End-stage kidney disease has a considerable human and social cost. Annual costs for medical care exceed $50,000 per person, and the 5-year survival is less than 50%. 1 There are few well-established risk factors for end-stage kidney disease. Diabetes and hypertension are strong predictors for the development and progression of chronic kidney diseases. 2–4 Experimental studies 5 have pointed to strong similarities between the process of systemic atherosclerosis and glomerulosclerosis (a common histologic end-point for chronic kidney diseases), suggesting analogous biologic mechanisms. We do not know whether glomerulosclerosis is caused by the intrarenal vascular disease associated with atherosclerosis, or whether it results from shared risk factors. However, it seems likely that the underlying risk factors for atherosclerosis, such as obesity, physical inactivity, smoking and alcohol consumption, might also predict the risk of chronic kidney disease.
Research on this question is sparse and inconclusive. Physical activity has been associated with a lower rate of nephropathy 6 and renal dysfunction 7 in diabetic patients, but its relation to the risk of chronic kidney disease in the general population is unknown. Smoking is a well-established risk factor for renal damage in diabetic patients, 8,9 but its role in other types of kidney disease is still debated. 10–14 Alcohol drinking has been associated with an increased risk of end-stage renal disease in the U.S. population. 15 The association of obesity (a major cause of hypertension) with abnormal kidney function 16 and focal segmental glomerulosclerosis 17,18 suggests that obesity may be a primary cause of chronic renal failure.
We analyzed data from a cohort study that passively followed a subset of participants in the second National Health and Nutrition Examination Survey (NHANES II) to test the hypotheses that physical inactivity, smoking, alcohol drinking and obesity are associated with the risk of chronic kidney disease in general, and with diabetic or hypertensive nephropathy in particular.
We constructed a population-based, nonconcurrent cohort study by linking several data sources. One was NHANES II, a cross-sectional survey conducted between 1976 and 1980, which collected extensive demographic, nutritional and health data in a probability sample of the U.S. civilian, noninstitutionalized population aged 6 months to 74 years. 19 Another was the Health Care Financing Administration’s end-stage renal disease Program Management and Medical Information System, also referred to as the Medicare end-stage renal disease registry. Finally, we used mortality data from the NHANES II Mortality Study, which ascertained subsequent vital status of 9,250 participants who were age 30-75 years at the time of their NHANES II examination. 20 We excluded 5 of these 9,250 individuals because they had treated end-stage kidney disease at baseline, according to the Medicare registry. We restricted the analysis to individuals identified as African-American or white, excluding 163 others because their chronic kidney disease risk was extremely heterogeneous. These exclusions left 9,082 subjects for analysis.
Exposure to Potential Risk Factors
The NHANES II survey included a structured interview, a standardized physical examination and laboratory tests. To assess physical activity, participants were asked whether they were getting “much exercise,” “moderate exercise” or “little exercise” for recreation, and whether, in their usual day, aside from recreation, they were “very active,” “moderately active” or “quite inactive.” We combined the activity categories in the two questions as follows. Persons getting “much exercise” or “very active” were classified as “very active” and used as the reference group; those reporting “little exercise” or being “quite inactive” were classified as “inactive;” all the others were considered “moderately active.” For smoking, participants were asked whether they had smoked at least 100 cigarettes during their entire life, whether they currently smoked, and the daily number of cigarettes smoked. We defined those who answered “yes” to the first two questions as current smokers, and those who responded “yes” to the first but “no” to the second as former smokers. Those who had smoked less than 100 cigarettes were categorized as nonsmokers. We then classified current smokers into two categories: those who smoked 20 or fewer cigarettes per day, and those who smoked more than 20 cigarettes per day. A food-frequency questionnaire provided information about the frequency of consumption of beer, wine and liquor. Participants were asked about their usual consumption of alcoholic beverages during the 3 months that preceded the interview. Consumption was classified as follows: never, seldom (less than once per week), weekly (one to six times per week) or daily (one or more times per day). The greatest frequency of reported alcohol consumption of any of the three types was used to assign an overall category of individual alcohol use. Body-mass index (BMI) was calculated from height and weight (kg/m2), and categorized according to the World Health Organization’s classification 21 as follows: thin (BMI < 18.5 kg/m2), normal weight (18.5 ≤ BMI < 25), overweight (25 ≤ BMI < 30), obese (30 ≤ BMI < 35) and morbidly obese (BMI ≥ 35).
We also recorded data about known renal risk factors that may be confounders or mediators for the relations we were studying. Seated blood pressure was measured twice during the physical examination, and averaged. Participants were considered hypertensive if they reported being told by a doctor that they had hypertension, or if they had a systolic blood pressure ≥140 mmHg or a diastolic blood pressure ≥90 mmHg. We considered participants to have diagnosed diabetes if they reported ever having been told by a doctor that they had diabetes, and a history of cardiovascular disease if they reported ever having had heart failure, a heart attack or a stroke. We analyzed serum cholesterol according to the protocol described for the Lipid Research Clinics program, 22 and measured serum creatinine with the Jaffé reaction.
We studied chronic kidney diseases, which we defined by either: (1) treatment of end-stage kidney disease due to any cause; or (2) death related to chronic kidney disease.
Treated cases of end-stage kidney disease were identified by obtaining Medicare end-stage renal disease registry data for NHANES II participants. The Medicare registry began on 1 January 1973 and contains records for 93% of all persons in the U.S. who have received renal replacement therapy since that date. Matching was conducted by the National Center for Health Statistics without knowledge of the participant’s exposure status through 31 December 1992, on the basis of the participant’s name (surname, first name, middle initial), date of birth, sex and race. The matching algorithm required exact or nearly exact matches on name and date of birth, except that in some instances, mismatches on surname were allowed for women. A quality-of-match variable was generated, and initial matches were classified into three categories: definite (exact match in all four fields or minor mismatch in one field, eg, different middle initial), probable (similar but inexact match in two fields, eg, surname spelled slightly differently and middle initial missing) and possible (eg, nonmatch for surname in women only, or similar but inexact match in three fields). Only individuals with a definite or a probable match were considered treated cases.
We identified chronic kidney-disease-related deaths using the NHANES II Mortality Study conducted by the National Center for Health Statistics. 20 Deaths were ascertained through 31 December 1992 by computerized matching to the Social Security Administration Death Master Files (1976–1988) and the National Death Index (1979–1992). Initial linkers were name, date of birth and sex. Matching was further reviewed manually according to the above criteria as well as race, state of birth, state of residence and marital status. Death certificates were abstracted to obtain the date, as well as the underlying and contributing causes of death. Deaths were classified as related to chronic kidney disease if any of the following codes of International Classification of Disease 9th Revision 23 was listed on the death certificate as the underlying or contributing cause: 250.4 (diabetes mellitus with nephropathy), 275.4 (nephrocalcinosis), 403 (hypertensive renal disease), 404 (hypertensive heart and renal disease), 580–589 (nephritis, nephrotic syndrome and nephrosis) and 593.9 (renal disease not otherwise specified).
Outcomes are mutually exclusive, so that patients who first received renal replacement therapy and then died of renal disease were classified as “treated end-stage kidney disease.”
Weighted prevalences of baseline characteristics and exposures were estimated with Software for Survey Data Analysis (SUDAAN) to account for the complex survey sampling. Incidence rates in person-years were calculated with weighted Poisson models in STATA 6.0. We used Kaplan–Meier survival methods to estimate the overall cumulative incidence in the cohort, 24 and constructed three Cox proportional hazards regression models to determine the relative risk of developing chronic kidney disease associated with exposures. 25 Model 1 included the four hypothesized exposures, ie, physical activity, smoking, alcohol consumption and BMI adjusted for age, sex and race; Model 2 included all the aforementioned variables, with further adjustment for serum cholesterol, systolic blood pressure and a history of diabetes, hypertension and cardiovascular disease; and Model 3 contained the variables from model 2, with adjustment for glomerular filtration rate, calculated with the prediction equation in Levey et al 26 based on age, sex, race and serum creatinine. In these models, age, systolic blood pressure, serum cholesterol and estimated glomerular filtration rate were analyzed continuously. With one exception, of all the variables used in this analysis, less than 0.5% were missing. The exception was serum creatinine, and consequently, calculated glomerular filtration rate, which was lacking for 30% of the participants.
To assess the robustness of the main results, we conducted several subsidiary analyses. First, to avoid possible intermingling of acute and chronic causes of renal failure, we conducted an analysis after excluding 23 incident cases with a diagnosis of acute renal failure on their death certificate; in the same way, because of possible misclassification of chronic kidney disease on death certificates, we conducted an analysis limited to incident cases that were treated for end-stage kidney disease, using Model 1 because of the small number of events. Second, to enhance the sensitivity of baseline assessment for diabetes, we limited our analyses to a subset of 3,182 subjects who underwent oral glucose-tolerance tests.
Third, to minimize the possibility of reverse causation (ie, the possibility that incipient chronic kidney disease at baseline would have modified exposure to the studied risk factors), we repeated the analyses after excluding subjects with a high serum creatinine at baseline, as defined by a value above 1.4 mg/dL (124 μmol/L) in men, and above 1.2 mg/dL (106 μmol/L) in women, and excluding incident cases that occurred within 5 years of follow-up. Finally, to enhance the clinical specificity of the outcomes, we divided chronic kidney disease into two broad categories: kidney disease due to diabetes or hypertension, and kidney disease due to other or unspecified causes.
Table 1 summarizes the baseline characteristics of the 9,082 NHANES II participants included in our study. Values represent U.S. prevalence estimates of lifestyle behaviors and health conditions from 1976 to 1980.
Incidence of Chronic Kidney Disease
Over the 12-16 years (13.2 years on average) covered by the study, 189 subjects developed chronic kidney disease (Table 2). Of these, 44 (23%) entered the Medicare end-stage renal disease program, 23 (12%) had kidney disease listed as an underlying cause of death, and 122 (65%) had it listed as a contributing cause. The overall cumulative incidence by age 75 was 8.9%. Diabetic and hypertensive nephropathies represented 50% of the assigned cases of treated end-stage kidney disease, and 26% and 12% of the assigned underlying and contributing causes of death, respectively.
Lifestyle Factors and Obesity as Risk Factors
After adjustment for age, sex, race, BMI, alcohol and smoking, the risk of chronic kidney disease was related to physical inactivity (Table 3). Compared with very active persons, those who were inactive had more than twice the risk of chronic kidney disease (Model 1). Further adjustment for self-reported diabetes, hypertension, systolic blood pressure, serum cholesterol, cardiovascular disease and baseline calculated glomerular filtration rate partly altered this association (Models 2 and 3). A graded relation was observed with smoking status (Model 1), even after controlling for the other variables (Models 2 and 3). Alcohol consumption was not related to the risk of chronic kidney disease, with the analysis by type of alcohol beverages showing similar patterns for beer, wine and liquor (data not shown).
The relation between BMI and the risk of chronic kidney disease was not graded, but those with morbid obesity had a risk of chronic kidney disease more than twice that of normal-weight persons, independent of age, sex, race, physical activity and smoking (Model 1). Adjustment for diabetes, hypertension and blood pressure notably weakened this association (Model 2).
When Model 2 was restricted to a subsample of the 6,341 subjects with a measurement of serum creatinine, the results were similar to those for the entire sample. This finding suggests that differences observed between Models 2 and 3 cannot be ascribed to differential subject selection. Adjustment for socioeconomic status, as defined by income and level of education, did not modify the above associations.
Excluding the 23 incident cases with acute renal failure did not alter the relations with physical inactivity (adjusted relative risk [RR] of inactive persons compared with very active persons in Model 1: 2.4 [95% confidence interval, 1.3–4.6]), smoking (RR of smokers of 1–20 cigarettes a day vs never smokers: 1.3 [0.7–2.5], and for those of >20 cigarettes a day: 2.3 [1.2–4.6]), alcohol consumption and BMI (RR of morbidly obese compared with normal weight persons: 2.2 [1.0–5.1]).This was also true when restricting the outcome to the 44 incident cases with treated end-stage kidney disease, except for the relation with smoking, which was more modest (Model 1: RR of smokers of 1–20 cigarettes a day compared with never smokers, 1.2 [0.3–4.3], and for those of more than 20 cigarettes a day, 1.2 [0.3–4.9]). Adjustment for biochemically defined diabetes did not modify the main study results. In particular, among the 3,182 participants who underwent the oral glucose-tolerance test, the RR for inactive persons compared with the RR for those who were very active was very similar, after adjusting for either self-reported diabetes (RR = 1.7 [0.7–4.5]) or biochemically defined diabetes (RR = 1.8 [0.7–4.9]). Exclusion neither of subjects with high serum creatinine at baseline, nor of incident cases that occurred within the first 5 years of follow-up modified any of the observed associations.
To test the hypothesis that the risk factors we studied were more strongly related to diabetic and hypertensive nephropathies than to the other types of kidney disease, we split the cases in two groups. Physical inactivity was more strongly associated with the risk of diabetic or hypertensive nephropathy than with that of other types of kidney disease (Table 4). BMI and alcohol consumption were similarly related to both types. The relation to smoking was weaker in those with diabetic or hypertensive nephropathy than in those with other types of kidney disease.
Our study suggests that physical inactivity, cigarette smoking and morbid obesity contribute to the development of chronic kidney disease in the general population. In contrast, we did not find that alcohol consumption was related to a higher risk of chronic kidney disease.
The most important obstacles to studying renal risk factors prospectively are the relatively low incidence and the long latency period of chronic kidney diseases. 27 Nonconcurrent cohort studies based on passive follow-up of individuals with previous extensive data collection, such as the NHANES II Mortality Study and its complementary renal morbidity study, provide a unique opportunity to examine the renal risk associated with numerous factors, and are also highly cost-effective. The large size of the initial cohort and the 12-to-16-year follow-up period produced a large number of cases.
Previous prospective studies of renal risk factors considered end-stage renal disease as defined by entry into the Medicare registry or death from renal disease as the underlying cause. 2,3 Here, we also studied the contributing causes of death to identify chronic kidney diseases in persons who may have died of another cause before reaching the end-stage and requiring renal replacement therapy. The use of multiple-cause mortality data has been shown to be particularly useful in studying long-term diseases that are not necessarily fatal, but are sufficiently serious to be listed on the death certificate. Renal disease is an excellent example of this situation, because it appears on U.S. death certificates 5 times more often as a contributing cause of death than as an underlying cause. 28 Consideration of both underlying and contributing causes of death strongly improves the sensitivity of case ascertainment and, therefore, the study power.
Few studies have related physical activity to the risk of chronic kidney disease. In one study, an association was observed between increased levels of leisure physical activity and a reduced occurrence of nephropathy in individuals with insulin-dependent diabetes mellitus. 6 Because of the cross-sectional design of that study, however, it was unclear whether that association reflected a protective effect of physical activity or simply physical limitations due to the disease itself. More interestingly, a prospective study of adults with diabetes found leisure physical activity to be negatively related to early renal function decline. 7 Animal studies are even scarcer. One study found that repeated exercise training increased glomerular filtration rate in mice with reduced renal mass. 29 We can postulate that physical activity has a protective effect against chronic kidney disease from the similar mechanisms that have been described for cardiovascular disease. Physical activity protects against cardiovascular disease by reducing risks of obesity, high blood pressure, adverse blood lipid profiles and noninsulin-dependent diabetes mellitus, but it may also afford protection independent of those factors. 30–31 In our study, the relation between physical activity and chronic kidney disease appeared partly mediated by diabetes and hypertension.
Cardiovascular diseases and chronic renal failure may, however, themselves cause low physical activity. The relation observed might thus reflect reduced activity due to the presence of these conditions at baseline, rather than causality. Controlling for a history of cardiovascular disease and estimated glomerular filtration rate, or excluding individuals with a high serum creatinine at baseline from the analysis, did not alter our finding, which suggests that reverse causation is unlikely. Our results were not substantially changed when we excluded incident cases that occurred within the first 5 years of follow-up (ie, subjects with any condition that might have led to reduced physical activity and early death). This also indicates that reverse causation is unlikely. Although we controlled for baseline diabetes and hypertension, the follow-up period of 12–16 years is sufficient for these diseases to occur and even progress to chronic kidney disease. The fact that low physical activity was more strongly associated with diabetic or hypertensive nephropathy than with other renal diseases is compatible with a possible role of these diseases in the causal pathway.
The renal risk from smoking is better documented than that of other lifestyle risk factors. 8–14,32–38 Our findings are consistent with those of the Multiple Risk Factor Intervention Trial study, in which a similar graded relation was observed in a large cohort of men. 12 They are also consistent with results from studies conducted in selected populations of older nondiabetic subjects, 36 and of patients with diabetic nephropathy, 9 primary glomerulonephritis 13,14 or polycystic kidney disease, 13 in which smoking was clearly related to the decline in renal function. Nevertheless, one large population-based case-control study failed to find that smoking was related to end-stage renal disease (from any cause). 10
When we restricted the analysis to treated end-stage kidney diseases, we no longer observed an association with smoking. Although it is possible that this finding is due to the limited power of this restricted analysis, we cannot rule out the possibility that the apparent relation of smoking with chronic kidney disease resulted from its relation with the other causes of death. An alternative explanation, however, is that smokers with chronic kidney disease are more likely to die before reaching end-stage renal failure and requiring replacement therapy than are their nonsmoking counterparts, ie, that smoking would be a risk factor for the severity of the disease through co-morbidities; this explanation would produce the same apparent discrepancies among epidemiological studies.
Several studies with different approaches suggest that smoking plays a causal role in renal damage. It has been related to increased microalbuminuria, proteinuria and serum creatinine, and to decreased glomerular filtration rates in patients with diabetes or hypertension, 37 as well as in healthy individuals. 32,33,36 It has been associated with transient but repeated increases of blood pressure, with renal hemodynamic dysfunction, 34–35 and with damage to small vessels. 38 Although the biologic mechanisms involved in smoking-induced renal damage are unknown, the latter may result from several nonexclusive mechanisms, including direct and indirect effects of smoking (and nicotine) on the kidney, as reviewed by Orth et al. 8
Epidemiological studies of the relation between alcohol consumption and chronic kidney disease are scarce. 15,29 In Perneger’s large population-based case-control study, 15 consumption of more than two alcoholic drinks per day was associated with a 4-fold increase in the risk of end-stage renal disease, whereas lower intake of alcohol did not appear to be harmful. In our study, alcohol consumption did not appear as a risk factor for chronic kidney disease. However, the small number of daily drinkers did not permit further stratification by number of drinks per day, specifying a subgroup of heavy drinkers.
Adiposity promotes an adverse profile of cardiovascular risk factors, including hypertension, type 2 diabetes, reduced high-density-lipoprotein cholesterol and elevated fibrinogen. Most of these factors have also been shown to be strongly related to an increased risk of chronic and end-stage renal failure. 2,3,40 In this study, BMI was associated with an increased risk of chronic kidney disease. However, the relation appeared to be limited to morbid obesity, and was largely mediated by hypertension and diabetes (since adjustment for these conditions strongly weakened the relative risk). Several mechanisms have been proposed to relate obesity to hypertension and glomerulosclerosis, and thus to chronic renal failure. 16 These include altered histology and compression of the renal medulla, leading to activation of the renin-angiotensin system, renal vasodilation and increased glomerular filtration rate and hypertension, as well as metabolic abnormalities (eg, lipids, glucose intolerance), ultimately resulting in hypertension and glomerulosclerosis. Thus, there is a biologic basis for the possible role of extreme obesity in the development of chronic kidney disease. This observation is also consistent with the overlapping time trends of obesity 42 and end-stage renal disease incidence 1 in the U.S., and the emergence of obesity-related glomerulopathy observed by pathologists. 17,18
Several limitations deserve comments. First, in our main analyses, we used the same International Classification of Disease codes to define chronic kidney diseases as in other similar cohorts, 2,3,12 which permits comparison of results, but has the disadvantage of including a few cases of acute renal disease. Subsidiary analysis excluding these cases, however, did not alter our findings. Second, inaccuracy of renal diagnosis on death certificates 43 may have also resulted in some degree of misclassification; approximately one third of the cases with underlying causes of death, and half of the contributing causes were assigned to renal failure not specified as acute or chronic. Although it is likely that the great majority of these cases of renal failure were chronic, we cannot rule out the possibility that some of them may have been acute. Nevertheless, restricting the outcome to treated end-stage kidney diseases (none due to acute renal failure) showed very similar relations to physical inactivity, obesity and alcohol consumption, strongly suggesting that the two groups of cases (treated and dead) were comparable as far as their association with these exposures were concerned. Third, the study had limited power in subgroup analysis. In particular, we were not able to study potential interactions with sex or race. Finally, most studies encounter difficulty in assessing baseline kidney disease to differentiate risk factors for initiation from those for progression of the disease. We addressed this limitation in two ways. First, prevalent end-stage kidney disease cases were excluded from the main analysis. Second, a subsidiary analysis was conducted that excluded subjects with hypercreatininemia at baseline, which might be considered overly conservative and could lead to a lack of power. Nevertheless, it remains difficult to determine whether a given risk factor plays a role in the initiation or in the progression of kidney disease.
Our results have two main implications. First, the increase in the incidence of kidney disease in the U.S. population 44 may be partly explained by the corresponding rise in obesity and physical inactivity. 42 Second, programs to promote physical activity, smoking cessation and weight reduction may reduce the growing burden of kidney disease in the U.S., as well as contribute to cardiovascular disease reduction.
1. NIH. U.S. Renal Data System 1999 Annual Data Report. Bethesda, MD: The National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 1999.
2. Klag MJ, Whelton PK, Randall BL, et al. Blood pressure and end-stage renal disease
in men. N Eng J Med. 1996; 334: 13–18.
3. Brancati FL, Whelton PK, Randall BL, et al. End-stage renal disease
risk in diabetes mellitus: a prospective cohort study
of 332, 544 men screened for MRFIT. JAMA. 1997; 278: 2069–2074.
4. Coresh J, Longenecker JC, Miller ER, Young HJ, Klag MJ. Epidemiology of cardiovascular risk factors in chronic renal disease. J Am Soc Nephrol. 1998; 9: S24–S30.
5. Diamond JR. Analogous pathobiologic mechanisms in glomerulosclerosis and atherosclerosis. Kidney Int Suppl. 1991; 31: 29–34.
6. Kriska AM, LaPorte RE, Patrick SL, et al. The association of physical activity
and diabetic complications in individuals with insulin-dependent diabetes mellitus: the epidemiology of diabetes complications study-VII. J Clin Epidemiol. 1991; 44: 1207–1214.
7. Krop JS, Coresh J, Chambless LE, et al. A community-based study of expanatory factors for the excess risk for early renal function decline in Blacks vs
Whites with diabetes. The Atherosclerosis Risk in Communities Study. Arch Intern Med. 1999; 159: 1777–1783.
8. Orth SR, Ritz E, Schrier RW. The renal risks of smoking
. Kidney Int. 1997; 51: 1669–1677.
9. Ritz E, Ogata H, Orth SR. Smoking
: a factor promoting onset and progression of diabetic nephropathy. Diabetes Metab. 2000; 26: 54–63.
10. Perneger TV. Case-control study of risk factors for end-stage renal disease
. Dissertation. Baltimore, MD: Johns Hopkins University, 1993.
11. Perry HM, Miller JP, Fornoff JR, et al. Early predictors of 15-years end-stage renal disease
in hypertensive patients. Hypertension. 1995; 25: 587–594.
12. Whelton PK, Randall B, Neaton J, et al. Cigarette smoking
and end-stage renal disease
incidence in men screened for MRFIT. J Am Soc Nephrol. 1995; 6: 408A.
13. Orth SR, Stöckmann A, Conradt C, et al. Smoking
as a risk factor for end-stage renal failure in men with primary renal disease. Kidney Int. 1998; 54: 926–931.
14. Stengel B, Couchoud C, Cénée S, et al. Effect of age and blood pressure on the relationship between smoking
and chronic renal failure in primary glomerular nephropathies: a case-control study. Kidney Int. 2000; 57: 2519–2526.
15. Perneger TV, Whelton PK, Puddey IB, et al. Risk of end-stage renal failure associated with alcohol
consumption. Am J Epidemiol. 1999; 150: 1275–1281.
16. Hall JH, Brands MW, Henegar JR, et al. Abnormal kidney function as a cause and a consequence of obesity
hypertension. Clin Exp Pharmacol Physiol. 1998; 25: 58–64.
17. Kambham N, Markowitz GS, Valeri AM, et al. Obesity
-related glomerulopathy: an emerging epidemic. Kidney Int. 2001; 59: 1498–1509.
18. Haas M, Spargo BH, Coventry S. Increasing incidence of focal-segmental glomerulosclerosis among adult nephropathies: a 20-year renal biopsy study. Am J Kidney Dis. 1995; 26: 740–750.
19. U.S. Department of Health and Human Services. Plan and operation of the Second National Health and Nutrition Examination Survey 1976-1980. Vital and Health Statistics. Series 11, No. 237, 1987.
20. U.S. Department of Health and Human Services. Plan and operation of the NHANES II Mortality Study, 1992. Vital and Health Statistics. Series 1, No. 38, 1999.
21. World Health Organization Consultation of Obesity
. Preventing and managing the global epidemic of obesity
. Geneva: WHO, 1997.
22. National Heart and Lung Institute. Manual of Laboratory Operations: Lipid Research Clinics Program, Lipid and Lipoprotein Analysis. Washington, D.C.: Government Printing Office. DHES publications No. (NIH) 75-628, 1974.
23. World Health Organization. International Classification of Diseases, 9th revision: ICD-9, Vol. 1. Geneva: WHO, 1977.
24. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958; 53: 457–481.
25. Cox DR. Regression models and life-tables. J R Stat Soc (B. 1972; 34: 187–202.
26. Levey AS, Greene T, Kusek JW, et al. A simplified equation to predict glomerular filtration rate from serum creatinine. J Am Soc Nephrol. 2000: 155A.
27. Perneger TV, Brancati FL, Whelton PK, et al. Studying the causes of kidney disease
in humans: a review of methodologic obstacles and possible solutions. Am J Kidney Dis. 1995; 25: 722–731.
28. Steenland K, Nowlin S, Ryan B, E. Use of multiple-cause mortality data in epidemiologic analyses: U.S. rate and proportion files developed by the National Institute for Occupational Safety and Health and the National Cancer Institute. Am J Epidemiol. 1992; 136: 855–862.
29. Averbukh Z, Marcus E, Berman S, et al. Effect of exercise training on glomerular filtration rate of mice with various degrees of renal mass reduction. Am J Nephrol. 1992; 12: 174–178.
30. Labarthe DR. Epidemiology and Prevention of Cardiovascular Diseases. A Global Challenge. Gaithersburg, MD: Aspen, 1998.
31. MRFIT Research Group. Leon AS, Connett J. Physical activity
and 10. 5 year mortality in the Multiple Risk Factor Intervention Trial (MRFIT). Int J Epidemiol. 1991; 20: 690–697.
32. Pinto-Sietsma SJ, Mulder J, Janssen WMT, et al. Smoking
is related to albuminuria and abnormal renal function in nondiabetic persons. Ann Intern Med. 2000; 133: 585–591.
33. Metcalf PA, Baker JR, Scragg RKR, et al. Albuminuria in people at least 40 years old: effect of alcohol
consumption, regular exercise, and cigarette smoking
. Clin Chem. 1993; 39: 1793–1797.
34. Gambaro G, Verlato F, Budakovic A, et al. Renal impairment in chronic cigarette smokers. J Am Soc Nephrol. 1998; 9: 562–567.
35. Ritz E, Benck U, Franek E, et al. Effects of smoking
on renal hemodynamics in healthy volunteers and in patients with glomerular disease. J Am Soc Nephrol. 1998; 9: 1798–1804.
36. Bleyer AJ, Shemanski LR, Burke GL, et al. Tobacco, hypertension, and vascular disease: risk factors for renal dysfunctional decline in an older population. Kidney Int. 2000; 57: 2072–2079.
37. Regalado M, Yang S, Wesson DE. Cigarette smoking
is associated with augmented progression of renal insufficiency in severe essential hypertension. Am J Kidney Dis. 2000; 35: 6876–6894.
38. Lhotta K, Rumpelt HJ, Konig P, et al. Cigarette smoking
and vascular pathology in renal biopsies. Kidney Int. 2002; 61: 648–654.
Parekh RS, Klag JK. Alcohol
: role in the development of hypertension and end-stage renal disease
. Curr Opin Nephrol Hypertens. 2001; 10: 385–390.
40. Muntner P, Coresh J, Smith JC, et al. Plasma lipids and risk of developing renal dysfunction: the Atherosclerosis Risk in Communities Study. Kidney Int. 2000; 58: 293–301.
Anastasio P, Spital L, Frangiosa A, et al. Glomerular filtration rate in severely overweight normotensive humans. Am J Kidney Dis. 2000; 35: 1144–1148.
42. Kuczmarski RJ, Flegal KM, Campbell SM, et al. Increasing prevalence of overweight among US adults. The National Health and Nutritional Examination Surveys, 1960 to 1991. JAMA. 1994; 272: 205–211.
43. Perneger TV, Klag MJ, Whelton PK. Cause of death in patients with end-stage renal disease
: death certificate vs registry reports. Am J Public Health. 1993; 83: 1735–1738.
44. Eberhardt MS, Wagener DK, Herman WH, et al. Trends in renal disease morbidity and mortality in the United States, 1979 to 1990. Am J Kidney Dis. 1995; 26: 308–320.