Routine upper GI series after gastric bypass does not reliably identify anastomotic leaks or predict stricture formation.

BACKGROUND: Many surgeons who perform Roux-en-Y gastric bypass (RYGB) for morbid obesity routinely obtain an upper gastrointestinal (GI) series in the early postoperative period to search for anastomotic leaks and signs of stricture formation at the gastrojejunostomy. We hypothesized that this practice is unreliable. METHODS: We analyzed 654 consecutive RYGBs, of which 63% were completed laparoscopically. An upper GI series was obtained in 634 (97%) patients. The radiographic findings (leak or delayed emptying) were compared with clinical outcomes (leak or stricture formation) to calculate the sensitivity and specificity. Univariate analysis identified risk factors for leaks or stricture formation; events were too few for multivariate analysis. RESULTS: Of 634 routine upper GI series, anastomotic leaks at the gastrojejunostomy were diagnosed in 5 (0.8%); 2 of these 5 were later reinterpreted as artifacts. Four leaks were not seen on the initial upper GI series, yielding an overall sensitivity of 43% and a positive predictive value (PPV) of 60%. Univariate analysis showed that cases done early (odds ratio [OR] 5.4 for the first 100 cases, p = 0.02) and prolonged operating time (OR 7.8 for cases >or= 300 min, p = 0.01) were associated with leaks. Emptying into the Roux-en-Y limb was delayed in 127 (20%) of the upper GI series. Strictures requiring dilatation developed in 16 (2.4%) patients. The PPV of delayed emptying for stricture formation was 6%. Risk factors for stricture formation included stapled anastomosis (OR 7.8, p = 0.002), surgeon inexperience (OR 2.9 for first 50 cases, p = 0.04), and delayed emptying (OR 3.3; p = 0.02). CONCLUSIONS: Because the incidence of anastomotic complications and the sensitivity of upper GI series were both low, routine upper GI series did not reliably identify leaks or predict stricture formation. A selective approach, whereby imaging is reserved for patients with clinical evidence of a leak or stricture, may be more appropriate.

Lack of effect of treatment for Helicobacter pylori on symptoms of nonulcer dyspepsia.

BACKGROUND: Prior studies have yielded conflicting results on whether or not Helicobacter pylori causes nonulcer dyspepsia. PATIENTS AND METHODS: We enrolled 100 consecutive patients with nonulcer dyspepsia into a randomized, double-blind, placebo-controlled trial. Patients with peptic ulcer disease, esophagitis, hepatobiliary disease, irritable bowel disease, or predominantly reflux-related symptoms were excluded by history and upper endoscopy. Helicobacter pylori infection was determined by biopsy and histologic examination. Serum H. pylori IgG antibodies and CagA status were determined by Western blot. Enrolled patients were randomized to a 14-day regimen of omeprazole (20 mg twice daily) and clarithromycin (500 mg three times daily) or placebo. Dyspeptic symptoms were assessed by use of a visual analog scale at baseline and at 1, 3, 6, and 12 months after treatment. Follow-up upper endoscopy with biopsy was performed 4 weeks after treatment. Compliance was measured by tablet counts. RESULTS: At 1 year, the change in dyspeptic symptoms was -24.0 (95% confidence interval, -69.0 to 21.0) in the omeprazole and clarithromycin group and -24.2 in the placebo group (95% confidence interval, -70.0 to 21.6). Furthermore, patients with persistent H. pylori infection demonstrated a greater, but not significant, improvement in symptoms (-40 +/- 144 [mean +/- SD], -65 +/- 142, -45 +/- 138, and -39 +/- 163) than those with successful eradication (-26 +/- 126, -26 +/- 148, -12 +/- 126, and -25 +/- 151) at months 1, 3, 6, and 12, respectively. CONCLUSION: Patients with nonulcer dyspepsia should not routinely be treated for H. pylori, since it is not a cause of this condition in most patients.

Teaching of primary care in an internal medicine residency program.

A teaching module in ambulatory internal medicine has been added to a residency program in a large referral hospital. The effort was directed to structure a program of instruction reflecting common problems seen by primary care internists. Patients were screened by supervising staff to make the rotation an efficient learning experience. Experienced nurse clinicians were used to monitor patients with stabilized medical problems, thus freeing the housestaff for problems more suited to their level of training. The nurses also served as role models with which the housestaff could interact and provide continuity and accessibility of care to patients being followed up by transient physicians-in-training. Various teaching conferences based on traditional inpatient models were conducted. Self-assessment and peer review techniques were structured around monthly conferences in which medical records were assessed by each participant.

Liquid caustic ingestion. Spectrum of injury.

Seventeen patients who ingested liquid caustics were reviewed for location, extent, severity, and outcome of the mucosal injury. Although many complained of glossopharyngeal pains and dysphagia (12 patients), and most had some oral mucosal burns (15 patients), the absence of severe oral burns or pharyngoesophageal symptoms did not exclude esophageal or gastric injury as determined by endoscopy. The location of the most severe mucosal injury was unpredictable by symptoms alone; seven of the 17 patients demonstrated gastric mucosal injury greater than esophageal. One patient died of extensive esophageal, gastric, and duodenal injury; esophageal strictures developed in three patients. Caustic ingestion is a serious medical condition whose severity can best be gauged by endoscopic findings and not by patient symptoms alone. Outcome is variable, ranging from an asymptomatic state to stricture formation or even death.

Pancreaticopleural fistula. Report of 7 patients and review of the literature.

Pancreaticopleural fistula is an uncommon clinical condition. Its presentation is often confusing because of the paucity of clues suggestive of pancreatic disease and the preponderance of pulmonary symptoms and signs. Most patients are alcoholics but only one-half will have a clinical history of previous pancreatitis. Pleural effusions are large, recurrent, and highly exudative in nature. Many patients go through extensive pulmonary evaluation before the pancreas is identified as the site of primary pathology. An elevated serum amylase may be the first clue to the diagnosis. However, the key to the diagnosis is a dramatically elevated pleural fluid amylase. Effusions in association with acute pancreatitis, esophageal perforation, and thoracic malignancy are important to consider in the differential diagnosis of an elevated pleural fluid amylase but are usually easy to exclude. Computed tomography is excellent in defining pancreatic abnormalities and should be the first abdominal imaging study in suspected cases. Endoscopic retrograde cholangiopancreatography (ERCP) is used as a diagnostic tool only in confusing cases. Although no systematic study evaluates medical versus surgical therapy, we recommend an initial 2 to 4-week trial of medical therapy, including allowance of no oral intake, total parenteral nutrition, chest tube thoracostomy, and possibly a regimen of somatostatin or its analogs. The major complication in these patients is superinfection, which results in significant morbidity and mortality. Failure of medical therapy should be considered failure of pleural effusion(s) to clear, recurrence after reinstatement of oral intake, or superinfection. For those patients who fail to benefit from medical therapy, surgery is indicated.(ABSTRACT TRUNCATED AT 250 WORDS)

De novo lipogenesis predicts short-term body-composition response by bioelectrical impedance analysis to oral nutritional supplements in HIV-associated wasting.

We studied the effects of enteral supplements on protein and energy intakes, body composition, energy expenditure, and gastrointestinal histology in 49 subjects with human immunodeficiency virus-associated weight loss (12.7 +/- 0.9% of body wt). We also determined whether a stable-isotope mass spectrometric measurement at baseline might predict the short-term response of fat-free mass (FFM) measured by bioelectrical impedance analysis. Thirty-nine subjects completed the study after being randomly assigned to receive either a whole-protein-based (n = 22) or a peptide-based (n = 17) formula. A nonsupplemented, nonrandomly assigned group (n = 13) was followed concurrently. Both formulas were well tolerated. Voluntary intakes of energy and protein from nonsupplement sources decreased significantly during supplementation [by 819-1638 kJ (196-382 kcal)/d and 5.6-14.4 g protein/d, respectively; P < 0.01] but to a lesser extent than the intake from the supplement [2300-2510 kJ(550-600 kcal)/d and 19-28 g protein/d, respectively], so that net increases in intakes of protein and energy (P < 0.03), as well as of several vitamins and trace elements were increased. Nevertheless, the mean FFM did not increase for the group as a whole, although there was considerable interindividual heterogeneity. Changes in FFM at 6 wk were significantly inversely correlated (r = 0.65, P < 0.01) with baseline synthesis of fat (de novo hepatic lipogenesis), but not with other potential measures of energy intake (insulin-like growth factor 1 or its binding protein) or inflammation (soluble tumor necrosis factor receptors I or II). The prospective identification of FFM response by measurement of de novo hepatic lipogenesis supported the hypothesis that the subset of wasting patients whose FFM is unresponsive to nutrient supplementation have altered nutrient metabolism.

Acquired immunodeficiency syndrome cholangiopathy: spectrum of disease.

PURPOSE: Reports of biliary tract abnormalities in patients with acquired immunodeficiency syndrome (AIDS) are infrequent. Nine patients with features of papillary stenosis and sclerosing cholangitis and one patient with primary bile duct lymphoma were previously described. Since those reports, an additional 16 AIDS patients with clinical and laboratory features suggestive of biliary tract disease have been evaluated. The findings in these 26 patients form the basis of this study. PATIENTS AND METHODS: The patients, 25 men and one woman, with right upper quadrant abdominal pain, fevers, or elevated serum alkaline phosphatase levels, or any combination of the three symptoms, underwent endoscopic retrograde cholangiopancreatography and ampullary biopsy. RESULTS: Twenty of these 26 (77%) patients were found to have markedly abnormal cholangiograms. Four patterns of cholangiographic abnormalities were identified: sclerosing cholangitis and papillary stenosis (10 patients), papillary stenosis alone (three patients), sclerosing cholangitis alone (four patients), and long extrahepatic bile duct strictures (three patients). An AIDS-associated pathogen or malignancy was identified in 11 patients (55%), with cytomegalovirus being the most common (five patients). Demographic, clinical, and laboratory features were not significantly different between AIDS patients with cholangiopathy and those without cholangiopathy. However, 15 of 20 patients with cholangiographic abnormalities (versus only one of six patients without such abnormalities) had ductal dilation on ultrasound or computed tomography, or both. CONCLUSION: Clinicians should be aware that profound biliary tract abnormalities may develop in a large number of human immunodeficiency virus-infected patients. Confirmation of the anatomic abnormalities, determination of correct etiology, and effective therapy may depend upon early endoscopic cholangiography.

Eosinophilic gastroenteritis--a complex disease entity.

Although usually regarded as merely a manifestation of a simple food allergy, eosinophilic gastroenteritis remains a poorly understood disorder. Characterized by peripheral eosinophilia, eosinophilic infiltration of the bowel wall to a variable depth and gastrointestinal complaints, the disease responds inconsistently to simple food withdrawal programs. Immunoglobulin E (IgE) antibodies to specific food substances have been found in a few patients leading to a postulated pathophysiologic mechanism involving tissue mast cells, release of slow-reacting substance of anaphylaxis (SRS-A) and chemotaxis of eosinophils. Oral steroids appear, in uncontrolled trials, to ameliorate the disease.

Long-term follow-up of endoscopic retrograde cholangiopancreatography sphincterotomy for patients with acquired immune deficiency syndrome papillary stenosis.

PURPOSE: To evaluate the long-term effects on biliary-type pain and changes in biochemical parameters in patients with AIDS-associated papillary stenosis who underwent endoscopic retrograde cholangiopancreatography (ERCP) sphincterotomy. PATIENTS AND METHODS: Twenty-five consecutive patients were diagnosed by cholangiography with AIDS-associated papillary stenosis using standard criteria. Patients underwent ERCP sphincterotomy and were followed prospectively in the Gastrointestinal or Liver Clinics, San Francisco General Hospital, and by their primary physicians. Post-procedure data was prospectively collected by chart review or in-person or telephone interview, and analyzed using statistical software. RESULTS: All patients presented with severe right upper quadrant and/or mid-epigastric abdominal pain and had marked elevations of serum alkaline phosphatase. Following ERCP sphincterotomy, pain scores decreased significantly for at least 9 months of follow-up. Serum alkaline phosphatase levels, however, remained essentially unchanged. Overall quality of life was difficult to assess, as patients suffered from other AIDS-associated debilitating diseases. CONCLUSIONS: ERCP sphincterotomy, while not without risks, provided significant reduction in pain in patients with AIDS-associated papillary stenosis.

Cholestasis and disseminated cytomegalovirus disease in patients with the acquired immunodeficiency syndrome.

A cholestatic syndrome caused by papillary stenosis and sclerosing cholangitis (PS/SC) has been recently described as a complication of the acquired immunodeficiency syndrome (AIDS). Cytomegalovirus has been implicated as an etiologic factor in this syndrome. Consecutive cases of disseminated cytomegalovirus disease and viremia were reviewed for evidence of cholestasis. In 12 of 36 (33 percent) patients with cytomegalovirus infection (gastrointestinal disease [eight of 16 patients] or retinitis [four of 20 patients]), cholestatic liver enzyme abnormalities developed within three months of the diagnosis of cytomegalovirus disease. Radiologic imaging demonstrated biliary dilation in three of nine (33 percent) patients studied. Review of serial liver function tests showed no effect of ganciclovir therapy on the course of cholestatic abnormalities. AIDS patients without CMV end-organ disease were also reviewed for evidence of cholestasis. Six of 15 AIDS patients with positive blood culture results for cytomegalovirus had cholestatic liver enzyme abnormalities compared with four of 36 AIDS patients with negative viral blood culture results (p = 0.023, Fisher's exact test). Patients with cytomegalovirus retinitis, gastrointestinal disease, or viremia appear to be at increased risk for the development of cholestasis. In 3 to 11 percent of patients with cytomegalovirus retinitis or gastrointestinal disease, PS/SC may develop. This clinical association supports an etiologic role for cytomegalovirus in the PS/SC syndrome.

Asymptomatic chronic gastrointestinal blood loss in patients taking aspirin or warfarin for cardiovascular disease.

OBJECTIVE: To determine whether low-dose aspirin or warfarin induces fecal occult blood loss. PATIENTS AND METHODS: A prospective, cross-over study, of 100 participants over 40 years of age in 1 of 3 groups, taking: (1) no aspirin or warfarin, (2) daily aspirin (either 81 or 325 mg), or (3) warfarin, but no aspirin. Stool samples were collected and analyzed for the presence of occult blood using HemoQuant and Hemoccult II. After collection of baseline samples, patients initially taking no aspirin (group 1) were asked to take regular aspirin (325 mg daily) for 2 months. Patients initially taking aspirin 81 mg daily (group 2) were switched to 325 mg daily for 2 months, and vice versa. RESULTS: Patients taking no aspirin had mean fecal blood of 0.68 +/- 0.05 mg hemoglobin/g stool, which increased to 1.41 +/- 0.36 mg/g after taking 325 mg of aspirin daily (P = 0.02). In contrast, patients in group 2, taking 81 mg and 325 mg of aspirin, had mean fecal blood of 0.82 +/- 0.08 mg/g (P = 0.57) and 1.04 +/- 0.23 mg/g (P = 0.13), respectively (comparisons with patients taking no aspirin). The mean blood loss in patients taking warfarin was 0.51 +/- 0.04 mg/g (P = 0.55), and fecal blood was not related to the degree of anticoagulation. There was no increase over normal in the rate of Hemoccult II-positive stool tests with aspirin or warfarin therapy. CONCLUSION: Aspirin, but not warfarin, caused a small but clinically insignificant increase in occult fecal blood. The small blood loss in patients taking aspirin or warfarin is unlikely to interfere with fecal occult blood test. Therefore, positive fecal occult blood tests, in patients taking either low-dose aspirin or warfarin, should be managed in the same fashion as patients not taking these medications.

Sucralfate suspension versus titrated antacid for the prevention of acute stress-related gastrointestinal hemorrhage in critically ill patients.

Seventy-four medical and surgical patients having a minimum of two risk factors for stress-related gastric mucosal bleeding were prospectively selected randomly to receive prophylaxis by antacid titration (to maintain a gastric pH of more than 4) or with sucralfate suspension (1 g/10 ml every four hours). Gastric aspirates were monitored every two hours for pH and overt and occult bleeding. Despite a significantly greater severity of illness in the sucralfate group (p less than 0.01), no significant difference in overt or occult bleeding between the groups could be demonstrated. Low-grade occult blood loss occurred frequently in both groups, but only one of the 74 patients (four risk factors, sucralfate group) had significant stress-related bleeding as defined by preset criteria and documented by endoscopy. The effectiveness of sucralfate appeared unrelated to acid neutralization in keeping with its classification as a cytoprotective agent. There were eight antacid-related side effects (four severe diarrhea, four hypermagnesemia), and none related to sucralfate. Sucralfate suspension was safe and effective and had fewer side effects than antacid titration for the prophylaxis of stress-related bleeding in critically ill patients.

Endoscopic sclerotherapy versus esophageal transection of Child's class C patients with variceal hemorrhage. Comparison with results of portacaval shunt: preliminary report.

Thirteen Child's class C patients with variceal hemorrhage requiring 6 or more units of blood were randomly assigned to groups receiving either flexible endoscopic sclerotherapy using 5% sodium morrhuate or esophageal transection/reanastomosis employing the EEA Auto Suture stapling instrument. One patient with a previous hiatus hernia repair was assigned to the transection group and had a mesocaval shunt. The randomization groups compared favorably with each other and with a retrospective shunt group (20 patients). Although transection patients required significantly less operative time and fewer blood transfusions, the outcome of the transection group was significantly worse than that of the shunt group. The trend in survival likewise favored the shunt group rather than the sclerotherapy group. Although the reported operative mortality rates of transection and sclerotherapy are less than that of portal decompression, the results of this study--the first controlled trial of these procedures--would warrant distinct caution in using either of these new techniques in high-risk patients. Portal decompression may still be the procedure of choice in class C patients with substantial variceal hemorrhage.

Gastrointestinal manifestations of HIV infection.

Gastrointestinal symptoms are pervasive in patients with AIDS. Organ-specific symptoms, particularly dysphagia, postprandial emesis, hematemesis, biliary colic, increasing abdominal girth, and small volume diarrhea, are usually easy to diagnose definitively and may respond to specific therapy. Thorough invasive and noninvasive investigation is warranted.

Factors influencing survival after therapeutic shunts. Results of a discriminant function and linear logistic regressions analysis.

Data on 100 consecutive patients undergoing portasystemic shunt at three hospitals of the University of California, San Francisco, were analyzed retrospectively to look for variables portending poor immediate and long-term outcome. As a determinant of early mortality after portacaval shunt, the Child's classification of the patient remains the single most important factor. If the patient is in Child's class C and has a hematocrit of less than 32 percent, he is even less likely to survive 30 days. The malnourished male patient who resumes drinking postoperatively is least likely to survive 1 year. Though short- and long-term mortality did not correlate with type of shunt, the prosthetic interposition mesocaval shunt was associated with an unacceptably high thrombosis rate of 20 percent in our institutions, and represented a technical failure to achieve the goal of preventing further variceal bleeding. No matter what type of shunt was performed, however, the 30 day mortality of Child's class C patients exceeded 50 percent. In the latter patients methods of treatment other than portasystemic shunts should be evaluated.

AIDS-Related biliary tract disease.

Increasingly, specialists caring for patients with AIDS have reported relatively small numbers of patients with biliary tract disease. These conditions fall into three general categories: (1) non-HIV-associated conditions of the bile duct, (2) acalculous cholecystitis, and (3) AIDS cholangiopathy. Whatever the cause, a sizable percentage of patients with AIDS are found to have abnormal biliary tract morphology. It is essential for the clinician first to exclude biliary tract conditions such as gallstone disease and then to clearly investigate those patients with clinical, biochemical, or radiographic features suggestive of papillary stenosis. Patients with AIDS-associated papillary stenosis do respond symptomatically to ERCP sphincterotomy.

The laparoscopic evaluation of ascites.

Laparoscopy is an invaluable technique for the evaluation of ascites in subgroups of patients with ascites. Indications for laparoscopic examination include determination of the causes of ascites when routine tests fail to disclose the source, evaluation for the presence of multiple causes of ascites formation, or histopathologic verification of malignancy within the peritoneal cavity. Several reported series have illustrated the efficacy of laparoscopy for the diagnosis of peritoneal carcinomatosis, tuberculous peritonitis, or unsuspected cirrhosis, securing its role in the management of selected patients with ascites.

Steatorrhea: a common manifestation in patients with HIV/AIDS.

Multiple infectious causes of diarrhea are known in patients with HIV/AIDS. Maldigestion and malabsorption have been reported in patients with HIV/AIDS and may be independent of infectious etiologies. Among ambulatory patients with HIV/AIDS, we examined the prevalence of fat malabsorption (steatorrhea). Sixty-one patients with unexplained diarrhea (defined as > 2 stools/d) and/or weight loss despite adequate caloric intake (and without clinical evidence of chronic pancreatitis) were evaluated in our outpatient Gastroenterology-Nutrition Clinic between March 1, 1993, and July 1994. Patients were instructed by a dietitian to follow a > or = 100 g/d fat diet for 24 h before submitting a stool sample for qualitative (or quantitative) fecal fat determination. Forty-five patients, 32 with ongoing diarrhea and 13 without diarrhea, submitted stool samples. Twenty-two of 45 patients (49%) had qualitative or quantitative steatorrhea, 16/32 with diarrhea (50%) and 6/13 patients without diarrhea (46%). Thirty of 32 patients with diarrhea had had extensive microbiologic and/or endoscopic evaluations. Only 9 patients had a detectable intestinal pathogen, 5 patients had cytomegalovirus (4 treated), 4 patients had cryptosporidia (3 treated), and 1 patient had microsporidia. Steatorrhea, as determined by abnormal qualitative fecal fat, is detectable in nearly 50% of patients with HIV/AIDS. Fat malabsorption appears to be a primary defect in these patients independent of detectable pathogens. Assessment of fat malabsorption should be considered in patients with unexplained weight loss or diarrhea before extensive evaluation for opportunistic infections.

Serum zinc and protein levels: lack of a correlation in hospitalized patients with AIDS.

Abnormally low serum zinc levels are associated with advanced states of malnutrition. Zinc levels are thought to parallel serum albumin, and repletion of zinc has reportedly led to increased albumin. We examined the correlation between zinc deficiency and serum proteins in hospitalized patients with AIDS. Over 500 inpatient consultations were performed by our Gastroenterology-Nutrition Consult Service from May 1992 to June 1994. We reviewed the medical records from all 228 AIDS patients in whom a serum zinc level was measured (by atomic absorption spectrophotometry). The correlation between serum zinc, albumin, prealbumin, and transferrin drawn on the same hospital day was analyzed by linear regression. The patients were stratified by the level of albumin, group A albumin < 2.0 g/dL, group B albumin 2.0-2.9 g/dL, and group C albumin > 3.0 g/dL, to allow comparison of the incidence of diarrhea and mean zinc level by chi square. Thirty-four patients had more than one serum zinc and albumin determination; the change over time was compared by linear regression. Serum zinc and albumin, prealbumin, and transferrin levels did not correlate strongly (r2 < or = 0.01). Furthermore, changes in zinc over time did not correlate with parallel changes in serum albumin (r2 < or = 0.01). In patients grouped by albumin, the proportion of patients with diarrhea did not differ significantly. The mean zinc levels were also not statistically significantly different. Serum zinc levels do not strongly correlate with serum proteins in hospitalized patients with AIDS. Serial measures of zinc over time also do not correlate strongly with changes in albumin. The incidence of diarrhea was not significantly different in patients with hypoalbuminemia or hypozincemia when compared with patients with near normal albumin or zinc. The relationship between zinc and serum proteins in AIDS patients is not linear. Zinc deficiency should be assessed in patients independent of the serum albumin.

Zinc levels and infections in hospitalized patients with AIDS.

Impaired cellular and humoral immunity and phagocytic function have been attributed to zinc deficiency. This study examined the association between low serum zinc concentration and opportunistic infections in hospitalized patients with the acquired immune deficiency syndrome (AIDS). We examined the records from all 505 inpatient consultations performed by our Nutrition Service from May 1992 through June 1994. The medical records from all 228 patients with AIDS with known serum zinc levels (determined by atomic absorption spectrophotometry) were reviewed. The length of HIV seropositivity, most recent CD4 count, presence of diarrhea, and degree of malnutrition were noted. The principal diagnosis accounting for the admission was grouped according to the type of infection: Pneumocystis carinii pneumonia (PCP), viral, fungal, bacterial, and other. Sixty-seven patients (29%) had abnormally low serum zinc levels (LSZ < 55 micrograms/dL), 49 patients (21%) had borderline low serum zinc (BSZ > or = 55 and < or = 65 microgram/dL), and 112 (49%) patients had normal serum zinc levels (NSZ > 65 micrograms/dL). There was no significant difference among the groups in CD4 count, length of HIV seropositivity, presence of diarrhea, or severity of malnutrition. Patients with zinc deficiency (LSZ) had a significantly higher incidence of bacterial infection than did patients with normal zinc. Patients with borderline zinc levels had an intermediate incidence of bacterial infection. There were no significant differences among the three groups in the incidence of PCP, viral, or fungal infections. Severe zinc deficiency was noted in 29% and borderline levels in an additional 21% of hospitalized AIDS patients. A low zinc level was not associated with the length of HIV seropositivity, CD4 count, or degree of malnutrition. Hypozincemia was associated with an increased incidence of concomitant systemic bacterial infections.