Surgeons' participation in continuing medical education: is it evidence-based?

BACKGROUND: Interactive forms of continuing medical education (CME) are more likely to improve clinical practice than traditional, passive approaches. This study investigated CME participation and preferences among surgeons. METHOD: Questionnaire survey of surgeons in New South Wales, Australia. RESULTS: On average, respondents (n = 418, 77% response rate) committed 364 hours (interquartile range 228-512 hours) to CME per year. Surgeons working at tertiary referral teaching hospitals were twice as likely as those working in other types of hospital to report spending more than 12 hours per month on CME (OR 2.1, 95% CI: 1.4-3.1). Overall, reading accounted for 17% of CME time and attending conferences a further 12%. Clinical audit accounted for significantly less CME time (3.5%) (both P < 0.001). Conferences were considered the single most useful form of CME by 28% (95% CI: 24-33%). Over half (55%, 95% CI: 50-59%) ranked reading as 1 of the 3 most useful types of CME, whereas significantly fewer so ranked clinical audit (6%, 95% CI: 4-9%) (chi2 = 230.8, 1 d.f., P < 0.001). CONCLUSION: Australian surgeons commit a considerable amount of time to CME, but much of this time is spent in passive educational activities. Development of acceptable and effective CME programmes will benefit both surgeons and their patients.

Pancreaticojejunostomy for chronic pancreatitis.

The treatment of chronic pancreatitis commonly yields disappointing results. Patients with chronic pancreatitis and a dilated pancreatic duct can be treated by longitudinal pancreaticojejunostomy. In order to evaluate the procedure, 20 patients undergoing pancreaticojejunostomy were followed for a median time of more than 5 years. Their clinical characteristics and outcomes have been compared with a group of 43 patients with chronic pancreatitis and small pancreatic ducts. There were no differences between the two groups in the major epidemiological parameters, except that calcification in the gland was more frequently noted in those with large ducts. The operation of longitudinal pancreaticojejunostomy could be accomplished with an acceptable morbidity. There was one death in the postoperative period. Seventy-six per cent of patients were found to have benefited clinically at five years, compared with 48% of those with small duct disease. This difference was statistically significant. Patients who benefited were defined by four factors; they were carrying out their usual occupation at the time of surgery, they were not narcotic dependent at the time of surgery, they had a pancreatic duct width greater than 7 mm and, they had totally abstained from alcohol from before the operation to the time of follow-up. Longitudinal pancreaticojejunostomy probably remains the best surgical treatment for suitable patients with chronic pancreatitis. The operation should only be performed when the pancreatic duct is greater than 7 mm in width. In such patients the operation produces considerable improvement of pain with minimal metabolic disturbance.

Non-operative management of blunt liver injuries.

Over a 10-year period 281 patients with liver injuries following blunt abdominal trauma were seen at Westmead Hospital. A decision to treat the liver injury without an operation was made in 55 (20 per cent) of the 281 patients. Two of these 55 patients required a subsequent laparotomy, one for continued bleeding and one for a bile leak. No patient died from a missed intra-abdominal or retroperitoneal injury or from concealed haemorrhage. These patients were compared with 181 patients with liver injuries who were operated on. The two groups of patients could be distinguished in terms of their clinical status on arrival at the hospital, the magnitude of their injury and the sequelae of the injury. Non-operative management proved a safe alternative to an operation in appropriate patients, but a laparotomy to exclude unsuspected intra-abdominal injuries is associated with little significant morbidity and offers some reassurance in managing patients with multiple injuries.

Impact of the CUSA and operative ultrasound on hepatic resection.

New technologies have been developed for liver surgery, and, like all new technologies, they have a glamour which makes them seem desirable. There is an understanding abroad that they make liver surgery easier and open up the field to those without special training. But there is no proof that the new devices are in any way cost-effective, and certainly no proof that liver surgery has become safer since their advent. Fifty consecutive elective liver resections have been studied, almost half performed with the aid of the ultrasonic dissector and aspirator and diagnostic intraoperative ultrasound. There was no mortality in the whole group, but a 24% morbidity. Operative diagnostic ultrasound was thought to allow more precise planning of surgery. Its use was not associated with any increase in operative time, nor was there any increase in postoperative morbidity. The ultrasonic dissector and aspirator improved technique, reflected in a lower blood loss for each case, in fewer transfusions required, in a shorter postoperative hospital stay and in an ability to achieve these benefits in older patients. Neither device could be said to offer an entree to instant liver surgery. The use of the two devices apparently offered savings measured by a fall in the median postoperative hospital stay of 4.5 days, by a saving of 700 mls in median blood requirement and by a fall in transfusion rate from 64% to 9%.

Post-traumatic bile fistulae.

Bile leakage after liver injury has been reported to be a complication associated with significant mortality. In a prospective study of 306 patients with liver injuries 13 (4%) developed a bile leak. There appear to be two groups of patients, those with a major bile duct injury (three patients) and those with a more peripheral injury (ten patients). Bile leaks may present either as bile peritonitis (six patients) or as a bile leak through a drain site (seven patients). All but one bile leak closed spontaneously at a median of 33 days (range, 3-110 days). There was no mortality but patients incurred a median of four complications each (range, 1-5). This may reflect the severity of the initial injury rather than the bile leak. Respiratory failure and/or infection and sepsis were the most common complications whilst abnormalities of liver function, although common, were seldom of clinical significance.

The accuracy of the first chest X-ray in the trauma patient.

Chest X-rays (CXR) are important in assessing chest trauma. This study assessed the accuracy of the first CXR in diagnosing moderate to severe chest injuries. One hundred patients presenting during 1985 and identified in the prospective Westmead Trauma Registry as having a chest injury with Abbreviated Injury Scale (AIS) scores of 3 or greater were included (males:females = 4.9:1; age range, 5-74 years). The officially reported interpretations of the first CXR were compared to the chest injuries finally diagnosed. In 19 of 77 survivors and 8 of 17 patients who had a CXR before dying, significant abnormalities (most commonly pneumothorax, haemothorax, spinal and sternal injuries) were not diagnosed on CXR. Although erect CXR with postero-anterior (PA) and lateral projections are said to show accurately most major intrathoracic pathology, CXR of trauma patients are often taken supine, suboptimally, and soon after arrival, whereas chest injury may take time to become apparent. The limitations of mobile CXR must be appreciated, and more formal CXR (antero-posterior and lateral) are advisable at a later stage if the clinical situation permits.

The role of hepatic resection in the management of blunt liver trauma.

Forty-two (14%) of 306 patients with liver injuries presenting to Westmead Hospital over a 10-year period required hepatic resection as their definitive treatment. Two types of resection were performed: Resectional debridement utilized the plane of injury as the line of resection while anatomical resection utilized anatomical planes. Resectional debridement was used in 35 patients. In 29, the major technical problem was bleeding and 21 of these patients had associated hepatic vein injuries. In 5, the major problem was devitalized parenchyma, and, in 1, it was an intrahepatic bile duct injury. Anatomical resection was performed in 7 patients: 3 with bleeding, 2 with devitalized parenchyma, and 2 with intrahepatic bile duct injuries. Overall, 15 patients died (36%). The most common cause of death was bleeding in 9 of the 15 patients. Survivors spent a median of 32 days in hospital (range: 11-162 days) and sustained a median of 2 complications (range: 0-6). The most common complications were respiratory infection and/or failure, coagulopathy, and sepsis. Resection successfully addressed bleeding, devitalized parenchyma, and intrahepatic bile duct injuries with an acceptable mortality in critically ill patients who would otherwise have died.

Hepatic incidentaloma: a modern problem.

As clinical skills give way to increased reliance on organ imaging, a new clinical problem is identified--the hepatic "incidentaloma." This may be defined as an unexpected solid filling defect in the liver of a well patient. Thirty-six such lesions have been seen in one practice over a period of 36 months. Twenty-nine (81%) were benign: 24 (67%) nonneoplastic conditions 5 (14%) benign tumors. The remaining 7 (19%) were malignant: 5 secondary tumors and 2 primary tumors. Patients with physical signs of liver mass or enlargement were more likely to harbor malignancies. An elevated serum alkaline phosphatase (SAP) was suggestive of malignancy. Hepatic hemangioma was the most common single diagnosis (20 patients, 56%). We propose a regimen of investigation which should allow diagnosis to be reached in about one-half of these patients without admission to the hospital. The rest will need at least a short hospital admission for angiography and fine-needle aspiration biopsy.

PIP: Hepatic incidentaloma, a new entity--an unexpected solid filling defect in the liver of a well patient discovered incidentally by scanning--was characterized in a series of 36 patients. These undiagnosed solid liver masses were referred between 1985-1988 after ultrasound or CAT scans. There were 27 women and 9 men with a median age of 44.5. 67% had nonneoplastic lesions, 20 of them hemangiomas, 14 in women. 19% had malignant lesions, 5 with metastatic carcinoma of colorectal origin, and 2 women with primary hepatocellular carcinomas. 13 women had taken oral contraceptives for 3 months in the last 5 years. All of 4 of the patients with hepatic cell adenoma, and the patient with focal nodular hyperplasia had taken orals, while only 8 of the remaining patients had done so. 25% of the group had diagnoses not known to be linked to oral contraceptives. Other than those with malignancies who had a poor prognosis, the patients with hepatic cell adenoma and leiomyoma must have the lesion resected because of the risk of progression to malignancy.

Hepatic venous injury after blunt abdominal trauma.

Data on 306 liver injuries were collected prospectively during a 10-year period. Of the 306 injuries, 281 were the result of blunt trauma. Major hepatic venous injuries were encountered in 41 of 306 patients (13%). Blunt trauma was responsible for 39 of the 41 venous injuries. Twenty-five (61%) of these patients died compared with an overall mortality rate of 31% (chi 2; p less than 0.001). Twenty-two of the 25 deaths were caused by blood loss. Two patterns of hepatic venous injury appeared to predominate: avulsion of the trunk of the right hepatic vein from the inferior vena cava and avulsion of the upper branch of the right hepatic vein. The trunk injury was seen in 15 patients, 12 of whom died. The branch injury was seen in 13 patients, only 4 of whom died (Fisher's exact test; p = 0.006). In what appears to be the largest series of blunt hepatic venous injuries published, the injuries have been classified according to the anatomic site of the injury. Such a classification correlates with prognosis. Differences between blunt and penetrating hepatic venous trauma have been discussed.

Obstructive jaundice in chronic pancreatitis.

Significant obstructive jaundice in chronic pancreatitis is generally considered to be rare. Eleven of 57 consecutive patients with proven chronic pancreatitis have developed significant obstructive jaundice of more than transient duration. Eight presented as jaundice complicating known pancreatitis and three as jaundice of unknown cause. Life table analysis showed a steady rise in the risk of developing jaundice up to the end of 10 years from the onset of chronic pancreatitis. Jaundice was found to occur in the presence of more "destructive" disease, and jaundiced patients had a higher incidence of pancreatic calcification, diabetes and malabsorption at the time of presentation with jaundice. Obstructive jaundice caused by chronic pancreatitis was found to carry a good prognosis for jaundice, for pain and for life. Only one of the 11 patients died in hospital. It is important to distinguish chronic pancreatitis from cancer in these patients. Pre-operative and intra-operative cytology have been helpful. Stent insertion is not an appropriate method of treatment for these patients because of the benign nature of the disease and the possibility of exacerbating the pancreatitis. It is important to be aware of another form of "malignant masquerade" causing obstructive jaundice.

Percutaneous feeding gastrostomy.

Percutaneous endoscopic gastrotomy tubes have been inserted in 50 patients with no mortality and minimal morbidity. The principal indications for their insertion were an inability to swallow (39 patients), a need for nutritional support (eight patients) and venting (three patients). The most common complication was peristomal wound infection (14 patients), the incidence of which was reduced significantly by the use of single-dose prophylactic antibiotic therapy. Percutaneous gastrostomies provide adequate nutrition with minimal morbidity in an acceptable way to both patients and those who care for them.

Clinical and histological sequelae of Roux-en-Y diversion.

Nine post-Polya gastrectomy patients with symptoms of reflux gastritis were studied immediately before and 6 months after Roux-en-Y diversion with a 60 cm jejunal loop. Endoscopy, histology, HIDA radionuclide studies and acid secretory studies were performed to determine the effect of Roux-en-Y diversion. Symptomatic results were pleasing in that all patients were graded as Visick I or II. There was a statistically significant improvement in endoscopic evidence of mucosal damage (median score of 6 reduced to median of 1) and visible bile reflux (median score of 10.5 reduced to 0). Foveolar hyperplasia, superficial chronic gastritis and histological precursors of malignancy all improved. There was a significant decrease in the reflux of 99mTc-labelled HIDA into the gastric remnant while acid secretion remained unchanged. Roux-en-Y diversion with a 60 cm jejunal loop provides not only symptomatic relief but scintigraphic, endoscopic and histological evidence of pancreatobiliary diversion.

Perihepatic packing: its role in the management of liver trauma.

Perihepatic packing was used in 25 of 197 (12.7%) patients presenting with liver trauma to Westmead Hospital over an 8 year period. Packing was used either to provide temporary haemostasis prior to transfer or as part of a definitive treatment plan at this hospital. Thirteen patients were packed prior to transfer. Only two were unstable on arrival, one of whom died. They were compared with 18 'comparison' patients with liver injuries of similar severity. In this group 10 were unstable on arrival (P = 0.027), nine of whom died (P = 0.015). Packing was used as part of a definitive treatment plan at Westmead on 17 occasions. Four patients were coagulopathic and five had also been packed prior to arrival. Eight of this group died. Packing is a convenient and safe way of controlling major hepatic haemorrhage prior to transfer to a tertiary referral centre. It may also be part of a definitive treatment plan to control hepatic bleeding especially as many patients arrive with a coagulopathy or develop a coagulopathy during the course of surgery to control bleeding. Packing will control haemorrhage until the coagulopathy has been corrected.

The management of injuries--a review of deaths in hospital.

A prospective review was undertaken of the management of 111 consecutive patients who died in hospital after admission for treatment of injuries. A standard set of data relating to each patient was reviewed by each member of a trauma death audit committee and then by the whole committee. Autopsy reports were available on all patients. Conclusions were drawn concerning defective aspects of patient management and possible avoidance of each death. Injury severity was assessed using the Trauma Score (TS) and Injury Severity Score (ISS). The possibly avoidable death (PAD) rate was 17%. The most common defects in management were related to inadequate fluid resuscitation and delays in definitive management. The greatest contributions to the PAD rate were from inadequate fluid resuscitation, delays and inadequate perception of the severity of injuries or significance of clinical deterioration. Increasing age was related to a higher frequency of PAD. PAD rate in the presence of severe head injury was 8%, but was 63% in the absence of a severe head injury. It is concluded that review of all trauma deaths is an achievable, beneficial and essential part of a hospital-based integrated trauma service. TS and ISS are not sufficiently sensitive to justify their use in selecting deaths for review. Improved blood volume replacement, earlier and more direct management and supervision by senior specialist staff, and elimination of causes of delay in patient management should all decrease the death rate from injuries particularly in patients without severe head injury.

Hepatocellular carcinoma in western Sydney.

All the cases of proven hepatocellular carcinoma seen at Westmead Hospital, Sydney between January 1980 and the end of 1987 were reviewed. Hepatitis B infection was the major predisposing condition. Six patients had taken significant doses of sex steroids. Seventeen of the patients were cirrhotic at the time of diagnosis and in seven of these there was a significant history of alcohol abuse. AFP was elevated in only 15 of the 34 patients. Multiple regression analysis revealed that the single, independent determinant of a raised AFP level was found to be presence of Hepatitis B infection. Resection was possible in 10 patients. In the last ten months, seven patients have been treated by embolisation of the tumour with Adriamycin bonded to lipidol. Survival was influenced by the presence or absence of cirrhosis but not by evidence of Hepatitis B infection. The prognosis for patients with hepatocellular carcinoma in Australia is as dismal as it is in any other country. Although a rare tumour its incidence may well increase as the community now contains relatively greater numbers of immigrants from areas where the risk of developing a hepatocellular carcinoma is higher and because of the number of drug addicts who are frequently exposed to Hepatitis B infection. With the exception of patients with Hepatitis B infection, screening with AFP holds little promise in the Caucasian community.

Parenteral aspirin and sodium salicylate are equally injurious to the rat gastric mucosa.

The effects of parenteral aspirin (ASA) or sodium salicylate (SA) on the gastric mucosa were investigated in anesthetized pylorus-ligated rats 3 h after a bolus intravenous injection of ASA or SA, 150 mg/kg, or NaCl (control). Aspirin or SA produced similar extensive gross mucosal hemorrhagic lesions and similar microscopic damage in the presence of luminal acid (luminal pH 1.3 +/- 0.05). Neither ASA nor SA produced gastric mucosal injury with intragastric instillation of saline (luminal pH 3.7 +/- 0.5). Pretreatment for 1 h with luminal or subcutaneous 16,16-dimethyl prostaglandin E2 completely prevented the formation of red streaks in ASA-treated rats but not in SA-treated rats, although prostaglandin E2 pretreatment significantly reduced the gross lesion area in SA-treated rats (p less than 0.05). We conclude the following: (a) Intravenous SA is as damaging as intravenous ASA as long as luminal acid is present. (b) 16,16-Dimethyl prostaglandin E2 completely protected the gastric mucosa from injury by intravenous ASA, and to a lesser extent by intravenous SA. (c) In view of the damaging effects of SA on the gastric mucosa and the rapid conversion of ASA to SA, the mechanism of the gastric mucosal injury by intravenous ASA is much more complex than simple inhibition of endogenous prostaglandin synthesis.

Chloride transport of frog gastric fundus: effects of omeprazole.

Omeprazole (10(-4) M) inhibited H+ secretion and increased potential difference (PD), resistance, and short-circuit current (Isc) in chambered bullfrog gastric mucosa, but the electrical changes developed only in tissues previously exposed to histamine. Net chloride transport (JnetCl) did not change after omeprazole under short-circuited conditions, and Isc increased to become equal to JnetCl. Under open-circuit conditions, JnetCl was reduced by 38%, the decrement attributable to the concomitant increase in PD, as evidenced by a linear relationship between JnetCl and PD in omeprazole-treated mucosae clamped to different PD (0-45 mV). The effect of omeprazole on PD and Isc could be blocked by metiamide and was absent in spontaneously resting tissues. HEPES nutrient solutions did not alter the electrical response or Cl- transport after omeprazole. In Na+-free solutions, omeprazole induced only a transient rise in PD and Isc. We conclude that omeprazole uncouples H+ and Cl- secretion. This Cl- secretion is electrogenic and dependent upon stimulation by histamine. Both Na+ and HCO3- seem to be involved in movement of Cl- across the basolateral membrane.