I Contrast Nephropathy in Azotemic Diabetic Patients Undergoing Coronary Angiography CONNIEL. MANSKE,M.D., J. MICHAELSPRAFKA,M.P.H., Ph.D., JOHNT. STRONY,M.D., YANGWANG, M.D., Minneapolis, Minnesota
PURPOSE: To evaluate the incidence of, risk factors for, and outcome of contrast nephropathy in azotemic diabetic patients undergoing coronary angiography.
PATIENTS AND METHODS: F i f t y - n i n e insulin-de-
pendent diabetics with a mean serum creatinine level of 522 #mol/L (5.9 mg/dL) u n d e r w e n t coronary angiography as p a r t of a pretrangplant evaluation. T w e n t y - f o u r azotemic diabetics undergoing inpatient evaluation not including angiography for transplantation formed the control group. Serum creatlnine measurements obtained at baseline and after radiocontrast exposure w e r e compared in patients and control subjects. Risk factors for contrast nephropathy w e r e evaluated in patients with a 25 % or greater increase in serum creatinine. RESULTS: Serum creatinine was significantly elevated 24 hours after radiocontrast exposure in patients (557 ± 141 #mol/L versus 522 ± 141 #tool/L, mean ± SD; p (0.001) but not in controis. Seven patients required dialysis within 6 days of coronary angiography and two additional patients required dialysis within 14 days. Contrast nephropathy, defined as a serum creatinine increase of g r e a t e r than 25% when measured 48 hours after radiocontrast exposure, occurred in 50% of patients and no controls. Univariate analysis of risk factors for contrast nephropathy revealed a significant association with dye q - a n tity (p = 0.002), mean arterial pressure less than 100 mm Hg (p = 0.02), and ejection fraction less
From the Departments of Medicine and Epidemiology, University of Minnesota School of Medicine and School of Public Health, Minneapolis, Minnesota. Dr. Manske was supported by the Division of Research Resources Grant MOI RRO04400 and National Institutes of Health Grant DK 13083. Requests for reprints should be addressed to Connie L. Manske, M.D., University of Minnesota, Department of Medicine, Box 736 UMHC, 516 Delaware Street S.E., Minneapolis, Minnesota 55455. Current addresses: Drs. Manske and Wang, University of Minnesota, Department of Medicine, Minneapolis, Minnesota; Dr. Strony, Department of Medicine, Cardiology Division, Medical College of Virginia, Richmond, Virginia; Dr. Sprafka, University of Minnesota School of Public Health, Division of Epidemiology, Minneapolis, Minnesota. Manuscript submitted May 7, 1990, and accepted in revised form July 23, 1990.
than 50% (p = 0.04). Stepwise logistic regression verified the independence of dye quantity and low mean arterial pressure but not low ejection fraction as risk factors for contrast nephropathy. Follow-up serum creatinlne values were not significantly different in patients and control subjects. CONCLUSIONS: A z o t e m i c patients with diabetes are at high risk of developing contrast nephropathy even when less than 100 mL of radiocontrast agent is used. The acute renal failure is reversible but precipitates the need for short-term dialysis in some patients. Radiocontrast q , a n t i t y is an important risk factor not previously noted. The incidence of contrast nephropathy can be mlnimiT£d by using less than 30 mL of radiocontrast agent.
cute renal failure associated with intravascular A administration of radiocontrast material occurs frequently in hospitalized patients and accounts for 10% of cases of hospital-acquired renal failure [1]. Retrospective uncontrolled studies suggest that diabetic patients with a serum creatinine level higher than 354 #mol/L (4 mg/dL) have a 45% to 95% incidence of acute renal failure after exposure to radiocontrast agents [2-7]. Standard textbooks continue to state that diabetics with chronic renal failure are a prohibitively high-risk group to undergo radiocontrast studies [8]. Yet, in the only prospective study evaluating diabetics, the incidence of contrast nephropathy was only 9%. This series included older patients with non-insulin-dependent diabetes who may not have had diabetic nephropathy as a cause of chronic renal failure, and only one diabetic with a serum creatinine greater than 398/~mol/L [9]. Therefore, the risk of contrast nephropathy in diabetic patients with advanced chronic renal insufficiency remains uncertain. Recent studies have revealed a high incidence of silent myocardial ischemia and premature cardiac death in azotemic diabetics [10-13]. Therefore, cardiac catheterization has become increasingly necessary for optimal management. Since there is a sharp increase in the prevalence of coronary artery dis-
November 1990
The American Journal of Medicine
Volume 89
615
CONTRAST NEPHROPATHYIN AZOTEMICDIABETICS/ MANSKE ET AL
ease during the first 6 years after onset of proteinuria, many patients are not yet undergoing dialysis when coronary angiography is needed [14]. However, angiography is often postponed until the patient is receiving dialysis treatments, because of the perception that this group is at high risk of developing contrast nephropathy. Because the incidence of contrast nephropathy in diabetics with advanced chronic renal failure is uncertain and the role for coronary angiography is increasing, we designed a prospective study to evaluate the incidence and outcome of acute renal failure after radiocontrast exposure. We compared changes in renal function of hospitalized azotemic diabetic patients undergoing elective coronary angiography with the changes in a control group of hospitalized azotemic diabetics not undergoing angiography. We then compared patients who did or did not develop contrast nephropathy to determine whether specific risk factors might be associated with a poor outcome.
diac catheterization with that obtained 3 hours after the procedure. Mean serum glucose was determined by averaging serum glucose values measured immediately before and after the procedure. Serum creatinine values were obtained at baseline and 24 hours after angiography in all patients and 24 hours following admission in control subjects. Follow-up creatinine measurements were obtained in patients and controls between 3 and 6 weeks after angiography. If a patient or control began dialysis during that period, the serum creatinine value obtained immediately prior to dialysis initiation was used. Follow-up creatinine values were also compared excluding patients who began dialysis. Serum creatinine values obtained at baseline and 24 hours following angiography were compared using a paired t-test. Follow-up creatinine values in patients and controls were compared using a twosample t-test to determine the long-term effect of radiocontrast agent exposure on renal function. A serum creatinine value measured 24 hours after angiography that was greater than 10% above baseline PATIENTSAND METHODS was considered significantly increased. The sensiAzotemic patients with insulin-dependent diabe- tivity and specificity of a significant creatinine intes referred to the University of Minnesota Hospi- crease 24 hours after angiography for prediction of tal between February 1987 and May 1989 for evalu- need for acute dialysis were assessed. ation for renal transplantation were eligible for this Patients with a serum creatinine value measured study. Patients were included in this study if they 48 hours after angiography that was greater than had developed insulin-dependent diabetes before 25% above baseline were considered to have develage 21 or had absent C-peptide, and if a 24-hour oped contrast nephropathy. Data were analyzed to urine creatinine clearance obtained on the first hos- define the incidence of contrast nephropathy and pital day was less than 30 mL/minute. Patients identify clinical factors related to its occurrence. were excluded if they were already undergoing dial- Discrete variables were analyzed for univariate asysis, had begun angiotensin-converting enzyme in- sociation by Fisher's exact test. Means and stanhibitor therapy within 1 week of study entry, or dard deviations for continuous variables were deunderwent coronary artery bypass graft surgery termined and examined for univariate association within 3 days of entry into the study. Patients un- by the two-sample t-test. When the distribution dergoing coronary angiography formed the study was not normal, median values were compared usgroup. The control group included 21 azotemic pa- ing the Wilcoxon rank-sum test. Stepwise logistic tients with insulin-dependent diabetes undergoing regression analysis was used to identify those varievaluation for renal transplantation between July ables independently associated with the occurrence 1985 and February 1987, when coronary angiogra- of contrast nephropathy. phy was not performed as part of the transplant evaluation, and three patients evaluated between RESULTS February 1987 and May 1989 refused angiography. From February 1987 to May 1989, 59 patients Coronary angiography was performed using bi- with insulin-dependent diabetes referred to the plane cineangiography to minimize radiocontrast University of Minnesota Hospital for transplant quantity. Left ventricular function was assessed evaluation underwent coronary angiography and with radionuclide scanning. All patients received a met the study criteria. Eleven additional insulinprehydration protocol consisting of 150 mL/hour of dependent diabetics were considered but could not 5% dextrose in 0.45% normal saline with 25 g of be included because of inadequate follow-up (sevmannitol beginning 2 hours prior to the procedure. en), emergent coronary artery bypass surgery (two), Nonionic radiocontrast agents (iohexol or iopami- or initiation of converting enzyme inhibitor therapy dol) were used for all patients. Mean arterial blood (two). Between July 1985 and May 1989, 24 addipressure was determined by averaging the mean tional patients with insulin-dependent diabetes unarterial pressure obtained immediately prior to car- derwent identical inpatient evaluation for renal 616
November 1990 The American Journal of Medicine Volume 89
CONTRAST NEPHROPATHY IN AZOTEMIC DIABETICS / MANSKE ET AL
transplantation but did not have coronary angiography. These patients formed the control group. The clinical profile of patients and controls is shown in Table I. The mean serum creatinine in patients was 522 #mol/L (range 309 to 911 #mol/L), corresponding to a mean creatinine clearance of 13.6 mL/minute. Radiocontrast dose was recorded in all patients; the mean dose was 31 mL (range 12 to 90 mL, median 25 mL). The serum creatinine obtained at 24 hours was significantly elevated over baseline in patients (557 ± 141 #mol/L versus 522 ± 141 #mol/L, mean ± SD; p
PURPOSE: To evaluate the incidence of, risk factors for, and outcome of contrast nephropathy in azotemic diabetic patients undergoing coronary angiography.
PATIENTS AND METHODS: F i f t y - n i n e insulin-de-
pendent diabetics with a mean serum creatinine level of 522 #mol/L (5.9 mg/dL) u n d e r w e n t coronary angiography as p a r t of a pretrangplant evaluation. T w e n t y - f o u r azotemic diabetics undergoing inpatient evaluation not including angiography for transplantation formed the control group. Serum creatlnine measurements obtained at baseline and after radiocontrast exposure w e r e compared in patients and control subjects. Risk factors for contrast nephropathy w e r e evaluated in patients with a 25 % or greater increase in serum creatinine. RESULTS: Serum creatinine was significantly elevated 24 hours after radiocontrast exposure in patients (557 ± 141 #mol/L versus 522 ± 141 #tool/L, mean ± SD; p (0.001) but not in controis. Seven patients required dialysis within 6 days of coronary angiography and two additional patients required dialysis within 14 days. Contrast nephropathy, defined as a serum creatinine increase of g r e a t e r than 25% when measured 48 hours after radiocontrast exposure, occurred in 50% of patients and no controls. Univariate analysis of risk factors for contrast nephropathy revealed a significant association with dye q - a n tity (p = 0.002), mean arterial pressure less than 100 mm Hg (p = 0.02), and ejection fraction less
From the Departments of Medicine and Epidemiology, University of Minnesota School of Medicine and School of Public Health, Minneapolis, Minnesota. Dr. Manske was supported by the Division of Research Resources Grant MOI RRO04400 and National Institutes of Health Grant DK 13083. Requests for reprints should be addressed to Connie L. Manske, M.D., University of Minnesota, Department of Medicine, Box 736 UMHC, 516 Delaware Street S.E., Minneapolis, Minnesota 55455. Current addresses: Drs. Manske and Wang, University of Minnesota, Department of Medicine, Minneapolis, Minnesota; Dr. Strony, Department of Medicine, Cardiology Division, Medical College of Virginia, Richmond, Virginia; Dr. Sprafka, University of Minnesota School of Public Health, Division of Epidemiology, Minneapolis, Minnesota. Manuscript submitted May 7, 1990, and accepted in revised form July 23, 1990.
than 50% (p = 0.04). Stepwise logistic regression verified the independence of dye quantity and low mean arterial pressure but not low ejection fraction as risk factors for contrast nephropathy. Follow-up serum creatinlne values were not significantly different in patients and control subjects. CONCLUSIONS: A z o t e m i c patients with diabetes are at high risk of developing contrast nephropathy even when less than 100 mL of radiocontrast agent is used. The acute renal failure is reversible but precipitates the need for short-term dialysis in some patients. Radiocontrast q , a n t i t y is an important risk factor not previously noted. The incidence of contrast nephropathy can be mlnimiT£d by using less than 30 mL of radiocontrast agent.
cute renal failure associated with intravascular A administration of radiocontrast material occurs frequently in hospitalized patients and accounts for 10% of cases of hospital-acquired renal failure [1]. Retrospective uncontrolled studies suggest that diabetic patients with a serum creatinine level higher than 354 #mol/L (4 mg/dL) have a 45% to 95% incidence of acute renal failure after exposure to radiocontrast agents [2-7]. Standard textbooks continue to state that diabetics with chronic renal failure are a prohibitively high-risk group to undergo radiocontrast studies [8]. Yet, in the only prospective study evaluating diabetics, the incidence of contrast nephropathy was only 9%. This series included older patients with non-insulin-dependent diabetes who may not have had diabetic nephropathy as a cause of chronic renal failure, and only one diabetic with a serum creatinine greater than 398/~mol/L [9]. Therefore, the risk of contrast nephropathy in diabetic patients with advanced chronic renal insufficiency remains uncertain. Recent studies have revealed a high incidence of silent myocardial ischemia and premature cardiac death in azotemic diabetics [10-13]. Therefore, cardiac catheterization has become increasingly necessary for optimal management. Since there is a sharp increase in the prevalence of coronary artery dis-
November 1990
The American Journal of Medicine
Volume 89
615
CONTRAST NEPHROPATHYIN AZOTEMICDIABETICS/ MANSKE ET AL
ease during the first 6 years after onset of proteinuria, many patients are not yet undergoing dialysis when coronary angiography is needed [14]. However, angiography is often postponed until the patient is receiving dialysis treatments, because of the perception that this group is at high risk of developing contrast nephropathy. Because the incidence of contrast nephropathy in diabetics with advanced chronic renal failure is uncertain and the role for coronary angiography is increasing, we designed a prospective study to evaluate the incidence and outcome of acute renal failure after radiocontrast exposure. We compared changes in renal function of hospitalized azotemic diabetic patients undergoing elective coronary angiography with the changes in a control group of hospitalized azotemic diabetics not undergoing angiography. We then compared patients who did or did not develop contrast nephropathy to determine whether specific risk factors might be associated with a poor outcome.
diac catheterization with that obtained 3 hours after the procedure. Mean serum glucose was determined by averaging serum glucose values measured immediately before and after the procedure. Serum creatinine values were obtained at baseline and 24 hours after angiography in all patients and 24 hours following admission in control subjects. Follow-up creatinine measurements were obtained in patients and controls between 3 and 6 weeks after angiography. If a patient or control began dialysis during that period, the serum creatinine value obtained immediately prior to dialysis initiation was used. Follow-up creatinine values were also compared excluding patients who began dialysis. Serum creatinine values obtained at baseline and 24 hours following angiography were compared using a paired t-test. Follow-up creatinine values in patients and controls were compared using a twosample t-test to determine the long-term effect of radiocontrast agent exposure on renal function. A serum creatinine value measured 24 hours after angiography that was greater than 10% above baseline PATIENTSAND METHODS was considered significantly increased. The sensiAzotemic patients with insulin-dependent diabe- tivity and specificity of a significant creatinine intes referred to the University of Minnesota Hospi- crease 24 hours after angiography for prediction of tal between February 1987 and May 1989 for evalu- need for acute dialysis were assessed. ation for renal transplantation were eligible for this Patients with a serum creatinine value measured study. Patients were included in this study if they 48 hours after angiography that was greater than had developed insulin-dependent diabetes before 25% above baseline were considered to have develage 21 or had absent C-peptide, and if a 24-hour oped contrast nephropathy. Data were analyzed to urine creatinine clearance obtained on the first hos- define the incidence of contrast nephropathy and pital day was less than 30 mL/minute. Patients identify clinical factors related to its occurrence. were excluded if they were already undergoing dial- Discrete variables were analyzed for univariate asysis, had begun angiotensin-converting enzyme in- sociation by Fisher's exact test. Means and stanhibitor therapy within 1 week of study entry, or dard deviations for continuous variables were deunderwent coronary artery bypass graft surgery termined and examined for univariate association within 3 days of entry into the study. Patients un- by the two-sample t-test. When the distribution dergoing coronary angiography formed the study was not normal, median values were compared usgroup. The control group included 21 azotemic pa- ing the Wilcoxon rank-sum test. Stepwise logistic tients with insulin-dependent diabetes undergoing regression analysis was used to identify those varievaluation for renal transplantation between July ables independently associated with the occurrence 1985 and February 1987, when coronary angiogra- of contrast nephropathy. phy was not performed as part of the transplant evaluation, and three patients evaluated between RESULTS February 1987 and May 1989 refused angiography. From February 1987 to May 1989, 59 patients Coronary angiography was performed using bi- with insulin-dependent diabetes referred to the plane cineangiography to minimize radiocontrast University of Minnesota Hospital for transplant quantity. Left ventricular function was assessed evaluation underwent coronary angiography and with radionuclide scanning. All patients received a met the study criteria. Eleven additional insulinprehydration protocol consisting of 150 mL/hour of dependent diabetics were considered but could not 5% dextrose in 0.45% normal saline with 25 g of be included because of inadequate follow-up (sevmannitol beginning 2 hours prior to the procedure. en), emergent coronary artery bypass surgery (two), Nonionic radiocontrast agents (iohexol or iopami- or initiation of converting enzyme inhibitor therapy dol) were used for all patients. Mean arterial blood (two). Between July 1985 and May 1989, 24 addipressure was determined by averaging the mean tional patients with insulin-dependent diabetes unarterial pressure obtained immediately prior to car- derwent identical inpatient evaluation for renal 616
November 1990 The American Journal of Medicine Volume 89
CONTRAST NEPHROPATHY IN AZOTEMIC DIABETICS / MANSKE ET AL
transplantation but did not have coronary angiography. These patients formed the control group. The clinical profile of patients and controls is shown in Table I. The mean serum creatinine in patients was 522 #mol/L (range 309 to 911 #mol/L), corresponding to a mean creatinine clearance of 13.6 mL/minute. Radiocontrast dose was recorded in all patients; the mean dose was 31 mL (range 12 to 90 mL, median 25 mL). The serum creatinine obtained at 24 hours was significantly elevated over baseline in patients (557 ± 141 #mol/L versus 522 ± 141 #mol/L, mean ± SD; p
