Acute kidney injury after trauma: Prevalence, clinical characteristics and RIFLE classification

Background: Acute kidney injury (AKI) is an uncommon but serious complication after trauma. The objective of this study was to evaluate the prevalence, clinical characteristics and outcome of AKI after trauma. Patients and Methods: This was a retrospective study performed from January 2006 to January 2008 in an emergency specialized hospital in Fortaleza city, northeast of Brazil. All patients with AKI admitted in the study period were included. Prevalence of AKI, clinical characteristics and outcome were investigated. Results: Of the 129 patients admitted to the intensive care unit (ICU), 52 had AKI. The mean age was 30.1 ± 19.2 years, and 79.8% were males. The main causes of AKI were sepsis in 27 cases (52%) and hypotension in 18 (34%). Oliguria was observed in 33 cases (63%). Dialysis was required for 19 patients (36.5%). Independent risk factors associated with AKI were abdominal trauma [odds ratio (OR) = 3.66, P = 0.027] and use of furosemide (OR = 4.10, P = 0.026). Patients were classified according to RIFLE criteria as Risk in 12 cases (23%), Injury in 13 (25%), Failure in 24 (46%), Loss in 1 (2%) and End-stage in 2 (4%). Overall in-hospital mortality was 95.3%. The main cause of death was sepsis (24%). Mortality was 100% among patients with AKI. Conclusions: AKI is a fatal complication after trauma, which presented with a high mortality in the studied population. A better comprehension of factors associated with death in trauma-associated AKI is important, and more effective measures of prevention and treatment of AKI in this population are urgently needed.


Introduction
Acute kidney injury (AKI) is a complex disorder common in critically ill patients, which has been reported to affect from 1 to 25% of intensive care unit (ICU) patients and has led to mortality rates ranging from 15 to 60%. [1][2][3][4] AKI is an uncommon but serious complication after trauma. In large trauma populations, the incidence of post-traumatic AKI varies from 0.098 to 8.4% in published series [5,6] with mortality ranging from 7 to 83%. [5,[7][8][9] The natural history of post-traumatic AKI is not well established. Recent publications cite decreased renal perfusion as the common cause of this complication, [10,11] while the early literature suggested that AKI was secondary to crush injuries and rhabdomyolysis. [12] Renal abnormality observed is death or damage of tubular cells due to the imbalance between oxygen supply and demand of energy by hypoperfusion. [13][14][15] Improvements in treatment, including the introduction of dialysis, have not changed the mortality rates of AKI. [16] The aim of this study was to investigate the prevalence, clinical manifestations and outcome of AKI in patients admitted to an ICU of trauma.

Patients and Methods
This was a retrospective study performed from January 2006 to January 2008 with patients consecutively admitted to the ICU of a trauma specialized hospital in Fortaleza city, northeast of Brazil. Data were collected from medical record review of trauma registry. Demographic characteristics and specifi c information, such as cause of AKI, co-morbidities presented by each patient, use of medications, time to develop AKI after ICU admission, length of hospital stay, need for surgery, mechanism of injury, time to beginning dialysis and the cause of death, were evaluated. All clinical signs and symptoms presented by each patient at hospital admission and laboratory data during hospital stay were analyzed.
The protocol of this study was approved by the ethical committee of the Dr. José Frota Institute.
AKI was defi ned according to the RIFLE criteria, based on creatinine, and patients were investigated for the presence of AKI during the hospital stay. [1] Hypotension was defi ned as mean arterial blood pressure (MAP) of <60 mmHg and therapy with vasoactive drugs was initiated when MAP remained lower than 60 mmHg. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) at admission were also analyzed. Sepsis was defi ned according to the American College of Chest Physicians/Society of Critical Care Medicine (ACCP/ SCCM) as "the systemic response to infection, manifested by two or more of the following conditions as a result of infection: (1) temperature > 38ºC or <36ºC; (2) heart rate > 90 beats/minute; (3) respiratory rate > 20 breaths/ minute or PaCO 2 < 32 mmHg; and white blood count > 2000/mm 2 , <4000/mm 2 or >10% immature (band) forms". [17] Hypovolemic shock was differentiated from septic shock when a patient without sepsis, i.e., those who did not fi ll the criteria for sepsis by ACCP/SCCM developed hypotension. Metabolic acidosis was defi ned as pH of <7.35 and arterial bicarbonate of <20 mEq/L; and coagulation abnormalities were defi ned a platelet count of <100 × 10 3 /mm 3 . Oliguria was considered to be present when the urinary volume was less than 400 mL/day despite appropriate fl uid replacement. Rhabdomyolysis was defi ned as creatine kinase (CK) level of >1000 IU/L. Other laboratory data evaluated were total blood count, aspartate aminotransaminase (AST), and alanine aminotransaminase (ALT). Renalspecifi c variables collected included admission, peak and discharge creatinine, levels of urea, potassium, sodium, and the main signs and symptoms at AKI diagnosis. Outcomes included length of stay in the ICU and hospital, use of antibiotics, need for dialysis and mortality.
RIFLE defi nes three grades of increasing severity of AKI -risk (class R), injury (class I) and failure (class F) -and two outcomes (Loss and End-stage kidney disease). [18] We used the change in the serum creatinine level to classify patients according to the RIFLE criteria. Patients who met any of the criteria of RIFLE classifi cation were classifi ed as AKI.
Patients were divided into four groups: patients with and without AKI, renal replacement therapy (RRT) and non-renal replacement therapy. We compared these groups in order to investigate the differences in clinical manifestations and laboratory features. Risk factors associated with AKI were investigated through a univariate and multivariate analysis.
The results were expressed through tables and summary measures (mean ± standard deviation) in the cases of quantitative variables. Data were analyzed with SPSS version 10.0 (SPSS Inc., Chicago, IL, USA) and Epi Info version 6.04b (Centers for Disease Control and Prevention) software. Comparison of parameters of the four groups (patients with and without AKI, RRT and non-RRT) was done with Student's t-test and Fischer's exact test. A logistic regression model was used for quantitative variables. Adjusted odds ratios (ORs) and 95% confi dence intervals (CIs) were calculated. A multivariate logistic regression was performed to analyze the possible risk factors for AKI. The factors included in the multivariate model were those that showed a significance level <10% in the univariate analysis (Mann-Whitney test and chi-square test). P values <0.05 were statistically signifi cant.
The frequency of hypotension was also not statistically different between AKI and non-AKI patients (5.8% vs. 15.6%, P = 0.10).
Of the 52 patients who developed AKI, 19 required RRT. This group represented 36.5% of patients with AKI and 14.7% of all patients admitted at ICU. This group consisted of 100% of men whose average age is 38 ± 14 years (21-66 years). A comparison between patients who required RRT and those who did not require is summarized in Table 2.

Discussion
The natural history of post-traumatic AKI is not well established. In retrospective studies on large trauma populations, a low incidence of AKI is generally reported (0.1-8.4%). [5,6] However, the incidence of AKI in the ICU varies from 1.5 to 24%. [2] In our population of patients admitted at an ICU, the incidence was higher (40.3%). We observed a higher prevalence of males with AKI (84% vs. 15%) because men are more frequently exposed to external injuries. Previous studies have shown similar fi ndings but some have shown that the mortality is similar for men and women. [3,4,20,21] This study was conducted in a reference center in Brazil where accidents involving motorcycles, cars and falls are the most common causes of trauma. In larger cities, where traffi c accidents and physical violence are more common, we can observe more serious injuries. It is important to note that males and younger individuals are more often involved in accidents. [21] The mean age of our patients was around 30 years, which is lower than that reported in literature (which ranges from 51 to 68 years). [4] The causes of AKI include sepsis, hypovolemia, preexisting renal impairment, and nephrotoxins such as aminoglycoside antibiotics and radiological contrast agents. [23,24] Causes of post-traumatic renal failure are likely to be multifactorial. [24][25][26][27][28][29][30][31] Additional risk posed by renal trauma itself must also be considered in this population, with subsequent outcomes varying because of both degree and type of injury occurring. [23] In the present study, the main causes of AKI were sepsis, hypotension and rhabdomyolysis. The low number of patients in whom CK level was assayed was responsible for only 6 patients detected with rhabdomyolysis; it probably would be higher if all patients had their CK level requested. We observed that 10 out of 11 patients with increased CK developed AKI. In the early literature, AKI in trauma patients was reported to be mainly secondary to crush injuries and rhabdomyolysis, [12] whereas more recently decreased renal perfusion has emerged as the most common cause of AKI. [5,10,11,32,33] The reported incidence of AKI in rhabdomyolysis ranges from 13% to approximately 50% and the prognosis in these cases is substantially worse. Among patients in the ICU who developed rhabdomyolysis, the mortality has been reported to be 59% when AKI is present and 22% when it is not present. [31,[34][35][36][37] The mechanisms involved in the pathogenesis of rhabdomyolysis are direct sarcolemmic injury (e.g., trauma) or depletion of ATP within the myocyte, leading to an unregulated increase in intracellular calcium. [38,39] In the case of patients with rhabdomyolysis caused by trauma, additional injury results from ischemia reperfusion and inflammation by neutrophils that infi ltrate damage muscle. [40] The value of CK above which the risk of AKI is signifi cantly higher is not set. AKI may be related to values as low as 500 IU/L, but this usually occurs when there are conditions such as sepsis, dehydration and metabolic acidosis, [31] similar to that observed in our study.
Brivet et al., [4] in a prospective study from 282 patients with AKI in an ICU, found sepsis in 48%, hemodynamic CU, intensive care unit, Mean ± SD; significant when P < 0.05, *CK in AKI diagnosis (RRT, n = 3; non-RRT, n = 7), , Values given in the parenthesis are in percentage dysfunction (excluding sepsis) in 32% and toxic injuries in 20%, similar to our results. Sepsis is one of the main causes of AKI in ICUs and is associated with worse prognosis. [3,[41][42][43][44][45][46] Patients with sepsis have generalized arterial vasodilatation, with an associated decrease in renal vascular resistance, which causes renal hypoperfusion and AKI. [45] Regel et al. [6] showed that during recent decades, the dramatic increase in intravenous fl uid administration to trauma patients in the first 24 hours after injury has markedly reduced the incidence of AKI and has improved the outcome.
The major fi ndings of the present study were that abdominal trauma and use of furosemide were independent risk factors for AKI in our population. During trauma, the consequences of elevated intraabdominal pressure are also signifi cant determinants for the impairment of renal function. However, experimental studies show that the impairment in renal function produced by increased intra-abdominal pressure is a local phenomenon caused by direct renal compression and is not related to cardiac output. [47] Furosemide is frequently used to facilitate fl uid and electrolyte management of AKI in many institutions although its potential benefi ts, adverse effects, and cost effectiveness to prevent or treat AKI remain uncertain. [23] Several small, randomized, controlled studies evaluating the use of furosemide to either prevent or treat AKI have produced negative results. [22,48,49] Furthermore, the use of diuretics for AKI has also been associated with an increased risk of non-recovery of renal function and mortality. [50] Furosemide, especially at high doses, is associated with important side effects. Previous observational studies have produced conflicting results about the association between furosemide and mortality. [22,50] Indeed it is not associated with any clinical benefi ts when used to prevent and treat acute renal failure in adults. [22] Efforts to establish the true incidence of AKI in trauma patients are complicated by alterations in the defi nition used to characterize renal dysfunction in various studies. To foster uniformity in both research and clinical practice, an expert group [Acute Dialysis Quality Initiative (ADQI)] developed a new classification of AKI that has been increasingly used, i.e., RIFLE classifi cation. It includes both biochemical measures of renal function and urine output as components of the defi nition. In a meta-analysis that included patient-level data of more than 71,000 patients in 13 studies, RIFLE criteria displayed a graded association with adverse outcomes. [51,52] The overall mortality in the study was 6.9% in the group without AKI and 31.2% in the AKI group. The RIFLE classifi cation was associated with increased risk of death and decreased likelihood of renal recovery. The majority of patients included in the present study were classifi ed as "Failure", according to RIFLE classifi cation. RIFLE was not an independent predictor of mortality in this ICU population with posttraumatic AKI, which may be due to the high mortality rate (almost all patients died, so it was not possible to fi nd specifi c risk factors for death). This fact may be due to the specifi c clinical conditions and severity of the trauma that override renal lesions in determining mortality. The majority of patients in the present study who required RRT were in "Failure" stage, but a high proportion of patients in this stage did not received RRT, which may be due to the severity of disease (most patients died before RRT could be initiated). In a recent study by Bagshaw et al. [53] with 9449 critically ill trauma patients, AKI was found in 18.1%, which was lower than that observed in the present study, and the factors associated with AKI were older age, female gender and the presence of co-morbidities. Another recent study with 3945 patients admitted to a Chinese hospital after road traffi c injury found AKI in a lower proportion of cases (10.7%), and the risk factors for AKI were use of vasopressor drugs for more than 4 hours, use of highdose diuretics and delayed transport time. [54] AKI is associated with a signifi cant risk of morbidity and mortality (rates as high as 78% in patients who require RRT). [22] In our study, the overall in-hospital mortality was 95.3% and it was 100% among patients with AKI. Both RRT and non-RRT group had no difference in mortality, despite differing in severity of renal failure. Gomes et al., [55] in a recent study with 436 trauma patients admitted to an ICU in Portugal, observed AKI in 50% of cases. Overall mortality was 30%, which was signifi cantly lower than that observed in the present study. Late mortality (2 days after admission) was 18% among AKI patients and 22% in non-AKI (P = 0.31) patients, showing that AKI was not associated with increased mortality in trauma patients, as observed in our study. Bagshaw et al., [53] however, found a signifi cantly higher mortality among trauma patients with AKI (16.7% vs. 7.8%, P < 0.001). Yuan et al. [54] also found higher mortality in patients with AKI, and this was higher according to the RIFLE criteria (mortality was 7.1% in patients without AKI, 37.4% in Risk, 52.9% in Injury and 79.2% in Failure). The mortality rate in our study was very high, but we do not know the actual reason for this. We believe that this was due to the severity of studied cases. The majority of patients had severe AKI (classifi ed as "Injury" of "Failure"), and this is associated with high mortality. Another possible reason for the high mortality could be related to a delay in the beginning of dialysis therapy, due to technical problems.
Because of the high mortality rates, prevention of AKI in severe trauma patients admitted to the ICU remains crucial. The risk factors for post-traumatic AKI identifi ed in the present study may help the provision of future strategies. The majority of patients were young males, and the high mortality is surprising since mortality was expected to be lower in this group of patients. We can attribute this to the fact that the patients were victims of severe trauma, which is very common in the hospital where they were admitted.

Limitations of the Study
The study limitations were its small population, insuffi cient resource to allow complete data collection and limitation to the required laboratory data. The small sample size is probably an important factor for not being able to show any effect of AKI on mortality even if our mortality rates had been different. We believe, however, that our results provide some information that will be useful in clinical practice.

Conclusion
The prevalence of AKI and overall mortality of our patients was higher than that reported in the literature. AKI is a frequent and fatal complication after trauma. RIFLE classification was not a predictor factor for mortality in our ICU post-traumatic AKI population. A better comprehension of risk factors associated with death in patients with trauma-associated AKI is important, and more effective measures of prevention and treatment of AKI in this population are urgently needed.