- Email: [email protected]
systemic inflammatory response syndrome: results of a prospective study
Clinical Characteristics of Patients Developing ARF Due to Sepsis/Systemic Inflammatory Response Syndrome: Results of a Prospective Study Itir Yegenaga, MD, PhD, Erik Hoste, MD, Wim Van Biesen, MD, PhD, Raymond Vanholder, MD, PhD, Dominique Benoit, MD, Gulcin Kantarci, MD, PhD, Annemieke Dhondt, MD, PhD, Francis Colardyn, MD, and Norbert Lameire, MD, PhD ● Background: Acute renal failure (ARF) in patients with sepsis provokes high mortality and financial cost. In this prospective study, we collected characteristics of patients in the intensive care unit (ICU) who developed sepsis/systemic inflammatory response syndrome (SIRS) to analyze differences between those who subsequently did or did not develop ARF. Methods: All patients admitted to the ICU of the University Hospital Gent, Belgium, between January 1, 2001, and December 31, 2001, who developed sepsis/SIRS were included if they had a serum creatinine level less than 2 mg/dL (<177 mol/L). Results: Of 2,442 patients admitted to the ICU, 257 patients developed sepsis/SIRS. Of those, 29 patients (11%) developed ARF. In a univariate analysis, age, central venous pressure (CVP), and serum creatinine and blood urea nitrogen levels were greater (P ⴝ 0.003, P ⴝ 0.006, P < 0.001, and P < 0.001, respectively), whereas mean arterial and diastolic blood pressures, 24-hour urinary output, arterial pH, bicarbonate level, thrombocyte count, albumin level, and prothrombin time were lower (P ⴝ 0.05, P ⴝ 0.004, P ⴝ 0.005, P ⴝ 0.03, P ⴝ 0.009, P ⴝ 0.037, P ⴝ 0.05, and P ⴝ 0.006, respectively) in the ARF group. Prevalence of diabetes, sex, and need for ventilation were not different between the ARF and no-ARF groups, but in the ARF group, diuretic use, vasopressor use, and presence of primary hepatic failure were more prevalent (P ⴝ 0.001 for each). In a multivariate analysis, age, serum creatinine level, CVP, and presence of liver failure significantly contributed to a logistic regression model for ARF. Conclusion: Several parameters already were disturbed at the first day of SIRS/sepsis in patients who later developed ARF. Older age, elevated serum creatinine level despite elevated CVP, and presence of hepatic failure are predictive for ARF in septic patients. Am J Kidney Dis 43:817-824. © 2004 by the National Kidney Foundation, Inc. INDEX WORDS: Acute renal failure (ARF); sepsis; systemic inflammatory response syndrome (SIRS); intensive care unit (ICU); clinical and laboratory evaluation.
S
EPSIS IS DEFINED as a clinical syndrome, is related to severe infection, and is characterized by systemic inflammation and injury to many organs and functional systems,1-3 including vasodilatation, increased microvascular permeability, and leukocyte accumulation.2,3 The term systemic inflammatory response syndrome (SIRS) is used to refer to a disregulated host inflammatory response when no evidence of infection is present.1,2 Sepsis and SIRS are among the main causes of acute renal failure (ARF). The mechanism by which sepsis and endotoxinemia might lead to ARF is not yet well understood. Systemic hypotension, direct renal vasoconstriction, and release of cytokines may contribute to renal injury, as well as diastolic dysfunction of the heart, attributable, at least in part, to interstitial myocarditis secondary to infection.4 Loss of sympathetic nervous system autoregulation is an additional factor causing renal ischemia.4,5 The reported prevalence of ARF in sepsis ranges from 9% to 40%.6 In 1998, Liano et al7 presented data from the Madrid Acute Renal
Failure Study Group indicating that sepsis caused acute tubular necrosis in 35% of patients treated in the intensive care unit (ICU) and 27% of non-ICU patients. In a retrospective study in the surgical ICU of our unit, the incidence of de novo ARF in patients with sepsis was 16.2%.8 These investigators also observed that the mortal-
From the Department of Nephrology, Internal Medicine, Kocaeli University Medical School, Izmit; Department of Nephrology, Internal Medicine, Marmara University Medical School, Istanbul, Turkey; Department of Intensive Care; and Department of Internal Medicine, Nephrology Division, University Hospital of Gent, Belgium. Received October 28, 2003; accepted in revised form December 29, 2003. Supported in part by a grant from the Nephrocore Program of Fresenius Medical Care, Bad Homburg/Germany (I.Y.). Address reprint requests to Norbert Lameire, MD, PhD, Department of Nephrology, University Hospital Ghent, De Pintelaan 185, 9000 Ghent, Belgium. E-mail: norbert. [email protected] © 2004 by the National Kidney Foundation, Inc. 0272-6386/04/4305-0007$30.00/0 doi:10.1053/j.ajkd.2003.12.045
American Journal of Kidney Diseases, Vol 43, No 5 (May), 2004: pp 817-824
817
818
YEGENAGA ET AL
ity rate in septic patients without ARF was 28.4% versus 56.7% in those with ARF. Early recognition of risk factors for ARF in septic patients might be of importance in preventing the development or attenuating the course of ARF and thus could result in a reduction in morbidity and mortality related to sepsis. The aim of this study is to identify clinical characteristics of septic patients developing ARF in a cohort of patients followed up prospectively. METHODS
Patients All patients who developed sepsis and SIRS in the medical and surgical ICUs of the University Hospital of Gent, Belgium, from January 2001 to December 2001 were included if they had a serum creatinine level less than 2 mg/dL (⬍177 mol/L) on the day they developed SIRS/sepsis. Sepsis and SIRS are defined according to the consensus conference of the American College of Chest Physicians/ Society of Critical Care Medicine.1 Patients who met criteria for SIRS with at least 1 positive bacteriological culture, with the exception of smear cultures, were defined as having sepsis. If patients were treated by vasopressor agents, they were considered to have septic shock, and if more then 2 organ systems were functionally impaired, their disease state was defined as multiorgan system dysfunction syndrome. All included patients were followed up clinically and biochemically in a prospective way for 2 weeks or until discharge or death. By definition, patients with known previous renal failure were not included in this study. Patients who were supposed to survive for less than 24 hours after their admission to the ICU, pediatric patients (age ⬍ 17 years), and patients no longer considered for reanimation measures were not included. Patients were considered to belong to the ARF group if they had a sudden increase in serum creatinine concentration to greater than 2 mg/dL (⬎177 mol/L) or when they developed oliguria, defined as urine output less then 400 mL/24 h.9 All other patients were allotted to the no-ARF group. Demographic characteristics of each patient, such as age and sex, were noted, together with his or her medical history, presence of chronic diseases, and recent intake of medications, administration of radiocontrast material, and total duration of ICU stay. Primary hepatic failure is defined as the presence of liver cirrhosis in the medical history. Chronic obstructive pulmonary disease is defined as the need to administer bronchodilators in the present or past. Cardiovascular disease is defined as the presence of cardiomyopathy, ischemic heart disease, or peripheral vascular disease. Diabetes mellitus is defined as the need for blood glucose–controlling drugs for at least 1 year before admission and/or the presence of diabetic retinopathy. When patients had an immunosuppressive disease (any cancer, including hematologic malignancy,
or human immunodeficiency virus infection) or were treated by corticosteroids or other immunosuppressive drugs, they were considered to be in an immunosuppressive status. In addition, the following parameters were noted every day from the first day of sepsis. 1. Parameters scoring severity of disease. The Acute Physiology and Chronic Health Evaluation II score was calculated based on parameters collected during the first 24 hours after admission.10 2. Clinical condition. The highest and lowest levels of body temperature, heart rate, systolic and diastolic blood pressure, central venous pressure (CVP), ventilation support, respiratory rate, fraction of inspired oxygen, and Glasgow Coma Score were noted for 2 weeks. 3. Biochemical parameters. Highest and lowest values for arterial blood pH; PCO2; bicarbonate level; base excess; PO2; white blood cell count; hematocrit; thrombocyte count; serum sodium and potassium levels; the lowest value for red blood cell count; levels of hemoglobin, chloride, calcium, phosphorus, and albumin; prothrombin time; and the highest value for glucose, blood urea nitrogen (BUN), creatinine, and bilirubin; activated partial thromboplastin time; and levels of alanine aminotransferase, aspartate aminotransferase, ␥-glutamyl transferase, lipase, and lactate were recorded. Twenty-four–hour urine output was recorded. Laboratory tests were performed according to standard methods. 4. Therapeutic parameters. Daily doses of diuretics (bumetanide/furosemide), vasopressors (noradrenalin, adrenalin, dopamine, dobutamine, and vasopressin), and fluid replacement (crystalloids, colloids, and mannitol) were registered. Quantities of administered human albumin, packed erythrocytes, and fresh frozen plasma also were noted.
Statistical Analysis For statistical calculations, SPSS software, version 10.00 (SPSS Inc, Chicago, IL) was used. For statistical univariate analysis, Student’s t-test was calculated for comparison of unpaired samples, and chi-square test was used for comparison of dichotomous variables. Values are expressed as mean ⫾ SD, median and range, or percentage, when appropriate. A logistic regression model for the development of ARF was constructed using parameters showing a difference between the ARF and no-ARF groups in univariate analysis. The 2-tailed significance level is set at P less than 0.05.
RESULTS
During the study period, 2,442 patients were admitted to the ICU. Of those, 257 patients (10.5%) met inclusion criteria and hence were defined to have SIRS or sepsis. Infection was documented by at least 1 positive culture in 217 of 257 patients (84%), and these patients were defined as septic. The remain-
ARF AND SEPSIS
819
Fig 1. Mortality rates in ICU patients (A) without sepsis or SIRS and without ARF; (B) with sepsis or SIRS, but without ARF; and (C) with sepsis or SIRS and who developed ARF. Closed bars, percentage of survivors; hatched bars, percentage of nonsurvivors. *P < 0.001.
ing patients were classified as having SIRS. Of 217 patients with documented sepsis, 105 patients (48%) had sepsis alone, 39 patients (18%) had septic shock, and 73 patients (34%) had multiorgan system dysfunction syndrome. Of 217 patients with documented infection, 41 of 217 patients (19%) had a positive hemoculture, 24 (58%) of which showed the presence of gram-negative germs. No significant differences between the ARF and no-ARF groups were found regarding type of germ causing sepsis. Twenty-nine of the patients with sepsis/SIRS (13%) developed ARF between days 2 and 10 of follow-up (median, 5 days), of whom 13 patients (45%) needed renal replacement therapy (RRT). The mortality rates were 72% (21 of 29 patients) in the septic ARF group and 24% (54 of 228 patients) in the septic group without ARF (P ⬍ 0.001). In comparison, the mortality rate was 8% in nonseptic ICU patients admitted during the same period (179 of 2,185 patients; P ⬍ 0.001; Fig 1). Mortality among patients with ARF with the need for RRT was not significantly different from those without RRT (9 of 13 versus 12 of 16 patients). The median time the ARF group stayed in the ICU was 23 days (range, 9 to 73 days), whereas the no-ARF group stayed 15 days (range, 2 to 132; P ⫽ not significant).
The main clinical characteristics of the ARF group compared with the no-ARF group are listed in Table 1. Median patient age was 58 years (range, 17 to 95 years). Age was significantly older in the ARF group compared with the no-ARF group (P ⫽ 0.003; Fig 2). Primary hepatic failure was present in 8 of 29 patients with ARF (27%) and 15 of 228 non-ARF patients (6%; P ⫽ 0.015). A vascular surgical intervention as a cause of admission to the ICU was more prevalent in the ARF group (4 of 29 [14%] versus 7 of 228 patients [3%]; P ⫽ 0.025). In the ARF group, more patients were administered diuretics (P ⫽ 0.001) or vasopressors (P ⫽ 0.001). On day 1, CVP (lowest of the day) was higher and urinary output, mean arterial blood pressure, and diastolic blood pressure were lower in the ARF group. Systolic blood pressure was not different. The most relevant laboratory parameters are listed in Table 2. At the time of development of SIRS/sepsis, serum creatinine and BUN values were greater in the ARF group. In addition, values for bicarbonate, albumin, prothrombin time, and thrombocyte count were lower. On day 1, the total dose of colloids administered was significantly greater in the ARF group (P ⫽ 0.021). Conversely, amounts of infused
820
YEGENAGA ET AL Table 1.
Clinical Characteristics of Patients With and Without ARF on the First Day of Sepsis/SIRS
Age (y) Sex (men) APACHE II score Admission for surgical indications Previous administration of nephrotoxic agents Primary hepatic failure Vascular surgical intervention Chronic obstructive pulmonary disease Use of diuretics Use of mechanical ventilation Cardiovascular disease Diabetes mellitus Immunosuppressive state Highest systolic blood pressure (mm Hg) Highest mean arterial blood (mm Hg) Highest diastolic blood pressure (mm Hg) Lowest CVP (cmH2O) Urinary output (mL/24 h) Use of vasopressors (septic shock)
ARF (n ⫽ 29)
No-ARF (n ⫽ 228)
P
65.2 ⴞ 13.3 21 (72) 19.2 ⫾ 7.2 17 (59) 10 (34) 8 (27) 4 (14) 5 (17) 21 (72) 25 (86) 3 (10) 4 (14) 8 (28) 153.0 ⫾ 26.5 99.3 ⴞ 13.3 74.3 ⴞ 9.5 9.4 ⴞ 4.4 1,347.1 ⴞ 649.4 25 (86)
55.4 ⴞ 17.2 153 (68) 17.1 ⫾ 6.6 118 (52) 119 (53) 15 (6) 7 (3) 25 (10) 43 (19) 184 (81) 38 (15) 26 (10) 32 (14) 158.0 ⫾ 27.3 104.9 ⴞ 16.6 80.6 ⴞ 15.5 5.2 ⴞ 3.6 1,849.8 ⴞ 916.0 102 (45)
0.003 NS NS NS NS 0.001 0.025 NS <0.001 NS NS NS NS NS 0.05 0.004 0.006 0.005 <0.001
NOTE. Data in bold indicate significant differences between ARF and no-ARF. Values expressed as mean ⫾ SD or number (percent). Abbreviations: NS, not significant; APACHE II, Acute Physiology and Chronic Health Evaluation II.
crystalloids, packed erythrocyte cells, fresh frozen plasma, human albumin, or mannitol were not different between groups. In multivariate analysis, age, serum creatinine level, CVP, and presence of liver failure contrib-
uted significantly to a logistic regression model for ARF when only parameters differing significantly between the ARF and no-ARF groups in univariate analysis were entered simultaneously (Table 3).
Fig 2. Incidence (percent) of ARF in different age groups.
ARF AND SEPSIS Table 2.
821 Laboratory Evaluation of Patients With and Without ARF on the First Day of Sepsis/SIRS
Creatinine (mg/dL) BUN (mg/dL) Glomerular filtration rate* (mL/min) pH PCO2 PO2 Base excess Bicarbonate Highest value of potassium (mEq/L) Highest value of sodium (mEq/L) Chloride (mEq/L) Calcium (mg/dL) Phosphorus (mg/dL) Red blood cell count (⫻106/L) Hematocrit (%) Hemoglobin (g/dL) White blood cell count (⫻103/L) Platelet count (⫻103/L) Albumin (g/dL) Bilirubin (mg/dL) Prothrombin time (%) Activated partial thromboplastin time (s)
ARF (n ⫽ 29)
no-ARF (n ⫽ 228)
P
1.3 ⴞ 0.4 30 ⴞ 20 46.6 ⴞ 19.1 7.41 ⴞ 0.0 45.5 ⫾ 10.1 85.0 ⫾ 23.7 2.3 ⫾ 4.0 20.4 ⴞ 6.7 4.0 ⫾ 0.5 139.5 ⫹ 4.5 107.8 ⫾ 5.8 7.9 ⫾ 0.9 3.6 ⫾ 1.4 3.4 ⫾ 0.6 32.0 ⫾ 5.0 9.6 ⫾ 2.2 11.4 ⫾ 7.5 151.2 ⴞ 92.6 2.2 ⴞ 0.6 1.9 ⫾ 2.7 57.5 ⴞ 19.1 51.8 ⫾ 53.6
1.0 ⴞ 0.3 20 ⴞ 10 69.8 ⴞ 29.5 7.44 ⴞ 0.0 44.6 ⫾ 12.5 85.0 ⫾ 29.4 2.5 ⫾ 4.4 23.0 ⴞ 4.3 4.0 ⫾ 0.5 138.7 ⫹ 5.0 105.6 ⫾ 6.2 8.2 ⫾ 0.8 3.2 ⫾ 1.2 3.3 ⫾ 0.8 32.3 ⫾ 6.3 11.0 ⫾ 7.4 11.1 ⫾ 7.9 201.1 ⴞ 123.3 2.7 ⴞ 1.4 1.3 ⫾ 2.7 68.8 ⴞ 20.1 4164 ⫾ 29.1
<0.001 <0.001 0.0001 0.03 NS NS NS 0.009 NS NS NS NS NS NS NS NS NS 0.037 0.050 NS 0.006 NS
NOTE. Data in bold indicate significant differences between the ARF and no-ARF groups. To convert creatinine in mg/dL to mol/L, multiply by 88.4; BUN in mg/dL to mmol/L, multiply by 0.357; potassium, sodium, and chloride in mEq/L to mmol/L, multiply by 1; calcium in mg/dL to mmol/L, multiply by 0.2495; phosphorus in mg/dL to mmol/L, multiply by 0.3229; red blood cells in ⫻106/L to ⫻1012/L, multiply by 1; hemoglobin from g/dL to g/L, multiply by 10; white blood cells and platelets in ⫻103/L to ⫻109/L, multiply by 1; albumin from g/dL to g/L, multiply by 10; bilirubin in mg/dL to mol/L, multiply by 17.1. Abbreviation: NS, not significant. *Calculated by means of the Modification of Diet in Renal Disease formula, given in22, Table 3, equation 7.
DISCUSSION
This study prospectively collected clinical and laboratory data for patients with sepsis and SIRS in the ICU, enabling us to describe differences in characteristics of patients who developed ARF (the ARF group) compared with those who did not (no-ARF group). To the best of our knowledge, to date, no such study has been performed prospectively in the ICU population with sepsis. Patients prone to develop ARF in combination with sepsis were older; had greater CVP values; had greater serum creatinine levels, although still in the normal range; and were more likely to have preexisting liver dysfunction. According to these results, factors predisposing patients with normal renal function and sepsis to develop ARF are present at the first day of sepsis and are difficult to modify. ARF is one of the most serious complications of sepsis and SIRS. Because its occurrence promi-
nently affects outcome, it is important to distinguish patients who eventually will develop ARF, making it possible to correct the factors predisposing to ARF. To our knowledge, there are only 2 studies available in which an attempt was made to define factors having a role in the development of ARF in septic ICU patients.8,11 Most other studies describing the occurrence of ARF were performed in the general population or specific patient groups, eg, after cardiac surgery or coronarography,12,13 whereas our patient population is restricted to patients with sepsis/SIRS who had normal renal function on inclusion. Vogelaers et al11 retrospectively found that septic patients who developed ARF had more associated organ failure, more need for vasoactive therapy, and a lower CVP despite more aggressive fluid therapy compared with patients with sepsis without ARF. Hoste at al8 recently reported in a retrospective study that on the first day of sepsis, the risk for
822
YEGENAGA ET AL Table 3.
Step forward model, with P to enter ⬍ 0.1 Bilirubin ⬎ 1.5 mg/dL (yes/no) Age (y) Creatinine (mg/dL) CVP (mm Hg) Constant
Multivariate Logistic Regression
Coefficient 
SE Coefficient 
Exp 
95% Confidence Interval
P
2.3 0.074 0.023 0.41 ⫺12.9
0.9 0.025 0.008 0.09 2.6
9.7 1.1 1.02 1.5 0.00
1.65-60.3 1.03-1.13 1.007-1.04 1.26-1.80
0.018 0.003 0.005 0.0001 0.0001
P to Enter
(103/L)
Thrombocytes Diuresis (mL) Mean arterial pressure pH Bicarbonate (mEq/L) Serum albumin (g/dL)
0.20 0.21 0.59 0.92 0.90 0.55
NOTE. To convert bilirubin in mg/dL to mol/L, multiply by 17.1; creatinine in mg/dL to mol/L, multiply by 88.4; bicarbonate in mEq/L to mmol/L, multiply by 1; albumin in g/dL to g/L, multiply by 10.
the development of ARF was increased 7.5 times when serum creatinine level was greater than 1.0 mg/dL (⬎88 mol/L) and 6 times when pH was less than 7.30. CVP also was significantly greater in the group with ARF and sepsis.8 These data corroborate our present prospective data. In the present study, ARF occurred in only 13% of the entire sepsis/SIRS population, but the incidence increased up to 20% (19 of 107 patients) in patients with septic shock. In another prospective study including a larger number of patients with sepsis and septic shock (n ⫽ 2,527), the incidence of ARF was 19% in patients with sepsis, 23% in those with severe sepsis, and 51% in those with septic shock.14 However, in the present study, patients with a known history of chronic renal failure or existing ARF at the time of admission to the ICU, as well as patients who developed sepsis in other hospitals and were transferred to our hospital, were excluded to avoid selection bias. The present population consisted of patients who developed ARF only during follow-up in the ICU. Our values, which are limited strictly to the primary development of sepsis in a given setting, probably are more realistic, making this population well suited to identify predictive parameters for the development of ARF. There are no significant differences in incidence of infection, type of infection, or type of positive culture between the ARF and no-ARF
groups. Apparently these factors seem to not affect the risk for ARF in this study, although previous studies have shown that the incidence of positive cultures increases mortality risk.14 Thirteen of 29 patients with ARF in this series (44%) experienced not only sepsis, but also other conditions enhancing the risk for renal failure. The most prominent of these conditions was the presence of cirrhosis in the medical history. It is well known that cirrhotic patients frequently develop renal failure for several reasons, such as sepsis, hypovolemia, intake of nephrotoxic drugs, glomerulonephritis, or hepatorenal syndrome.15-17 It previously was shown that renal functional impairment causes a prominent increase in the mortality rate of patients with hepatic failure.18 Interestingly, such renal function parameters as serum creatinine or BUN levels were increased and urinary volume was decreased in the ARF population at the time of inclusion in this study (Table 2). These data indicate that structural damage of the kidneys or at least hemodynamic changes leading to this damage were present before overt sepsis developed. These data stress the importance of intensive follow-up and timely transfer to the ICU in infected patients before they develop sepsis, especially if they have other risk factors, such as older age, serum creatinine level in the high-normal or elevated range, decreasing urinary output, and/or the presence of concomitant liver disease.
ARF AND SEPSIS
Of note, there were no differences in the incidence of diabetes mellitus between the 2 groups. It might be that the low number of patients with diabetes (11%) caused a type 2 error. However, at least at our institution, the label “diabetic” causes increased awareness and scrutiny, so that preventive treatment and transfer might have occurred in a more timely manner in this subgroup of patients. Conversely, if problems occur, these fragile patients develop deterioration of their kidney function rapidly, resulting in exclusion from our study because of a serum creatinine level greater than 2 mg/dL (177 mol/L) at the time they came to the ICU. The finding that diabetes is not a risk factor for ARF in sepsis thus most likely cannot be generalized to the general population in real-life conditions. In the ARF group, diastolic blood pressure tended to be lower from the start (Table 1), and this difference became more prominent during follow-up (data not shown) despite greater use of vasopressors, whereas a larger quantity of colloids was administered, as well. These findings also indicate that additional aggressive fluid loading probably is not likely to reverse renal impairment after ARF has been initiated in this type of septic patient. Moreover, recent studies suggest that attempts to increase cardiac output and oxygen delivery with the administration of large volumes of fluids and inotropic agents can actually increase mortality.19 It can be argued that CVP does not reflect volume status correctly in critically ill patients with sepsis. For example, in patients with severe acute respiratory distress syndrome and diastolic dysfunction, CVP can be elevated because of high pressure in the right atrium and ventricle, although they are not fluid overloaded.19 In septic patients, high CVP thus can be a marker of severity of disease, indicating pulmonary and/or cardiac involvement, rather than volume status. In a recent study, Mehta et al20 showed that diuretic use in ICU patients with ARF was associated with greater in-hospital mortality and nonrecovery of renal function. In the present study, use of diuretics was greater in the ARF group. The underlying mechanism of this observation is unclear, but might be related to toxic effects of diuretics or the delay in consultation of a nephrologist while awaiting the effect of the diuretics.
823
It could be argued that the number of patients with ARF is relatively small. This can be attributed in part to the strict inclusion criteria, in which secondary transfers, patients who developed SIRS/sepsis before admission to the ICU, and patients with existing renal failure were excluded. It was believed that all these factors would lead to selection bias and unclear clinical conditions, potentially giving rise to incorrect conclusions. Despite these strict inclusion criteria, highly significant differences were found for a large number of factors. It can be argued that at a serum creatinine level greater than 2 mg/dL (⬎177 mol/L), a substantial part of kidney function already is lost. In this regard, we recalculated our data with exclusion of patients with a serum creatinine level greater than 1.5 mg/dL (⬎133 mol/L). However, this resulted in the exclusion of only 3 additional patients and did not alter the other results. Therefore, the original cutoff value of 2 mg/dL was maintained. Our data support a timely referral to the ICU and underscore that therapeutic measures should be applied more aggressively and earlier than is done currently in the very early stages before the vicious cascade of septic events is started.21 In conclusion, ARF is still an important mortality risk in septic patients. In view of the factors related to ARF in this patient group, such as age, high CVP, liver disease, and elevated serum creatinine level at the start of sepsis, it appears important to prevent ARF before sepsis/SIRS develops by careful monitoring and aggressive treatment of patients at risk. ACKNOWLEDGMENT The authors thank J. Vienken for his kind support and the nurses and secretaries of the medical and surgical ICUs of the University Hospital of Gent for secretarial assistance.
REFERENCES 1. American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference: Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med 20:864-874, 1992 2. Bone RG: The pathogenesis of sepsis. Ann Intern Med 115:457-469, 1991 3. Bone RG: Immunologic dissonance: A continuing evolution in our understanding of the systemic inflammatory response syndrome (SIRS) and the multiple organ dysfunction syndrome (MODS). Ann Intern Med 125:680-687, 1996
824
4. Schor N: Acute renal failure and sepsis syndrome. Kidney Int 61:764-776, 2002 5. Thijs A, Thijs LG: Pathogenesis of renal failure in sepsis. Kidney Int Suppl 66:S34-S37, 1998 6. Brivet FG, Kleinknecht DJ, Loirat P, Landais PJ: Acute renal failure in intensive care units—Causes, outcome, and prognostic factors of hospital mortality; A prospective, multicenter study. French Study Group on Acute Renal Failure. Crit Care Med 24:192-198, 1996 7. Liano F, Junco E, Pascual J, Madero R, Verde E: The spectrum of acute renal failure in the intensive care unit compared with that seen in other settings. The Madrid Acute Renal Failure Study Group. Kidney Int Suppl 66:S16-S24, 1998 8. Hoste EAJ, Lameire NH, Vanholder RC, Benoit DD, Decruyenaere JMA, Colardyn FA: Acute renal failure in patients with sepsis in surgical ICU; predictive factors, incidence, comorbidity, and outcome. J Am Soc Nephrol 14:1022-1030, 2003 9. Liano F, Garcia-Martin F, Gallego A, et al: Easy and early prognosis in acute tubular necrosis: A forward analysis of 228 cases. Nephron 51:307-313, 1989 10. Knaus WA, Drapper EA, Wagner DP, Zimmerman JE: APACHE II: A severity of disease classification system. Crit Care Med 13:818-829, 1985 11. Vogelaers D, Vanholder R, Colardyn F, Lameire NH: Acute renal failure in an ICU population with sepsis: Retrospective analysis on risk factors for outcome. Presented at the Third International Symposium on Acute Renal Failure, Halkidiki, Greece, June 20-23, 1993, p 131 (abstr) 12. Coritsidis GN, Guru K, Ward L, Bashir R, Feinfeld DA, Carvounis CP: Prediction of acute renal failure by “bedside formula” in medical and surgical intensive care patients. Ren Fail 22:235-244, 2000
YEGENAGA ET AL
13. Chertow G, Lazarus J, Cristiansen C, et al: Prospective renal risk stratification. Circulation 95:878-884, 1997 14. Rangel-Frausto MS, Pittet D, Costigan M, Hwang T, Davis CS, Wenzel RP: The natural history of the systemic inflammatory response syndrome (SIRS). JAMA 273:117123, 1995 15. Paganini EP, Halstenberg WK, Goormastic M: Risk modelling in acute renal failure requiring dialysis: The introduction of a new model. Clin Nephrol 46:206-211, 1996 16. Sanyal AJ: Hepatorenal syndrome. J Gastroenterol Hepatol 17:S248-S252, 2002 (suppl 13) 17. Dagher L, Moore K: The hepatorenal syndrome. Gut 49:729-737, 2002 18. Sort P, Navasa M, Arroyo V, et al: Effect of intravenous albumin on renal impairment and mortality in patients with cirrhosis and spontaneous bacterial peritonitis. N Engl J Med 341:403-409, 1999 19. Esson ML, Schrier RW: Diagnosis and treatment of acute tubular necrosis. Ann Intern Med 137:744-752, 2002 20. Mehta RL, Pascual MT, Soroko S, Chertow GM, for the PICARD Study Group: Diuretics, mortality, and nonrecovery of renal function in acute renal failure. JAMA 288:2547-2553, 2002 21. Lameire N, Vanholder R, Van Biesen W: Loop diuretics for patients with acute renal failure, helpful or harmful? JAMA 288:2599-2601, 2002 22. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D: 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 130:461-470, 1999
S
EPSIS IS DEFINED as a clinical syndrome, is related to severe infection, and is characterized by systemic inflammation and injury to many organs and functional systems,1-3 including vasodilatation, increased microvascular permeability, and leukocyte accumulation.2,3 The term systemic inflammatory response syndrome (SIRS) is used to refer to a disregulated host inflammatory response when no evidence of infection is present.1,2 Sepsis and SIRS are among the main causes of acute renal failure (ARF). The mechanism by which sepsis and endotoxinemia might lead to ARF is not yet well understood. Systemic hypotension, direct renal vasoconstriction, and release of cytokines may contribute to renal injury, as well as diastolic dysfunction of the heart, attributable, at least in part, to interstitial myocarditis secondary to infection.4 Loss of sympathetic nervous system autoregulation is an additional factor causing renal ischemia.4,5 The reported prevalence of ARF in sepsis ranges from 9% to 40%.6 In 1998, Liano et al7 presented data from the Madrid Acute Renal
Failure Study Group indicating that sepsis caused acute tubular necrosis in 35% of patients treated in the intensive care unit (ICU) and 27% of non-ICU patients. In a retrospective study in the surgical ICU of our unit, the incidence of de novo ARF in patients with sepsis was 16.2%.8 These investigators also observed that the mortal-
From the Department of Nephrology, Internal Medicine, Kocaeli University Medical School, Izmit; Department of Nephrology, Internal Medicine, Marmara University Medical School, Istanbul, Turkey; Department of Intensive Care; and Department of Internal Medicine, Nephrology Division, University Hospital of Gent, Belgium. Received October 28, 2003; accepted in revised form December 29, 2003. Supported in part by a grant from the Nephrocore Program of Fresenius Medical Care, Bad Homburg/Germany (I.Y.). Address reprint requests to Norbert Lameire, MD, PhD, Department of Nephrology, University Hospital Ghent, De Pintelaan 185, 9000 Ghent, Belgium. E-mail: norbert. [email protected] © 2004 by the National Kidney Foundation, Inc. 0272-6386/04/4305-0007$30.00/0 doi:10.1053/j.ajkd.2003.12.045
American Journal of Kidney Diseases, Vol 43, No 5 (May), 2004: pp 817-824
817
818
YEGENAGA ET AL
ity rate in septic patients without ARF was 28.4% versus 56.7% in those with ARF. Early recognition of risk factors for ARF in septic patients might be of importance in preventing the development or attenuating the course of ARF and thus could result in a reduction in morbidity and mortality related to sepsis. The aim of this study is to identify clinical characteristics of septic patients developing ARF in a cohort of patients followed up prospectively. METHODS
Patients All patients who developed sepsis and SIRS in the medical and surgical ICUs of the University Hospital of Gent, Belgium, from January 2001 to December 2001 were included if they had a serum creatinine level less than 2 mg/dL (⬍177 mol/L) on the day they developed SIRS/sepsis. Sepsis and SIRS are defined according to the consensus conference of the American College of Chest Physicians/ Society of Critical Care Medicine.1 Patients who met criteria for SIRS with at least 1 positive bacteriological culture, with the exception of smear cultures, were defined as having sepsis. If patients were treated by vasopressor agents, they were considered to have septic shock, and if more then 2 organ systems were functionally impaired, their disease state was defined as multiorgan system dysfunction syndrome. All included patients were followed up clinically and biochemically in a prospective way for 2 weeks or until discharge or death. By definition, patients with known previous renal failure were not included in this study. Patients who were supposed to survive for less than 24 hours after their admission to the ICU, pediatric patients (age ⬍ 17 years), and patients no longer considered for reanimation measures were not included. Patients were considered to belong to the ARF group if they had a sudden increase in serum creatinine concentration to greater than 2 mg/dL (⬎177 mol/L) or when they developed oliguria, defined as urine output less then 400 mL/24 h.9 All other patients were allotted to the no-ARF group. Demographic characteristics of each patient, such as age and sex, were noted, together with his or her medical history, presence of chronic diseases, and recent intake of medications, administration of radiocontrast material, and total duration of ICU stay. Primary hepatic failure is defined as the presence of liver cirrhosis in the medical history. Chronic obstructive pulmonary disease is defined as the need to administer bronchodilators in the present or past. Cardiovascular disease is defined as the presence of cardiomyopathy, ischemic heart disease, or peripheral vascular disease. Diabetes mellitus is defined as the need for blood glucose–controlling drugs for at least 1 year before admission and/or the presence of diabetic retinopathy. When patients had an immunosuppressive disease (any cancer, including hematologic malignancy,
or human immunodeficiency virus infection) or were treated by corticosteroids or other immunosuppressive drugs, they were considered to be in an immunosuppressive status. In addition, the following parameters were noted every day from the first day of sepsis. 1. Parameters scoring severity of disease. The Acute Physiology and Chronic Health Evaluation II score was calculated based on parameters collected during the first 24 hours after admission.10 2. Clinical condition. The highest and lowest levels of body temperature, heart rate, systolic and diastolic blood pressure, central venous pressure (CVP), ventilation support, respiratory rate, fraction of inspired oxygen, and Glasgow Coma Score were noted for 2 weeks. 3. Biochemical parameters. Highest and lowest values for arterial blood pH; PCO2; bicarbonate level; base excess; PO2; white blood cell count; hematocrit; thrombocyte count; serum sodium and potassium levels; the lowest value for red blood cell count; levels of hemoglobin, chloride, calcium, phosphorus, and albumin; prothrombin time; and the highest value for glucose, blood urea nitrogen (BUN), creatinine, and bilirubin; activated partial thromboplastin time; and levels of alanine aminotransferase, aspartate aminotransferase, ␥-glutamyl transferase, lipase, and lactate were recorded. Twenty-four–hour urine output was recorded. Laboratory tests were performed according to standard methods. 4. Therapeutic parameters. Daily doses of diuretics (bumetanide/furosemide), vasopressors (noradrenalin, adrenalin, dopamine, dobutamine, and vasopressin), and fluid replacement (crystalloids, colloids, and mannitol) were registered. Quantities of administered human albumin, packed erythrocytes, and fresh frozen plasma also were noted.
Statistical Analysis For statistical calculations, SPSS software, version 10.00 (SPSS Inc, Chicago, IL) was used. For statistical univariate analysis, Student’s t-test was calculated for comparison of unpaired samples, and chi-square test was used for comparison of dichotomous variables. Values are expressed as mean ⫾ SD, median and range, or percentage, when appropriate. A logistic regression model for the development of ARF was constructed using parameters showing a difference between the ARF and no-ARF groups in univariate analysis. The 2-tailed significance level is set at P less than 0.05.
RESULTS
During the study period, 2,442 patients were admitted to the ICU. Of those, 257 patients (10.5%) met inclusion criteria and hence were defined to have SIRS or sepsis. Infection was documented by at least 1 positive culture in 217 of 257 patients (84%), and these patients were defined as septic. The remain-
ARF AND SEPSIS
819
Fig 1. Mortality rates in ICU patients (A) without sepsis or SIRS and without ARF; (B) with sepsis or SIRS, but without ARF; and (C) with sepsis or SIRS and who developed ARF. Closed bars, percentage of survivors; hatched bars, percentage of nonsurvivors. *P < 0.001.
ing patients were classified as having SIRS. Of 217 patients with documented sepsis, 105 patients (48%) had sepsis alone, 39 patients (18%) had septic shock, and 73 patients (34%) had multiorgan system dysfunction syndrome. Of 217 patients with documented infection, 41 of 217 patients (19%) had a positive hemoculture, 24 (58%) of which showed the presence of gram-negative germs. No significant differences between the ARF and no-ARF groups were found regarding type of germ causing sepsis. Twenty-nine of the patients with sepsis/SIRS (13%) developed ARF between days 2 and 10 of follow-up (median, 5 days), of whom 13 patients (45%) needed renal replacement therapy (RRT). The mortality rates were 72% (21 of 29 patients) in the septic ARF group and 24% (54 of 228 patients) in the septic group without ARF (P ⬍ 0.001). In comparison, the mortality rate was 8% in nonseptic ICU patients admitted during the same period (179 of 2,185 patients; P ⬍ 0.001; Fig 1). Mortality among patients with ARF with the need for RRT was not significantly different from those without RRT (9 of 13 versus 12 of 16 patients). The median time the ARF group stayed in the ICU was 23 days (range, 9 to 73 days), whereas the no-ARF group stayed 15 days (range, 2 to 132; P ⫽ not significant).
The main clinical characteristics of the ARF group compared with the no-ARF group are listed in Table 1. Median patient age was 58 years (range, 17 to 95 years). Age was significantly older in the ARF group compared with the no-ARF group (P ⫽ 0.003; Fig 2). Primary hepatic failure was present in 8 of 29 patients with ARF (27%) and 15 of 228 non-ARF patients (6%; P ⫽ 0.015). A vascular surgical intervention as a cause of admission to the ICU was more prevalent in the ARF group (4 of 29 [14%] versus 7 of 228 patients [3%]; P ⫽ 0.025). In the ARF group, more patients were administered diuretics (P ⫽ 0.001) or vasopressors (P ⫽ 0.001). On day 1, CVP (lowest of the day) was higher and urinary output, mean arterial blood pressure, and diastolic blood pressure were lower in the ARF group. Systolic blood pressure was not different. The most relevant laboratory parameters are listed in Table 2. At the time of development of SIRS/sepsis, serum creatinine and BUN values were greater in the ARF group. In addition, values for bicarbonate, albumin, prothrombin time, and thrombocyte count were lower. On day 1, the total dose of colloids administered was significantly greater in the ARF group (P ⫽ 0.021). Conversely, amounts of infused
820
YEGENAGA ET AL Table 1.
Clinical Characteristics of Patients With and Without ARF on the First Day of Sepsis/SIRS
Age (y) Sex (men) APACHE II score Admission for surgical indications Previous administration of nephrotoxic agents Primary hepatic failure Vascular surgical intervention Chronic obstructive pulmonary disease Use of diuretics Use of mechanical ventilation Cardiovascular disease Diabetes mellitus Immunosuppressive state Highest systolic blood pressure (mm Hg) Highest mean arterial blood (mm Hg) Highest diastolic blood pressure (mm Hg) Lowest CVP (cmH2O) Urinary output (mL/24 h) Use of vasopressors (septic shock)
ARF (n ⫽ 29)
No-ARF (n ⫽ 228)
P
65.2 ⴞ 13.3 21 (72) 19.2 ⫾ 7.2 17 (59) 10 (34) 8 (27) 4 (14) 5 (17) 21 (72) 25 (86) 3 (10) 4 (14) 8 (28) 153.0 ⫾ 26.5 99.3 ⴞ 13.3 74.3 ⴞ 9.5 9.4 ⴞ 4.4 1,347.1 ⴞ 649.4 25 (86)
55.4 ⴞ 17.2 153 (68) 17.1 ⫾ 6.6 118 (52) 119 (53) 15 (6) 7 (3) 25 (10) 43 (19) 184 (81) 38 (15) 26 (10) 32 (14) 158.0 ⫾ 27.3 104.9 ⴞ 16.6 80.6 ⴞ 15.5 5.2 ⴞ 3.6 1,849.8 ⴞ 916.0 102 (45)
0.003 NS NS NS NS 0.001 0.025 NS <0.001 NS NS NS NS NS 0.05 0.004 0.006 0.005 <0.001
NOTE. Data in bold indicate significant differences between ARF and no-ARF. Values expressed as mean ⫾ SD or number (percent). Abbreviations: NS, not significant; APACHE II, Acute Physiology and Chronic Health Evaluation II.
crystalloids, packed erythrocyte cells, fresh frozen plasma, human albumin, or mannitol were not different between groups. In multivariate analysis, age, serum creatinine level, CVP, and presence of liver failure contrib-
uted significantly to a logistic regression model for ARF when only parameters differing significantly between the ARF and no-ARF groups in univariate analysis were entered simultaneously (Table 3).
Fig 2. Incidence (percent) of ARF in different age groups.
ARF AND SEPSIS Table 2.
821 Laboratory Evaluation of Patients With and Without ARF on the First Day of Sepsis/SIRS
Creatinine (mg/dL) BUN (mg/dL) Glomerular filtration rate* (mL/min) pH PCO2 PO2 Base excess Bicarbonate Highest value of potassium (mEq/L) Highest value of sodium (mEq/L) Chloride (mEq/L) Calcium (mg/dL) Phosphorus (mg/dL) Red blood cell count (⫻106/L) Hematocrit (%) Hemoglobin (g/dL) White blood cell count (⫻103/L) Platelet count (⫻103/L) Albumin (g/dL) Bilirubin (mg/dL) Prothrombin time (%) Activated partial thromboplastin time (s)
ARF (n ⫽ 29)
no-ARF (n ⫽ 228)
P
1.3 ⴞ 0.4 30 ⴞ 20 46.6 ⴞ 19.1 7.41 ⴞ 0.0 45.5 ⫾ 10.1 85.0 ⫾ 23.7 2.3 ⫾ 4.0 20.4 ⴞ 6.7 4.0 ⫾ 0.5 139.5 ⫹ 4.5 107.8 ⫾ 5.8 7.9 ⫾ 0.9 3.6 ⫾ 1.4 3.4 ⫾ 0.6 32.0 ⫾ 5.0 9.6 ⫾ 2.2 11.4 ⫾ 7.5 151.2 ⴞ 92.6 2.2 ⴞ 0.6 1.9 ⫾ 2.7 57.5 ⴞ 19.1 51.8 ⫾ 53.6
1.0 ⴞ 0.3 20 ⴞ 10 69.8 ⴞ 29.5 7.44 ⴞ 0.0 44.6 ⫾ 12.5 85.0 ⫾ 29.4 2.5 ⫾ 4.4 23.0 ⴞ 4.3 4.0 ⫾ 0.5 138.7 ⫹ 5.0 105.6 ⫾ 6.2 8.2 ⫾ 0.8 3.2 ⫾ 1.2 3.3 ⫾ 0.8 32.3 ⫾ 6.3 11.0 ⫾ 7.4 11.1 ⫾ 7.9 201.1 ⴞ 123.3 2.7 ⴞ 1.4 1.3 ⫾ 2.7 68.8 ⴞ 20.1 4164 ⫾ 29.1
<0.001 <0.001 0.0001 0.03 NS NS NS 0.009 NS NS NS NS NS NS NS NS NS 0.037 0.050 NS 0.006 NS
NOTE. Data in bold indicate significant differences between the ARF and no-ARF groups. To convert creatinine in mg/dL to mol/L, multiply by 88.4; BUN in mg/dL to mmol/L, multiply by 0.357; potassium, sodium, and chloride in mEq/L to mmol/L, multiply by 1; calcium in mg/dL to mmol/L, multiply by 0.2495; phosphorus in mg/dL to mmol/L, multiply by 0.3229; red blood cells in ⫻106/L to ⫻1012/L, multiply by 1; hemoglobin from g/dL to g/L, multiply by 10; white blood cells and platelets in ⫻103/L to ⫻109/L, multiply by 1; albumin from g/dL to g/L, multiply by 10; bilirubin in mg/dL to mol/L, multiply by 17.1. Abbreviation: NS, not significant. *Calculated by means of the Modification of Diet in Renal Disease formula, given in22, Table 3, equation 7.
DISCUSSION
This study prospectively collected clinical and laboratory data for patients with sepsis and SIRS in the ICU, enabling us to describe differences in characteristics of patients who developed ARF (the ARF group) compared with those who did not (no-ARF group). To the best of our knowledge, to date, no such study has been performed prospectively in the ICU population with sepsis. Patients prone to develop ARF in combination with sepsis were older; had greater CVP values; had greater serum creatinine levels, although still in the normal range; and were more likely to have preexisting liver dysfunction. According to these results, factors predisposing patients with normal renal function and sepsis to develop ARF are present at the first day of sepsis and are difficult to modify. ARF is one of the most serious complications of sepsis and SIRS. Because its occurrence promi-
nently affects outcome, it is important to distinguish patients who eventually will develop ARF, making it possible to correct the factors predisposing to ARF. To our knowledge, there are only 2 studies available in which an attempt was made to define factors having a role in the development of ARF in septic ICU patients.8,11 Most other studies describing the occurrence of ARF were performed in the general population or specific patient groups, eg, after cardiac surgery or coronarography,12,13 whereas our patient population is restricted to patients with sepsis/SIRS who had normal renal function on inclusion. Vogelaers et al11 retrospectively found that septic patients who developed ARF had more associated organ failure, more need for vasoactive therapy, and a lower CVP despite more aggressive fluid therapy compared with patients with sepsis without ARF. Hoste at al8 recently reported in a retrospective study that on the first day of sepsis, the risk for
822
YEGENAGA ET AL Table 3.
Step forward model, with P to enter ⬍ 0.1 Bilirubin ⬎ 1.5 mg/dL (yes/no) Age (y) Creatinine (mg/dL) CVP (mm Hg) Constant
Multivariate Logistic Regression
Coefficient 
SE Coefficient 
Exp 
95% Confidence Interval
P
2.3 0.074 0.023 0.41 ⫺12.9
0.9 0.025 0.008 0.09 2.6
9.7 1.1 1.02 1.5 0.00
1.65-60.3 1.03-1.13 1.007-1.04 1.26-1.80
0.018 0.003 0.005 0.0001 0.0001
P to Enter
(103/L)
Thrombocytes Diuresis (mL) Mean arterial pressure pH Bicarbonate (mEq/L) Serum albumin (g/dL)
0.20 0.21 0.59 0.92 0.90 0.55
NOTE. To convert bilirubin in mg/dL to mol/L, multiply by 17.1; creatinine in mg/dL to mol/L, multiply by 88.4; bicarbonate in mEq/L to mmol/L, multiply by 1; albumin in g/dL to g/L, multiply by 10.
the development of ARF was increased 7.5 times when serum creatinine level was greater than 1.0 mg/dL (⬎88 mol/L) and 6 times when pH was less than 7.30. CVP also was significantly greater in the group with ARF and sepsis.8 These data corroborate our present prospective data. In the present study, ARF occurred in only 13% of the entire sepsis/SIRS population, but the incidence increased up to 20% (19 of 107 patients) in patients with septic shock. In another prospective study including a larger number of patients with sepsis and septic shock (n ⫽ 2,527), the incidence of ARF was 19% in patients with sepsis, 23% in those with severe sepsis, and 51% in those with septic shock.14 However, in the present study, patients with a known history of chronic renal failure or existing ARF at the time of admission to the ICU, as well as patients who developed sepsis in other hospitals and were transferred to our hospital, were excluded to avoid selection bias. The present population consisted of patients who developed ARF only during follow-up in the ICU. Our values, which are limited strictly to the primary development of sepsis in a given setting, probably are more realistic, making this population well suited to identify predictive parameters for the development of ARF. There are no significant differences in incidence of infection, type of infection, or type of positive culture between the ARF and no-ARF
groups. Apparently these factors seem to not affect the risk for ARF in this study, although previous studies have shown that the incidence of positive cultures increases mortality risk.14 Thirteen of 29 patients with ARF in this series (44%) experienced not only sepsis, but also other conditions enhancing the risk for renal failure. The most prominent of these conditions was the presence of cirrhosis in the medical history. It is well known that cirrhotic patients frequently develop renal failure for several reasons, such as sepsis, hypovolemia, intake of nephrotoxic drugs, glomerulonephritis, or hepatorenal syndrome.15-17 It previously was shown that renal functional impairment causes a prominent increase in the mortality rate of patients with hepatic failure.18 Interestingly, such renal function parameters as serum creatinine or BUN levels were increased and urinary volume was decreased in the ARF population at the time of inclusion in this study (Table 2). These data indicate that structural damage of the kidneys or at least hemodynamic changes leading to this damage were present before overt sepsis developed. These data stress the importance of intensive follow-up and timely transfer to the ICU in infected patients before they develop sepsis, especially if they have other risk factors, such as older age, serum creatinine level in the high-normal or elevated range, decreasing urinary output, and/or the presence of concomitant liver disease.
ARF AND SEPSIS
Of note, there were no differences in the incidence of diabetes mellitus between the 2 groups. It might be that the low number of patients with diabetes (11%) caused a type 2 error. However, at least at our institution, the label “diabetic” causes increased awareness and scrutiny, so that preventive treatment and transfer might have occurred in a more timely manner in this subgroup of patients. Conversely, if problems occur, these fragile patients develop deterioration of their kidney function rapidly, resulting in exclusion from our study because of a serum creatinine level greater than 2 mg/dL (177 mol/L) at the time they came to the ICU. The finding that diabetes is not a risk factor for ARF in sepsis thus most likely cannot be generalized to the general population in real-life conditions. In the ARF group, diastolic blood pressure tended to be lower from the start (Table 1), and this difference became more prominent during follow-up (data not shown) despite greater use of vasopressors, whereas a larger quantity of colloids was administered, as well. These findings also indicate that additional aggressive fluid loading probably is not likely to reverse renal impairment after ARF has been initiated in this type of septic patient. Moreover, recent studies suggest that attempts to increase cardiac output and oxygen delivery with the administration of large volumes of fluids and inotropic agents can actually increase mortality.19 It can be argued that CVP does not reflect volume status correctly in critically ill patients with sepsis. For example, in patients with severe acute respiratory distress syndrome and diastolic dysfunction, CVP can be elevated because of high pressure in the right atrium and ventricle, although they are not fluid overloaded.19 In septic patients, high CVP thus can be a marker of severity of disease, indicating pulmonary and/or cardiac involvement, rather than volume status. In a recent study, Mehta et al20 showed that diuretic use in ICU patients with ARF was associated with greater in-hospital mortality and nonrecovery of renal function. In the present study, use of diuretics was greater in the ARF group. The underlying mechanism of this observation is unclear, but might be related to toxic effects of diuretics or the delay in consultation of a nephrologist while awaiting the effect of the diuretics.
823
It could be argued that the number of patients with ARF is relatively small. This can be attributed in part to the strict inclusion criteria, in which secondary transfers, patients who developed SIRS/sepsis before admission to the ICU, and patients with existing renal failure were excluded. It was believed that all these factors would lead to selection bias and unclear clinical conditions, potentially giving rise to incorrect conclusions. Despite these strict inclusion criteria, highly significant differences were found for a large number of factors. It can be argued that at a serum creatinine level greater than 2 mg/dL (⬎177 mol/L), a substantial part of kidney function already is lost. In this regard, we recalculated our data with exclusion of patients with a serum creatinine level greater than 1.5 mg/dL (⬎133 mol/L). However, this resulted in the exclusion of only 3 additional patients and did not alter the other results. Therefore, the original cutoff value of 2 mg/dL was maintained. Our data support a timely referral to the ICU and underscore that therapeutic measures should be applied more aggressively and earlier than is done currently in the very early stages before the vicious cascade of septic events is started.21 In conclusion, ARF is still an important mortality risk in septic patients. In view of the factors related to ARF in this patient group, such as age, high CVP, liver disease, and elevated serum creatinine level at the start of sepsis, it appears important to prevent ARF before sepsis/SIRS develops by careful monitoring and aggressive treatment of patients at risk. ACKNOWLEDGMENT The authors thank J. Vienken for his kind support and the nurses and secretaries of the medical and surgical ICUs of the University Hospital of Gent for secretarial assistance.
REFERENCES 1. American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference: Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med 20:864-874, 1992 2. Bone RG: The pathogenesis of sepsis. Ann Intern Med 115:457-469, 1991 3. Bone RG: Immunologic dissonance: A continuing evolution in our understanding of the systemic inflammatory response syndrome (SIRS) and the multiple organ dysfunction syndrome (MODS). Ann Intern Med 125:680-687, 1996
824
4. Schor N: Acute renal failure and sepsis syndrome. Kidney Int 61:764-776, 2002 5. Thijs A, Thijs LG: Pathogenesis of renal failure in sepsis. Kidney Int Suppl 66:S34-S37, 1998 6. Brivet FG, Kleinknecht DJ, Loirat P, Landais PJ: Acute renal failure in intensive care units—Causes, outcome, and prognostic factors of hospital mortality; A prospective, multicenter study. French Study Group on Acute Renal Failure. Crit Care Med 24:192-198, 1996 7. Liano F, Junco E, Pascual J, Madero R, Verde E: The spectrum of acute renal failure in the intensive care unit compared with that seen in other settings. The Madrid Acute Renal Failure Study Group. Kidney Int Suppl 66:S16-S24, 1998 8. Hoste EAJ, Lameire NH, Vanholder RC, Benoit DD, Decruyenaere JMA, Colardyn FA: Acute renal failure in patients with sepsis in surgical ICU; predictive factors, incidence, comorbidity, and outcome. J Am Soc Nephrol 14:1022-1030, 2003 9. Liano F, Garcia-Martin F, Gallego A, et al: Easy and early prognosis in acute tubular necrosis: A forward analysis of 228 cases. Nephron 51:307-313, 1989 10. Knaus WA, Drapper EA, Wagner DP, Zimmerman JE: APACHE II: A severity of disease classification system. Crit Care Med 13:818-829, 1985 11. Vogelaers D, Vanholder R, Colardyn F, Lameire NH: Acute renal failure in an ICU population with sepsis: Retrospective analysis on risk factors for outcome. Presented at the Third International Symposium on Acute Renal Failure, Halkidiki, Greece, June 20-23, 1993, p 131 (abstr) 12. Coritsidis GN, Guru K, Ward L, Bashir R, Feinfeld DA, Carvounis CP: Prediction of acute renal failure by “bedside formula” in medical and surgical intensive care patients. Ren Fail 22:235-244, 2000
YEGENAGA ET AL
13. Chertow G, Lazarus J, Cristiansen C, et al: Prospective renal risk stratification. Circulation 95:878-884, 1997 14. Rangel-Frausto MS, Pittet D, Costigan M, Hwang T, Davis CS, Wenzel RP: The natural history of the systemic inflammatory response syndrome (SIRS). JAMA 273:117123, 1995 15. Paganini EP, Halstenberg WK, Goormastic M: Risk modelling in acute renal failure requiring dialysis: The introduction of a new model. Clin Nephrol 46:206-211, 1996 16. Sanyal AJ: Hepatorenal syndrome. J Gastroenterol Hepatol 17:S248-S252, 2002 (suppl 13) 17. Dagher L, Moore K: The hepatorenal syndrome. Gut 49:729-737, 2002 18. Sort P, Navasa M, Arroyo V, et al: Effect of intravenous albumin on renal impairment and mortality in patients with cirrhosis and spontaneous bacterial peritonitis. N Engl J Med 341:403-409, 1999 19. Esson ML, Schrier RW: Diagnosis and treatment of acute tubular necrosis. Ann Intern Med 137:744-752, 2002 20. Mehta RL, Pascual MT, Soroko S, Chertow GM, for the PICARD Study Group: Diuretics, mortality, and nonrecovery of renal function in acute renal failure. JAMA 288:2547-2553, 2002 21. Lameire N, Vanholder R, Van Biesen W: Loop diuretics for patients with acute renal failure, helpful or harmful? JAMA 288:2599-2601, 2002 22. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D: 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 130:461-470, 1999