Pyogenic hepatic abscess. Changing trends over 42 years.

OBJECTIVE: The authors document changes in the etiology, diagnosis, bacteriology, treatment, and outcome of patients with pyogenic hepatic abscesses over the past 4 decades. SUMMARY BACKGROUND DATA: Pyogenic hepatic abscess is a highly lethal problem. Over the past 2 decades, new roentgenographic methods, such as ultrasound, computed tomographic scanning, direct cholangiography, guided aspiration, and percutaneous drainage, have altered both the diagnosis and treatment of these patients. A more aggressive approach to the management of hepatobiliary and pancreatic neoplasms also has resulted in an increased incidence of this problem METHODS: The records of 233 patients with pyogenic liver abscesses managed over a 42-year period were reviewed. Patients treated from 1952 to 1972 (n = 80) were compared with those seen from 1973 to 1993 (n = 153). RESULTS: From 1973 to 1993, the incidence increased from 13 to 20 per 100,000 hospital admissions (p < 0.01. Patients managed from 1973 to 1993 were more likely (p < 0.01) to have an underlying malignancy (52% vs. 28%) with most of these (81%) being a hepatobiliary or pancreatic cancer. The 1973 to 1993 patients were more likely (p < 0.05) to be infected with streptococcal (53% vs. 30%) or Pseudomonas (30% vs. 9%) species or to have mixed bacterial and fungal 26% vs. 1%) infections. The recent patients also were more likely (p < 0.05) to be managed by percutaneous abscess drainage (45% vs. 0%). Despite having more underlying problems, overall mortality decreased significantly (p < 0.01) from 65% (in 1952 to 1972 period) to 31% (in 1973 to 1993 period). The reduction was greatest for patients with multiple abscesses (88% vs. 44%; p < 0.05) with either a malignant or a benign biliary etiology (90% vs. 38%; p < 0.05). Mortality was increased (p < 0.02) in patients with mixed bacterial and fungal abscesses (50%). From 1973 to 1993, mortality was lower (p = 0.19) with open surgical as opposed to percutaneous abscess drainage (14% vs. 26%). CONCLUSIONS: Significant changes have occurred in the etiology, diagnosis, bacteriology, treatment, and outcome patients with pyogenic hepatic abscesses over the past 4 decades. However, mortality remains high, and proper management continues to be a challenge. Appropriate systemic antibiotics and fungal agents as well as adequate surgical, percutaneous, or biliary drainage are required for the best results.

Training of care givers for our aging population.

We cannot rely on geriatricians, internists, and family practitioners alone in the medical community to provide all of the geriatric care. Even though there are alternatives to the use of specialists, we cannot afford to ignore the largest group of current physician trainees who will provide a great deal of geriatric medical care in the future. We need to help make the basic principles of geriatric care part of every training program for every resident, whether in general or specialty programs.

On the supply of physicians.

The congressionally created Council of Graduate Medical Education is conducting a study to assess the adequacy of the current and future supply of physicians in six specialties. The study is revising and updating the needs-based method used by the Graduate Medical Education National Advisory Committee and using this method to project the requirements for physicians for the next two decades. Projections on the supply of physicians will be based on the latest data from the federal government's Bureau of Health Profession's health manpower model. Previous physician manpower studies are reviewed to provide a historic perspective and a frame of reference for the variables, data sources, and assumptions being considered by the Council of Graduate Medical Education study.

Preventing postoperative acute bleeding of the upper part of the gastrointestinal tract.

Two hundred and ninety-eight critically ill patients at risk for the development of postoperative stress ulcers and bleeding were randomized into three groups. The first group comprised 85 patients who received meciadanol, a new bioflavonoid, 500 milligrams every six hours through a nasograstric tube; the second group comprised 100 patients who received sucralfate (crushed tablets), 1,000 milligrams every six hours through a nasogastric tube, and the third group comprised 113 patients who received an antacid (Maalox [magnesium aluminum hydroxide gel]) through a nasogastric tube at an initial dose of 15 milliliters every hour. The gastric pH was measured hourly and titrated to a pH greater than or equal to 4.0 in patients in the group receiving the antacid. The gastric pH was measured every two hours in the other two groups. Bleeding in the upper part of the gastrointestinal tract was determined visually (frank blood in gastric contents) or by guaiac testing. Bleeding occurred in seven patients receiving meciadanol, nine receiving sucralfate and six receiving the antacid. The difference in rates of bleeding was not statistically significant. Correlation between the severity of illness index and the development of bleeding was poor, at least in the low and intermediate index range. In contrast, there was a strong correlation between the age of the patient and the development of bleeding. Only one patient younger than 50 years had bleeding develop. Apparently, meciadanol exerts its action by a mechanism other than pH control. It may, therefore, fill an important gap in the ability to prevent postoperative stress ulcers and bleeding.

Closed-suction versus Penrose drainage after cholecystectomy. A prospective, randomized evaluation.

Closed-suction drainage was compared prospectively to open, passive drainage (Penrose drains) in 128 patients undergoing cholecystectomy. Patients were randomized at the time of operation to receive either closed-suction drains (Group I, 67 patients) or Penrose drains (Group II, 61 patients). The preoperative clinical parameters of the two groups were similar. The patients in Group I when compared with those in Group II had a shorter duration of drainage (3.3 days and 4.1 days, respectively, p less than 0.01), a lesser volume of drainage in the first 48 hours postoperatively (78 ml and 132 ml, respectively, p less than 0.001), a decreased incidence of fever on the night of operation (24 of 67 patients and 39 of 61 patients, respectively, p less than 0.05) and on the first postoperative day (26 of 67 patients and 32 of 61 patients, respectively, p less than 0.05), and a lower leukocyte count on the first postoperative day (12,000 cells/mm3 and 14,100 cells/mm3, respectively, 0.05 less than p less than 0.1). Patients in Group I tended to have a lower rate of wound infection (1 of 67 patients versus 5 of 61 patients in Group II, 0.05 less than p less than 0.1) and had a much lower incidence of drain site tenderness (8 of 67 patients in Group I versus 24 of 61 patients in Group II, p less than 0.05). This study demonstrates the superiority of closed-suction drains over open, passive drains after cholecystectomy.

Experimental pancreatic allotransplantation in large animals. The role of donor kidney and cyclosporine in modifying rejection.

Experiments were carried out in outbred dogs and pigs to evaluate the relative immunogenicity of pancreatic islets and segmental pancreas grafts, and whether these could be ameliorated by transplanting a kidney simultaneously from the same donor animal. Various immunosuppressive regimens were also studied. Pancreatic islet allografts never normalized blood glucose in totally pancreatectomized recipients despite the use of cyclosporine (CsA) in high doses (40 mg/kg/day) and the simultaneous transplantation of a kidney from the same donor. These grafts which never "took" contrast sharply with the experience of pancreatic islet autografts prepared in the same way and inoculated into the spleen, which in all nine instances normalized blood glucose in pancreatectomized recipients. Segmental transplants were performed in swine with duct drainage into the jejunum. Totally pancreatectomized pigs died at 7.8 +/- 1.0 days. In recipients suppressed with low-dose azathioprine (Az) and prednisone (Pred) pancreas grafts alone were rejected in 12.9 +/- 10 days. Synchronous pancreas and kidney transplants treated similarly extended the mean survival of pancreatic grafts to 20 +/- 10 days--which, however, was not significant (P less than 0.1 greater than 0.05). Mean survival time of pancreatic grafts in recipients receiving CsA at 20 mg/kg/day and prednisone 1 mg/kg/day was 14 +/- 6.3 days. The combination of CsA 20 mg/kg/day, Az 2 mg/kg/day, and Pred 1 mg/kg/day prolonged the mean survival time to 39.8 +/- 22 days. These results allow us to conclude that: crude preparations of islet tissue invariably capable of normalizing blood sugar at day 4 when used as autografts failed to "take" despite the existence of alternative sources of antigen present in a well vascularized kidney from the same donor, and despite very high dosages of CsA; triple immunosuppressive therapy had synergistic effects on pancreatic allograft survival; and simultaneous transplantation of kidney and pancreas had little effect on survival times of the pancreas or the kidney.

Covert splenic abscess: a continuing challenge.

Although splenic abscess is a rare cause of intra-abdominal sepsis, the mortality rate remains high especially in patients with silent or covert lesions. The clinical presentation and course of five patients with overt splenic abscess and seven patients with covert splenic abscess seen during a thirty year period were analyzed. Average age of patients with overt lesions was 44.5 years. Direct extension from a contiguous source, hematogenous spread from a distant site (metastatic) and trauma comprised the known etiologies. Clinical features of localized left upper quadrant sepsis were commonly present but only one patient exhibited multiple organ failure. The clinical diagnosis was established preoperatively in four patients (80%) and all underwent splenectomy without mortality. All resected spleens contained solitary abscesses. In contrast, patients with covert lesions tended to be older (average age 56.1 years), uniformly exhibited multiple organ failure and rarely demonstrated local clinical findings of left upper quadrant sepsis. Trauma was a less common etiology than metastatic infection and direct extension. Four patients died without operation. Three patients underwent exploration for unrelated reasons, but the diagnosis of splenic abscess was made intraoperatively in only one patient. Mortality among patients with covert lesions was 86%. Multiple splenic abscesses were demonstrated in all patients with covert lesions. Splenic abscess presents as a spectrum of clinical disease. Solitary lesions can be readily diagnosed and treated by splenectomy. Multiple abscesses are usually covert, associated with multiple organ failure and highly lethal. The role of splenectomy in patients with covert lesions remains unknown.

The effect of vascular fellowships on general surgical residency training.

To answer the controversial question of whether or not vascular fellowships detract from general surgical training, questionnaires were sent to directors of 41 approved vascular fellowship programs, 41 residency directors in the same institutions, and 40 residency directors in university programs without approved fellowships. Overall response rate was 74% (93% of vascular fellowship directors, 63% of same-institution residency directors, and 65% of residency directors without vascular fellowships). Thirty-four per cent of fellowship directors and 38% of same-institution residency directors indicated that the fellowship has reduced the vascular surgery case load of residents. In institutions with fellowships, general surgery residents performed an average of 71 major vascular procedures and first assisted on 44, whereas residents performed 65 major vascular procedures and assisted on 47 in institutions without fellowships. Overall, 79% of fellowship directors and 62% of same-institution general surgical directors indicated that the fellowship improved the quality of vascular surgical training. Only 15% of same-institution residency directors and 3% of fellowship directors felt that the fellowship detracted from the general surgical experience. Fifteen per cent of institutions without approved fellowships have now initiated vascular fellowship programs, and an additional 23% plan to begin such fellowships. No fellowship directors plan to abolish their programs, although 8% plan to decrease the number of fellows in order to increase general surgery resident participation. Among the majority of institutions studied, vascular fellowships have not adversely affected general surgical training, and often enhanced it.

Documentation of care and prospective payment. One hospital's experience.

Hospitals are now being reimbursed by Diagnosis Related Group (DRG) for Medicare patients. The Johns Hopkins Hospital has worked successfully under this system for the past 5 years, with cost increases being maintained well below the national average. Allowable revenue varies considerably by diagnosis depending on such factors as secondary diagnoses, procedure, and patient age. Failure to document accurately may result in substantial loss of hospital income. More worrisome is the use of data by outside agencies to evaluate quality of care. Recent reports of mortality rates for surgery in Maryland hospitals and of permanent pacemaker use are illustrative. Conclusions were inaccurate because of inadequate documentation of diagnoses and procedures by physicians and inaccurate coding by quality assurance coordinators. Surgeons need to be aware that in the prospective payment era, accurate and complete documentation is essential and that their data are likely to be used for purposes other than monitoring fiscal performance.

Management of hemobilia associated with transhepatic internal biliary drainage catheters.

Hemobilia developed as a complication in nine patients with transhepatic biliary drainage catheters. Arteriography revealed pseudoaneurysms of the hepatic artery in five patients and hepatic arterio-portal venous fistulas in three patients. In one patient no arteriographic abnormality was seen. Seven of eight patients were treated successfully with angiographic embolization of the bleeding site. Hemobilia stopped spontaneously in one patient without therapeutic intervention.

The use of intraperitoneal xenon for early diagnosis of acute mesenteric ischemia.

We evaluated the technique of intraperitoneal use of xenon Xe 133, previously described for the diagnosis of early intestinal strangulation obstruction in rats and dogs, for the recognition of acute mesenteric vascular occlusion in these animals. 133Xe was injected intraperitoneally into five groups of six rats: control, sham operation, superior mesenteric artery (SMA) ligation, superior mesenteric vein ligation, and portal vein ligation. Residual gamma-activity was monitored by external counting and camera imaging. At 30 minutes after injection, the activity was significantly higher in the rats from the three groups with vascular ligation than in the control and sham operation animals (P less than 0.001). gamma-Camera images reflected these findings, with positive images only in the rats that underwent vascular ligation. "Blinded" readings of the 30 sets of scans confirmed the diagnostic accuracy of the images. Results were essentially the same in a second series of experiments in eight control dogs and six dogs with balloon occlusion of the SMA. Concentrations of isotope in ischemic intestine ranged from 10(3) to 10(5) times the levels in adjacent normal bowel. These levels and the positive images appeared early, prior to the development of tissue necrosis. The intraperitoneal use of 133Xe therefore continues to show promise for the recognition of patients with early intestinal ischemia.

Reasons for delay of the diagnosis of acute appendicitis.

The essence of the problem, as previously reported, indicated that few complications of acute appendicitis occur as long as the infection is contained within the appendix, but once the invading bacteria have penetrated the peritoneal appendicular surface or have invaded the regional circulation, any one or more of a series of serious complications can develop. Thus, rightfully, emphasis has been placed upon early removal of the inflamed appendix before penetration has occurred as the best method of preventing complications. We have shown that early appendectomy is predicated on early diagnosis and that diagnostic delay is not limited to extremes of age. The diagnosis may be obscured by an accurate, although misleading, history of prior acute attacks, by precident acute disease, such as viral gastroenteritis and by unimpressive symptoms blunted by intercurrent chronic illness, such as diabetes mellitus. If the elements of periumbilical pain, anorexia, nausea or vomiting and the migration of pain to the right lower abdominal quadrant are contained within the clinical history, one must suspect transmural progression of acute appendicitis; frequent inpatient examinations will allow earliest diagnosis and, thereby, fewest perforations and their attendant serious complications. Misdiagnosis is common. Any patient observed for an ostensibly nonsurgical acute condition of the abdomen who fails to improve markedly during a brief course of appropriate specific or supportive therapy must be thoroughly re-evaluated as a potential surgical candidate. Despite the proliferation of accessible laboratory tests and imaging procedures, the early diagnosis of appendicitis rests upon the clinical skills of the physician. A high index of suspicion is crucial. As Doctor Warfield M. Firor, former senior surgeon commented: "Pain and tenderness at any point where the appendix can lie must raise the diagnostic possibility of appendicitis."

Intraperitoneal xenon for the detection of early intestinal ischemia: effect of ascites, adhesions, and misdirected injections.

Significant delay in the washout of intraperitoneal xenon (133Xe) in rats and dogs with decreased splanchnic blood flow (bowel strangulation, superior mesenteric artery and vein occlusion) has been previously demonstrated as the basis for radionuclide imaging to detect early (prenecrotic) intestinal ischemia. In this study, the effect of ascites, adhesions, and misdirected injections on the validity of this technique is evaluated. Xenon-133 (0.6 mCi) in 3 ml saline was injected into the peritoneal cavity of anesthetized rats and the washout of gamma activity monitored externally for 90 min. Gamma camera images were obtained at 30-min intervals. After 60 min, only 12 +/- 2% of injected activity remained in the controls. Sham operation (13 +/- 1%) and simple obstruction (12 +/- 2%) had been previously shown not to significantly slow washout, but segmental strangulation had done so dramatically (32 +/- 2%, P less than 0.0001). In these experiments, ascitic fluid (Ringer's lactate) in volumes of 10 ml (13 +/- 1%), 20 ml (13 +/- 1%), and 40 ml (13 +/- 1%), did not significantly slow washout in nonischemic rats. Sixty and eighty milliliters produced very tense ascites and slight but significant delay in washout (14 +/- 1%, 17 +/- 1%, respectively, P less than 0.05). Moderate (11 +/- 1%) and severe (11 +/- 1%) adhesions produced by serosal scarification did not delay washout nor affect imaging. Injections of isotope intentionally misdirected into the abdominal wall (32 +/- 2%), bowel wall (18 +/- 1%), and bowel lumen (19 +/- 2%), each significantly (P less than 0.001) slowed washout. However, such misdirected injections were easily recognizable as such on the 1-min gamma camera images and could thereby be excluded as artifactual. Therefore, no false positive readings were obtained. It is concluded that the intraperitoneal xenon technique is not invalidated by mild to moderate ascites nor by moderate to severe adhesions. Misdirected injections produce invalid studies that are recognizable as such and thus are not misinterpreted. This approach should therefore be applicable to most patients with suspected intestinal ischemia.