Chronic kidney disease prevalence in a UK residential care home population

Transcription

1 Nephrol Dial Transplant (2008) 23: doi: /ndt/gfm792 Advance Access publication 19 November 2007 Original Article Chronic kidney disease prevalence in a UK residential care home population Joanne L. Carter 1, Shelagh E. O Riordan 2, Gillian L. Eaglestone 3, Michael P. Delaney 3 and Edmund J. Lamb 1 1 Departments of Clinical Biochemistry, 2 Health Care of the Older Person and 3 Renal Medicine, east Kent Hospitals NHS Trust, Canterbury, Kent, CT1 3NG, United Kingdom Abstract Background. Chronic kidney disease (CKD) is common ( 30%) in non-institutionalized older people but little is known about the prevalence of CKD amongst older people living in residential care. Methods. An observational study of older subjects [n = 250, median age 86 (range ) years, 79% female, 100% Caucasian, 16% diabetic, 48% hypertensive, 5% known renal disease, mean number of medications 7] who were recruited over a 9-month period from 155 residential care homes in east Kent (total population 3811) using a randomization process. The estimated glomerular filtration rate (egfr, ml/min/1.73 m 2 ) was calculated using the Cockcroft and Gault equation corrected for the body surface area and the simplified Modification of Diet in Renal Disease (MDRD) Study equation. Serum cystatin C concentration was also measured. Results. Using the MDRD equation 18% had egfr 60, 39% stage 3A CKD (egfr 45 59), 34% stage 3B CKD (egfr 30 44) and 10% stage 4 CKD (egfr 15 29). By the Cockcroft Gault equation the equivalent figures were 3%, 18%, 48% and 31%, respectively. Agreement between the equations for staging of CKD was poor (κ = 0.07). However, >80% of residents were categorized as having stage 3 CKD (>40% stage 3B) or worse whichever equation was used. Serum cystatin C concentration was increased in 92% of the population. Increasing age and higher body mass index were predictive of decreased renal function. Conclusion. Significant CKD is prevalent and unrecognized in this population. This may have important management implications particularly for treatment with renally excreted drugs, fracture prevention or managing cardiovascular risk. Keywords: chronic kidney disease; cystatin C; glomerular filtration rate; older people; residential care homes Correspondence and offprint requests to: Dr Edmund J. Lamb, Clinical Biochemistry, east Kent Hospitals NHS Trust, Kent and Canterbury Hospital, Ethelbert Road, Canterbury, Kent CT1 3NG, United Kingdom. Tel: ext ; Fax: ; Edmund.lamb@ekht.nhs.uk Introduction Kidney disease is predominantly a disease of older people [1,2]; the prevalence of chronic kidney disease (CKD) has been estimated to be 30% in the non-institutionalized older (>80 years) population [3]. To date only non-uk studies have estimated the prevalence of CKD in residential/nursing home settings; little is known about the prevalence of CKD in this population in the United Kingdom. Patients in UK residential care homes need help with activities of daily living but do not predominantly need nursing care. There is a high prevalence of dementia [4]. The average number of activities of daily living ( Barthel ) score in residential homes in east Kent is 14, indicating requirements for assistance in at least some activities of daily living: approximately 25% of residents require aids or assistance to transfer or feed. Kidney function is typically assessed by measuring the glomerular filtration rate (GFR) and the internationally accepted classification of CKD is largely predicated by GFR [5]. Serum creatinine has poor sensitivity for kidney disease, especially amongst older people [6 8], and healthcare organizations in the western world have recommended that GFR should be estimated using creatinine-based equations that take into account the influences of age, body size and gender [9 13]. The most widely recommended of the GFRestimating equations is that derived from the Modification of Diet in Renal Disease (MDRD) Study [14 16], although the Cockcroft and Gault equation [17] has been used in clinical practice for many years. We report the prevalence of CKD in a UK residential home population obtained using both the MDRD equation and the Cockcroft and Gault equation. Methods Recruitment of participants Older subjects (n = 250) were recruited over a 9-month period (March to November 2006) from residential care C The Author [2007]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org

2 1258 J. L. Carter et al. homes in east Kent (total population 3811) using a randomization process. A list of all residential care homes in East Kent was obtained and permission was sought from the Kent County Council to contact them. From an alphabetical list of 155 homes, 52 were randomly selected (every third home on the list was chosen). A letter was sent to each home outlining the study. No financial incentives for participation were offered to either homes or individual residents. When a home refused to participate in the study they were replaced with the next home on the alphabetical list. In total, 23 homes declined to participate. The 52 participating homes housed 1240 residents. On arrival at the home a list of residents was requested from the person in charge. To minimize bias the type of list was not specified (e.g. could have been a dinner, fire, birthday or room list) and from this list the residents were randomly chosen again using systematic sampling. The number of residents needed to be sampled from each home (n) was determined as follows: n = [(number of places in home being visited/1240)] 250. Using n, it was possible to work out which number resident had to be chosen from the list. For example a home with 25 residents with five needed for the study and with the first resident being chosen would mean that every subsequent fifth resident was then chosen. Where a selected resident did not wish to take part or where they were deemed to be unable to give informed consent (either by the researcher or home manager) the next resident on the list was chosen. Sixteen residents declined to take part and 10 were felt unable to give informed consent. The study had full ethical approval from the east Kent Local Research Ethics Committee (reference 04/Q1803/47). Informed consent was obtained; residents with impaired cognitive function were not specifically excluded but those with severe cognitive impairment unable to give informed consent were. Nursing homes were excluded. There were no other exclusions. Residents were weighed and a detailed clinical history was recorded, including the presence of co-morbid conditions, medications, smoking and alcohol habits. For the purposes of statistical analysis, the comorbidity record was in some cases adjusted following a careful review of the individual s medication by a consultant geriatrician. East Kent is a semi-rural area of southeastern England with an above average prevalence of older people. Laboratory analyses Blood was collected into vacutainer devices and upon receipt in the laboratory analysed for full blood count, serum urea, creatinine and electrolytes. Samples were transported to the laboratory within 4 h of venepuncture and analysed on the same day, with the exception of the cystatin C measurements, where samples were stored for a maximum of 3 months at 80 C; the stability of cystatin C under these conditions is well established [18]. Creatinine was measured using a compensated rate Jaffe method on an Integra 800 analyser (Roche Diagnostics plc, Lewes, East Sussex, UK) with standardization traceable to isotopedilution mass spectrometry. The reference ranges for serum creatinine concentration were µmol/l in women and µmol/l in men [19]. GFR was estimated using the simplified MDRD [15] and the Cockcroft and Gault [17] equations; Cockcroft and Gault GFR estimates were corrected to a standardized body surface area of 1.73 m 2 [20]. Prior to GFR estimation using both equations a correction was applied so that the method yielded comparable values to those produced by the laboratory serving the MDRD study [21]. Cystatin C was measured by a particle-enhanced nephelometric immunoassay on a BN Prospec analyser (Dade Behring Ltd, Milton Keynes, UK). The laboratory reference range was mg/l and between-day imprecision was 3.5% at a concentration of 2.3 mg/l. Statistics and data analysis Data were analysed using Analyse-It (Analyse-it Software Ltd, Yorkshire, UK) and InStat (GraphPad.com). P less than 0.05 is considered significant. On the basis of either the MDRD or Cockcroft and Gault estimated GFR (egfr), individuals were classified as having either GFR >60 ml/min/1.73 m 2, stage 3 CKD (GFR ml/min/1.73 m 2 ) or stage 4 CKD (GFR ml/min/1.73 m 2 ), according to the internationally accepted staging system [5], and the prevalence of CKD in the population was estimated accordingly. Additionally, the population was studied in relation to a recently proposed stratification of CKD in which stage 3 is broken down into two sub-stages: 3A (GFR ml/min/1.73 m 2 ) and 3B (GFR ml/min/1.73 m 2 ) [22]. Comparison between the CKD classifications derived from the two GFR estimates was undertaken using the weighted kappa test for agreement. Demographic and clinical variables were compared among patients with GFR >60 ml/min/1.73 m 2, stages 3A, 3B and 4 CKD by using the Kruskal Wallis test (nonparametric analysis of variance) for continuous variables and the Chi-square test for trend for categorical variables. Spearman s rank analysis was used to test for univariate relationships between egfr and other continuous variables [age, body mass index (BMI), number of medications, length of time in residential care]. The Mann Whitney U-test was used to assess whether egfr differed between patients with and without dementia. Multiple regression modelling was undertaken using the MDRD egfr estimate to establish the relationship between egfr and other clinical variables. Log egfr was the dependent outcome, and other clinical variables [age, gender, BMI, diabetes mellitus, vascular disease (comprising cardiovascular disease, stroke and hypertension), smoking history, dementia, thyroid disease, joint replacement, osteoporosis and length of residence] were included as either continuous or dichotomous data. Residuals were found to be normally distributed when the dependent outcome, egfr, was log10 transformed. Multicollinearity was not detected in any models used. Manual backward elimination was performed; clinical variables that were not significant (P < 0.05) were excluded from the analysis. A further,identical analysis was undertaken substituting log cystatin C for log egfr.

3 CKD in residential care 1259 Table 1. Demographic and clinical details of the population, stratified by MDRD egfr; data available in all patients unless otherwise indicated Total cohort >60 ml/min/1.73 m 2 CKD stage 3A CKD stage 3B CKD stage 4 P for trend n (%) 250 (100%) 44 (17.6%) 98 (39.2%) 84 (33.6%) 24 (9.6%) Age (year) 86 (82 90) 84.0 ( ) 85.0 ( ) 87.0 ( ) 86.5 ( ) Female, n (%) 197 (78.8%) 24 (54.5%) 80 (81.6%) 72 (85.7%) 21 (87.5%) Height (m) 1.62 ( ) 1.65 ( ) 1.62 ( ) 1.62 ( ) 1.60 ( ) Weight (kg) 64 (55 74) 60 (51 75) 65 (55 76) 64 (56 71) 66 (61 73) BSA (m 2 ) 1.69 ( ) 1.68 ( ) 1.70 ( ) 1.70 ( ) 1.67 ( ) BMI, weight (kg)/height (m) ( ) 21.7 ( ) 24.2 ( ) 23.8 ( ) 25.7 ( ) Haemoglobin (g/dl), 12.3 (±1.5) 12.8 (±1.5) 12.3 (±1.4) 12.2 (±1.4) 11.3 (±1.6) mean (±SD) Serum bicarbonate (mmol/l) 29 (27 31) 29 (26 31) 29 (27 31) 29 (27 31) 25 (24 29) Serum creatinine (µmol/l) 87 (68 109) 62 (56 73) 75 (67 83) 105 (95 120) 147 ( ) Serum creatinine, n (%) 127 (50.8%) 0 (0.0%) 19 (7.6%) 84 (100.0%) 24 (100.0%) exceeding reference range Serum cystatin C (mg/l) 1.36 ( ) 1.05 ( ) 1.27 ( ) 1.71 ( ) 2.24 ( ) < Serum cystatin C, n (%) 230/249 (92.4%) 34 (77.0%) 89/97 (91.8%) 83 (98.8%) 24 (100.0%) exceeding reference range MDRD egfr 47.3 ( ) 67.8 ( ) 52.8 ( ) 37.2 ( ) 25.6 ( ) (ml/min/1.73 m 2 ) Cockcroft and Gault egfr 35.3 ( ) 48.8 ( ) 38.7 ( ) 28.4 ( ) 20.8 ( ) (ml/min/1.73 m 2 ) Medications, mean (range) 7 (4 9) 6 (5 9) 6 (3 9) 6 (4 9) 7 (6 10) Co-morbidities, n (%) b Hypertension 128 (51.2%) 21 (47.7%) 52 (53.1%) 44 (52.4%) 11 (45.8%) Cardiovascular disease 108 (43.2%) 17 (38.6%) 41 (41.8%) 38 (45.2%) 12 (50.0%) Stroke 65 (26.0%) 14 (31.8%) 24 (24.5%) 22 (26.2%) 5 (20.8%) Vascular disease c 196 (78.4%) 32 (72.7%) 76 (77.6%) 69 (82.1%) 19 (79.2%) Dementia 64 (25.6%) 13 (29.5%) 23 (23.5%) 23 (27.4%) 5 (20.8%) Joint replacement 53 (21.2%) 11 (25.0%) 22 (22.4%) 16 (19.0%) 4 (16.7%) Osteoporosis 50 (20.0%) 6 (13.6%) 19 (19.5%) 24 (28.6%) 1 (4.2%) Thyroid disease 50 (20.0%) 7 (15.9%) 22 (22.4%) 16 (19.0%) 5 (20.8%) Diabetes (types 1 and 2) 44 (17.6%) 5 (11.4%) 15 (15.3%) 14 (16.7%) 10 (41.7%) a Values for continuous variables expressed as median (interquartile range) unless stated otherwise. b Additionally 10.4% of patients had bowel disease, 5.2% known kidney disease, 3.6% inflammatory diseases and 1.6% liver disease. c Includes hypertension, cardiovascular disease and stroke. Results Subjects ranged in age from 68 to 100 years, were exclusively Caucasian, predominantly female and had a high prevalence of co-morbidities (Table 1). Drugs were widelyprescribed in this population with, on average, each resident taking seven different classes of agent; of note, 25% were taking either angiotensin converting enzyme inhibitors or angiotensin receptor blockers, 55% diuretics and 7% non-steroidal anti-inflammatory drugs. Seventeen percent of our cohort were obese (BMI 30 kg/m 2 ) and 18% were underweight (BMI 20 kg/m 2 ). Approximately half of the residents exceeded the upper limit of the reference range for serum creatinine concentration (54% for females, 42% for males) compared with >90% exceeding the laboratory reference range for cystatin C. Using the MDRD equation 18% had egfr 60 ml/min/1.73 m 2, 73% had stage 3 CKD (39% stage 3A, 34% stage 3B) and 10% stage 4 CKD. By the Cockcroft and Gault equation the equivalent figures were 3%, 66% (18%, 48%) and 31%, respectively (Figure 1). Agreement between the equations for staging of CKD was poor, irrespective of whether the group was stratified into three groups (i.e. 60 ml/min/1.73 m 2, stages 3 and 4; κ = 0.17) or four groups (i.e. 60 ml/min/ 1.73 m 2, stages 3A, 3B and 4; κ = 0.07). However, >80% of residents were categorized as having stage 3 CKD (>40% stage 3B) or worse no matter whichever equation was used. Haemoglobin and serum bicarbonate concentrations were significantly lower in residents with stage 4 CKD (P = and P = respectively). Median (interquartile range) egfr (ml/min/1.73 m 2 ) did not differ (P = 0.76) between residents with [47.8 (21.4)] and without [47.2 (19.7)] dementia. Univariate analyses using Spearman s rank statistic demonstrated negative relationships between egfr and age (r s 0.22, P < ) and BMI (r s 0.19, P = 0.003), and positive relationships between cystatin C and age (r s 0.18, P = ) and BMI (r s 0.14, P = 0.03). Neither egfr nor cystatin C demonstrated an association with either number of medications or length of time in residential care. In a multiple regression model age (P = ), BMI (P = ), presence of diabetes (P = ) and female gender (P = ) were found to be significantly associated with declining egfr. The model explained 11.4% of the variance in egfr (P < ) (Table 2). Age (P = ), BMI (P = ) and vascular disease (P = ) were found to be significant predictors of increased serum cystatin C concentration; the model explained 7.2% of the variance in cystatin C concentration (P = ) (Table 3).

4 1260 J. L. Carter et al. Table 2. Independent effect of clinical variables on the estimated glomerular filtration rate (egfr) by multiple linear regression analysis; egfr data were log-transformed Significant variables Unstandardized Standard error Standardized Multiplicative change in median egfr β coefficient β coefficient for an increase in one unit or presence of independent variable Age (year) Body mass index, weight (kg)/height (m) Presence of diabetes Female gender There were no cases with missing data for the significant independent variables (n = 250). The initial model included age, gender, body mass index, diabetes mellitus, vascular disease (comprising cardiovascular disease, stroke and hypertension), smoking history, dementia, thyroid disease, joint replacement, osteoporosis and length of residence. Table 3. Independent effect of clinical variables on the serum cystatin C concentration by multiple linear regression analysis; cystatin C data were log-transformed Significant variables Unstandardized Standard error Standardized Multiplicative change in median cystatin C β coefficient β coefficient concentration for an increase in one unit or presence of independent variable Age (year) Body mass index, weight (kg)/height (m) Presence of vascular disease Only cases with complete data for the significant independent variables were included (n = 249). The initial model included age, gender, body mass index, diabetes mellitus, vascular disease (comprising cardiovascular disease, stroke and hypertension), smoking history, dementia, thyroid disease, joint replacement, osteoporosis and length of residence. Fig. 1. Distribution of CKD stages using MDRD (solid bars) and Cockcroft and Gault (hatched bars) equations. Significantly different estimates of CKD staging, in particular of CKD stage 4, are obtained depending on which estimate is used. Agreement between the equations for staging of CKD was poor, irrespective of whether the group was stratified into three groups (a, i.e. 60 ml/min/1.73 m 2, stages 3 and 4; κ = 0.17) or four groups (b, i.e. 60 ml/min/1.73 m 2, stages 3A, 3B and 4; κ = 0.07). Discussion How robust are our estimates of CKD prevalence in this population? In this UK study of people in residential care, we report a high prevalence of CKD (GFR <60 ml/min/1.73 m 2 ) regardless whichever GFR-estimating equation was used. We observed poor agreement between the Cockcroft and Gault and MDRD estimates of GFR, with the Cockcroft and Gault estimate predicting worse kidney function than the MDRD equation as previously noted in older people [3,23 25]. In the absence of a reference GFR procedure it is difficult to draw firm conclusions as to which GFR-estimating equation is most accurate in this population. We were careful to ensure traceability of our creatinine assay to that of the laboratory that supported the MDRD study [21]. We have previously shown the MDRD equation to be unbiased compared to a reference GFR measurement in an older population, whereas the Cockcroft and Gault estimate showed significant negative bias [26]. On this basis, we would suggest that the MDRD estimates of CKD prevalence are more likely to be accurate in this population and further analyses were undertaken using these data. Using the MDRD equation 82% of residents had CKD, whereas 51% had serum creatinine concentrations that exceeded the reference range; cystatin C concentration was increased amongst 92% of the population. It is well known that GFR estimating equations will detect more CKD in this population than serum creatinine alone [7,8]. Nevertheless, it is perhaps notable that over half of this population had increased serum creatinine concentrations. The

5 CKD in residential care 1261 reference range for serum creatinine does not increase with age, except amongst nonagenarians [27]. The MDRD equation was not initially developed or validated in older people and there have been concerns raised that it is identifying non-disease. However, even using a more conservative approach to CKD, in which it has been suggested that clinically relevant higher risk really only begins as GFR falls below 45 ml/min/1.73 m 2 [22], application of the equation unveils a high burden (43%) of disease in this population. Only 5% of this cohort were previously known to have kidney disease. How does the CKD prevalence we have observed compare with other published data? The CKD prevalence we have observed appears to exceed that observed in epidemiological studies of other, predominantly non-institutionalized, populations. A 26% prevalence of MDRD-estimated GFR <60 ml/min/1.73 m 2 was observed amongst US adults 70 years [3]. In a Norwegian study, 15%, 28% and 42% of the population aged years, years and 90 years respectively had MDRD-estimated GFR <60 ml/ min/1.73 m 2 [28]. Australian data, based upon the Cockcroft and Gault equation, describes a prevalence of GFR <60 ml/min/1.73 m 2 of 55% in the non-institutionalized population 65 years [29]. In the United Kingdom, a large population survey that will have captured both institutionalized and non-institutionalized adults reports an MDRDeGFR <60 ml/min/1.73 m 2 to be present in 33% and 45% of males aged and 85 years respectively and in 42% and 49% of females aged and 85 years respectively [30]. We are aware of two other studies exclusively from older people in long-term care. Garg et al. [25] studied a Canadian population in many ways similar to our own: predominantly Caucasian, female (73%), mean medications 6, and with a high prevalence of hypertension and heart failure (>50%) and diabetes mellitus (16%). However, they observed a much lower prevalence of CKD; nearly 40% had an MDRD-eGFR <60 ml/min/1.73 m 2 compared to 82% in our study and 3.5% of men and 4.0% of women had MDRD-eGFRs <30 ml/min/1.73 m 2 compared to 8.4% in our study. As noted above, it is hard to explain these differences on the basis of the population demographics, although our cohort was slightly older (mean age 86 years compared to 82 years). Zochling et al. [31] studied an Australian population of older (mean age 86 years) individuals living in hostels and nursing homes. They used the Cockcroft and Gault equation to estimate GFR but, in keeping with our data, observed a prevalence of egfr <60 ml/min/1.73 m 2 of 90% amongst females. Using serum cystatin C concentration as a renal marker in this cohort identified more than 90% of residents as having impaired kidney function supporting, and extending, our conclusion that kidney disease has a high prevalence in the residential care home population. Finney et al. [32] also observed almost universally increased serum cystatin C concentrations in older people compared to a reference range derived in younger adults. Further, concentrations were significantly higher (12%) in older people living in institutions, probably reflecting the presence of more comorbidity in such individuals. Cystatin C is known to more sensitively detect declining GFR in older people than serum creatinine, mainly due to the decline in muscle mass that occurs with ageing and to which serum creatinine concentration is related [6]. GFR estimating equations adjust for this by including a factor that takes into account the changing relationships between GFR, serum creatinine concentration and ageing. In patients living in institutions it is plausible that physical inactivity will disturb this relationship, with muscular atrophy [33] leading to relatively lower serum creatinine concentrations for the same level of GFR, and a consequent underestimation of the true degree of renal impairment. This could explain our observation that the apparent prevalence of kidney disease using cystatin C is even higher than that observed with the MDRD equation. Why is CKD so prevalent in this cohort? We used multiple regression analysis to explore factors independently associated with both reduced egfr and increased cystatin C. The models explained only a small amount of the variation in both outcomes. Increasing age was associated with worsening kidney function, whether expressed as egfr or cystatin C concentration. This finding is consistent with all other longitudinal and cross-sectional studies, irrespective of how kidney function has been measured [2,3,28,29,34,35], although debate continues as to whether this represents normal ageing or pathology [36,37]. High BMI was also found to be an independent predictor of worsening kidney function, consistent with other studies [38 40]. Female gender was an independent predictor of reduced egfr but not of increased cystatin C concentration. Coresh et al. [3] observed a higher prevalence of CKD amongst women than men but report that this difference disappeared after adjustment for age. Conversely, Stevens et al. report an increased CKD prevalence amongst women even after age-standardization [30] and Chadban et al. [29] report female gender to be independently predictive of reduced kidney function. Clearly, egfr includes a gender term in the equation and there is some concern that this term could be inaccurate. Although Froissart et al. [41] found the MDRD equation to be similarly accurate compared to a reference GFR procedure in older men and women, Cirrillo et al. [42] found that the MDRD equation significantly underestimated GFR in women. The absence of an independent effect of gender on serum cystatin C concentration is consistent with other published data [18,43]. We were somewhat surprised that diabetes and cardiovascular disease did not emerge as more consistent independent predictors of reduced kidney function in this cohort. Diabetes was independently predictive of reduced egfr but not increased cystatin C, whereas a composite of vascular disease was predictive of increased cystatin C but not reduced egfr. The relationship between reduced GFR and increased prevalence of diabetes, hypertension [3] and cardiovascular disease [30] is well known, although in the AusDiab study only diabetes remained

6 1262 J. L. Carter et al. an independent predictor of reduced egfr on multivariable regression analysis [29]. A relatively large (>1200) study of the non-institutionalized elderly population in Finland also found increased cystatin C to be predicted by older age, increased BMI (females only), hypertension and cardiovascular disease, but not by the presence of diabetes [43]. Inclusion of hypertension, stroke and cardiovascular disease separately in our multiple regression models did not yield significant associations with either cystatin C or egfr (data not shown). It is possible that selective mortality of individuals, with both more advanced kidney disease plus diabetes and/or cardiovascular disease, is masking a true effect of these risk factors for reduced kidney function in this cross-sectional study of this population. Additionally, the relatively small size of our study limits the statistical power to detect such effects. Others [44,45] have reported an association between cognitive impairment and mild to moderate CKD in older people. In our study dementia was not an independent predictor of declining kidney function although the relatively low prevalence of dementia in our study (26%) reflects the exclusion of individuals with severe cognitive impairment. What are the implications of the high CKD prevalence for care of residents? The majority of residents we have identified as having reduced kidney function can probably continue to be managed outside of secondary care. Nevertheless, the identification of CKD may have significant implications for their management. Reduced GFR [46,47] and increased cystatin C [48,49] in older people are predictive of increased mortality, particularly from cardiovascular disease, and identification of CKD should, where appropriate, act as a spur to optimize risk management. Many people in care homes in the United Kingdom are very frail, have multiple medical problems and may have cognitive impairment. It could be debated that optimizing cardiovascular risk by the use of antihypertensives and cholesterol-lowering drugs could cause more harm than good. However, we need to weigh up the benefits of preventing, for instance, a major stroke against the risks of pharmacotherapy. Management decisions should be made on a case-by-case basis, but in the clear knowledge of risk, including the presence or absence of kidney disease. It may be that those with more significant CKD should be jointly managed by nephrologists in collaboration with geriatricians and primary care physicians who are more used to performing comprehensive geriatric assessments. Once patients with CKD are identified their risk of renal progression should be assessed, including measurement of urinary protein excretion, and balanced clinical decisions about the risks and benefits of further CKD management should be made. The presence of diabetes mellitus and GFR <30 ml/min/1.73 m 2 would appear to identify older people at particular risk of renal progression [50]. As a minimum, attention should be paid to the risks of prescribing renally excreted and/or potentially nephrotoxic drugs to patients with renal impairment. We suspect that optimizing the osteoporosis-related fracture reduction and management of anaemia [51] in this population may also benefit from improved knowledge of an individual s renal function. Limitations of the study We attempted to avoid biases inherent in the recruitment process and have therefore described our selection process in some detail; nevertheless, it is possible that some bias remains. Undertaking research in this population is clearly not straightforward. However, the CKD prevalence we have observed is also supported by the almost universal demonstration of increased serum cystatin C concentrations in this cohort. Further, we did not measure urinary protein in our cohort, which could have lead to the identification of CKD in some individuals with relatively preserved GFR. It is known that recent cooked meat intake can significantly lower creatinine-based GFR estimates [52]. Our patients were not sampled in the fasting state and this potentially introduces bias into our data. However, only 41 (16%) of our subjects were sampled after lunch, the remainder having been sampled in the morning. It therefore seems unlikely that our results were biased by significant meat consumption. A diagnosis of CKD stage 3 or worse requires the presence of reduced kidney function to be documented over at least 3 months [5] whereas our analysis is based on data from a single time point. Undoubtedly some patients with egfr <60 ml/min/1.73 m 2 would have demonstrated a higher egfr on repeat testing, and vice versa. However, epidemiological studies against which we have compared our data have used a similar approach [3,28 30]. Conclusion In conclusion, we have identified a previously unrecognized high prevalence of CKD in the residential care home population. This finding has important implications for the management and care of this vulnerable population. Acknowledgements. The authors are grateful to the managers and residents of east Kent care homes for participating in this study, Dr S Burns for initial help with the study and the staff of the Clinical Biochemistry Department for their cooperation and help. The study received statistical advice at the protocol development stage from Dr C Cryer, Centre for Health Services Studies, University of Kent. The study received financial support from the British Renal Society (grant reference ). Conflict of interest statement. None declared. References 1. Jungers P, Chauveau P, Descamps-Latscha B et al. Age and genderrelated incidence of chronic renal failure in a French urban area: a prospective epidemiologic study. Nephrol Dial Transplant 1996; 11: Magnason RL, Indridason OS, Sigvaldason H et al. Prevalence and progression of CRF in Iceland: a population-based study. Am J Kidney Dis 2002; 40: Coresh J, Astor BC, Greene T et al. Prevalence of chronic kidney disease and decreased kidney function in the adult US population: Third

7 CKD in residential care 1263 National Health and Nutrition Examination Survey. AmJKidneyDis 2003; 41: Matthews FE, Dening T. Prevalence of dementia in institutional care. Lancet 2002; 360: Levey AS, Eckardt KU, Tsukamoto Y et al. Definition and classification of chronic kidney disease: a position statement from Kidney Disease: Improving Global Outcomes (KDIGO). Kidney Int 2005;67: O Riordan SE, Webb MC, Stowe HJ et al. Cystatin C improves the detection of mild renal dysfunction in older patients. Ann Clin Biochem 2003; 40: Swedko PJ, Clark HD, Paramsothy K et al. Serum creatinine is an inadequate screening test for renal failure in elderly patients. Arch Intern Med 2003; 163: Duncan L, Heathcote J, Djurdjev O et al. Screening for renal disease using serum creatinine: who are we missing? Nephrol Dial Transplant 2001; 16: Department of Health. National Service Framework for Renal Services. Part Two: Chronic Kidney Disease, Acute Renal Failure and End of Life Care, Available at: /en/publicationsandstatistics/publications/publicationspolicyandguidance/dh_ (accessed 1 October 2007) 10. Department of Health. Estimated glomerular filtration rate (egfr), Available at: /index.htm?searchterms=egfr (accessed 1 October 2007) 11. National Kidney Foundation. Clinical practice guidelines for chronic kidney disease: evaluation classification and stratification. Am J Kidney Dis 2002; 39: S1 S Joint Specialty Committee on Renal Medicine of the Royal College of Physicians and the Renal Association and the Royal College of General Practitioners. Chronic Kidney Disease in Adults: UK Guidelines for Identification, Management and Referral. London: Royal College of Physicians, Available at: 1 October 2007) 13. Mathew TH. Chronic kidney disease and automatic reporting of estimated glomerular filtration rate: a position statement. Med J Aust 2005; 183: Levey AS, Bosch JP, Lewis JB et al. 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: Levey AS, Greene T, Kusek JW et al. A simplified equation to predict glomerular filtration rate from serum creatinine. J Am Soc Nephrol 2000; 11(Suppl): 155A 16. Levey AS, Coresh J, Greene T et al. Using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate. Ann Intern Med 2006; 145: Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976; 16: Newman DJ, Cystatin C Ann Clin Biochem 2002; 39: Mazzachi BC, Peake MJ, Ehrhardt V. Reference range and method comparison studies for enzymatic and Jaffe creatinine assays in plasma and serum and early morning urine. Clin Lab 2000; 46: Du Bois E, Du Bois D. A formula to estimate the approximate surface area if height and weight be known. Arch Intern Med 1916; 17: Vickery S, Stevens PE, Dalton RN et al. Does the ID-MS traceable MDRD equation work and is it suitable for use with compensated Jaffe and enzymatic creatinine assays? Nephrol Dial Transplant 2006; 21: Archibald G, Bartlett W, Brown A et al. UK Consensus Conference on Early Chronic Kidney Disease, 6 and 7 February Nephrol Dial Transplant 2007; 22: Van Den Noortgate NJ, Janssens WH, Delanghe JR et al. Serum cystatin C concentration compared with other markers of glomerular filtration rate in the old old. J Am Geriatr Soc 2002; 50: Pedone C, Corsonello A, Incalzi RA. Estimating renal function in older people: a comparison of three formulas. Age Ageing 2006; 35: Garg AX, Papaioannou A, Ferko N et al. Estimating the prevalence of renal insufficiency in seniors requiring long-term care. Kidney Int 2004; 65: Lamb EJ, Webb MC, O Riordan SE. Using the modification of diet in renal disease (MDRD) and Cockcroft and Gault equations to estimate glomerular filtration rate (GFR) in older people. Age Ageing 2007;in press. PMID: doi: /ageing/afm Tietz NW, Shuey DF, Wekstein DR. Laboratory values in fit aging individuals sexagenarians through centenarians. Clin Chem 1992;38: Hallan SI, Dahl K, Oien CM et al. Screening strategies for chronic kidney disease in the general population: follow-up of cross sectional health survey. BMJ 2006; 333: Chadban SJ, Briganti EM, Kerr PG et al. Prevalence of kidney damage in Australian adults: the AusDiab kidney study. J Am Soc Nephrol 2003; 14: S131 S Stevens PE, O Donoghue DJ, de Lusignan S et al. Chronic kidney disease management in the United Kingdom: NEOERICA project results. Kidney Int 2007; 72: Zochling J, Chen JS, Seibel M et al. Calcium metabolism in the frail elderly. Clin Rheumatol 2005; 24: Finney H, Bates CJ, Price CP. Plasma cystatin C determinations in a healthy elderly population. Arch Gerontol Geriatr 1999; 29: Nakamura H, Fukushima H, Miwa Y et al. A longitudinal study on the nutritional state of elderly women at a nursing home in Japan. Intern Med 2006; 45: Lindeman RD, Tobin J, Shock NW. Longitudinal studies on the rate of decline in renal function with age. J Am Geriatr Soc 1985; 33: Rowe JW, Andres R, Tobin JD et al. The effect of age on creatinine clearance in men: a cross-sectional and longitudinal study. J Gerontol 1976; 31: Lindeman RD. Is the decline in renal function with normal aging inevitable? Geriatr Nephrol Urol 1998; 8: Lamb EJ, O Riordan SE, Delaney MP. Kidney function in older people: pathology, assessment and management. Clin Chim Acta 2003; 334: Kramer H, Luke A, Bidani A et al. Obesity and prevalent and incident CKD: the Hypertension Detection and Follow-Up Program. Am J Kidney Dis 2005; 46: Gelber RP, Kurth T, Kausz AT et al. Association between body mass index and CKD in apparently healthy men. AmJKidneyDis2005; 46: Hallan S, de Mutsert R, Carlsen S et al. Obesity, smoking, and physical inactivity as risk factors for CKD: are men more vulnerable? Am J Kidney Dis 2006; 47: Froissart M, Rossert J, Jacquot C et al. Predictive performance of the modification of diet in renal disease and Cockcroft Gault equations for estimating renal function. J Am Soc Nephrol 2005; 16: Cirillo M, Anastasio P, De Santo NG. Relationship of gender, age, and body mass index to errors in predicted kidney function. Nephrol Dial Transplant 2005; 20: Wasen E, Isoaho R, Mattila K et al. Serum cystatin C in the aged: relationships with health status. Am J Kidney Dis 2003; 42: Kurella M, Chertow GM, Fried LF et al. Chronic kidney disease and cognitive impairment in the elderly: the health, aging, and body composition study. J Am Soc Nephrol 2005; 16: Seliger SL, Siscovick DS, Stehman-Breen CO et al. Moderate renal impairment and risk of dementia among older adults: the Cardiovascular Health Cognition Study. J Am Soc Nephrol 2004; 15: Maaravi Y, Bursztyn M, Hammerman-Rozenberg R et al. Glomerular filtration rate estimation and mortality in an elderly population. QJM 2007; 100:

8 1264 J. L. Carter et al. 47. Maaravi Y, Bursztyn M, Hammerman-Rozenberg R et al. Moderate renal insufficiency at 70 years predicts mortality. QJM 2006; 99: Shlipak MG, Katz R, Fried LF et al. Cystatin-C and mortality in elderly persons with heart failure. J Am Coll Cardiol 2005; 45: Shlipak MG, Sarnak MJ, Katz R et al. Cystatin C and the risk of death and cardiovascular events among elderly persons. NEnglJMed2005; 352: Hemmelgarn BR, Zhang J, Manns BJ et al. Progression of kidney dysfunction in the community-dwelling elderly. Kidney Int 2006; 69: National Institute for Health and Clinical Excellence. Anaemia management in people with chronic kidney disease, Available at: (accessed 1 October 2007) 52. Preiss DJ, Godber IM, Lamb EJ et al. The influence of a cooked-meat meal on estimated glomerular filtration rate. Ann Clin Biochem 2007; 44: Received for publication: Accepted in revised form: