Does Renal Dysfunction Predict Mortality After Acute Stroke?
A 7-Year Follow-Up Study
Background and Purpose— The purpose of this study was to investigate renal function as a long-term predictor of mortality in patients hospitalized for acute stroke.
Methods— This was a cohort study done in a Scottish tertiary teaching hospital. Participants included 2042 (993 male) unselected consecutive stroke patients (mean age, 73 years) admitted to hospital within 48 hours of stroke between1988 and 1994. Follow-up was up to 7 years. Main outcome measure was all-cause mortality.
Results— The total number of deaths at the end of follow-up was 1026. Most subjects (1512) had creatinine <124 μmol/L. The mean calculated creatinine clearance was 54.8 mL/min (SD, 23 mL/min). Renal function indexes were analyzed by quartiles with Cox proportional-hazards model. Stroke survivors had higher calculated creatinine clearance and lower serum creatinine, urea, and ratios of urea to creatinine. Calculated creatinine clearance ≥51.27 mL/min significantly predicted better long-term survival in these stroke patients even after adjustment for confounders (age, neurological score, ischemic heart disease, hypertension, smoking, and diuretic use). Similarly, creatinine ≥119 μmol/L “relative risk (RR), 1.59; 95% confidence interval (CI), 1.32 to 1.92”, urea 6.8 to 8.9 mmol/L (RR, 1.34; 95% CI, 1.09 to 1.65) or ≥9 mmol/L (RR, 1.74; 95% CI, 1.42 to 2.13), and ratio of urea to creatinine ≥0.08573 mmol/μmol (RR, 1.24; 95% CI, 1.03 to 1.50) remained significant predictors of mortality after adjustment for confounders.
Conclusions— After acute stroke, patients with reduced admission calculated creatinine clearance, raised serum creatinine and urea concentrations (even within conventional reference intervals), and raised ratio of urea to creatinine had a higher mortality risk. This finding may be used to stratify risk and target interventions, eg, the use of angiotensin-converting enzyme inhibitors.
- cerebrovascular disorders
- kidney function tests
- renal disease, end-stage
Renal dysfunction, even to a subtle degree, has been noted to be a prognostic indicator of overall mortality in the presence of comorbidities such as diabetes.1 Renal function has proved to be prognostic in many patients groups. Raised urea concentration predicted mortality in patients with pneumonia and gastrointestinal bleeding and after coronary artery bypass graft.2 Creatinine concentration among cohorts of post–myocardial infarction survivors3 and hypertensive patients4 also predicted mortality. More recently, calculated creatinine clearance has been shown to be an independent predictor of morbidity and mortality in patients with left ventricular dysfunction5,6⇓ and among patients admitted to a coronary care unit. 7 The latter cohort had a graded increase in the relative risk of cardiac problems and earlier mortality in patients with decreasing renal function.7
An association between renal function and stroke has previously been noted in a study designed to assess the relationship between hypertension and fatality rates and its determinants in black patients with recent stroke. Survivors of acute stroke had lower plasma urea on admission than those who died.8 Friedman 9 found that among stroke survivors (mean age, 75±7.5years) in New Zealand, serum creatinine concentration independently predicted mortality even after adjustment for confounders. However, more direct measures of renal function were not done (eg, creatinine clearance), and that study was small, involving 492 patients followed up for a mean of only 18 months.
In this study, we further investigated renal function as a long-term predictor of mortality with a large cohort of acute stroke patients using several markers of renal function obtained on admission to hospital.
Subjects and Methods
The study was approved by the Tayside Committee on Medical Research Ethics and was carried out at Ninewells Hospital and Medical School, a major teaching hospital in Scotland. The study team saw 2042 unselected consecutive patients within 48 hours of acute stroke between 1988 and 1994.
All patients had a clinical diagnosis of stroke confirmed in hospital and were admitted to hospital within 48 hours of symptom onset.
Patients were excluded if they were already receiving dialysis for chronic renal failure.
Details of patient demography, neurological status and disabilities, and current drug use were recorded. Other relevant clinical data, including ECG and CT scan results, were also collected. Laboratory data, including serum urea, creatinine, and glucose concentration on admission, were collected with the use of a hospital database system. Estimates of renal function included serum urea and creatinine concentrations, calculated creatinine clearance, and ratio of urea to creatinine. Creatinine clearance (mL/min) was calculated by use of a recent version of the Cockcroft-Gault equation as follows10: (140−age)×2.12×weight (kg)×k/(plasma creatinine×BSA), where k=1 if male or 0.85 if female and body surface area (BSA) is calculated as follows: BSA= 0.20247×height (m)0.725×weight (kg)0.425.
The primary outcome was total death from any cause during the 7-year follow- up period (minimum, 1 year; maximum, 7 years). The certified cause of death was obtained by record linkage with data from the Registrar General in Scotland. The accuracy of this data has been shown to be 98%.11
Urea, creatinine concentration, ratio of urea to creatinine, and creatinine clearance were all analyzed after categorization into quartile groups. Cox proportional-hazards models were used to estimate the relative risk of mortality (RR) with adjustment for age, neurological score, and comorbidities, including ischemic heart disease (including previous acute myocardial infarction, admission ischemic ECG, nitrates among their admission medication), hypertension, and the use of a diuretic on admission. RRs are given with 95% confidence intervals (CIs). For clinical relevance, analysis was repeated for creatinine as <124, 124 to 199, and ≥200 μmol/L.12 Total and subgroup survival was analyzed by Kaplan-Meier survival analysis and log-rank test; a log-rank value of P<0.05 was significant. SPSS was used for statistical analyses.
During the follow-up period, 1026 patients died. Baseline characteristics of this cohort are described in Table 1. Most patients in this cohort had normal renal function on admission according to creatinine concentration. The calculated creatinine clearance ranged from 6 to 235 mL/min. We found that 1512 patients had normal renal function defined by serum creatinine <124 μmol/L, 345 had mildly impaired renal function with creatinine 124 to 199 μmol/L, and 66 had clinically significantly impaired renal function with creatinine ≥200 μmol/L.12
This model showed that age, neurological score, renal function parameters on admission (calculated creatinine clearance, creatinine and urea concentrations, and ratio of urea to creatinine), presence of ischemic heart disease, and use of diuretics were all significantly adversely associated with outcome after acute stroke. More interesting is the finding that renal indexes remained significant predictors of mortality after taking into account significant confounders grouped as nonmodifiable risk factor (age), major neurological determinant (neurological score), and modifiable risk factor (ischemic heart disease, treated hypertension, smoking history, and diuretic use; Table 2). Preserved creatinine clearance correlated with better survival, and the second quartile (≤51.27 mL/min) of creatinine clearance significantly predicted mortality even after adjustment for confounders (Table 2). The Kaplan-Meier curve shows the association of quartile groups of calculated creatinine clearance with mortality (the Figure). Similarly, after adjustment for confounders, creatinine above quartile 3 (≥119 μmol/L) significantly predicted increased mortality. Assessing creatinine groups 124 to 199 and ≥ 200 μmol/L and comparing them with normal levels (<124 μmol/L) showed that creatinine also remained a significant predictor of long-term mortality after acute stroke (creatinine 124 to 199 μmol/L: RR, 1.65; 95% CI, 1.41 to 1.93; creatinine ≥200 μmol/L: RR, 1.63; 95% CI, 1.21 to 2.19). Urea above the median (≥6.8 mmol/L) also significantly predicted increased mortality after acute stroke after adjustment for all confounders (Table 2). The highest quartile of the ratio of urea to creatinine (>0.086 mmol/μmol) significantly predicted death at follow-up (Table 2). Known diabetes on admission (RR, 1.03; 95% CI, 0.83 to 1.28) and sex (RR, 0.91; 95% CI, 0.81 to 1.03) did not significantly influence prognosis. Renal function remained highly significant in predicting mortality among subgroups: those who died at <30 days compared with those who died at final follow-up, those who had CT showing ischemic compared with hemorrhagic strokes, those with a first stroke compared with recurrent strokes, and stroke subtypes as classified by the Bamford13 classification. Renal function was not significant in predicting outcome in hemorrhagic strokes in which the number of fatal events was too small to achieve significance (Table 3).
The prognostic value of renal dysfunction has been seen in many other patient groups.1–7⇓⇓⇓⇓⇓⇓ This prospective, long-term study demonstrated that admission renal function as assessed by calculated creatinine clearance, serum concentration of creatinine, urea, and ratio of urea to creatinine all predicted mortality even when values were within conventional normal reference intervals. This is consistent with the previous incidental findings that survivors of acute stroke had a lower urea concentration on admission8 and that creatinine concentration had prognostic value9 even within conventional population-based reference intervals.14
Stroke outcome in the first 30 days is determined largely by age, level of consciousness, and type of stroke. Comorbidities such as diabetes mellitus or hyperglycaemia on admission,15 cardiovascular disease, including congestive heart failure, and hypertension also predicted an elevated mortality in stroke patients in longer-term follow-up.15,16⇓ In this study, we confirmed that age was a profound long-term predictor of death after stroke. Patients >85 years of age were >8 times more likely to die compared with a patient with a stroke at ≤59 years of age. This probably suggests frailty and a higher burden of chronic disorders in those with advancing age (>80 years of age) in whom the most frequent conditions tend to be heart failure, chronic obstructive pulmonary disease, stroke, and pneumonia.17 Similarly, the major neurological determinant of outcome, neurological score on admission, significantly predicted long-term mortality.
In the Oxford Community Stroke Project, long-term stroke survivors of >30 days had an increased risk of dying over the next few years compared with the general population. The cause of death during the first 30 days after stroke was stroke-related complications (eg, pneumonia). Nonstroke cardiovascular disease becomes the most common cause of death after the first year.16 Renal failure was a very rare primary cause of death in our cohort. However, renal dysfunction was a significant predictor of increased mortality in both the short and long term. The degree of renal dysfunction present in these stroke patients may simply be a marker of end-organ damage from undetected preexisting untreated hypertension, eg, left ventricular hypertrophy, a potent predictor of mortality.18 Renal dysfunction may also indicate a higher comorbidity burden, especially atherosclerotic disease. Indeed, a prospective general population study of British men who were followed up for >14 years found that high serum creatinine, even within the reference interval, was a marker for increased risk of cerebrovascular disease among both the normotensive and hypertensive population.14 Our results are consistent with analyses of the Heart Outcomes and Prevention Evaluation (HOPE)12 and Hypertension Optimal Treatment (HOT)19 studies in which mild renal impairment independently predicted increased cardiovascular mortality, total mortality, and incidence of myocardial infarction and stroke in their cohort of high-cardiovascular-risk and hypertensive patients. Conversely, renal dysfunction may correlate with endothelial dysfunction because patients with end-stage renal disease have increased arterial stiffness independent of other risk factors for atherosclerosis. Even among patients with mild to moderate renal insufficiency, there was increased central artery stiffness, suggesting that renal dysfunction adversely affected small and large arteries, 20 which may contribute to cardiovascular morbidity and mortality.
This study demonstrated that renal dysfunction predicted long-term mortality and therefore can be used to stratify risk for stroke patients. A stroke management guideline has already suggested that urea and creatinine measurements are important in the acute situation.21 In stroke patients, whether the outcome improves with intervention to lower urea and creatinine is not yet known. In this cohort of patients, management was routinely intravenous fluid replacement therapy and early nasogastric feeding. Although a disproportionally high urea to creatinine ratio may reflect dehydration, that ratio in fact predicted mortality throughout the range of results. A subgroup analysis in the HOPE study comparing patients with mild renal impairment with those with normal renal function showed they were more likely to have cardiovascular events and die even after adjustment for other cardiovascular risk factors and treatments. Ramipril decreased the hazard ratio for these end points in both groups.12 The use of angiotensin-converting enzyme inhibitors in high-cardiovascular-risk patients (eg, previous stroke) is further supported by the results of the Perindopril Protection Against Recurrent Stroke (PROGRESS) study, which showed that blood pressure lowering was important in secondary prevention.22
Glomerular filtration rate is the gold standard for measuring renal function. Serum creatinine concentration is widely interpreted as a measure of the glomerular filtration rate and is used as an index of renal function in clinical practice, which is far easier than and as accurate a guide to glomerular filtration rate as 24-hour urine collection in patients with stable renal function.23 The glomerular filtration of creatinine, however, is only one of the variables that determine its concentration in the serum, which could be altered by increasing age (change in renal handling), diabetes, renal failure, creatinine metabolism, and methodological interference in its measurement. More accurate measures of renal function are frequently necessary. More specific measures of renal function other than glomerular filtration rate such as serum cystatin C10 could be used in the future to assess more intimately the relationship between renal dysfunction and morbidity or mortality in patients who with stroke.
A limitation of our study is that community hospital– or home-treated stroke patients were naturally excluded from the study. These patients are likely to have had less severe strokes, and it is not certain whether renal dysfunction in this group will similarly predict mortality.
Raised admission urea and creatinine concentrations even within conventional reference intervals, raised ratio of urea to creatinine, and decreased calculated creatinine clearance in patients with acute stroke predicted long-term mortality independently of age, neurological score, ischemic heart disease, hypertension, smoking, and diuretic use. Renal dysfunction therefore can help stratify risk in stroke patients for further interventions such as cardiovascular investigations or use of angiotensin-converting enzyme inhibitor in the absence of contraindications.
Funding for this study was provided by Tayside Health Board, Tayside University Hospitals NHS Trust, and University of Dundee. S.Y.S.W. is supported by an educational grant from Pharmacia; K.Y.K.W., by the British Heart Foundation. We thank S. Kendrick and the ISD, National Health Service, Edinburgh, Scotland, for providing the mortality data; Helen Brewster and Wendy Cassels, who collected the data in this study; and Valerie Bruce, who entered the data.
- Received December 13, 2001.
- Revision received February 4, 2002.
- Accepted February 14, 2002.
- ↵Shulman NB, Ford CE, Hall WD, Blaufox MD, Simon D, Langford HG, Schneider KA. Prognostic value of serum creatinine and effect of treatment of hypertension on renal function: results from the hypertension detection and follow-up program: the Hypertension Detection and Follow-up Program Cooperative Group. Hypertension. 1989; 13 (suppl I): I80–I93.
- ↵Hillege HL, Girbes AR, de Kam PJ, Boomsma F, de Zeeuw D, Charlesworth A, Hampton JR, van Veldhuisen DJ. Renal function, neurohormonal activation, and survival in patients with chronic heart failure. Circulation. 2000; 102: 203–210.
- ↵Finney H, Newman DJ, Price CP. Adult reference ranges for serum cystatin C, creatinine and predicted creatinine clearance. Ann Clin Biochem. 2000; 37: 49–59.
- ↵Macintyre K, Stewart S, Chalmers J, Pell J, Finlayson A, Boyd J, Redpath A, McMurray J, Capewell S. Relation between socioeconomic deprivation and death from a first myocardial infarction in Scotland: population based analysis. BMJ. 2001; 322: 1152–1153.
- ↵Wannamethee SG, Shaper AG, Perry IJ. Serum creatinine concentration and risk of cardiovascular disease: a possible marker for increased risk of stroke. Stroke. 1997; 28: 557–563.
- ↵Sacco RL, Shi T, Zamanillo MC, Kargman DE. Predictors of mortality and recurrence after hospitalized cerebral infarction in an urban community: the Northern Manhattan Stroke Study. Neurology. 1994; 44: 626–634.
- ↵Dennis MS, Burn JP, Sandercock PA, Bamford JM, Wade DT, Warlow CP. Long-term survival after first-ever stroke: the Oxfordshire Community Stroke Project. Stroke. 1993; 24: 796–800.
- ↵Kannel WB. Left ventricular hypertrophy as a risk factor in arterial hypertension. Eur Heart J. 1992; 13 (suppl D): 82–88.
- ↵Zanchetti A, Hansson L, Dahlof B, Elmfeldt D, Kjeldsen S, Kolloch R, Larochelle P, McInnes GT, Mallion JM, Ruilope L, Wedel H. Effects of individual risk factors on the incidence of cardiovascular events in the treated hypertensive patients of the Hypertension Optimal Treatment Study: HOT Study Group. J Hypertens. 2001; 19: 1149–1159.
- ↵SIGN (Scottish Intercollegiate Guidelines Network). Management of patients with stroke, part I: assessment, investigation, immediate management and secondary prevention. SIGN Guideline No. 13. Edinburgh, SIGN, RCPE; 1997.
- ↵Waller DG, Fleming JS, Ramsey B, Gray J. The accuracy of creatinine clearance with and without urine collection as a measure of glomerular filtration rate. Postgrad Med J. 1991; 67: 42–46.