Ajai K. Malhotra

Blunt bowel and mesenteric injuries: the role of screening computed tomography.

BACKGROUND Early generation scanners have demonstrated poor sensitivity detecting blunt bowel/mesenteric injuries (BBMI). This study was aimed at determining the accuracy and role of helical scanners in BBMI. METHODS Retrospective chart review of patients with BBMI, or computed tomographic scans suspicious of BBMI, from August of 1995 to December of 1998. RESULTS One hundred of 8,112 scans (1.2%) were suspicious of BBMI. Of these suspicious scans, 53 patients had BBMI (true positive-TP) and 47 patients did not (false positive-FP). Seven patients with negative scans had BBMI (false negative-FN). Computed tomography contributed toward early surgery in 77% of patients who may have been delayed. Six patients developed intra-abdominal abscess. The abscess group had a significantly longer time interval from injury to surgery. Multiple findings were seen in 57% of true positive scans, whereas in 13% of false positive scans (p < 0.0001). An algorithm for management of BBMI is presented. CONCLUSION Helical scanners have high accuracy in detecting BBMI. Single versus multiple findings are useful in managing these injuries.

Minimal Aortic Injury: A Lesion Associated with Advancing Diagnostic Techniques

BACKGROUND With the increasing use of high-resolution diagnostic techniques, minimal aortic injuries (MAI) are being recognized more frequently. Recently, we have used nonoperative therapy as definitive treatment for patients with MAI. The current study examines our institutional experience with these patients from July 1994 to June 2000. METHODS All patients suspected of blunt aortic injury (BAI) by screening helical CT (HCT) underwent confirmatory aortography with or without intravascular ultrasound (IVUS). MAI was defined as a small (<1 cm) intimal flap with minimal to no periaortic hematoma. RESULTS Of the 15,000 patients evaluated by screening HCT, 198 (1.3%) were suspected of having BAI. BAI was confirmed in 87 (44%), and 9 (10%) of these had MAI. The initial aortogram was considered normal in five of the MAI patients. The correct diagnosis was made by IVUS (four patients), and video angiography (one patient). One MAI patient had surgery, and two (22%) died of causes not related to the aortic injury. Follow-up studies were done on the six MAI patients that were discharged. In two, the flap had completely resolved, and in one it remained stable. The remaining three patients formed small pseudoaneurysms. CONCLUSION Ten percent of BAI diagnosed with high resolution techniques have MAI. These intimal injuries heal spontaneously and hence may be managed nonoperatively. However, the long-term natural history of these injuries is not known, and hence caution should be exercised in using this form of treatment.

Associated injuries in blunt solid organ trauma: implications for missed injury in nonoperative management.

BACKGROUND During the past decade, nonoperative management (NOM) of hemodynamically stable blunt trauma patients with liver (L) or spleen (S) injury has become the standard of care. This trend has led to concerns over missed associated intra-abdominal injuries with concomitant morbidity. To better understand the incidence and risk of missed injury, patterns of associated intra-abdominal injury were examined in all patients with blunt liver and spleen injuries, and missed injuries were reviewed in patients undergoing NOM. METHODS Patients were identified from the registry of a Level I trauma center over a 3-year period. Records were reviewed for demographics, injury characteristics, and associated injuries. Indications for primary operation were hemodynamic instability or significant associated intra-abdominal injury. Missed injury was defined as unsuspected intra-abdominal injury requiring laparotomy in patients otherwise undergoing NOM for liver or spleen injury. RESULTS Eight hundred three patients (338 in the L group, 345 in the S group, and 120 in the L + S group) were treated between December 1995 and December 1998. Rates of planned NOM were 89% (L group), 78% (S group), and 75% (L + S group). On examination of all patients with blunt liver or spleen injuries, the incidence of associated intra-abdominal injury was higher in the L group at 5% as compared with 1.7% in the S group (p = 0.02). The associated intra-abdominal injury rate in the L + S group was similar to the L group at 4.2%. Although in the L and S groups, rates of diaphragm (0.5% vs. 1%, p = 0.45) and intra-abdominal bladder injury (0.3% vs. 0.3%, p = 0.99) were similar, bowel injury was more common in the L group (11% vs. 0%, p = 0.0004), as was pancreatic injury (7% vs. 0%, p = 0.007). In NOM patients, missed injury occurred in seven (2.3%) L patients versus zero S patients (p = 0.012). No L + S patient had unexpected injuries. Missed injuries included two small bowel, three diaphragm, one pancreas, and one mesenteric tear. CONCLUSION Damage to the pancreas and bowel is significantly associated with liver as opposed to spleen injuries. Actual missed intra-abdominal injury with NOM mirrors this pattern, occurring more often with liver than with spleen injuries. However, the overall incidence of missed injury is quite low, and should not influence decisions concerning eligibility for NOM. We speculate that the greater amount and/or different vector of energy transfer needed to injure the liver versus the spleen accounts for the greater rate of associated injuries to the pancreas/small bowel.

Patient safety in trauma: maximal impact management errors at a level I trauma center.

BACKGROUND The Division of Research at JCAHO developed a taxonomy (common terminology and classification schema) to promote consistency in reporting and facilitate root cause analysis. We undertook a review of trauma management errors at our institution with maximal impact (death). The analysis was based on the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) taxonomy. METHODS Trauma deaths between 2001 and 2006 at our Level I trauma were peer-reviewed to identify errors in management. The errors are classified according to type, domain, and cause. RESULTS Seventy-six (9.9%) of 764 deaths had management errors contributing to potentially preventable deaths in 60 (errors in management might have contributed to death) and preventable deaths (management errors definitely contributed to death) in 16 patients. Questionable resuscitation was the commonest type and involved poor treatment in the majority. Errors were made in all domains but most commonly in the emergency department and the operating room and in the resuscitative phase. Human errors predominated. CONCLUSIONS Management errors in the basics of trauma care continue even in established trauma centers, despite guidelines, protocols, and continuous performance improvement. Standardized reporting such as the taxonomy may result in progressive collection of patient safety data and lead to innovations to minimize these errors.

Effects of fiberoptic bronchoscopy on intracranial pressure in patients with brain injury: a prospective clinical study.

BACKGROUND Fiberoptic bronchoscopy (FB) plays an important role in making the diagnosis of nosocomial pneumonia and resolving lobar atelectasis in critically injured trauma patients. It has been shown to be a safe procedure with only occasional complications. However, in patients with head injuries, FB can lead to intracranial hypertension. Sustained increases in intracranial pressure (ICP) leads to poor outcome in these patients. Because of this, a prospective study was done not only to assess the effect of FB on ICP and cerebral perfusion pressure (CPP) in patients with brain injuries, but also to identify a regimen of sedation and anesthesia that could prevent significant increases in ICP during FB. METHODS Twenty-six FB were performed in 23 patients with ICP monitors or ICP monitors and ventriculostomy drains in place for Glasgow Coma Scale score < 8 or management of postcraniotomy trauma. FB was performed to aid in the diagnosis of nosocomial pneumonia or to aid in resolving lobar atelectasis. Before FB, all patients received a standard anesthetic regimen consisting of vecuronium (10 mg), morphine sulfate (4 mg), and midazolam (2.5 mg). Patients with diminished cranial compliance, defined as ICP > 10 mm Hg, also received a nebulizer treatment of 3 mL of 4% lidocaine before FB. All patients were preoxygenated with FIO2 = 1.0 for 10 minutes. Intracranial pressure, mean arterial pressure, and CPP were monitored continuously throughout the procedure. These same variables were also recorded at baseline and at 2-minute intervals during the procedure. The time to return to baseline ICP was also recorded. RESULTS The mean ICP at baseline (immediately before FB) was 12.6 mm Hg. After introduction of the bronchoscope, the ICP rapidly increased in 21 procedures (81%) and the mean highest ICP was 38.0 mm Hg. There was also a concomitant increase in mean arterial pressure such that there was no substantial change in CPP. The mean lowest CPP was 73.1 mm Hg. The average time for return of ICP to baseline was 13.9 minutes. In the subgroup of patients with ICP > 10, attempting to blunt the tracheal stimulation by anesthetizing the trachea with 4% nebulized lidocaine did not seem to be successful. The mean highest ICP in this subgroup was 41.8 mm Hg. The CPP changed in a similar manner, as the mean lowest CPP was 74.0 mm Hg. The mean time to return to baseline was 12.5 minutes. No patient had acute neurologic deterioration secondary to FB. CONCLUSIONS Although FB is an important procedure in the pulmonary care of head injured patients, it produces substantial, but transient, increases in ICP and should be used with caution in patients with diminished cranial compliance. Sedation, analgesia, paralysis, and topical tracheal anesthesia did not completely prevent the rise in ICP. Although no acute deterioration in condition occurred, secondary brain injury caused by localized cerebral ischemia is certainly possible. Because of the substantial increases in ICP, herniation may be precipitated in an occasional patient. Further study is needed to identify a regimen that will confer protection.

Early enteral nutrition does not decrease hypermetabolism associated with burn injury

BACKGROUND A prospective, randomized study was performed to compare the effects of early versus late enteral feeding on postburn metabolism. METHODS Burn patients were randomized to receive enteral feedings either within 24 hours (early) or 7 days (late) of injury. Basal energy expenditure (BEE) was calculated from Harris-Benedict equations and resting energy expenditure (REE) was obtained from indirect calorimetry. The average daily energy expenditure (DEE) was expressed as REE/BEE. RESULTS Average age, burn size, infections, and length of stay were similar between groups. Mortality between groups was similar (early, 28%; late, 38%) and not significantly influenced by inhalation injury. When controlled for percentage of total body surface area burn, inhalation injury, and age, the early group had an increased rather than decreased DEE, with a mean DEE calorie 0.17 more than the late group (p = 0.07). CONCLUSION Early enteral feeding does not decrease the average energy expenditure associated with burn injury.

Resuscitation with a novel hemoglobin-based oxygen carrier in a Swine model of uncontrolled perioperative hemorrhage.

BACKGROUND Systemic and pulmonary hypertension, possibly related to nitric oxide scavenging by free hemoglobin (Hb), is often seen during resuscitation with hemoglobin-based oxygen carriers (HBOCs). Recently, a second-generation HBOC, rHb2.0 for Injection (rHb), has been developed using recombinant human Hb that has reduced reactivity with nitric oxide. The current study evaluates the efficacy of this novel compound for resuscitation in a swine model of uncontrolled perioperative hemorrhage. METHODS After instrumentation, animals underwent splenectomy and rapid hemorrhage to a systolic blood pressure of 35 mm Hg and isoelectric electroencephalography. 15 minutes of shock was followed by resuscitation over 30 minutes. In phase I, 18 animals were randomized into three resuscitation groups: (1) lactated Ringers (LR) equal to three times the shed blood, the negative control group; (2) heterologous blood (BL) equal to Hb 2 g/kg, the positive control group; and (3) rHb equal to 2 g/kg, the treatment group. In phase II, six animals underwent the same experiment with a first-generation HBOC, diaspirin cross-linked Hb (DCLHb) equal to 2 g/kg, an additional control group. On day 0 after 2 hours of observation, spontaneously breathing animals were returned to their cages. Surviving animals were redosed on days 1, 2, and 3 (rHb/DCLHb 1 g/kg; LR/BL-LR 500 mL). Survivors were killed on day 5 and organs harvested for histologic examination. Group comparisons were performed using Students t test, repeated-measures analysis of variance, and chi2 test. Significance was set at 95% confidence intervals. RESULTS After resuscitation, systemic mean arterial pressure (MAP) (baseline = 107 +/- 15 mm Hg) was 128 +/- 34 and 108 +/- 15 mm Hg in rHb and BL animals, respectively, and remained stable. In LR and DCLHb animals, after normalization, MAP declined to 67 +/- 13 and 84 +/- 34 mm Hg, respectively. The rHb group maintained higher MAP than the LR and BL groups (p < 0.05 vs. both). With resuscitation, mean pulmonary arterial pressure (PAP) (baseline = 25 +/- 5 mm Hg) increased in rHb (40 +/- 4 mm Hg), BL (34 +/- 3 mm Hg), and DCLHb (40 +/- 3 mm Hg) groups, but stayed elevated only in the DCLHb group (36 +/- 3 mm Hg). PAP in the rHb group was similar to the BL group (p > 0.05), and both rHb and BL groups showed a higher PAP than the LR group (p < 0.05 vs. both). PAP was highest in the DCLHb group (p < 0.05 vs. rHb). Cardiac output of rHb and BL groups was similar (p > 0.05) throughout the observation period. Arterial lactate increased to 5.6 +/- 2.5 mmol/L with shock and then normalized to < 2.0 mmol/L in the rHb, BL, and LR groups within 30 minutes of resuscitation. It remained elevated to > 3.5 mmol/L and showed a delayed increase in the DCLHb group (p < 0.05). Causes and number of deaths were as follows: rHb, zero of six; BL-transfusion reaction, one of six; LR-irreversible shock, four of six; and DCLHb-ventricular failure, six of six. There was no significant increase in plasma methemoglobin (rHb) and no difference in liver or cardiac enzymes (rHb vs. BL). No histologic abnormalities were seen in the rHb group except for cytoplasmic vacuolation, a process thought to be related to metabolism of the test article. CONCLUSION rHb2.0 for Injection, a second-generation recombinant human HBOC, performs as well as heterologous blood for resuscitation after perioperative blood loss, does not cause sustained pulmonary hypertension, maintains adequate cardiac output and oxygen delivery, and is superior to either LR or DCLHb.

Preservation of splenic immunocompetence after splenic artery angioembolization for blunt splenic injury.

BACKGROUND Splenic artery angioembolization (SAE) is increasingly being used as an adjunct to nonoperative management for stable patients with blunt splenic injury (BSI). However, little is known about splenic immunocompetence after SAE. This study aims at assessing splenic immunocompetence after SAE for BSI. METHODS Peripheral blood was obtained from BSI patients (n = 8) who had SAE >6 months prior. Splenic immunocompetence was assessed by isolating mononuclear cells and incubating with CD4 and CD45RA and CD45RO antibody to analyze the proportion of T-cells expressing CD4 receptor, or coexpressing CD4 and either CD45RA or CD45RO receptors. Cells were counted by fluorescence-activated cell sorting and compared with trauma patients that had splenectomy for BSI also >6 months prior (n = 4, negative controls) and normal healthy volunteers with intact spleens (n = 4, positive controls). RESULTS The test was discriminatory for the asplenic state. %CD4 cells were significantly lower in splenectomized patients (16 ± 1) versus normal (40 ± 2), p < 0.05. This was due to significant decrease (8 ± 2 vs. 26 ± 4, p < 0.05) in %CD4CD45RA cells whereas the proportion of CD4CD45RO cells were maintained similar to normal. SAE patients had values (CD4, 36 ± 2, and CD4CD45RA, 24 ± 2) comparable to normal (p > 0.05) and significantly higher than splenectomized patients (p < 0.05). When the SAE group was subdivided into main (total, n = 4) and branch vessel (partial, n = 4) SAE, results were the same for both types of SAE. CONCLUSION Splenic immune function, measured by T-cell subset, generated only in the presence of an immunocompetent spleen, is preserved after SAE for BSI, main or partial.

Findings of a randomized controlled trial using limited transthoracic echocardiogram (LTTE) as a hemodynamic monitoring tool in the trauma bay

BACKGROUND We hypothesize that limited transthoracic echocardiogram (LTTE) is a useful tool to guide therapy during the initial phase of resuscitation in trauma patients. METHODS All highest-level alert patients with at least one measurement of systolic blood pressure less than 100 mm Hg, a mean arterial pressure less than 60 mm Hg, and/or a heart rate greater than 120 beats per minute who arrived to the trauma bay (TB) were randomized to have either LTTE performed (LTTEp) or not performed (non-LTTE) as part of their initial evaluation. Images were stored, and results were reported regarding contractility (good vs. poor), fluid status (empty inferior vena cava [hypovolemic] vs. full inferior vena cava [not hypovolemic]), and pericardial effusion (present vs. absent). Time from TB to operating room, intravenous fluid administration, blood product requirement, intensive care unit admission, and mortality were examined in both groups. RESULTS A total of 240 patients were randomized. Twenty-five patients were excluded since they died upon arrival to the TB, leaving 215 patients in the study. Ninety-two patients were in the LTTEp group with 123 patients in the non-LTTE group. The LTTEp and non-LTTE groups were similar in age (38 years vs. 38.8 years, p = 0.75), Injury Severity Score (ISS) (19.2 vs. 19.0, p = 0.94), Revised Trauma Score (RTS) (5.5 vs. 6.0, p = 0.09), lactate (4.2 vs. 3.6, p = 0.14), and mechanism of injury (p = 0.44). Strikingly, LTTEp had significantly less intravenous fluid than non-LTTE patients (1.5 L vs. 2.5 L, p < 0.0001), less time from TB to operating room (35.6 minutes vs. 79.1 min, p = 0.0006), higher rate of intensive care unit admission (80.4% vs. 67.2%, p = 0.04), and a lower mortality rate (11% vs. 19.5%, p = 0.09). Mortality differences were particularly evident in the traumatic brain injury patients (14.7% in LTTEp vs. 39.5% in non-LTTE, p = 0.03). CONCLUSION LTTE is a useful guide for therapy in hypotensive trauma patients during the early phase of resuscitation. LEVEL OF EVIDENCE Therapeutic study, level II.

Ventilator-associated pneumonia in injured patients: Do you trust your Gram's stain?

BACKGROUND The results of sputum or bronchoalveolar lavage (BAL) fluid Grams stain have been used to guide presumptive antibiotic therapy for ventilator-associated pneumonia (VAP) in injured patients, despite reported variability in sensitivity, specificity, positive predictive value, negative predictive value, and accuracy. Our aim was to evaluate the utility of Grams stain of BAL fluid in the diagnosis of VAP. METHODS We conducted a retrospective chart review of all mechanically ventilated trauma patients who developed pneumonia over a 5-year period in whom Grams stain and final culture data were available. RESULTS One hundred fifty-five records with complete data sets were reviewed. VAP was diagnosed by Centers for Disease Control and Prevention criteria and confirmed by BAL and quantitative culture in all patients. Overall accuracy of Grams stain in diagnosing VAP for any organism was 88% (137 true-positives). When assessed for the ability to predict pneumonia caused by a specific organism, the accuracy decreased significantly, with only 63% of Gram-negative VAPs and 72% of Gram-positive VAPs accurately identified by Grams stain. However, the absence of Gram-positive organism of Grams stain excludes Gram-positive VAP in 80% of patients. CONCLUSION All trauma patients should be covered presumptively for gram-negative organisms, as they encompass 70% of infections, but are not reliably identified by Grams stain. As 88% of VAP can be identified by the presence of any organism on Grams stain, it may be useful in the early diagnosis of VAP but cannot reliably be used to guide presumptive therapy.