapid eGFR decline, owing to the discrete nature of the data.

In conclusion, in this community-based cohort with normal kidney function, body weight fluctuations were associated with a significant increase in the risk of rapid kidney function decline. This risk increase was independent of traditional risk factors and overall body weight gain or loss over time. Further studies are warranted to elucidate the mechanism underlying the relationship between body weight fluctuations and rapid kidney function deterioration.

ACKNOWLEDGMENTS

The epidemiologic data used in this study were obtained from the Korean Genome and Epidemiology Study (KoGES; 4851–302) of the National Research Institute of Health, Centers for Disease Control and Prevention, Ministry for Health and Welfare, Republic of Korea.

CONFLICT OF INTEREST

The authors declared no conflict of interest.

AUTHOR CONTRIBUTIONS YSJ and JTP conceived and designed the study; YSJ, KHN, HRY, SL, JHJ, and JTP acquired the data; YSJ, KHN, HRY, SL, JHJ, SHH, THY, JTP, and SWK analyzed the data; YSJ and KHN created the figures; YSJ, KHN, and JTP drafted the manuscript; SHH, THY, and SWK supervised the study. All authors contributed to important intellectual content during manuscript drafting or revision and accepted accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work were appropriately investigated and resolved.

From: 'Bodyweight fluctuation is associated with rapid kidney function decline' by Young Su Joo, et al

---Obesity (Silver Spring). 2022;30:257–267.

DOI: 10.1002/oby.23326

REFERENCES

1. Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med. 2004;351(13):1296-1305.
2. Jha V, Garcia-Garcia G, Iseki K, et al. Chronic kidney disease: global dimension and perspectives. Lancet. 2013;382(9888):260-272.
3. Fox CS, Larson MG, Leip EP, Culleton B, Wilson PW, Levy D. Predictors of new-onset kidney disease in a community-based population. JAMA. 2004;291(7):844-850.
4. Park S, Lee S, Kim Y, et al. Reduced risk for chronic kidney disease after recovery from metabolic syndrome: a nationwide population-based study. Kidney Res Clin Pract. 2020;39(2):180-191.
5. Ryu S, Chang Y, Woo H-Y, et al. Changes in body weight predict CKD in healthy men. J Am Soc Nephrol. 2008;19(9):1798-1805.
6. Morales E, Valero MA, León M, Hernández E, Praga M. Beneficial effects of weight loss in overweight patients with chronic proteinuric nephropathies. Am J Kidney Dis. 2003;41(2):319–327.
7. Bonnet F, Deprele C, Sassolas A, et al. Excessive body weight as a new independent risk factor for clinical and pathological progression in primary IgA nephritis. Am J Kidney Dis. 2001;37(4):720-727.
8. Kidney Disease: Improving Global Outcomes CKD Work Group. KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int. 2013;3:73-90.
9. Bangalore S, Fayyad R, Laskey R, DeMicco DA, Messerli FH, Waters DD. Body-weight fluctuations and outcomes in coronary disease. N Engl J Med. 2017;376(14):1332-1340.
10. Zhang Y, Yatsuya H, Li Y, et al. Long-term weight-change slope, weight fluctuation and risk of type 2 diabetes mellitus in middle-aged Japanese men and women: findings of Aichi Workers' Cohort Study. Nutr Diabetes. 2017;7(3):e252. doi:10.1038/nutd.2017.5
11. Lissner L, Odell PM, D'Agostino RB, et al. Variability of body weight and health outcomes in the Framingham population. N Engl J Med. 1991;324(26):1839-1844.
12. Delahanty LM, Pan Q, Jablonski KA, et al. Effects of weight loss, weight cycling, and weight loss maintenance on diabetes incidence and change in cardiometabolic traits in the Diabetes Prevention Program. Diabetes Care. 2014;37(10):2738-2745.
13. Kim Y, Han BG. Cohort profile: the Korean Genome and Epidemiology Study (KoGES) consortium. Int J Epidemiol. 2017;46:e20. doi:10.1093/ije/dyv316
14. Ahn Y, Kwon E, Shim JE, et al. Validation and reproducibility of food frequency questionnaire for Korean genome epidemiologic study. Eur J Clin Nutr. 2007;61(12):1435-1441.
15. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med. 1999;130(6):461-470.
16. Levey AS, Coresh J, Greene T, et al. Expressing the Modification of Diet in Renal Disease study equation for estimating glomerular filtration rate with standardized serum creatinine values. Clin Chem. 2007;53(4):766-772.
17. Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604-612.
18. 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.
19. Rifkin DE, Shlipak MG, Katz R, et al. Rapid kidney function decline and mortality risk in older adults. Arch Intern Med. 2008;168(20):2212-2218.
20. Proust-Lima C, Letenneur L, Jacqmin-Gadda H. A nonlinear latent class model for joint analysis of multivariate longitudinal data and a binary outcome. Stat Med. 2007;26(10):2229-2245.
21. Lennon H, Kelly S, Sperrin M, et al. Framework to construct and interpret latent class trajectory modeling. BMJ Open. 2018;8(7):e020683. doi:10.1136/BMJ open-2017-020683
22. Lai YJ, Hu HY, Lee YL, Ku PW, Yen YF, Chu D. Association between obesity and risk of chronic kidney disease: a nationwide cohort study in Taiwan. Nutr Metab Cardiovasc Dis. 2017;27(11):1008-1014.
23. Ejerblad E, Fored CM, Lindblad P, Fryzek J, McLaughlin JK, Nyren O. Obesity and risk for chronic renal failure. J Am Soc Nephrol. 2006;17(6):1695-1702.
24. Ogna A, Forni Ogna V, Bochud M, et al. Association between obesity and glomerular hyperfiltration: the confounding effect of smoking and sodium and protein intakes. Eur J Nutr. 2016;55(3):1089-1097.
25. Imam TH, Fischer H, Jing B, et al. Estimated GFR before and after bariatric surgery in CKD. Am J Kidney Dis. 2017;69(3):380-388.
26. Straznicky NE, Grima MT, Lambert EA, et al. Exercise augments weight loss-induced improvement in renal function in obese metabolic syndrome individuals. J Hypertens. 2011;29(3):553-564.
27. Moncrieft AE, Llabre MM, McCalla JR, et al. Effects of a multi-component lifestyle intervention on weight, glycemic control, depressive symptoms, and renal function in low-income, minority patients with type 2 diabetes: results of the community approach to lifestyle
modification for diabetes randomized controlled trial. Psychosom Med. 2016;78(7):851-860.
28. Gilardini L, Zulian A, Girola A, Redaelli G, Conti A, Invitti C. Predictors of the early impairment of renal disease in human obesity. Int J Obes (Lond). 2010;34(2):287-294.
29. Cook SA, MacLaughlin H, Macdougall IC. A structured weight management program can achieve improved functional ability and significant weight loss in obese patients with chronic kidney disease. Nephrol Dial Transplant. 2008;23(1):263-268.
30. Zhang H, Tamakoshi K, Yatsuya H, et al. Long-term body weight fluctuation is associated with metabolic syndrome independent of current body mass index among Japanese men. Circ J. 2005;69(1):13-18.
31. Field AE, Manson JE, Laird N, Williamson DF, Willett WC, Colditz GA. Weight cycling and the risk of developing type 2 diabetes among adult women in the United States. Obes Res. 2004;12(2):267-27