Relationship with health status and medications

Serum creatinine levels in older adults: relationship with health status and medications

Marcel E. Salive


Ageing has been associated with numerous physiological changes, some of which are caused by the cumulative effects of pathological conditions such as hypertension, atherosclerosis, and diabetes, and others that may be due to ‘normal ageing’ in otherwise healthy individuals. Some of the structural and functional alterations of the kidney that have been shown to occur with ageing include decreases in glomerular filtration rate, renal blood and plasma flow rates, renal concentrating and diluting ability, tubular transport and excretory capacity, renal acid excretion, and increases in the number of sclerotic glomeruli [1-8].

Serum creatinine concentration is commonly used as a measure of renal function, although there are many pitfalls in interpreting its relationship to glomerular filtration rate, particularly among older adults [9]. Creatinine is a non-enzymatic breakdown product of the muscle protein creatine [10, 11]. Creatinine is excreted only by the kidneys. It is freely filtered through the glomerulus into the urine and is not reabsorbed. In addition, a small amount of creatinine may be secreted into the urine by the renal tubules. When kidney function is not rapidly changing, creatinine excretion in the urine is equivalent to creatinine production. The serum creatinine level in healthy adults thus represents the balance between dietary creatinine ingestion, creatinine synthesis, and elimination of creatine through the kidney. Although there are wide individual variations, decreases in glomerular filtration rate in the kidney are associated with rising serum creatinine levels.

There is little information on the distribution of measures of kidney function in a representative sample of the older population. National cross-sectional or observational studies which measured serum creatinine generally have limited age ranges [12], and technical problems with the laboratory measurements precluded reporting creatinine data from the first two National Health and Nutrition Examination Surveys (NHANES). Although serum creatinine levels are currently being measured in the NHANES III, the data are not yet available. The Baltimore Longitudinal Study on Aging (BLSA), established in 1958, has followed over 1000 participants with biennial evaluations that included creatinine clearance determinations [5]. However, the BLSA sample was selected for excellent health status, and therefore may not be representative of the general population.

This paper reports the distribution and correlates of serum creatinine in a group of community-dwelling elderly individuals, aged 65 years and older, who survived at least 6 years after their enrolment into the Established Populations for Epidemiologic Studies of the Elderly (EPESE).


Study sample: The data for this report are from three EPESE communities: East Boston, MA; Iowa and Washington counties, IA; and New Haven, CT. Details of the methods used in the longitudinal study have been reported previously [13-16]. Between 1981 and 1983, participants aged 65 years and older were enrolled by full community surveys in East Boston and rural Iowa, and by a stratified random sample based on housing status with an over-sampling of men in New Haven. Initial interviews were completed with 80-84% of the eligible community residents. Trained interviewers conducted the in-person household baseline survey and six annual follow-up interviews, by telephone or in person. Blood samples were collected shortly after the sixth annual follow-up interview, when participants were aged 71 years and older.

The present study groups were drawn from consenting participants for the sixth-year interview who had usable creatinine values. Consent rates for the blood sampling among interviewed participants varied between sites, from 55% in East Boston, to 63% in New Haven, to 76% in rural Iowa. Usable specimens for chemistry analysis were available for 90%, 87% and 99% of those who consented to blood sampling by site, respectively. This resulted in a study group of 3999 adults distributed as follows: 1157 in East Boston, 915 in New Haven and 1927 in Iowa.

Blood sampling and laboratory methods: Blood specimens were obtained using venepuncture from participants in a sitting position. The blood sample was spun, the serum removed and shipped to a commercial laboratory (Nichols Institute, San Juan Capistrano, CA) on dry ice for chemical [TABULAR DATA FOR TABLE I OMITTED] analysis. A Technicon SMAC (Tarrytown, NY) instrument, calibrated to the manufacturer’s specifications, analysed creatinine using the alkaline picrate reaction (Jaffe reaction) without Lloyd’s reagent to separate non-creatinine chromogens. Creatinine was measured using the Jaffe reaction, in which the picrate ion reacts with creatinine under alkaline conditions to form a complex which can be measured colorimetrically. Other substances, such as serum proteins, ketoacids, and drugs (cephalosporins and cimetidine) also react with pierate under these conditions, potentially overestimating the creatinine level by as much as 20% [17, 18]. Cimetidine may also reduce tubular secretion of creatinine, although this effect is most pronounced at low levels of renal function. Some compounds such as glucose, vitamin C, and uric acid can decrease the colorimetric measurement [11]. Creatinine clearance was estimated by the Cockcroft-Gault equation [140-age (years)] x weight (kg)]/[72 x serum creatinine (mg/dl)]; this equation was multiplied by 0.85 for women [19].

During the initial interview, information on age, sex, race, chronic health conditions diagnosed by a physician (and those diagnosed by other health care providers, in East Boston), health habits and functional status were obtained. Blood pressure was measured two or three times during the baseline interview using the Hypertension Detection and Followup Program protocol [20], and the average of two readings was used for analysis as reported previously [21]; mean arterial pressure was computed. Heart attack and diabetes were categorized as present or absent each year based on the participant’s report of whether a definite diagnosis was given by a physician. Information from the sixth follow-up visit also included self-reported height, weight, limitations in activities of daily living and current meditations, and measured blood pressure. Self-reported weight and height are quite accurate [22]; we confirmed this for weight in the Iowa community with a correlation coefficient of 0.97 (n = 1720, p [less than] 0.0001). Limitations in activities of daily living (walking, bathing, dressing, eating, transferring from bed to chair and using the toilet) were coded as present if the individual was unable to perform any of them or required help from another person [23, 24]. Body mass index was computed as weight (kg)/height [(m).sup.2] and classified into approximate tertiles. Current prescription and over-the-counter medications (those taken in the past 2 weeks) were recorded, based upon information collected by the interviewer from the medication label when available, otherwise based upon the participant’s report. Any use of diuretics (thiazide, potassium-sparing or frusemide), cimetidine, and heart disease medications (beta blockers, cardiac glycosides, antilipaemic agents, hypotensive agents, ACE inhibitors, nitrates, and calcium channel blockers) was determined for this analysis.

Statistical analysis: Analyses were stratified by sex in all cases, because of known differences in creatinine levels and lean body mass [25]. Mean creatinine concentration was determined according to demographic characteristics and [TABULAR DATA FOR TABLE II OMITTED] health status, for the entire group and also stratified by site. Few differences in relationships with creatinine were noted by site; consequently we report only pooled analyses. After excluding persons taking cimetidine, which is known to raise creatinine measurement artefactually, the creatinine distribution was examined with stratification on the basis of a history of any of the following: diagnosis of a heart attack or diabetes or the use of diuretics.

Multiple linear regression analysis was performed to examine simultaneously the association of a number of variables with creatinine as the dependent variable. All models included demographic characteristics and blood pressure measurements. Variables for health status and other measurements were added using a stepwise procedure if they were statistically relevant (p [less than] 0.1) on entry, but they were removed from the model if they did not remain statistically relevant (p [less than] 0.2). Dichotomous variables were used for several characteristics, to permit interpretation of the coefficients as adjusted differences in creatinine level between the indicator group (level = 1) and the reference group (level = 0).

Table III. Mean serum creatinine (mg/dl) according to selected

characteristics, by sex, EPSE 1988-89

Because of a substantial number of refusals to allow blood drawing, only 58% of surviving EPESE participants had creatinine values measured in this study. The participants for whom blood testing was done were similar in racial distribution and height to those who declined; however, they tended to be slightly younger, heavier and less disabled than the group that declined the blood testing. Disease status did not differ consistently between those who participated in blood sampling and those who refused, but more male participants had a history of hypertension and fewer women had a history of heart attack.

Salive and Havlik recently discussed the effect of blood pressure at baseline and at the 6-year follow-up on creatinine levels in this population using a linear regression model [21]. They reported that the baseline systolic and diastolic blood pressure were significantly associated with level of creatinine at the 6-year follow-up, while at that follow-up only the systolic blood pressure showed an association with creatinine, and that association was of lesser magnitude. Perneger et al. reported on 1339 middle-aged residents of Washington County, MD, who volunteered to have their blood pressure measured in 1974 and again in 1988-9, along with a serum creatinine value in 1988-9. They found that the serum creatinine value was better predicted by past than by the current blood-pressure values, even for blood pressures that were in the normal range in 1974. They also reported that in the middle-aged participants, the sex-adjusted association of blood pressure with creatinine was stronger for the past diastolic blood pressure rather than the systolic [28]. Hypertension is also associated with progression of nephropathy in patients with diabetes [29, 30], renal artery stenosis, polycystic kidney disease [31], and many other conditions that affect the kidney [32].

In the current study, creatinine level tended to rise with increasing age, while the diastolic blood pressure fell with age in both men and women and systolic blood pressure fell with age in men only. These blood pressure changes with age have been reported in other studies [33], and may be a result of selective mortality of persons with higher blood pressures over time, such that the older participants are the survivors of a cohort in which persons with elevated blood pressure have already died, or the result of increased prevalence of coronary heart disease or congestive heart failure with low output states. It is less likely to be due to increased rigour in treating blood pressure at the oldest ages, since these data were collected in 1981-9, before publication of the Systolic Hypertension in the Elderly Project (SHEP) [34] which demonstrated the benefit derived from treating isolated systolic hypertension in older adults. Only 5% of all participants in this study were black, but in the New Haven community, 22% of participants with creatinine values were black. A pattern of higher creatinine was seen in black men and women, independent of the effects of age and other illness-related variables, as has been reported in younger populations [25, 35], possibly related to differences in muscle mass. Previous studies in younger populations suggest that black participants tend to have higher creatinine levels, that past blood-pressure level is associated with creatinine levels, and that control of blood pressure has different effects on subsequent creatinine levels in black versus white patients [25, 36, 37].

Diabetes is known to be associated with high risk of developing nephropathy [38]. Diabetic nephropathy first manifests as a higher than normal glomerular filtration rate, followed by microalbuminuria, a declining glomerular filtration rate, and elevation in serum creatinine [39]. The incidence of non-insulin dependent diabetes increases with age, is higher in females than in males up to age 64 years, and is more prevalent in obese individuals [40]. Among patients with non-insulin dependent diabetes, it is estimated that up to 40% will develop microalbuminuria, up to 30% will develop dipstick-positive proteinuria, and up to 8% will develop end-stage renal disease requiring dialysis [41].

Myocardial infarction maybe associated with decreased cardiac output and low renal perfusion pressure, leading to decreased glomerular filtration and elevated creatinine. Patients who have had a myocardial infarction, most commonly due to coronary atherosclerosis, may also have renal vascular atherosclerotic disease. Renal ischaemia from either cause could result in elevations in creatinine. Current use of cardiovascular medications is probably a proxy for prevalent atherosclerotic vascular disease.

Use of diuretics has not been associated with increased serum creatinine levels in the absence of volume depletion. However, persons with kidney disease manifesting with elevated creatinine levels might be more likely to require a diuretic to help eliminate ingested water and salt. Therefore, the weaker association of thiazides with creatinine may reflect their use more as an initial therapy for hypertension than for control of salt and water. Frusemide represents a medication useful in salt and water diuresis in patients with impaired renal function; thus its association with significantly higher creatinine values is expected.

There are certain limitations to this study. Only cross-sectional information on serum creatinine is available because serum creatinine levels at baseline were not determined. Also, serum creatinine level was available for only 58% of the participants at the 6-year follow-up. The level of serum creatinine is influenced by many different factors, and muscle atrophy as well as decreased dietary intake of meat will tend to decrease the level of serum creatinine. Conversely, worsening kidney function will be associated with increases in serum creatinine. These forces may balance out, so that a creatinine value in the ‘normal range’ in an older person may indicate substantial loss of kidney function (as much as 50-60% or more, as suggested by Table II). Thus, single creatinine values, and serial creatinine values in the face of changing diet or body composition may not be that helpful in following kidney function.

Creatinine levels greater than 1.7mg/dl have been associated with a doubling of the risk of death in 8 years in hypertensive individuals [25]. Thus, selective mortality of EPESE participants who had higher creatinine levels at baseline may affect the reported distribution of creatinine values. However, this highlights the need for careful study of the oldest old, as has been done in the present analysis. Further, there is a potential bias due to study dropout and refusal to have blood drawn at the time of the 6-year follow-up. Because measured weight was available only from one community, we used self-reported weight for this analysis. Although body mass index was not associated with serum creatinine, we have no better measure of lean body mass to use in the analysis.

Many of the equations that use serum creatinine and weight to estimate the creatinine clearance as a proxy for the level of kidney function were developed using young male populations; few were developed using older free-living individuals. Malmrose et al. [9] compared the estimated and measured creatinine clearance values in a healthier subset of EPESE participants who completed 12-hour urine collections and found very little consistency. One of the most frequently used equations to estimate creatinine clearance, the Cockcroft-Gault equation (which we used for Table II), did no better than any of the other 15 equations in estimating creatinine clearance. The equations were associated with under- and over-estimation of creatinine clearances of up to 40%. None was judged to be acceptable by the authors of that paper, who suggested that drug levels be used instead of estimated creatinine clearance, to adjust therapy in elderly patients [9]. However, serum creatinine and estimated creatinine clearance is still used by most clinicians as the first approximation of kidney function. Table II demonstrates the dramatic decline in estimated creatinine clearance in the older population that results from age-associated increases in creatinine levels and decreases in body weight.

While this paper only approximates the distributions of creatinine values in the general older population because of the limitations mentioned above, we believe that the associations with creatinine levels demonstrated are valid for chronic conditions and medication use. Site differences were not detectable between the three EPESE communities, implying that at least in the three areas studied the results can be generalized.

Longitudinal information is needed to document changes in renal function with ageing in the elderly population. These studies must account for medications and diseases such as hypertension, diabetes, atherosclerotic renal vascular disease, coronary artery disease, congestive heart failure and past urinary tract infections, and include a measure of lean body mass.


This study was supported in part by contracts N01-AG-02105, N01-AG-02106 and N01-AG-02107 from the National Institute on Aging.


1. Lewis WH Jr, Alving AS. Changes with age in the renal function in adult men. Am J Physiol 1938;153:500-15.

2. Davies DF, Shock NW. Age changes in glomerular filtration rate, effective renal plasma flow, and tubular excretory capacity in adult males. J Clin Invest 1949;29:496-507.

3. Friedman SA, Raizner AE, Rosen H, Solomon NA, Sy W. Functional defects in the aging kidney. Ann Int Med 1972;76:41-5.

4. Kafetz K. Renal impairment in the elderly: a review. J R Soc Med 1983;76:398-402.

5. Lindeman RD, Tobin JC, Shock NW. Association between blood pressure and the rate of decline in renal function with age. Kidney Int 1984;26:861-8.

6. Anderson KS, Brenner BM. Effects of aging on the renal glomerulus. Am J Med 1986;80:435-42.

7. Moser M. Physiological differences in the elderly: are they clinically important? Eur Heart J 1988;9(suppl D):55-61.

8. Lindeman RD. Changes in renal function with aging: implications for treatment. Drugs Aging 1992;2:423-31.

9. Malmrose LC, Gray SL, Pieper CF, et al. Measured versus estimated creatinine clearance in a high-functioning elderly sample: MacArthur Foundation study of successful aging. J Am Geriatr Soc 1993;41:715-21.

10. Narayanan S, Appleton HD. Creatinine: a review, Clin Chem 1980;26:1119-26.

11. Perrone RD, Madias NE, Levey AS. Serum creatinine as an index of renal function: new insights into old concepts. Clin Chem 1992;38:1933-53.

12. National Center for Health Statistics. Plan and operation of the second National Health and Nutrition Examination Survey, 1976-1980. Vital Health Star [1] No. 15, DHHS Pub. No. (PHS) 81-1317. Washington, DC.: US Government Printing Office, July 1981.

13. Cornoni-Huntley J, Ostfeld AM, Taylor JO, et al. Established populations for epidemiologic studies of the elderly: study design and methodology. Aging Clin Exp Res 1993;5:27-37.

14. Cornoni-Huntley JC, Foley DJ, White LR, et al. Epidemiology of disability in the oldest old: methodologic issues and preliminary findings. Milbank Mem Fund Q 1985;63:350-76.

15. Salive ME, Cornoni-Huntley J, Phillips CL, et al. Serum albumin in older persons: relationship with age and health status. J Clin Epidemiol 1992;45:213-21.

16. Salive ME, Cornoni-Huntley J, Guralnik JM, et al. Anemia and hemoglobin levels in older persons: relationship with age, gender, and health status. J Am Geriatr Soc 1992;40:489-96.

17. Young DS, Pestaner LC, Gibberman V. Effects of drugs on clinical laboratory tests. Clin Chem 1975;21:1-432D.

18. Murray RL. Creatinine. In: Pesce AJ, Kaplan LA, eds. Methods in clinical chemistry. St Louis: Mosby, 1987;10-17.

19. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976;16:31-41.

20. Hypertension Detection and Followup Program Cooperative Group. Blood pressure studies in 14 communities; a two-stage screen for hypertension. JAMA; 1977;237:2385-91.

21. Salive ME, Havlik RJ. Blood pressure and serum creatinine [Letter]. JAMA 1993;269:2983-4.

22. Heaney RP, Ryan R. Relation between measured and recalled body height [Letter]. N Engl J Med 1988;319:795-6.

23. Katz S, Ford AB, Moskovitz AW, Jackson BA, Jaffe MW. Studies of illness in the aged: the index of ADL: a standardized measure of biological and psychosocial function. JAMA 1963;185:914-19.

24. Branch LG, Katz S, Kniepmann K, Papsidero JA. A prospective study of functional status among community elders. Am J Public Health 1984;74:266-8.

25. Shulman NB, Ford CE, Hall WD, et al. Hypertension Detection and Follow-up Program Cooperative Group: Prognostic value of serum creatinine and effect of treatment of hypertension on renal function – results from the Hypertension Detection and Follow-up Program. Hypertension 1989;13(suppl I):180-93.

26. National Center for Health Statistics, Harlan WR, Hill AL, Schmouder RP, Thompson FE, Larkin FA, Landis JR. Dietary intake and cardiovascular risk factors: Part I. Blood pressure correlates. Vital Health Slat [11] No. 226. DHHS Pub. No. (PHS) 83-1676. Washington DC: US Government Printing Office, February 1983.

27. Prohaska T, Mermelstein R, Miller B, Jack S. Functional status and living arrangements. In: Van Nostrand JR, Furner SE, Suzman R, eds. Health data on older Americans: United States 1992. National Center for Health Statistics. Vital Health Stat 1993;3(27):23-39.

28. Perneger TV, Nieto J, Whelton PK, Klag MJ, Comstock GW, Szklo M. A prospective study of blood pressure and serum creatinine. JAMA 1993;269:488-93.

29. Mogensen CE. Long-term antihypertensive treatment inhibiting progression of diabetic nephropathy. Br Med J 1982;285:685.

30. Parving HH, Andersen AR, Smidt UM, et al. Effect of antihypertensive treatment on kidney function in diabetic nephropathy. Br Med J 1987;294:1443.

31. Lieske JC, Toback FG. Autosomal dominant polycystic kidney disease. J Am Soc Nephrol 1993;3:1442-50.

32. Jacobson HR, Striker GE, Klahr. S. The principles and practice of nephrology. Philadelphia, PA; B C Decker Inc. 1991.

33. Bots ML, Grobbee DE, Hofman A. High blood pressure in the elderly. Epidemiol Rev 1991;13:294-314.

34. SHEP Cooperative Research Group. Prevention of stroke mortality by antihypertensive drug treatment in older persons with isolated systolic hypertension. JAMA 1991;265:3255-64.

35. Tierney WM, McDonald CJ, Luft FC. Renal disease in hypertensive adults: effect of race and Type II diabetes mellitus. Am J Kidney Dis 1989;12:485-93.

36. Walker WG, Neaton JD, Cutler JA, Neuwirth R, Cohen JD. Renal function change in hypertensive members of the Multiple Risk Factor Intervention Trial. JAMA 1992;268:3085-91.

37. Rostand SG, Brown G, Kirk K, Rutsky EA, Dustan HP. Renal insufficiency in treated essential hypertension. N Engl J Med 1989;320:684-8.

38. Selby JV, FitzSimmons SC, Newman JM, et al. The natural history and epidemiology of diabetic nephropathy: implications for prevention and control. JAMA 1990;263:1954-60.

39. Omachi R. The pathogenesis and prevention of diabetic nephropathy. West J Med 1986;145:222.

40. Harris MI, Hadden WC, Knowler WC, Bennett PH. Prevalence of diabetes and impaired glucose tolerance and plasma glucose levels in U.S. population aged 20-74 yr. Diabetes 1987;36:523-34.

41. Tung P, Levin SR. Nephropathy in non-insulin dependent diabetes mellitus. Am J Med 1988;85(suppl 5A):131-6.

Author’s address

M. E. Salive, J. M. Guralnik, M. Pahor Epidemiology, Demography and Biometry Program, National Institute on Aging

C. Jones, L. Agodoa Division of Kidney, Urologic and Hematologic Diseases, National Institute of Diabetes, Digestive Disease and Kidney Diseases,

National Institutes of Health, 7201 Wisconsin Avenue, Gateway Building, Suite-3C309, Bethesda, MD 20892, USA

R. B. Wallace Department of Preventive and Environmental Health, University of Iowa, Iowa City, Iowa, USA

COPYRIGHT 1995 Oxford University Press

COPYRIGHT 2004 Gale Group