van der Wilden GM, Velmahos GC, Joseph DK, Jacobs L, DeBusk MG, Adams CA, Gross R, Burkott B, Agarwal S, Maung AA, Johnson DC, Gates J, Kelly E, Michaud Y, Charash WE, Winchell RJ, Desjardins SE, Rosenblatt MS, Gupta S, Gaeta M, Chang Y, de Moya MA. Successful Nonoperative Management of the Most Severe Blunt Renal InjuriesA Multicenter Study of the Research Consortium of New England Centers for Trauma. JAMA Surg. 2013;148(10):924-931. doi:10.1001/jamasurg.2013.2747
Severe renal injuries after blunt trauma cause diagnostic and therapeutic challenges for the treating clinicians. The need for an operative vs a nonoperative approach is debated.
To determine the rate, causes, predictors, and consequences of failure of nonoperative management (NOM) in grade IV and grade V blunt renal injuries (BRIs).
Retrospective case series.
Twelve level I and II trauma centers in New England.
A total of 206 adult patients with a grade IV or V BRI who were admitted between January 1, 2000, and December 31, 2011.
Main Outcomes and Measures
Failure of NOM, defined as the need for a delayed operation or death due to renal-related complications during NOM.
Of 206 patients, 52 (25.2%) were operated on immediately, and 154 (74.8%) were managed nonoperatively (with the assistance of angiographic embolization for 25 patients). Nonoperative management failed for 12 of the 154 patients (7.8%) and was related to kidney injury in 10 (6.5%). None of these 10 patients had complications because of the delay in BRI management. The mean (SD) time from admission to failure was 17.6 (27.4) hours (median time, 7.5 hours; range, 4.5-102 hours), and the cause was hemodynamic instability in 10 of the 12 patients (83.3%). Multivariate analysis identified 2 independent predictors of NOM failure: older than 55 years of age and a road traffic crash as the mechanism of injury. When both risk factors were present, NOM failure occurred for 27.3% of the patients; when both were absent, there were no NOM failures. Of the 142 patients successfully managed nonoperatively, 46 (32.4%) developed renal-related complications, including hematuria (24 patients), urinoma (15 patients), urinary tract infection (8 patients), renal failure (7 patients), and abscess (2 patients). These patients were managed successfully with no loss of renal units (ie, kidneys). The renal salvage rate was 76.2% for the entire population and 90.3% among patients selected for NOM.
Conclusions and Relevance
Hemodynamically stable patients with a grade IV or V BRI were safely managed nonoperatively. Nonoperative management failed for only 6.5% of patients owing to renal-related injuries, and three-fourths of the entire population retained their kidneys.
Up to 10% of patients with blunt abdominal trauma have renal injuries.1 Over the last few decades, nonoperative management (NOM) has become increasingly popular, especially for low-grade (I-III) blunt renal injuries (BRIs). The published evidence is unclear about the role of NOM for higher grades (IV and V). In short, grade IV injuries indicate lacerations that extend to the renal pelvis, and grade V injuries indicate a shattered kidney or renal artery injuries with parenchymal devascularization. Santucci and Fisher2 showed in a meta-analysis of 16 studies that approximately 90% of grade IV BRIs could be managed without an operation. Even for grade V injuries, there is evidence that NOM can be successful.3- 6 Because grade IV and V injuries are uncommon, most of these studies lack an appropriate sample size to uncover statistical significances and make valid recommendations.
The Research Consortium of New England Centers for Trauma includes level I and II trauma centers in the region and has produced, so far, 4 studies.7- 10 In particular, it has evaluated injuries of low prevalence, which can only be analyzed by a multicenter collaboration. The objective of the present study is to determine the rate and predictors of failure of NOM (f-NOM) of patients with severe BRIs. We hypothesize that such injuries can be safely managed without an operation but that the rate of complications following operative management and NOM of BRIs and the need for subsequent interventions are high.
This is a retrospective case series including all adult patients with a grade IV or V BRI who were admitted between January 1, 2000, and December 31, 2011, to a New England trauma center. Renal injuries were graded based on computed tomographic (CT) findings and according to the American Association for the Surgery of Trauma Organ Injury Scale (Table 1).11Quiz Ref IDTen centers were verified by the American College of Surgeons Committee on Trauma as level I (9 centers) or II (1 center) trauma centers; 2 additional centers were accredited as level II trauma centers by their state but not the American College of Surgeons. Exclusion criteria were being younger than 15 years of age, undergoing an urgent operation in an outside hospital, and dying before or on arrival at the receiving hospital.
The patients were divided into those who received an immediate operation (IO) and those managed nonoperatively (ie, NOM). Hemodynamic instability and peritonitis were indications for an IO. Quiz Ref IDPatients were included in the NOM group if there was a clear note in the medical record to that effect or if an operation was booked more than 3 hours after the diagnosis of BRI. The decision to use the cutoff point of 3 hours in order to define NOM has been used in previous studies by the Research Consortium of New England Centers for Trauma.8,10 Because the decision-making process concerning an operation vs NOM was not always clear from the review of the medical records, we were obliged to make some arbitrary assumptions. Based on the infrastructure of our trauma centers, we believe that longer than 3-hour intervals, between the diagnosis of BRI and the booking of an operation, indicated that the patient was initially managed nonoperatively. It would be extremely unlikely that a decision for an IO was made 3 hours after diagnosis. On the other hand, it is possible that shorter periods of NOM (eg, 2 hours) were labeled as an IO in this study. Failure of NOM was defined as the need for surgery after a trial of NOM or as death due to renal-related complications during NOM; if none of this happened, NOM was considered successful (s-NOM).
The following data were collected: demographics, mechanism of injury (road traffic crash, fall, assault, or other), associated injuries, admission hemodynamics, Injury Severity Score (ISS), CT findings, renal injury grade (IV or V), presence of free abdominal blood on CT scan (recorded as diffuse or confined around the kidney), type of management (NOM or IO), indication for operative intervention, operative procedures, operative findings, any other interventions required, intensive care unit (ICU) and hospital stay, morbidity, and mortality.
The primary outcome was f-NOM, classified as renal-specific and non–renal-specific according to the reason for which the NOM of the patient failed. The secondary outcome was complications of NOM and the interventions required to treat those complications.
Patients who received an IO were compared with those who were managed nonoperatively. In addition, f-NOM and s-NOM patients were compared. We dichotomized certain continuous variables across clinically meaningful values. Age was dichotomized at 55 years, the ISS at 25, systolic blood pressure at 100 mm Hg, heart rate at 100 beats per minute, and hematocrit level at 30%. Continuous variables were summarized using mean values with standard deviations and compared using 2-sample t tests, or summarized using median values with interquartile ranges and compared using Wilcoxon rank sum tests. Categorical variables (reported as counts and proportions) were compared using the χ2 test or the Fisher exact test. Logistic regression was performed to identify independent predictors of f-NOM. Odds ratios (ORs) and 95% CIs were reported for each predictor. The incidence of f-NOM based on different combinations of independent predictors of f-NOM was examined. P ≤ .05 indicated statistical significance. SAS version 9.2 (SAS Institute) was used for the entire analysis. Our study was approved by the institutional review boards of all participating hospitals.
Quiz Ref IDIn total, 206 patients were identified, 154 (74.8%) with a grade IV BRI who received an IO and 52 (25.2%) with grade V BRI who were managed nonoperatively. The mean (SD) age of the population was 36.2 (18.3) years (median age, 30 years; range, 15-90 years), with 173 of the 206 patients (84.0%) younger than 55 years of age. Of the 206 patients, 154 (74.8%) were men, with a mean (SD) ISS of 25.7 (13.1) (median ISS, 24; range, 4-75), and 131 (63.6%) were in road traffic crashes. Other injuries besides the BRI were found in 164 patients (79.6%), including splenic injuries (62 patients [30.1%]), liver injuries (57 patients [27.7%]), and other abdominal injuries (29 patients [14.1%]).
Among the 160 patients (77.7%) who required intensive care, the mean (SD) ICU stay was 8.0 (17.8) days (median ICU stay, 2.5; range, 0-204 days). The mean (SD) hospital stay for the entire population was 15.7 (23.6) days (median hospital stay, 8; range, 1-210 days). Overall, 17 patients (8.3%) died, 3 (1.5%) of whom had deaths that were related to the kidney injury. The renal unit (ie, the kidney) was preserved in 157 (76.2%) patients, including 18 of 52 IO patients (34.6%), 135 of 142 s-NOM patients (95.1%), and 4 of 12 f-NOM patients (33.3%).
Of the 52 IO patients, 30 (57.7%) had a nephrectomy, and 1 (1.9%) had a partial nephrectomy. Four patients had a nephrectomy too, but 3 of them died within 24 hours, and 1 had a nonperfused kidney, which was left in situ after a damage control operation. In total, 35 kidneys were lost.
Of the 154 NOM patients, 33 (21.4%) underwent angiography during the acute stage, 25 (16.2%) proceeded to angiographic embolization for bleeding control, and 2 (1.3%) required another embolization. In addition, 13 patients (8.4%) received a ureteral stent early after the injury.
As expected, there were multiple differences in injury characteristics and physiologic stability between IO and NOM patients (Table 2). Morbidity and mortality were higher among IO patients than NOM patients, and ICU and hospital stays were longer for IO patients than NOM patients. There were 4 independent predictors of IO: ISS greater than 25 (OR, 1.06 [95% CI, 1.03-1.09]), the presence of diffuse intraperitoneal blood (OR, 5.86 [95% CI, 2.38-14.41]), active vascular extravasation on CT scan (OR, 3.45 [95% CI, 1.45-8.15]), and associated abdominal injuries (OR, 8.73 [95% CI, 2.97-25.67]).
Quiz Ref IDFailure of NOM for 12 of 154 NOM patients (7.8%) was observed, without a significant difference in failure rates between grade IV patients (7.6%) and grade V patients (7.8%). For 10 of 12 f-NOM patients (83.3%), NOM failed for renal-related reasons (renal-related failure rate, 6.5%). Failure of NOM for the 2 remaining patients was due to their liver injuries. Operative intervention took place a mean (SD) 17.6 (27.4) hours after admission (median time, 7.5 hours after admission; range, 4.5-102 hours after admission). Hemodynamic instability was the indication for operation for 10 of 12 f-NOM patients (83.3%); 1 patient underwent an operation because of peritonitis, and 1 patient underwent an operation because of abdominal compartment syndrome. Eight patients underwent a nephrectomy, 1 patient underwent a partial nephrectomy, and 1 patient underwent renal repair. The remaining 2 patients had no renal procedures performed; 1 patient had a decompressive laparotomy, and the other patients received a liver resection and a cholecystectomy. One patient (8.3%) died 2 days after operative intervention owing to multisystem organ failure. Failure of NOM for this patient occurred 7 hours later, and thus the patient received a nephrectomy and an adrenalectomy.
Compared with patients with s-NOM, patients with f-NOM were older, had active extravasation more frequently based on CT findings, and had a higher rate of general complications (Table 3). Quiz Ref IDTheir hospital and ICU stays were longer than those of patients with s-NOM. Two independent predictors of f-NOM were identified: older than 55 years of age (OR, 5.99 [95% CI, 1.58-22.61]) and a road traffic crash as the mechanism of injury (OR, 5.62 [95% CI, 1.09-28.86]). When both risk factors were present, f-NOM occurred in 27.3% of the patients; when both were absent, the s-NOM rate was 100%.
Renal-related complications developed in 12 of 52 IO patients (23.1%) and in 49 of 154 NOM patients (31.8%; P = .23) (Table 2). Persistent or recurrent hematuria was the most common complication and occurred in 26 of 206 patients (12.6%), including 1 of the 52 IO patient (1.9%) and 25 of the 154 NOM patients (16.2%) (Table 4). For the majority of patients, the hematuria was self-limited, but 5 of the 26 patients (19.2%) required angiographic embolization. The second most common complication was urinoma. It developed in 21 of 206 patients (10.2%), including 5 of the 52 IO patients (9.6%) and 16 of the 154 NOM patients (10.4%). The IO patients were managed with a ureteral stent (3 patients) or a percutaneous nephrostomy (2 patients), whereas the NOM patients received a stent (8 patients), underwent a percutaneous nephrostomy (2 patients), or received no intervention (6 patients). Thirteen of 206 patients (6.3%) with a urinary tract infection were managed with antibiotics. A perirenal abscess was identified in 3 of 206 patients (1.4%) and drained percutaneously. Renal dialysis was required for 3 of 206 patients (1.4%); an additional 7 patients had transient serum creatinine elevations, which returned to normal in all patients. Of other general complications, pneumonia and deep venous thrombosis were the most frequent.
Of 206 patients, 82 (39.8%) returned for a clinic visit, 15 of 52 IO patients (28.8%) and 67 of 154 NOM patients (43.5%). Four NOM patients were reported to have persistent hematuria at follow-up, while 1 NOM patient still had poor function of the kidney due to hydronephrosis. An additional patient was hypertensive after a partial infarction of the kidney. None of the IO patients that were seen in the clinic had long-term sequelae. Of the 82 patients who returned to the clinic, 7 IO patients and 40 NOM patients had another CT scan, and only 2 patients (both NOM patients) had a small urinoma. No action was taken on any of these patients, based on the CT findings.
Nonoperative management has become the preferred way of managing BRIs, even for high-grade injuries.2,3,12- 24 Of the 206 patients with either a grade IV or a grade V BRI analyzed in our multicenter study, 154 (74.8%) were offered NOM, which was successful for 142 of the 154 NOM patients (92.2%). The failure rate related to the BRI was 6.5% (ie, 10 of 154 patients). These findings agree, in general, with the existing evidence in the literature, consisting typically of limited case series, often with disparate results. Altman et al3 managed 13 patients who had grade V BRIs, offering NOM successfully to 6 patients. All of them had functioning renal parenchyma on additional CT scans. Buckley and McAnninch25 reported the s-NOM of one-third of patients with grade IV blunt and penetrating renal injuries; none of them required a delayed nephrectomy. Shariat et al26 successfully managed nonoperatively 41 of 51 patients with a grade IV BRI (80%). In a systematic review of the literature, Umbreit et al27 examined 95 children with a grade IV BRI. No intervention was possible for 86 of the 95 children (72%), and partial renal preservation was possible for 90 of the 95 children (95%). In contrast to that, Rogers et al28 reported on 10 children with a grade IV BRI and 10 children with a grade V BRI. Whereas only 8 of the 10 children (80%) with a grade IV BRI were successfully managed nonoperatively, all 10 children with a grade V BRI required an IO, and only 3 (30%) achieved long-term preservation of renal function. In our study, of the 52 patients who received an IO, 18 (34.6%) underwent a renal-salvage procedure (defined as a functioning renal unit of 50% or more25). Of the 154 NOM patients, 139 (90.3%) preserved the renal function of the injured kidney; of the 142 s-NOM patients, 135 (95.1%) preserved the renal function of the injured kidney.
Angiographic embolization has been a valuable adjunct of NOM and was used in 25 of our 154 NOM patients (16.2%), with a 92% initial success rate for bleeding control and 100% success rate after another attempt. Similarly, Hagiwara et al17 has shown excellent bleeding control in 7 of 8 patients with a high-grade BRI and active extravasation. The effectiveness of angiographic embolization was lower in a study of 22 patients with BRI by Menaker et al,16 who experienced 7 failures of NOM (27%). Similarly, Hotaling et al29 analyzed the National Trauma Data Bank and reported that 29% of patients who underwent angiographic embolization required another embolization.
Multiple predictors of f-NOM have been identified in the literature, most of them expressing common sense. Toutouzas et al12 found ISS, fluid and blood requirements, and diagnostic findings to be predictive of f-NOM. Associated abdominal injuries, older than 55 years of age, ISS greater than 25, ongoing blood transfusions, and worsening acidosis are additional predictors of f-NOM in other studies.2,5,19,25,30 An age of older than 55 years and a road traffic crash as the mechanism of injury were the 2 independent predictors in our study, and they both increased the odds of f-NOM by nearly 6-fold. Although an older age does not come as a surprise, the specific mechanism of injury requires further exploration. In a study by Kuan et al31 of 115 patients with BRI from the Crash Injury Research and Engineering Network, 36 were found to have grade III to grade V BRIs. In frontal collisions, the seat belt was deemed to be related to the injury, whereas side-door intrusion and the armrest seemed to be the causative factors in lateral collisions. Therefore, it may be true that road traffic crashes with or without seat belts predispose patients to BRI more than do other mechanisms of injury.
The complications following a high-grade BRI are significant32 even when the patients are managed nonoperatively. Starnes et al13 showed that the incidence of complications in severe BRI (grades III-V) were similar between patients who underwent renal exploration (7.3%) and NOM patients (7.9%). Nephrorrhaphy was associated with the highest rate of complications. Similarly, Shariat et al26 showed nonsignificantly different rates of complications between NOM patients (28%) and IO patients (13%) with grade IV BRIs. Obviously, these results may be misleading because renal-related complications are harder to develop after nephrectomy, and one could argue that the ultimate complication is the loss of the kidney. In our study, persistent hematuria and urinoma were the most frequent complications. An intervention was required for 19.2% of patients with persistent hematuria (angiographic embolization) and 71.4% of patients with urinoma (ureteral stenting, percutaneous nephrostomy, and percutaneous drainage of collection).
Long-term follow-up is the Achilles’ heel of every trauma study. We only had data on 39.8% of patients who returned to follow-up and on even fewer who had additional CT scans. Our study was not designed to address the speed of the healing process for patients with high-grade BRI or the common questions of patients returning to work or athletic activities. We could not comment on long-term sequelae, such as nephrogenic hypertension, in this population. Other limitations of this retrospective review included the lack of standardized protocols across all participating centers for the management of BRI, which resulted in significantly different practices. Analyzing these differences made no statistical sense because the number per center became prohibitively low. Finally, as it happens in such unfunded, multicenter collaborations, we compromised certain fields in order to improve participation. Not all desirable details were requested because the primary intent was to allow centers to collect data without an unreasonable effort.
Despite these limitations, the present study, to our knowledge, presents the highest number of grade IV and grade V BRIs to date. Nonoperative management was safely applied without risking the unnecessary loss of renal units or disastrous complications. Almost three-fourths of the population was offered NOM, with an f-NOM rate of less than 8% that was similar between grade IV and grade V BRIs. Two independent predictors of f-NOM were detected: older than 55 years of age and a road traffic crash; if both predictors were absent, there was a 100% success rate of NOM. Renal-related complications were significant, primarily in the form of hematuria or urinoma, but successfully managed by additional interventions. At the end, 76.2% of patients with a grade IV or grade V BRI (and 90.3% of those treated by NOM) preserved renal function. Patients with the most severe kidney injuries can be confidently managed nonoperatively.
Accepted for Publication: January 28, 2013.
Corresponding Author: George C. Velmahos, MD, PhD, Department of Surgery, Massachusetts General Hospital, 165 Cambridge St, Ste 810, Boston, MA 02114 (email@example.com).
Author Contributions:Study concept and design: van der Wilden, Velmahos, Gross.
Acquisition of data: van der Wilden, Joseph, Jacobs, DeBusk, Adams, Gross, Burkott, Agarwal, Maung, Johnson, Gates, Kelly, Michaud, Charash, Winchell, Desjardins, Rosenblatt, Gupta, Gaeta.
Analysis and interpretation of data: van der Wilden, Velmahos, Gates, Winchell, Chang, de Moya.
Drafting of the manuscript: van der Wilden, Velmahos, Burkott, Maung, Michaud, Desjardins, Chang.
Critical revision of the manuscript for important intellectual content: Velmahos, Joseph, Jacobs, DeBusk, Adams, Gross, Agarwal, Maung, Johnson, Gates, Kelly, Charash, Winchell, Rosenblatt, Gupta, Gaeta, de Moya.
Statistical analysis: van der Wilden, Maung, Chang.
Administrative, technical, and material support: van der Wilden, Velmahos, DeBusk, Gross, Agarwal, Kelly, Desjardins.
Study supervision: Velmahos, Jacobs, Gross, Gates, Kelly, Gupta, de Moya.
Published Online: August 14, 2013. doi:10.1001/jamasurg.2013.2747.
Conflict of Interest Disclosures: None reported.
Previous Presentation: This paper was presented at the 93rd Annual Meeting of the New England Surgical Society; September 23, 2012, Rockport, Maine; and is published after peer review and revision.