Rationale & objective
As the prevalence of obesity continues to rise in the United States, it is important to understand its impact on the lifetime risk of chronic kidney disease (CKD).
The CKD Health Policy Model was used to simulate the lifetime risk of CKD for those with and without obesity at baseline. Model structure was updated for glomerular filtration rate (GFR) decline to incorporate new longitudinal data from the Chronic Renal Insufficiency Cohort (CRIC) study.
Setting and population
The updated model was populated with a nationally representative cohort from National Health and Nutrition Examination Survey (NHANES).
Model, perspective, & timeframe
Simulation model following up individuals from current age through death or age 90 years.
Lifetime risk of any CKD stage was 32.5% (95% CI 28.6%–36.3%) for persons with normal weight, 37.6% (95% CI 33.5%–41.7%) for persons who were overweight, and 41.0% (95% CI 36.7%–45.3%) for persons with obesity at baseline. The difference between persons with normal weight and persons with obesity at baseline was statistically significant (p<0.01). Lifetime risk of CKD stages 4 and 5 was higher for persons with obesity at baseline (Stage 4: 2.1%, 95% CI 0.9%–3.3%; stage 5: 0.6%, 95% CI 0.0%–1.1%), but the differences were not statistically significant (stage 4: p = 0.08; stage 5: p = 0.23).
Due to limited data, our simulation model estimates are based on assumptions about the causal pathways from obesity to CKD, diabetes, and hypertension.
Citation: Yarnoff BO, Hoerger TJ, Shrestha SS, Simpson SK, Burrows NR, Anderson AH, et al. (2018) Modeling the impact of obesity on the lifetime risk of chronic kidney disease in the United States using updated estimates of GFR progression from the CRIC study. PLoS ONE 13(10): e0205530. https://doi.org/10.1371/journal.pone.0205530
Editor: Tatsuo Shimosawa, International University of Health and Welfare, School of Medicine, JAPAN
Received: July 10, 2018; Accepted: September 26, 2018; Published: October 19, 2018
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This research was supported by funding (Contract #200-2008-27958 to TJH) from the Centers for Disease Control and Prevention (CDC). The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of CDC. Funding for the CRIC Study was obtained under a cooperative agreement from National Institute of Diabetes and Digestive and Kidney Diseases (U01DK060990, U01DK060984, U01DK061022, U01DK061021, U01DK061028, U01DK060980, U01DK060963, and U01DK060902). In addition, this work was supported in part by: the Perelman School of Medicine at the University of Pennsylvania Clinical and Translational Science Award NIH/NCATS UL1TR000003, Johns Hopkins University UL1 TR-000424, University of Maryland GCRC M01 RR-16500, Clinical and Translational Science Collaborative of Cleveland, UL1TR000439 from the National Center for Advancing Translational Sciences (NCATS) component of the National Institutes of Health and NIH roadmap for Medical Research, Michigan Institute for Clinical and Health Research (MICHR) UL1TR000433, University of Illinois at Chicago CTSA UL1RR029879, Tulane COBRE for Clinical and Translational Research in Cardiometabolic Diseases P20 GM109036, Kaiser Permanente NIH/NCRR UCSF-CTSI UL1 RR-024131.
Competing interests: The authors have declared that no competing interests exist.
The prevalence of obesity (body mass index [BMI] ≥ 30 kg/m2) in the United States has increased significantly since the 1980s, contributing to higher risk for chronic diseases and mortality.[1–3] As a major risk factor for diabetes, hypertension, and cardiovascular disease, obesity is an important contributor to the risk for chronic kidney disease (CKD).[4–8] Furthermore, the risk of CKD onset and progression may be directly increased by obesity through adaptive glomerular changes that frequently evolve into pathologic alterations. These observations suggest that obesity may have a significant impact on the lifetime risk of CKD. However, the evidence for an independent association (i.e. apart from the impact of obesity on hypertension and diabetes) between obesity and CKD is mixed.[8, 10–13] Further, data must contain observations of persons over a long time period to assess the impact of risk factors such as obesity on the lifetime risk of CKD.
The long-term dynamics of these pathways on the lifetime risk of CKD is best assessed with a simulation model, because the length of time required to estimate lifetime risk with observed data is not feasible. The CKD Health Policy Model, a microsimulation model of CKD progression, has been used previously to examine the burden of CKD, and has been extensively validated. Here we used this model to examine the impact of obesity on the lifetime risk of CKD.
This study used the CKD Health Policy Model, a microsimulation model of CKD progression.[14–17] Briefly, the model simulates progression of CKD and its complications in a nationally representative cohort drawn from the National Health and Nutrition Examination Survey (NHANES) through age 90 years or death. The model includes eight states: no CKD, CKD stages 1 through 5 (with stage 3 divided into 3a and 3b), and death. CKD stages are defined by estimated glomerular filtration rates (eGFR) and kidney damage as measured by the presence of albuminuria (urinary albumin to creatinine ratio ≥30 mg/g). CKD stages are largely sequential based on GFR, but stages 1 and 2 also require kidney damage. So, in the simulations, persons who reach stage 3b must first go through stage 3a; similarly, those who reach stage 4 must go through stages 3a and 3b, and those who reach stage 5 must go through stages 3a, 3b, and 4. However, it is possible to reach stage 3a or higher without going through stages 1 and 2 if a person never has kidney damage as measured by elevated albuminuria. The model concomitantly simulates the natural history of complications from CKD. Model parameters are derived from the epidemiological literature, clinical trials, and new analysis of eGFR progression from the Chronic Renal Insufficiency Cohort (CRIC) study.
Change in eGFR
Previously, the eGFR component of the model was based on another cost-effectiveness study by Boulware et al. This structure used the mean change in eGFR over time for persons with different sets of risk factors (diabetes, hypertension, and proteinuria). For this study, we updated this component of the model, because model parameters were potentially out of date and the structure was deterministic with no inclusion of the variation in eGFR change across persons with the same risk factors. Variation in eGFR change is potentially an important driver of outcomes, because progression to severe stages of CKD such as stage 5 and ESRD is likely driven by the tail of the distributions rather than the mean.
To update the eGFR component of the model, we analyzed data on eGFR change from the CRIC study. The CRIC study is a longitudinal study of persons with CKD that began in 2001 with regular follow-up over 7 to 14 years. It was established to help inform the understanding of CKD and related cardiovascular disease. Importantly, one of its primary goals is to examine risk factors for decline in eGFR. The analysis sample from the CRIC study consisted of 3,302 persons. Table 1 presents descriptive statistics for the sample.
We used a mixed-effect regression model to estimate the association of clinical risk factors (hypertension, diabetes, proteinuria, and obesity) and demographic factors (sex and race) with annual change in eGFR and the variance of annual change in eGFR. We estimated the regression equations separately for persons with eGFR < 60 at baseline and eGFR ≥ 60 at baseline. We took coefficients from these regression equations as model parameters for the annual change in eGFR associated with sex, race, hypertension, diabetes, proteinuria, and obesity. We took the estimated variance of the error term in the mixed-effect model as the standard deviation of the distribution of annual change in eGFR. Table 2 lists model parameters for annual change in eGFR. After updating model parameters, we performed model calibration by comparing simulated output to CKD prevalence estimates in NHANES and comparing simulated and actual eGFR decline for a cohort from the Atherosclerosis Risk in Communities (ARIC) study. We found that simulated estimates of the ARIC cohort lined up well with actual eGFR decline for persons with baseline eGFR < 60, but overestimated the decline for persons with baseline eGFR ≥ 60. This is a logical outcome, because the sample of persons in the CRIC study with baseline eGFR ≥ 60 was small and not designed to be representative of that group in the broader population. Therefore, estimates of the variance in eGFR decline for that group in the sample was much larger than observed in the ARIC cohort. We calibrated the model to reflect this with a multiplier to bring the variance term for persons with baseline eGFR ≥ 60 in line with the variance observed for this group in the ARIC cohort.
Overweight and obesity
We updated the model by adding an obesity module that modeled the impact of overweight and obesity on diabetes, hypertension, and cardiovascular disease (CVD) and annual changes in BMI (Table 3). Parameters for the impact on risk factors and CVD are drawn from Wilson et al. which used data from the Framingham study to estimate the relationship between obesity and other conditions. We considered including parameters for the effect of obesity on non-CVD mortality, but found that studies with a full set of controls estimated no effect on non-CVD mortality. Finally, we updated the model to simulate changes in BMI over time. We conducted a literature review to identify studies that estimated average annual change in BMI. From this literature review we identified parameters for average annual change in BMI based on age, sex, and race (Table 3). We calibrated the model based on simulated versus actual change in BMI for the ARIC cohort.
Lifetime incidence of CKD
The lifetime incidence of CKD was estimated for persons with normal weight, overweight, and obesity at baseline. We conducted simulations overall and by baseline age (30 to 49, 50 to 64, and ≥65 years). Simulations were conducted with cohorts from NHANES 1999–2010, using a person’s starting age, sex, race/ethnicity, eGFR, albuminuria status (normal, moderately increased albuminuria, or severely increased albuminuria), diabetes status, hypertension status, cardiovascular disease status, and BMI. eGFR was computed using the CKD-Epi equation. Because CKD staging is based on persistent albuminuria and NHANES only includes one observation of albuminuria per person, we adjusted observed moderately increased albuminuria using an algorithm proposed by Coresh et al.
We simulated each person’s progression through the model until death or age 90. The starting year for the simulation is 2010. During each year, a person can develop diabetes, hypertension, overweight, obesity, or albuminuria. We estimated the lifetime risk of CKD as the probability of reaching any stage of CKD (stages 1–5). We also estimated lifetime risk of reaching each stage by death or age 90. We used bootstrapping to estimate 95% confidence intervals, simulating the lifetime risk 100 times. From this sample, we used the 2.5 and 97.5 percentile results to calculate confidence intervals.
To test the sensitivity of our results and conclusions to the choice of parameters for risks associated with obesity, we conducted six one-way sensitivity analyses by varying key model parameters by ±25%: the relative risk of diabetes for obesity, the relative risk of hypertension for obesity, and the annual change in eGFR for persons with obesity. For each test, we examined the difference in lifetime risk of CKD between persons with obesity and normal weight at baseline. These parameters relate to the impact of obesity on CKD progression and other risk factors for CKD progression, so varying them tests the sensitivity of results to these risks.
Fig 1 presents results of the lifetime risk of any-stage CKD and the highest CKD stage reached for persons with normal weight, overweight, or obesity at baseline. Lifetime risk of any-stage CKD was 32.5% (95% CI 28.6%–36.3%) for normal weight, 37.6% (95% CI 33.5%–41.7%) for overweight, and 41.0% (95% CI 36.7%–45.3%) for persons with obesity at baseline. The difference in lifetime risk of CKD between persons with normal weight and obesity at baseline was statistically significant (p<0.01). Persons with baseline obesity had a higher estimate of lifetime risk of CKD stages 4 (2.1%, 95% CI 0.9%–3.3%) and 5 (0.6%, 95% CI 0.0% -1.1%) than persons with normal weight (stage 4: 1.2%, 95% CI 0.3%–2.0%; stage 5: 0.3%, 95% CI 0.0%–0.7%). While these differences are relatively large (75% higher for stage 4 and 100% higher for stage 5), they were not statistically significant due to the challenge of getting sufficient precision with such small initial percentages (stage 4: p = 0.08; stage 5: p = 0.23).
The difference in lifetime risk or any CKD between persons with obesity and normal weight was statistically significant at the 1% level. No other differences were statistically significant. Abbreviations: CKD = Chronic Kidney Disease, BMI = Body Mass Index.
Fig 2 presents the lifetime risk of any-stage CKD by baseline BMI and age categories. Lifetime risk of CKD increased with age, and within age groups estimates increased with BMI category. Obesity was significantly associated with a higher risk of CKD only among persons 50–64 years old [47.1%, 95% CI 43.3%–51.0% with obesity vs. 37.8%, 95% CI 34.1%–41.5% with normal weight, p<0.01).
The difference in lifetime risk of any CKD between persons age 50–64 with obesity and normal weight was statistically significant at the 1% level. No other differences were statistically significant. Abbreviations: CKD = Chronic Kidney Disease, BMI = Body Mass Index.
Fig 3 shows the results of one-way sensitivity analysis for 25% changes in parameter estimates on the difference in lifetime risk of CKD between persons with obesity and normal weight at baseline. These parameters relate to the impact of obesity on CKD progression and other risk factors for CKD progression. Varying the parameters for the impact of obesity on hypertension and diabetes had the greatest impact on the estimated difference in lifetime risk of CKD between persons with obesity and normal weight at baseline. However, differences were relatively modest, demonstrating that results are not sensitive to the assumptions of these model parameters. Model parameters would have to differ by much more than the 25% tested here to substantially change results.
In the main analysis, the difference in lifetime risk between persons with normal weight and obesity was 8.5%. Bars to the left of this value in the figure show the change in results if parameters are decreased by 25% and bars to the right of this value show the change in results if the parameters are increased by 25%. Abbreviations: eGFR = estimated glomerular filtration rate.
This study presents the first simulation modeling of the impact of obesity on lifetime risk of CKD. This is an important innovation, because simulation modeling is essential to evaluate a time period sufficient for estimating lifetime risk. We found that compared with normal weight, obesity significantly increases the lifetime risk of any-stage CKD. Obesity was also found to be associated with lifetime risk of CKD stages 4 and 5. In the 50–64 age group, the lifetime risk of any-stage CKD was 9.3 percentage points higher in those with obesity than in those with normal weight. This is likely because persons in the younger age group with normal weight at baseline have a higher likelihood of developing obesity over time, mitigating differences with the group of persons with obesity at baseline, and because persons in the older age group have less time to develop CKD. While results were not statistically significant at the 5% level, the lifetime risk of CKD stages 4 and 5 could be doubled for those with baseline obesity, which might be translated to future high health and cost burden. For example, the estimated excess lifetime risk of stage 5 CKD would increase the cases of stage 5 CKD by approximately 350,000 at the current rate of adult obesity (36.5%), which could be expected to cost approximately $21 billion annually.
The literature indicates that this association may be driven either through obesity’s impact on diabetes, hypertension, and cardiovascular disease,[4–8] or directly through obesity’s impact on GFR.9 However, some studies have found that after controlling for diabetes and hypertension the direct effect of obesity on GFR is minimal.[8,10–13] We tested both a direct effect of obesity on CKD and an indirect effect through diabetes and hypertension. The results from this study of long-term measures of CKD are in keeping with the literature on short-term and intermediate measures. Namely, there was a significant association between obesity and lifetime risk of CKD, but sensitivity analysis showed that even when the direct effect of obesity on CKD is significantly higher than in the literature the relationship is driven almost exclusively by the effect on diabetes and hypertension.
This study is limited by the availability of parameters in the scientific literature. Although model parameters were based on current scientific literature, they may be imperfect or may omit additional unknown factors. For example, despite much research being done on the causal pathways from obesity to CKD, diabetes, and hypertension, much remains uncertain about the true impact. As this research develops, the model can be updated to include any new evidence that is generated.
The results of this study indicate that obesity may have an impact on the lifetime risk of CKD. These differences may be large and may have important implications for the expected future healthcare costs associated with CKD from rising rates of obesity. Results further suggest that obesity prevention may have important additional health and cost benefits through reducing the risk of CKD. This is important information for policymakers seeking to set priorities and targets for CKD prevention and treatment. The knowledge of factors that do and do not impact the burden of CKD allows policymakers to direct prevention efforts in the most impactful manner.
The CRIC Study Investigators are Harold I. Feldman, MD, MSCE, University of Pennsylvania (Principal investigator, firstname.lastname@example.org); Lawrence J. Appel, MD, MPH, Johns Hopkins University; Alan S. Go, MD, Kaiser Permanente of Northern California; Jiang He, MD, PhD, University of California, San Fransisco; John W. Kusek, PhD, National Institute of Diabetes, Digestive, & Kidney Diseases, NIH; James P. Lash, MD, University of Illinois at Chicago; Panduranga S. Rao, MD, University of Michigan; Mahboob Rahman, MD, University Hospitals Case Medical Center; and Raymond R. Townsend, MD, University of Pennsylvania.
- 1. National Center for Health Statistics:Health, United States, 2015: With Special Feature on Racial and Ethnic Health Disparities, Government Printing Office, 2016.
- 2. Flegal KM, Kit BK, Orpana H, Graubard BI: Association of all-cause mortality with overweight and obesity using standard body mass index categories: a systematic review and meta-analysis. JAMA, 309: 71–82, 2013. pmid:23280227
- 3. The Global BMI Mortality Collaboration: Body-mass index and all-cause mortality: individual-participant-data meta-analysis of 239 prospective studies in four continents. The Lancet, 388: 776–786, 2016.
- 4. Wilson PW, D’Agostino RB, Sullivan L, Parise H, Kannel WB: Overweight and obesity as determinants of cardiovascular risk: the Framingham experience. Archives of Internal Medicine, 162: 1867–1872, 2002. pmid:12196085
- 5. Garrison RJ, Kannel WB, Stokes J, Castelli WP: Incidence and precursors of hypertension in young adults: the Framingham Offspring Study. Preventive Medicine, 16: 235–251, 1987. pmid:3588564
- 6. Ford ES, Williamson DF, Liu S: Weight change and diabetes incidence: findings from a national cohort of US adults. American Journal of Epidemiology, 146: 214–222, 1997. pmid:9247005
- 7. Resnick HE, Valsania P, Halter JB, Lin X: Relation of weight gain and weight loss on subsequent diabetes risk in overweight adults. Journal of Epidemiology and Community Health, 54: 596–602, 2000. pmid:10890871
- 8. Hall ME, do Carmo JM, da Silva AA, Juncos LA, Wang Z, Hall JE: Obesity, hypertension, and chronic kidney disease. International Journal of Nephrology and Renovascular Disease, 7: 75–88, 2014. pmid:24600241
- 9. D’Agati VD, Chagnac A, de Vries AP, Levi M, Porrini E, Herman-Edelstein M, et al.: Obesity-related glomerulopathy: clinical and pathologic characteristics and pathogenesis. Nature Reviews Nephrology, 2016.
- 10. Elsayed EF, Sarnak MJ, Tighiouart H, Griffith JL, Kurth T, Salem DN, et al.: Waist-to-hip ratio, body mass index, and subsequent kidney disease and death. American Journal of Kidney Diseases, 52: 29–38, 2008. pmid:18511168
- 11. Stengel B, Tarver–Carr ME, Powe NR, Eberhardt MS, Brancati FL: Lifestyle factors, obesity and the risk of chronic kidney disease. Epidemiology, 14: 479–487, 2003. pmid:12843775
- 12. de Boer IH, Katz R, Fried LF, Ix JH, Luchsinger J, Sarnak MJ, et al.: Obesity and change in estimated GFR among older adults. American Journal of Kidney Diseases, 54: 1043–1051, 2009. pmid:19782454
- 13. Foster MC, Hwang S-J, Larson MG, Lichtman JH, Parikh NI, Vasan RS, et al.: Overweight, obesity, and the development of stage 3 CKD: the Framingham Heart Study. American Journal of Kidney Diseases, 52: 39–48, 2008. pmid:18440684
- 14. Hoerger TJ, Simpson SA, Yarnoff BO, Pavkov ME, Rios Burrows N, Saydah SH, et al.: The future burden of CKD in the United States: a simulation model for the CDC CKD Initiative. American Journal of Kidney Diseases, 65: 403–411, 2015. pmid:25468386
- 15. Hoerger TJ, Wittenborn JS, Segel JE, Burrows NR, Imai K, Eggers P, et al.: A health policy model of CKD: 2. The cost-effectiveness of microalbuminuria screening. American Journal of Kidney Diseases, 55: 463–473, 2010. pmid:20116910
- 16. Hoerger TJ, Wittenborn JS, Segel JE, Burrows NR, Imai K, Eggers P, et al.: A health policy model of CKD: 1. Model construction, assumptions, and validation of health consequences. American Journal of Kidney Diseases, 55: 452–462, 2010. pmid:20116911
- 17. Hoerger TJ, Wittenborn JS, Zhuo X, Pavkov ME, Burrows NR, Eggers P, et al.: Cost-effectiveness of screening for microalbuminuria among African Americans. Journal of the American Society of Nephrology, 23: 2035–2041, 2012. pmid:23204444
- 18. Kopple JD: National kidney foundation K/DOQI clinical practice guidelines for nutrition in chronic renal failure. American Journal of Kidney Diseases, 37: S66–70, 2001. pmid:11158865
- 19. Boulware LE, Jaar BG, Tarver-Carr ME, Brancati FL, Powe NR: Screening for proteinuria in US adults: a cost-effectiveness analysis. JAMA, 290: 3101–3114, 2003. pmid:14679273
- 20. Lewis CE, Jacobs DR, McCreath H, Kiefe CI, Schreiner PJ, Smith DE, et al.: Weight gain continues in the 1990s: 10-year trends in weight and overweight from the CARDIA study. American Journal of Epidemiology, 151: 1172–1181, 2000. pmid:10905529
- 21. Botoseneanu A, Liang J: Social stratification of body weight trajectory in middle-age and older Americans: results from a 14-year longitudinal study. Journal of Aging and Health: 0898264310385930, 2010.
- 22. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF, Feldman HI, et al.: A new equation to estimate glomerular filtration rate. Annals of Internal Medicine, 150: 604–612, 2009. pmid:19414839
- 23. Coresh J, Selvin E, Stevens LA, Manzi J, Kusek JW, Eggers P, et al.: Prevalence of chronic kidney disease in the United States. JAMA, 298: 2038–2047, 2007. pmid:17986697
- 24. Ogden, CL, Carroll, MD, Fryar, CD, Flegal, KM: Prevalence of obesity among adults and youth: United States, 2011–2014, US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics, 2015.
- 25. Saran R, Robinson B, Abbott KC, Agodoa LY, Albertus P, Ayanian J,et al.: US Renal Data System 2016 annual data report: epidemiology of kidney disease in the United States. American Journal of Kidney Diseases, 69: A7–A8, 2017. pmid:28236831