Single dose cephalosporin prophylaxis in high-risk patients undergoing surgical treatment of the biliary tract

During June 1985 through October 1986, 292 patients considered to be at high risk for having postoperative complications develop underwent cholecystectomy and were evaluated in a multicenter, randomized, prospective, double-blind study. Risk factors included age greater than 70 years, acute cholecystitis within the previous six months, obstructive jaundice, obesity and diabetes mellitus. One gram of cefamandole was administered intravenously to 144 patients and 148 patients received 1 gram of cefotaxime intravenously 30 minutes prior to skin incision. Culture-proved bactibilia was found in 55 patients and 11 of the patients had choledocholithiasis. Of the risk factors considered to place patients at high risk for postoperative infectious complications, obesity and acute cholecystitis proved to be the more common. However, age greater than 70 years, diabetes mellitus and obstructive jaundice were more significant risk factors predisposing to bactibilia. The most common organisms isolated from the bile and gallbladder intraoperatively were Staphylococcus, Streptococcus and Klebsiella species along with enterococcus, Escherichia coli and diphtheroids. Clinically significant postoperative infections occurred in eight patients, including six patients in the cefamandole group and two patients in the cefotaxime group. Antibiotic concentrations were measured in the serum, muscle, subcutaneous fat, gallbladder and bile, with cefamandole showing statistically significant greater concentrations in bile, gallbladder and muscle tissue. There was no statistical significance between the postoperative infection rates, total period of hospitalization or total hospital charges for each group. Therefore, there is no significant advantage between a single prophylactic dose of cefamandole versus cefotaxime for high-risk patients undergoing biliary tract operation.

Autotransfusion of potentially culture-positive blood (CPB) in abdominal trauma: preliminary data from a prospective study

Increased use of autotransfusion for traumatic hemorrhage may reduce amounts of banked blood needed for severe injuries. Autotransfusion is standard for traumatic hemothorax, but has been limited for abdominal injuries. This prospective study used microbiologic data from 152 patients with intestinal injuries. Where anticipated blood loss was greater than 1,000 mL, blood from the peritoneal cavity was cultured, washed, concentrated, and recultured before reinfusion. Infection rates were stratified using the Penetrating Abdominal Trauma Index (PATI). Fifty patients with PATI greater than 20 who received banked blood (group I) (mean: 1,800 mL) were compared with 20 patients (group II) who received autotransfused, potentially culture-positive blood (CPB) (mean: 3,900 mL). Wound infection rates were identical in both groups (25%). No statistically significant increase was found in site-specific infection risk when severity of injury was stratified according to PATI. Bacteremias, pulmonary infections, and urinary infections were not caused by bacteria cultured from autotransfused blood. We conclude that washed CPB may be autotransfused without significantly increased risk of infection in patients with severe abdominal injuries.

Decreased mesenteric blood flow in experimental multiple organ failure

Disorders of mucosal blood flow may contribute to gut barrier dysfunction in multiple organ failure (MOF). We evaluated alterations of mesenteric microcirculation in a rat model of MOF. Male Wistar rats received 1 mg/g body weight of zymosan A in 4 ml of mineral oil by intraperitoneal injection and were studied on Days 1, 3, and 5 following injection. A control group received no zymosan. The experimental group also received a fluid resuscitation regimen of 0.9 N saline subcutaneously equal to 0.1 ml/g body weight on the day of zymosan injection and 0.05 ml/g body weight daily thereafter. Day 1 animals tended toward a statistically significantly lower mean arterial pressure versus controls (86.6 +/- 8 mm Hg versus 106 +/- 5 mm Hg, F = 0.09 by ANOVA). Significant arteriolar vasoconstriction occurred on Days 1 and 3 versus control (70 +/- 4 microns and 57 +/- 8 microns versus 96 +/- 3 microns, F = 0.0002). Laser doppler velocity, indicating red blood cell velocity, expressed as a percentage of control paralleled this vasoconstriction (70 +/- 9 and 72 +/- 7%, respectively). We conclude that mesenteric arteriolar vasoconstriction occurs accompanied by decreased red blood velocity. We believe that this is indicative of decreased mesenteric perfusion in this zymosan model of MOF and that survival to Day 5 is associated with a reversal of these microcirculatory changes.

Pneumonia: incidence, risk factors, and outcome in injured patients

One hundred thirty (44.2%) of 294 patients hospitalized for trauma and admitted to the Surgical Intensive Care Unit for mechanical ventilation developed hospital-acquired bacterial pneumonia. The predominant pathogens isolated were gram-negative enteric bacilli (72%), but there was not an increase in mortality associated with gram-negative pneumonia compared with similar patients without pneumonia. Of the seven admission risk factors univariately associated with the development of acquired bacterial pneumonia, only emergent intubation (p less than 0.001), head injury (p less than 0.001), hypotension on admission (p less than 0.001), blunt trauma as the mechanism of injury (p less than 0.001), and Injury Severity Score (p less than 0.001) remained significant after stepwise logistic regression. Not surprisingly, as mechanical ventilation is continued, the probability of pneumonia emerging increases. The consequences of hospital-acquired bacterial pneumonia are a significant seven-, five-, and two-fold increase in mechanically ventilated days, intensive care, and hospital stay, respectively. We conclude that the incidence of hospital-acquired pneumonia in injured patients admitted to the ICU for mechanical ventilation occurs in nearly half the patients, is associated with specific risk factors, and significantly increases morbidity but does not increase mortality.

Early tracheostomy for primary airway management in the surgical critical care setting

During a 12-month period, 264 patients with multiple injuries who required mechanical ventilation were admitted to the surgical intensive care unit. One hundred twenty patients (46%) were disengaged from the ventilator, and 38 patients (14%) died. Of the remaining 106 patients (40%) 51 patients (group I) were to receive tracheostomy within 1 to 7 days, and 55 patients (group II) underwent late (8 or more days after admission) tracheostomy. Multiple variables in four categories (admission, operative, ventilatory, and outcome) were analyzed prospectively to define the impact that early tracheostomy had on duration of mechanical ventilation, intensive care stay, and hospital stay. Morbidity and mortality rates of the procedures were assessed. Early tracheostomy, in a homogeneous group of critically ill patients, is associated with a significant decrease in duration of mechanical ventilation, as well as shorter intensive care unit and hospital stays, compared with translaryngeal endotracheal intubation. There were no deaths attributable to tracheostomy, and overall morbidity of the procedures was 4%. We conclude that early tracheostomy has an overall risk equivalent to that of endotracheal intubation. Furthermore, early tracheostomy shortens days on the ventilator and intensive care unit and hospital days and should be considered for patients in the intensive care unit at risk for more than 7 days of intubation.

Pancreaticoduodenectomy. University experience and resident education

A radical pancreaticoduodenectomy offers the best chance for survival in patients with periampullary and pancreatic malignant neoplasms. A pancreaticoduodenectomy has educational value since complex pancreatic operations are demanding and important to the training of surgical residents. Increased pancreaticoduodenectomy experience (per surgeon) has been associated with improved outcomes. We examined the hypothesis that residents who are supervised by faculty surgeons can perform pancreaticoduodenectomies with acceptable outcomes. From 1976 to 1987, 127 pancreatic resections were performed by 81 residents who were supervised by 15 faculty surgeons in four teaching hospitals. A pancreaticoduodenectomy was performed on 61 patients. All residents served as an operating surgeon on a pancreatic resection, and 58 (82%) performed pancreaticoduodenectomies. The mortality for the pancreaticoduodenectomies was 8%, with a 36% major complication rate. A pancreaticoduodenectomy can be performed safely by residents under supervision. A review of the results identifies the means of improving outcomes. These results justify the preservation of a pancreaticoduodenectomy as an important experience for residents.

Indomethacin treatment of human adult respiratory distress syndrome

The purpose of this investigation was to evaluate the magnitude and duration of changes in lung function and oxygen transport in patients with adult respiratory distress syndrome (ARDS) receiving indomethacin. Ten patients with ARDS were randomized to receive intravenously either a single 50 mg dose of indomethacin or placebo. Comparing 1 hr postinfusion levels to baseline observations in the indomethacin group, PaO2 increased to 125 +/- 13 torr from 93 +/- 8 torr, PaO2/FIO2 ratio increased to 223 +/- 24 from 160 +/- 5, and Qs/Qt dropped to 0.20 +/- 0.03 from 0.27 +/- 0.03 (all P less than 0.05). These alterations in oxygenation gradually returned to baseline levels over the ensuing 8 hr. No such changes were noted in the placebo group.

Surgical management of complications of endoscopic sphincterotomy with precut papillotomy

We reviewed 574 endoscopic sphincterotomy procedures. Fifty-six precut papillotomies were performed. Presenting conditions included choledocholithiasis, cholangitis, benign and malignant papillary strictures, and stenosing papillitis. Complications were identified in 16 percent: perforation in 9 percent, pancreatitis in 5 percent, bleeding in 2 percent, and pancreatic abscess in 2 percent. One patient died. Six patients required operation for complications. Perforation of the duodenum or common bile duct seen within 8 hours was managed with drainage and closure of the perforation with minimal complications. Duodenal perforations operated on later than 8 hours required more extensive procedures. All these patients had significant post-operative complications. Three patients were managed nonoperatively. Precut papillotomy carries a significantly higher complication rate than conventional sphincterotomy. Our experience suggests that there is no place for conservative management of duodenal perforation.

Organ injury scaling: spleen, liver, and kidney

The Organ Injury Scaling (O.I.S.) Committee of the American Association for the Surgery of Trauma (A.A.S.T.) was appointed by President Trunkey at the 1987 Annual Meeting. The principal charge was to devise injury severity scores for individual organs to facilitate clinical research. The resultant classification scheme is fundamentally an anatomic description, scaled from 1 to 5, representing the least to the most severe injury. A number of similar scales have been developed in the past, but none has been uniformly adopted. In fact, this concept was introduced at the A.A.S.T. in 1979 as the Abdominal Trauma Index (A.T.I.) and has proved useful in several areas of clinical research. The enclosed O.I.S.'s for spleen, liver, and kidney represent an amalgamation of previous scales applied for these organs, and a consensus of the O.I.S. Committee as well as the A.A.S.T. Board of Managers. The O.I.S. differs from the Abbreviated Injury Score (A.I.S.), which is also based on an anatomic scale but designed to reflect the impact of a specific organ injury on ultimate patient outcome. The individual A.I.S.'s are, of course, the basic elements used to calculate the Injury Severity Score (I.S.S.) as well as T.R.I.S.S. methodology. To ensure that the O.I.S. interdiffuses with the A.I.S. and I.C.D.-9 codes, these are listed alongside the respective O.I.S. Both the currently used A.I.S. 85 and proposed A.I.S. 90 are provided because of the obligatory transition period. Indeed, A.I.S. 90 contains the identical descriptive text as the current O.I.S.'s. The Abdominal Trauma Index and other similar indices using organ injury scoring can be easily modified by replacing older scores with the O.I.S.'s.(ABSTRACT TRUNCATED AT 250 WORDS)

Effect of pyruvate on regional ventricular function in normal and stunned myocardium

The prolonged ventricular dysfunction following brief periods of coronary artery occlusion that does not produce irreversible damage has been termed the "stunned" myocardium. Although ventricular function returns to preischemic values by 1 to 7 days after reperfusion is established, inotropic therapy may be necessary to enhance contractility in the stunned heart. The purpose of this study was to determine the effect of pyruvate on ventricular function in normal and stunned myocardium. Eight chloralose/urethane anesthetized dogs were instrumented with ultrasonic crystals to measure systolic wall thickening in the left anterior descending artery (LAD) and left circumflex artery perfused regions of the left ventricle. Pyruvate (1 ml/min of 150 mM sodium pyruvate, pH 7.4) was infused directly into the LAD prior to and 30 minutes after a 10 minute LAD occlusion. Prior to LAD occlusion, LAD pyruvate infusion increased systolic wall thickening in the LAD-perfused region from 16.2% +/- 4.3% to 23.4% +/- 5.1% (p less than 0.05). Thirty minutes after LAD occlusion, regional wall thickening was depressed (3.3% +/- 2.6%; p less than 0.05), which is indicative of stunned myocardium. Subsequent LAD pyruvate infusion increased wall thickening in the stunned myocardium to 12.7% +/- 2.5%. The improvement of regional ventricular function was maintained only during the pyruvate infusion, as function returned to prepyruvate levels within 20 minutes after cessation of pyruvate infusion. These data indicate that pyruvate exerts a positive inotropic effect in normal and stunned myocardium. If pyruvate, a key intermediate in energy-producing pathways, exerts its inotropic effect through an enhancement of the energy state of the heart, it may have advantages over traditional inotropic agents in the treatment of postischemic contractile dysfunction.

The role of pediatric surgery in the medical school curriculum

In most medical schools, exposure to pediatric surgery is presented as a subspecialty elective. We have offered it as an integral part of the surgical clerkship for 10 years in the belief that it provides an excellent educational environment. To confirm this concept, the quizzes (Q), final examinations (FE), and grades of students assigned to the pediatric surgical service were prospectively studied. All students (N = 139) in the surgical clerkship entered the study. Thirty-two students were randomly selected and assigned to the surgical service of a major pediatric hospital (P-Surg) for 50% of their clerkship. The other students (N = 107) were assigned to a variety of adult surgical services (G-Surg) and served as the control group. All students attended the same seminars, used the same educational materials, were examined with the same test items, and were evaluated by the same oral examiners. Test items were electronically scored and the database was analyzed on an IBM computer. The statistical analysis was performed using a Student's t test and chi 2 analysis. There was no significant difference in the demonstrated cognitive performance and grades awarded to the two groups of students. We conclude that a pediatric surgical service provides an atmosphere that is educationally comparable to the adult general surgical service.

Early postoperative acute abdominal complications

The detection and management of acute postoperative conditions of the abdomen present a challenge to the surgeon. Many of these conditions threaten life because of their association with disorders of oxygenation, circulatory stability, and infection leading to multiple organ failure. This discussion highlights features of rapid diagnosis and focuses on problems relating to the decision to reoperate for diagnosis and therapy.

Penetrating arterial trauma. Analysis of missed vascular injuries

From 1976 to 1981, 677 patients with potential vascular injuries from penetrating wounds to the upper and lower extremities and the neck were treated. Surgical exploration was used to evaluate 237 patients; 440 patients were evaluated by arteriography alone, with negative results. In the group that underwent surgery, there were 137 arterial injuries detected. Follow-up studies were performed to assess the rate of missed vascular injuries using each diagnostic modality. Short-term follow-up was obtained in 81% of all patients. Long-term follow-up, averaging 5.1 years, was obtained in 33% of the patients. Vascular injuries, which went undetected at the time of initial evaluation, were present in both the group evaluated by vascular exploration and the arteriography group. The combination of arteriography and exploration detected no missed injuries on follow-up.

Skeletal microcirculatory responses to hyperdynamic Escherichia coli sepsis in unanesthetized rats

To determine the microvascular site of vasodilation during hyperdynamic sepsis, we measured arteriolar and venular responses to live Escherichia coli bacteremia in the rat cremaster muscle by direct in vivo videomicroscopy. Our data indicate that cardiac output (by thermodilution) increased, systemic vascular resistance decreased, and a differential arteriolar response occurred, with constriction of large arterioles and dilation of small terminal arterioles. We conclude that dilation of small terminal arterioles in skeletal muscle could contribute to decreased systemic vascular resistance during hyperdynamic sepsis. This may be an appropriate response to increased oxygen demand or decreased tissue utilization of oxygen. Alternatively, small-arteriole dilation may be an inappropriate response and secondary to release of vasoactive inflammatory mediators. If the latter is true, there is a potential therapeutic role for selective manipulation of the tone of small terminal arterioles in hyperdynamic sepsis.

Determinants of outcome in pancreatic trauma

Pancreatic injury remains a major source of morbidity and mortality. A previous review of this injury from our department confirmed that the failure to recognize major ductal injuries was a key element in poor outcome of these patients. We evaluated a treatment protocol in 72 consecutive patients that categorized injuries into one of four grades: grade I, pancreatic contusion or minor hematoma with an intact capsule and no parenchymal injury; grade II, parenchymal injury without major ductal injury; grade III, parenchymal disruption with presumed ductal injury; and grade IV, severe crush injury. All grade I and most grade II injuries were treated by drainage alone; the grade III and IV injuries were treated by pancreatic resection. Fifty-seven patients survived longer than 24 hours. There were 23 grade I patients. There were only minimal pancreatic complications and no deaths in this group. Of 18 patients with parenchymal injuries (grade II), only one death occurred, which was due to an inaccurate estimation of the degree of injury and delay in proper treatment. Sixteen patients with grade III and IV injuries were treated by resection with only one death, although the complications rose with increasing severity of the pancreatic injury. The mortality rate from pancreatic causes was 3% (2/57), a reduction from that in a previous report from our institution (19%). Our present study confirms the use of a vigorous diagnostic approach to pancreatic injuries, stresses the value of recognition of major ductal injury, and supports the utility of a treatment protocol in which clinical decisions are based on the severity of the pancreatic injury encountered.

Epithelial lining methods in esophageal repair: a comparative study using pedicle flaps in cats

Methods of restoring luminal lining in repair of partial-circumferential esophageal defects were evaluated to compare incidences of strictures and fistulas and quality of surface lining. In cats 50% and 67% circumferential esophageal defects were repaired by direct closure or pedicle flaps (latissimus dorsi) that were unlined (lining restored by epithelialization from wound margins) or that carried lining of normal skin (myocutaneous flaps), skin grafts, or mucosal grafts. Repairs were evaluated for esophagocutaneous fistulas, luminal stricture, flap luminal surface area, and quality of epithelial surface 6 weeks after surgery. Direct closure of 50% circumferential defects was as satisfactory as any flap repair method. Direct closure of 67% circumferential defects caused high incidences of fistulas and strictures, which were lessened by flap reconstructions. Among flap lining methods, normal skin (myocutaneous flaps) gave the lowest incidence of fistulas and strictures and the highest surface quality, but a high incidence of skin paddle loss occurred in this model. Split-thickness epithelial grafts were nearly as satisfactory as myocutaneous flaps, and less lining loss occurred. Epithelialization of unlined flaps gave the poorest results since lining was thin and often incomplete, and wound contraction produced loss of surface area and strictures. The findings are discussed from a perspective of wound healing physiology, and implications for clinical application are presented.

Effects of tissue acidosis on skeletal muscle microcirculatory responses to hemorrhagic shock in unanesthetized rats

Arteriolar dilatation, loss of venous tone, and uptake of shed blood characterize decompensated hemorrhagic shock. The loss of compensatory constrictor responses to hemorrhage mainly occurs in the skeletal muscle microcirculation. Tissue acidosis may be an important mediator of this phenomenon. Using a decerebrate in vivo rat cremaster muscle preparation, we observed the microcirculatory responses to hemorrhagic hypotension with cremaster bath conditions of pH 7.4 and pH 7.0. Our data indicate that tissue acidosis attenuates constrictor responses of larger arterioles (100-170 micron) and venules to hemorrhagic hypotension but has no effect on the dilator responses of small arterioles (10-30 micron). We conclude that tissue acidosis contributes significantly to loss of arteriolar resistance and to decreased venous return in the decompensatory phase of hemorrhagic shock.

Effective palliation of malignant biliary duct obstruction

The efficacy of palliative biliary decompression by operative and percutaneous methods was evaluated in 106 patients with unresectable malignancies obstructing the biliary system. Seventy patients had operative and 36 had percutaneous decompression. Primary malignancies of the pancreas and bile ducts were most common. Percutaneous transhepatic decompression was achieved beyond the site of obstruction in 72% of patients. Overall hospital mortality was 25% for patients having percutaneous catheter decompression and 17% for those patients operated upon. Early postoperative death was significantly related to: (1) age greater than 70 years; (2) preadmission weight loss greater than 15 pounds; (3) prothrombin time prolonged more than 2.5 seconds; and (4) hepatic metastases. Major complications were encountered in 56% of survivors of percutaneous drainage and 36% of those surviving operation. Intubation of the bile ducts due to inability to bypass the obstruction at operation was associated with the highest mortality (50%) and morbidity (86%). Mean survival was 14 months after operation, compared to 5 months after percutaneous decompression. The authors concluded that percutaneous decompression of the biliary tree is useful palliative treatment for those patients with proximal biliary obstruction due to malignancy when estimated operative risk is high, but operative decompression offers most patients the opportunity for longer survival with lower ultimate mortality and morbidity.

Emergency pneumonectomy for penetrating and blunt trauma

Emergency pneumonectomy for penetrating and blunt trauma has an attendant high mortality. Patients with major lung injuries presenting with prolonged shock followed by control of bleeding, resuscitation with or without aortic cross-clamping and pneumonectomy have had uniformly unsatisfactory results. From 1972 to 1982, eight patients at the University of Louisville Hospital underwent emergency pneumonectomy. All patients underwent expeditious evaluation, resuscitation, and thoracotomy with pneumonectomy. Three patients died of exsanguination (2 patients had major associated intra-abdominal injuries). Three other patients died due to pulmonary edema and right ventricular failure 2 to 3 hours after hemorrhage had been controlled and intravascular volume restored. Aortic cross-clamping was employed in four patients due to persistent hypovolemia with 100 per cent mortality. Of the two surviving patients, one presented with stable blood pressure and had pneumonectomy for tracheobronchial disruption, while the other had pneumonectomy for tangential laceration of the lung at the hilum. Pulmonary edema and right ventricular failure were responsible for mortality following emergency pneumonectomy and control of hemorrhage and restoration of blood volume. The addition of aortic cross-clamping did not seem to alter survival and may, indeed, hinder therapy due to increased vascular afterload and increased heart failure and pulmonary edema.

Serum level monitoring of aminoglycoside antibiotics. Limitations in intensive care unit-related bacterial pneumonia

Serum aminoglycoside assays have been accepted as useful methods of enhancing therapeutic efficacy in the treatment of intensive care unit-acquired pneumonia and in avoiding aminoglycoside nephrotoxicity. We prospectively studied 68 surgical patients with normal renal function and gram-negative bacterial pneumonia who were treated with aminoglycosides. Serum levels indicated subtherapeutic levels in 47 patients and verified optimum levels in 13 patients. Toxic trough levels developed in six patients and, despite immediate dosage adjustment, five patients suffered nephrotoxicity. Six additional patients also had nephrotoxicity. Five of these patients never had toxic peak or trough levels and rising trough levels developed in one patient after serum creatinine levels began to rise. We conclude that routine monitoring of serum levels effectively detects subtherapeutic antibiotic levels. This modality is useful for optimizing dosage schedules, but does not serve to predict or avoid nephrotoxicity in critically ill surgical patients.