American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference definitions of the systemic inflammatory response syndrome and allied disorders in relation to critically injured patients

OBJECTIVES

To determine the frequency of the proposed definitions for the systemic inflammatory response syndrome (SIRS), sepsis and septic shock, and to further define severe SIRS and sterile shock as determined at 24 hrs of admission to an intensive care unit (ICU) in critically ill trauma patients without head injury, and their relationships to mechanism of injury, Acute Physiology and Chronic Health Evaluation (APACHE) II score, risk of death, Injury Severity Score (ISS), number of organ failures, and mortality rate.

DESIGN

Prospective, inception cohort analysis.

SETTING

Sixteen-bed surgical ICU in a teaching hospital.

PATIENTS

Four hundred fifty critically injured patients without associated head trauma. Penetrating trauma accounted for 70% (gunshot 202; stab 113) and nonpenetrating trauma for 30% (motor vehicle collision 103; blunt 32) of admissions. Three hundred ninety-four (88%) patients underwent surgical procedures.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Infective and noninfective insults were distinguished by the need for therapeutic or prophylactic antibiotics, respectively, based on an established antibiotic policy. Three hundred ninety-five (87.8%) patients fulfilled a definition of the SIRS criteria. The frequency of the definitive categories was SIRS 21.8%, sepsis 14.4%, severe SIRS 8.4%, severe sepsis 13.6%, sterile shock 9.3%, and septic shock 20.2%. Patients with penetrating trauma had a significantly higher frequency of sepsis, severe sepsis, and septic shock (p < .01). The APACHE II score, risk of death, and number of organ failures increased significantly in both infective and noninfective groups with increasing severity of the inflammatory response. Sterile shock was associated with a significantly higher APACHE II score (p < .02), risk of death (p < .01), and number of organ failures (p = .03) compared with septic shock. Only sterile shock was associated with a significantly higher ISS (p < .01). Organ system failure was significantly (p < .001) higher in nonsurvivors compared with survivors in all categories. The only significant (p < .001) difference in mortality rate was found between patients in shock and all other categories.

CONCLUSIONS

The current definitions of SIRS, sepsis, and related disorders in critically injured patients without head trauma show a significant association with physiologic deterioration and increasing organ dysfunction. The only significant association with mortality, however, is the presence of shock. The definitions require refinement, with the possible inclusion of more objective gradations of organ system failure, if they are to be used for stratifying severity of illness in seriously injured patients.

Selective conservative management of abdominal gunshot wounds: a prospective study

A prospective study of 111 patients with low velocity gunshot wounds of the abdomen was conducted to determine whether a policy of selective conservative management based on repeated physical examination is a safe form of treatment. Laparotomy was undertaken in 89 patients (80 per cent), seven of which were negative. Of the patients 22 (20 per cent), eight of whom were considered to have peritoneal penetration, underwent conservative management. None required delayed laparotomy. Eight patients (7 per cent) died, all deaths occurring in the positive laparotomy group. The incidence of significant intra-abdominal injury if the peritoneal cavity had been penetrated was 89 per cent. Selective conservative management may be applied safely to a limited group of patients with gunshot wounds of the abdomen.

Abdominal content containment: practicalities and outcome

BACKGROUND

Alternatives to fascial closure of the abdominal wall are increasingly used in critically ill patients. They pose practical and logistical problems in management which are described in this single-institution study.

METHODS

Between January 1994 and December 1997, 157 predominantly young male patients (100 trauma and 57 non-trauma) underwent temporary abdominal content containment (t-ACC) using plastic bags or polyglactin mesh. Indications for t-ACC were severe sepsis requiring reoperation, abdominal compartment syndrome, abdominal wall tissue loss or a combination of these. A total of 385 laparotomies were performed.

RESULTS

Two t-ACC procedures failed owing to technical error and two were complicated by enteric fistulas. Six patients underwent early definitive abdominal closure within 15 days. The remaining survivors had a protracted hospital stay (mean(s.d.) 44.6(26.6) days) and all developed incisional hernias which were challenging to repair. The overall mortality rate was 44 per cent.

CONCLUSION

Plastic bags were cheaper and as effective as polyglactin mesh for t-ACC. Survivors require a multidisciplinary approach in management, undergo a protracted hospital stay and later need complex incisional hernia repairs.

Intensive care unit morbidity and mortality from eclampsia: an evaluation of the Acute Physiology and Chronic Health Evaluation II score and the Glasgow Coma Scale score

OBJECTIVE

To determine the maternal morbidity and mortality in patients with eclampsia admitted to an intensive care unit (ICU), and to establish the efficacy of the Acute Physiology and Chronic Health Evaluation (APACHE) II score, the organ system failure score as defined by Knaus, and the Glasgow Coma Scale (GCS) score in predicting outcome.

DESIGN

Retrospective analysis of a 3.5-yr period.

SETTING

Surgical ICU in a university hospital.

PATIENTS

A total of 105 patients who were admitted with a diagnosis of eclampsia were studied.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

The data captured included the reason for admission, maternal age, gestational age, parity, number of seizures, duration of ICU stay, anticonvulsant therapy, drug therapy, GCS score, APACHE II score, and the occurrence of organ failure. Of the 126 patients with eclampsia who were admitted to the ICU, records of 105 patients (83%) were found. The overall mortality was 10.5% (n = 11). The mean age, gestation, parity, number of preadmission seizures, and duration of stay were similar in survivors and nonsurvivors. Although the APACHE II score was significantly higher in nonsurvivors, multiple logistic regression analysis suggested that the goodness-of-fit scores for GCS and APACHE II were similar (38.29 vs. 38.01). The GCS scores of survivors were significantly higher than those of nonsurvivors (10.61 vs. 5.0; p<.001). Respiratory failure was the most common organ failure in both groups. The mean number of organ failures was higher in nonsurvivors compared with survivors (2.9 vs. 1.3; p<.001). An occurrence of more than two organ failures that persisted for >48 hrs was invariably associated with a fatal outcome. Anticonvulsant therapy consisted of magnesium sulfate or phenytoin and a midazolam infusion. Only one patient (0.9%) had a seizure, and this occurred en route to the ICU. No seizures occurred after admission to the ICU.

CONCLUSIONS

The organ system failure score and the GCS score are good predictors of outcome in eclampsia. Apart from the GCS score, other variables in the APACHE II score are not valuable for outcome prediction. The low GCS score in nonsurvivors suggests that closer attention to the neurologic management may be beneficial. A prospective study is indicated to validate these findings.

Penetrating visceral injuries of the neck: results of a conservative management policy

A study of the conservative treatment of 109 patients with penetrating neck injuries was carried out over 3 years. Patients with clinical or radiological evidence of injury to the oesophagus or trachea were included in the study while nine patients with major vascular trauma were explored immediately and excluded. Three late vascular operations were performed. The remaining 106 patients were treated conservatively. There were two deaths, both from associated injuries. The remaining 104 patients were treated successfully with only three cases of minor wound sepsis. We conclude that oesophageal and tracheal injuries after stab injuries and low velocity gunshot wounds can be treated successfully by non-operative treatment.

Routine daily chest radiography is not indicated for ventilated patients in a surgical ICU

OBJECTIVE

To determine a) if clinical examination can accurately predict radiological change and b) if routine chest radiography is efficacious.

DESIGN

All mechanically ventilated patients admitted to the Surgical Intensive Care Unit over a 4-week period were enrolled into the study. A physical examination was undertaken by two clinicians to predict significant (radiographic features which alter management) and insignificant (radiographic features which do not alter management) changes. The radiographs were then reviewed by a radiologist who noted any changes from previous radiographs. The clinical findings were then correlated with the radiographical findings.

SETTING

The study was conducted in a 16-bedded Surgical Intensive Care Unit which admits approximately 800 patients per year. The majority of these patients require mechanical ventilation.

PATIENTS AND PARTICIPANTS

All patients who required mechanical ventilation were included. Thirty-four patients were studied. The patients were young adults admitted primarily following trauma with a low incidence of pre-existing disease.

INTERVENTIONS

There were no interventions in this study.

MEASUREMENTS AND RESULTS

One hundred sixty-four radiographs were evaluated. Both examiners were efficient in predicting significant changes (sensitivity of 93 and 97%), but less efficient at predicting insignificant changes (sensitivity of 74 and 70%). Two significant radiographical changes were missed on clinical examination: one catheter malposition and one pneumothorax, representing a yield from radiography of 1%. A 52% reduction in the number of radiographs would have resulted if the need for radiography had been determined by clinical examination.

CONCLUSIONS

Clinical examination can effectively predict the need for radiography. Routine chest radiography is, therefore, not indicated for ventilated patients in our Surgical Intensive Care Unit.

Does HIV status influence the outcome of patients admitted to a surgical intensive care unit? A prospective double blind study

OBJECTIVES

(a) To assess the impact of HIV status (HIV negative, HIV positive, AIDS) on the outcome of patients admitted to intensive care units for diseases unrelated to HIV; (b) to decide whether a positive test result for HIV should be a criterion for excluding patients from intensive care for diseases unrelated to HIV.

DESIGN

A prospective double blind study of all admissions over six months. HIV status was determined in all patients by enzyme linked immunosorbent assay (ELISA), immunofluorescence assay, western blotting, and flow cytometry. The ethics committee considered the clinical implications of the study important enough to waive patients' right to informed consent. Staff and patients were blinded to HIV results. On discharge patients could be advised of their HIV status if they wished.

SETTING

A 16 bed surgical intensive care unit.

SUBJECTS

All 267 men and 135 women admitted to the unit during the study period.

INTERVENTIONS

None.

MAIN OUTCOME MEASURES

APACHE II score (acute physiological, age, and chronic health evaluation), organ failure, septic shock, durations of intensive care unit and hospital stay, and intensive care unit and hospital mortality.

RESULTS

No patient had AIDS. 52 patients were tested positive for HIV and 350 patients were tested negative. The two groups were similar in sex distribution but differed significantly in age, incidence of organ failure (37 (71%) v 171 (49%) patients), and incidence of septic shock (20 (38%) v 54 (15%)). After adjustment for age there were no differences in intensive care unit or hospital mortality or in the durations of stay in the intensive care unit or hospital.

CONCLUSIONS

Morbidity was higher in HIV positive patients but there was no difference in mortality. In this patient population a positive HIV test result should not be a criterion for excluding a patient from intensive care.

The spectrum and outcome of burns at a regional hospital in South Africa

INTRODUCTION

Burns remain a major cause of morbidity and mortality in Southern Africa. The more vulnerable of our population, namely the urban poor, children and epileptics, are most often affected. This audit documents our experience with burns in a busy regional hospital in Southern Africa.

PATIENTS AND METHODS

A prospective data base was maintained from September 2006 to February 2008 of all burn wound patients admitted in Edendale hospital. Standard demographic data, detailed description of the burn, surgical intervention, outcome and length of stay are recorded. The size and depth of the burn, as well as the initial fluid management are also recorded.

RESULTS

A total of 450 patients were admitted. Two hundred and thirty-five were male. There were 203 burnt children with an average age of 3 years (range 6 months to 12 years). Average age for adults was 40 years (range 13-82 years). The average surface area burnt in children was 7.5% versus 23% in adults. Of those who died, the average surface area burnt was 54%. In adults the average burn depth was superficial in 30%, deep dermal in 20% and full thickness in 50%. The aetiology of the burn was flame 70%, hot water 25% and miscellaneous 5%. In children the breakdown of burn depth was superficial in 77%, deep dermal in 15% and full thickness in 8%. The aetiology was hot water 83%, fire 6%, electrical 6% and miscellaneous 5%. The last mentioned included hot oil or porridge (15), electrical (10), chemical (6), flash burns (8) and lightning (4). Fifty percent of adults were epileptic and had sustained their burn wound during a seizure. In this group, over 40% had previously sustained burns. Fifteen percent had a delayed presentation on an average of 11 days. Hospital stay averaged 68 days (3.5 days per percent burn: range 1-161 days). Two hundred and two (45%) patients required skin grafting. The average time from burn to graft was 51 days (range 12-138). There were 40 deaths (9%) with an average age of 50 years (range 6 months to 82 years) and an average total burn surface area of 50% (range 14-85%). Aetiology of the burn in the deaths was fire in 30, lightning 4 and hot water 6. Cause of death was burn wound sepsis in 38 and inadequate resuscitation in 2.

CONCLUSION

Young children and epileptics are particularly vulnerable to sustaining burns. Our hospital sees a large number of burns predominantly involving smaller surface areas. Patients with small burns have a prolonged hospital stay and delayed grafting due to a conservative surgical approach and lack of resources. Large burns are fatal in our hands.

Obstetric and gynaecological patients in an intensive care unit: a 1 year review

Management of the critically ill patient forms a significant proportion of obstetric and gynaecological (O & G) practice. There have however, been very few reports on the management of such patients in intensive care units (ICU). We review all O & G patients admitted to the surgical ICU at King Edward VIII Hospital, Durban, South Africa, and make recommendations regarding management of such patients. The medical records of all O & G patients admitted to the surgical ICU between the period January-December 1992 were analysed. Of all admissions to the ICU 13.6% (n = 122) were O & G patients. Eclampsia was the most common diagnosis accounting for 66% of all obstetric admissions. Of all eclamptics in the study period 24% were admitted to the ICU. The overall maternal mortality was 21%. O & G patients form a major workload of surgical ICUs and the majority of these patients are women with eclampsia. Management of such patients requires an understanding of the physiological changes of normal and abnormal pregnancies. Therefore, all large obstetrical units in developing countries should establish their own ICU in order that patient care, health personnel training and continuing health care education may be improved.

Undetected injuries: a preventable cause of increased morbidity and mortality

A prospective audit of trauma patients managed at the discretion of six different general surgical units was performed over a 6-month period. Eighteen patients were identified in whom diagnostic delay or injuries undetected at operation contributed to increased morbidity and mortality. Failure to perform investigations as indicated by the nature of the trauma was the main reason for delay in diagnosis in seven patients. Incomplete exploration at laparotomy resulted in seven undetected injuries, while unexplored retroperitoneal hematomas accounted for the remaining four. Fourteen patients (78%) required management in the intensive care unit. Eight patients died (44%) as a result of ongoing sepsis and multiple organ failure. Seven of the deaths occurred in patients in whom surgical treatment was inadequate. Delays in diagnosis and undetected injuries, although uncommon, are a readily preventable cause of phase 3 trauma deaths. Strict adherence to standard surgical protocols as employed in dedicated trauma care centers does much to reduce unnecessary morbidity and mortality.

Management of sympathetic overactivity in tetanus with epidural bupivacaine and sufentanil: experience with 11 patients

OBJECTIVE

To determine the efficacy and safety of epidural bupivacaine and sufentanil for the management of sympathetic overactivity in tetanus.

DESIGN

Retrospective case review.

SETTING

Sixteen-bed surgical intensive care unit in a tertiary care centre.

PATIENTS

All patients referred to the unit during a 63-month period with the diagnosis of tetanus were included in the study.

MEASUREMENTS AND MAIN RESULTS

All patients (n = 11) had severe tetanus and developed sympathetic overactivity, which was managed by epidural blockade. Three patients died, but there were no fatalities directly attributable to sympathetic overactivity. Before epidural blockade, the average difference between the mean maximum and mean minimum systolic blood pressures was 78 +/- 28 (so) mm Hg. After blockade, this difference was reduced to 38 +/- 15 (so) mm Hg (p < .0001). Similar significant reductions in diastolic blood pressure and heart rate were observed. The mean hourly infusion doses of bupivacaine and sufentanil were 17 mg and 21 microg, respectively. Midazolam was the principal adjunctive sedative agent and was used in all patients (mean dose, 9 mg/hr). Additional pharmacologic agents were necessary in two patients in whom epidural blockade alone was insufficient to control sympathetic overactivity. One patient developed renal failure and there were no instances of pneumothorax. One patient developed an epidural abscess of probable hematogenous origin, which was successfully treated without neurologic sequelae.

CONCLUSIONS

Epidural blockade is effective in controlling sympathetic overactivity and the associated complications (renal failure, cardiac injury, and sudden death). Although a serious complication occurred in one patient, the efficacy of the technique deserves further validation.

Management of penetrating cervicomediastinal venous trauma

OBJECTIVES

to evaluate the results of management of penetrating cervicomediastinal venous trauma.

DESIGN

retrospective study. Materials forty-nine consecutive patients with cervical and thoracic venous injuries treated at a tertiary hospital between 1991 and 1997. Method patients identified from a computerised database and data extracted from case records.

RESULTS

forty-five patients were male and the mean age was 25.3 years. Forty injuries were due to stabs and 9 to gunshots. 22 patients were shocked, 25 actively bleeding and 31 were anaemic. Veins injured were internal jugular in 25, subclavian in 15, brachiocephalic in 6, and superior vena cava in 3. Injured veins were ligated in 25 cases and repaired by lateral suture in 22. No complex repairs were performed. There were 8 perioperative deaths and 5 cases of transient postoperative oedema. Venous ligation was not associated with increased risk of postoperative oedema.

CONCLUSIONS

ligation is an acceptable form of treatment of cervicomediastinal venous injuries in the presence of haemodynamic instability, or where complex methods of repair would otherwise be necessary.

Prediction of acute renal failure following soft-tissue injury using the venous bicarbonate concentration

Sixty-four patients with soft-tissue injuries were studied prospectively to determine whether an initial venous bicarbonate concentration (VBC) of less than 17 mmol/L would predict the development of myoglobin-induced acute renal failure. The VBC was > 17 mmol/L in 59 patients, seven of whom had myoglobinuria. All recovered without renal complications. The remaining five patients all had VBC < 17 mmol/L and four had myoglobinuria. Acute renal failure developed in four patients (p < 0.001). The VBC on hospital arrival was the most accurate predictor of these patients' risk for the development of acute renal failure following soft-tissue injury.

Prediction of the risk of death by APACHE II scoring in critically ill trauma patients without head injury

The ability of the Acute Physiology And Chronic Health Evaluation (APACHE) II scoring system to predict outcome in 199 critically ill patients with trauma but without head injury was assessed prospectively over an 8-month period. Some 178 patients (89.4 per cent) underwent operation and 50 (25.1 per cent) died, 44 after operation and six without undergoing surgery. The mean(s.d.) APACHE II score was 8.0(5.2) for survivors and 14.5(5.5) for non-survivors (P < 0.001). In patients who underwent surgery the mean(s.d.) scores were 7.7(4.6) and 13.4(5.5) (P < 0.001) and for those managed without operation 11.1(7.2) and 14.7(6.3) (P = 0.31) in survivors and non-survivors respectively. The predicted risk of death and observed mortality rate were 5.1 and 25.1 per cent respectively for the entire group, 5 and 25 per cent for patients undergoing surgery, and 7 and 29 per cent for those not operated on. Although the APACHE II system correctly identified all survivors (specificity 100 per cent), it failed to predict death in any patient (sensitivity 0 per cent). The results suggest that this objective prognostic scoring system is not applicable to the patient with trauma who does not have concurrent head injury.

Delayed pneumothorax and haemothorax following observation for stab wounds of the chest

In a prospective study, 85 patients underwent simple observation for minor injuries of the pleural cavity as a result of penetrating injury of the chest. Seven patients (8 per cent) required delayed drainage, because of an increase in the size of the pneumothorax and/or haemothorax, but in all cases drainage was instituted within 24 hours. Haemopneumothorax is the most likely underlying disorder to require delayed drainage. Subcutaneous emphysema is of no value for identifying those cases likely to need delayed drainage. Twenty-four hours is adequate for the observation of patients with slight effects of penetrating wounds of the chest.

Unreliability of standard quantitative criteria in diagnostic peritoneal lavage performed for suspected penetrating abdominal stab wounds

Thirty-five patients with abdominal stab wounds in whom clinical examination was equivocal on 2 separate occasions underwent diagnostic peritoneal lavage (DPL) prior to laparotomy. The red and white blood cell counts (cells/mm3) of the lavage effluent were compared with the operative findings. There were 26 positive and 9 unnecessary laparotomies, the latter consisting of 4 negative and 5 non-therapeutic operations. Use of the standard quantitative criteria for red cells in DPL failed to identify significant injury in eight patients (31%), while the standard white cell count missed six injuries (23%). Their combined use resulted in three missed injuries (12%). Two false-positive results occurred using the red cell count alone and four using the white cell count alone, producing a combined false-positive result in four patients (11%). Reducing the cell threshold level to exclude missed injuries would increase dramatically the rate of unnecessary laparotomies. Although the standard quantitative criteria for DPL are superior to clinical assessment in patients with equivocal findings, their use in penetrating trauma does not achieve the same diagnostic accuracy as in blunt abdominal trauma.

Penetrating injuries of the pleural cavity

Two hundred and fifty one cases of penetrating wounds of the chest were studied prospectively. Clinical evidence is presented to show that: basal intercostal drains are adequate to remove both air and fluid from within the pleural cavity; frequent chest radiographs are unnecessary and intercostal drains may be removed on clinical grounds alone; long term antibiotic prophylaxis is unnecessary; eight per cent of those undergoing initial observation will develop a delayed haemothorax or pneumothorax of sufficient size to require drainage; subcutaneous emphysema is of no prognostic significance in the symptomless patient with minimal intrapleural damage on admission; and outpatient follow up is not required.

Validation of an outcome prediction model for critically ill trauma patients without head injury

BACKGROUND

The Acute Physiology and Chronic Health Evaluation (APACHE) II system is inaccurate in predicting the risk of death in trauma patients, especially those without head injury. Using multivariate analysis of the APACHE II system in a development set, a new predictive equation was modeled. The four variables that were independently associated with mortality were PaO2/FiO2 ratio, mean arterial pressure, temperature, and the need for inotropic support. This model was tested prospectively in an independent validation set of 300 patients.

METHODS

Risk of death was calculated using the APACHE II system with the diagnostic category of multiple trauma and weighting for operative intervention as required. The new model was similarly assessed using the four predictor variables and their beta-coefficients for each mechanism of injury and the entire group. The predicted risk of death derived by both models was compared with the observed mortality rate. Discrimination was calculated using a 2 x 2 decision matrix with a decision threshold of r = 0.5 and receiver operating characteristic curves. Calibration was assessed graphically and by statistical correlation.

RESULTS

The observed mortality rate was 28.3% and the predicted mortality risk was 27.4% for the model and 6.26% for APACHE II. The sensitivity and specificity of the model were 58.8 and 90.7%, and the sensitivity and specificity of APACHE II were 1.2 and 100%. The areas under the receiver operating characteristic curves were 0.84 and 0.78 for the model and the APACHE II system, respectively. Calibration of the model was superior within all deciles of risk (model, R2 = 0.93, p < 0.001; APACHE II, R2 = 0.82, p = 0.02).

CONCLUSION

The model accurately predicted the risk of death for the entire group. It is superior to the APACHE II system and is the highest reported sensitivity for 24-hour intensive care unit predictive models that have been applied to the critically injured.

Protein electrophoresis of diagnostic peritoneal lavage in penetrating abdominal trauma: a marker of significant injury

Protein electrophoresis of diagnostic peritoneal lavage fluid has been shown in experimental work to be a marker of hollow viscus injury. In order to ascertain the value of this investigation in penetrating abdominal trauma, we conducted a prospective study, over a 4-month period, involving 35 patients with stab wounds to the abdomen whose initial clinical examination was equivocal. The results of electrophoresis were compared with the standard quantitative criteria for diagnostic lavage and the operative findings. Protein electrophoresis was found to be a more sensitive and specific marker of significant intra-abdominal injury than the standard quantitative criteria.

Evaluation of diagnostic peritoneal lavage in suspected penetrating abdominal stab wounds using a dipstick technique

The total protein content (g/l) and white blood cell count (cells/mm3) diagnostic peritoneal lavage was assessed using a urine dipstick in 46 patients with suspected penetrating abdominal stab wounds and equivocal physical examination. Those patients with a protein content greater than or equal to 1 g/l and white blood cell count of greater than 500 cells/mm3 were submitted to laparotomy while those with lower values underwent observation and repeat physical examination. In all, 26 patients had a positive lavage and significant injuries were found in 23 of these. Of 18 patients with a negative lavage, 17 were managed successfully without operation while one patient died from complications related to central venous catheterization. In two patients the lavage results were equivocal. One underwent a negative laparotomy and the remaining patient recovered uneventfully. The test has a 100 per cent sensitivity and 86 per cent specificity and provides an accurate, cheap, and rapid means of diagnosis of intra-abdominal injury in penetrating trauma.

Pancreatic pseudocyst haemorrhage presenting as a bleeding duodenal ulcer

We present a case of upper gastrointestinal haemorrhage where the preoperative endoscopic findings suggested a duodenal ulcer as the cause. Although at operation this proved to be the site of bleeding, the source was found to be the splenic artery in the base of a pancreatic pseudocyst.