THE ROLE OF CHOLECYSTECTOMY IN ACQUIRED IMMUNODEFICIENCY SYNDROME

David R. Flum, MD, Scott D. Steinberg, MD, Antoine Y. Sarkis, MD, James R. Pacholka, MD, and Marc K. Wallack, MD, FACS

[From the Department of Surgery, St. Vincent's Hospital and Medical Center, New York, NY. Correspondence address: David R. Flum, MD, 130 West 12th Street, 11G, New York, NY 10011.]

BACKGROUND
Hepatobiliary disease is a common manifestation of acquired immunodeficiency syndrome, although the role of surgical intervention in the spectrum of therapy is unclear.

STUDY DESIGN
A retrospective review was designed to evaluate the characteristics of patients given a diagnosis of human immunodeficiency virus infection or acquired immunodeficiency syndrome and undergoing cholecystectomy between January 1, 1986, and November 1, 1995.

RESULTS
The study included 40 patients (35 men, 5 women; mean age, 42±9 years), 33 (82.5 percent) with acquired immunodeficiency syndrome; their mean preoperative T-helper (CD4) cell count was 163/mL3. Gross pathologic findings included acute (n=9, 22.5 percent) and chronic (n=31, 77.5 percent) cholecystitis. Gallbladder specimens were positive for cholelithiasis in 28 (70 percent), Cryptococcus organisms in 5 (12.5 percent), cytomegalovirus in 3 (7.5 percent), and lymphoma in 2 (5 percent). The median follow-up time was 48 months (range, 6 to 63 months). The percentage survival was 92.5 percent (n=37) at 30 days, and 57.5 percent (n=23), 37.5 percent (n=15), and 25 percent (n=10) at 12, 24, and 36 months, respectively. The mean survival time was 25.1 months. The likelihood of survival was directly linked to the CD4 cell count. The mean survival period was 25 months for patients with CD4 cell counts less than 200/mL3 compared with 48 months for those with CD4 cell counts greater than 200/mL3.

CONCLUSIONS
Although the pathologic changes identified in patients with acquired immunodeficiency syndrome may occasionally be atypical, the clinical presentation, indications for operation, and pathologic findings identified are quite common. Patients tolerate cholecystectomy well with good long-term outcome and minimal infectious complications. Even in patients with the most compromised immune status, a 2-year survival after operation is acceptable. Cholecystectomy has a clear role in the spectrum of treatment for biliary disease relative to acquired immunodeficiency syndrome. J. Am. Coll. Surg., 1997, 184: 233-239.

DESPITE GROWING CLINICAL EXPERIENCE with acquired immunodeficiency syndrome (AIDS), the role of operative treatment in the spectrum of this disease has only recently been critically evaluated. Recent clinical reviews have indicated that, although abdominal pain and gastrointestinal complaints are common in patients infected with human immunodeficiency virus (HIV) and AIDS, relatively few require an operation (1, 2). The gastrointestinal manifestations of AIDS are multiple (3) and offer diagnostic and therapeutic challenges to the surgeon. Unique aspects of making clinical decisions about treating patients with AIDS require special consideration (1-3). As we become more familiar with the manifestations of abdominal disease in patients with AIDS, it is paramount that critical reviews evaluate the applicability of operative management.

Hepatobiliary disease is a common manifestation of AIDS, with a full spectrum of pathologic entities grouped together as AIDS cholangiopathy (4, 5). Cholecystitis that is often acalculous and infectious has been reported periodically (3); but, the reported clinical experience with operative intervention for cholecystitis is limited. Making critical appraisals of the natural history of cholecystitis and cholecystectomy in patients with AIDS and determining predictors of surgical outcome based on the limited experience have been difficult.

Our clinical experience suggests that the prevalence of cholecystitis that requires cholecystectomy is much higher than identified in the literature. We have retrospectively studied this clinical experience to address the following questions: 1. What are the typical demographic, clinical, and chemical features that describe the population undergoing surgery for biliary disease? 2. What types of biliary disease are found in patients with AIDS? 3. What is the role of operative treatment in the spectrum of this disease? 4. Does a subgroup of patients exist for whom benefit from operative treatment can be predicted?

METHODS
A retrospective review was conducted in the department of medical records at St. Vincent's Hospital and Medical Center, New York City. The study included 40 consecutive patients given a diagnosis of AIDS or HIV infection and undergoing "open" or laparoscopic cholecystectomy between January 1, 1986, and November 1, 1995. The medical records of these patients were reviewed by a team of examiners. Data obtained included demographics, clinical history, laboratory and radiologic findings, diagnostic considerations, operative procedure, pathologic findings, postoperative course, and the disposition of the case. At the time of data aquisition, follow- up surveys were completed through telephone contact with the patient or next available source. Follow-up questionnaires included disposition of the case and further clinical or operative information.

All data were evaluated for statistical relationships using the EPI Info, Version 6.0, and Number Cruncher, Version 5.0 (Numbercruncher Statistical Software, Kaysville, KY) statistical software packages. The Student's t test, chi-square test, and Fisher's exact test were applied as appropriate. Kaplan-Meier product limit survival distributions were used for outcome analysis.. A p value of less than .05 was considered statistically significant. Numeric values are represented as the mean plus or minus the standard error of the mean unless otherwise noted.

RESULTS
Of the 40 patients whose records were evaluated, 35 (87.5 percent) were men and 5 (12.5 percent) were women, with a mean age of 42±9 years. The most common, self-reported risk factors for HIV infection were homosexuality (n=33, 82 percent) and a history of intravenous drug abuse (n=4, 10 percent). The AIDS diagnosis was based on a history of opportunistic infection or absolute CD4 cell count (<200 cells/mL3) in 33 patients (82.5 percent). A diagnosis of HIV infection without stigmata of AIDS was noted in 7 patients (17.5 percent).

In all patients, surgical consultation was obtained by primary care physicians or gastroenterologists. Preoperative biliary disease was described as chronic cholecystitis in 14 patients (35 percent) or acute cholecystitis in 26 patients (65 percent). The most predominant initial complaintswere abdominal pain (n=37, 92.5 percent), nausea and vomiting (n=17, 42.5 percent), and diarrhea (n=11, 27.5 percent). Clinical findings included tenderness in the right upper quadrant, abdominal tenderness (n= 28, 70 percent), fever (n=17, 42.5 percent), and jaundice (n=2, 5 percent). The mean preoperative CD4 cell count was 142/mL3 (range, 17 to 293). Other pertinent preoperative laboratory values are listed in Table I.

Some form of diagnostic testing was performed in all patients. Abdominal ultrasonography obtained in 100 percent of the patients demonstrated cholelithiasis (n=28, 70 percent), thickened gallbladder wall (n=24, 60 percent), and gallbladder distention (n=13, 32.5 percent). Pericholecystic fluid was identified in two patients. Hepatic 2,6-dimethyl iminodiacetic acid radiolabeled with technetium 99m (HIDA) (DuPont Radiopharmaceutical Division, Billirica, Mass) scanning was performed in 19 patients. Of these, cystic duct obstruction was suggested in patients (36.8 percent) by a failure to visualize the gallbladder radiographically. In all studies, the duodenum was promptly visualized. Computed tomography (CT) performed in 6 (15 percent) of 40 patients identified disease in all but 1 patient. The CT findings reinforced the ultrasonographic findings in all patients. Endoscopic retrograde cholangiopancreatography performed in 1 (2.5 percent) patient did not identify common bile duct disease.

Of the patients, 12 (30 percent) were electively admitted, whereas 28 (70 percent) were hospitalized before the cholecystectomy. Twenty-one laparoscopic cholecystectomies and 19 open procedures were performed. No common duct explorations were performed. Although most open procedures (n=16, 84 percent) were performed before 1992, two operations after 1992 began as laparoscopic procedures and were converted to open because of intraoperative findings. A trend toward longer hospital stays was noted after open procedures (13±9 days) compared with laparoscopic (8±10 days); however, this difference was not statistically significant (p=.2). Furthermore, these groups were not historically matched, and the length of stay seemed to be related to the year that the operation was performed. The outcome was not statistically linked to the type of procedure (ie, open or laparoscopic) performed. No significant differences were found in preoperative laboratory values or clinical manifestations between the patients who underwent open procedures compared with those who underwent a laparoscopic operation. Abdominal drains were placed in 12 patients at the completion of the operation. The indications for drain placement included purulent drainage, bile staining, and inflammatory dissection.

Gross pathologic findings in specimens included chronic cholecystitis (78.5 percent) and evidence of acute cholecystitis (22.5 percent). Gallbladder specimens were positive for cholelithiasis in 28 patients (70 percent), Cryptococcus organisms in 5 patients (12.5 percent), cytomegalovirus (CMV) in 3 (7.5 percent), lymphoma in 2 (5 percent), and nonspecific acalculous cholecystitis in 5 (12.5 percent). Overlap of some pathologic findings was noted. Specifically, 2 patients with calculous cholecystitis also had evidence of CMV or Cryptosporidium organisms.

No deaths occurred in the immediate perioperative period; but, the 30-day operative mortality rate was 7 percent (3 of 40). Two of these patients were ventilator-dependent postoperatively and required prolonged therapy in the intensive care unit. One of the patients with postoperative respiratory failure had been intubated and mechanically ventilated preoperatively. These patients died on postoperative days 6 and 13, respectively. The third early death occurred in a patient who died at home on postoperative day 17. In total, 5 (12.5 percent) of 40 patients died within the hospital stay during which the cholecystectomy was performed. All early deaths were secondary to respiratory failure. The average CD4 cell count of this subgroup was 149.1/mL3, which was not significantly different from the mean CD4 cell count of the entire group. A postoperative wound infection, which resolved with routine outpatient wound care, developed in one patient after open cholecystectomy.

Follow-up was conducted at varying intervals (median, 48 months; range, 6 to 63 months). Although an attempt was made to contact all patients, only 19 (48 percent) were alive at the time of postoperative follow up (mean 26.25±8.8 months), and these patients were considered as a censored data group. In this censored group, the mean survival time was 25.1 months. Survival for the group as a whole was calculated using the Kaplan-Meier survival distribution (Fig. 1). Of the 40 patients, 50 percent were alive 12 months after the cholecystectomy. Survivorship was statistically linked to CD4 cell counts (Fig. 2), with better survival noted in patients with higher CD4 cell counts. The largest discrepancy in survivorship was evident when comparing patients with CD4 cell counts higher or lower than 200/mL3. The mean survival was 25 months for patients with CD4 counts less than 200/mL3 and 48 months for those with CD4 counts greater than 200/mL3. At the follow-up evaluation, symptoms of biliary tract disease were absent in all patients, and no patients had been rehospitalized for biliary tract disease. Of 8 patients with persistent gastrointestinal complaints, 5 noted diarrhea during the follow-up evaluation.

Several factors were not statistically significant predictors of survival. These included age, gender, preoperative blood chemistry results, type of preoperative opportunistic infections (ie, Pneumocystis carinii pneumonia, Kaposi's sarcoma, Cryptosporidium, CMV, and Mycobacterium tuberculosis), and operative and pathologic findings.

DISCUSSION
In large clinical reviews, the incidence of abdominal operations for patients with AIDS is low. LaRaja and colleagues identified abdominal surgical procedures in only 4.2 percent of 904 patients with AIDS (2). Barone and associates confirmed this low rate of operative treatment and reported a 2.1 percent rate in a similar population (1). Whether the low rates of surgical intervention represent the true incidence of surgical pathology remains to be seen. Gastrointestinal manifestations of AIDS and, more specifically, biliary tract complaints of abdominal pain and tenderness remain frequent reasons for surgical consultation (3). However, in surgical series, the incidence of cholecystectomy is low (1, 2). In one study, only 12 patients had undergone cholecystectomy (2). In 9 of the 12, acute acalculous cholecystitis was identified. None of these patients had chemical or radiologic evidenceof biliary tract disease. The result of the physical examination was a large determinant of operative intervention because ultrasonography and cholecystography were not diagnostic. A postoperative mortality rate of 33.3 percent was noted in this group (2); however, limited follow-up was available. Therefore, these authors advocated conservative therapy rather than operative treatment for these patients.

Wind and coworkers described eight patients who had clinical manifestations of acute cholecystitis and who underwent cholecystectomy (6). Of these, 25 percent had evidence of cholelithiasis, whereas 50 percent had associated cholangitis. The indications for surgical intervention in this group were dictated by deterioration in the clinical condition. Little benefit was found in the use of radiologic and chemical testing to determine the timing for operative intervention. At the time of the operation, 50 percent of patients had multiple organ failure, and 25 percent died within 20 days of the operation. The remainder died within 14 months. The high postoperative mortality was directly linked to the progression of the biliary disease. This high postoperative mortality was confirmed by Hinnant and colleagues who described two patients with Cryptosporidium and CMV cholecystitis who underwent operative treatment and progressed poorly (7).

Adolph and associates retrospectively reviewed the English language medical literature and identified 12 patients from various reports who underwent cholecystectomy for acalculous cholecystitis (8). The clinical manifestation in this diverse group of patients was episodic and variable. The timing of intervention was based on deterioration in the patient's clinical condition or failure of antibiotic therapy. In general, this review found a discordance between the clinical manifestations and the radiologic and pathologic features identified during the operation. Operative therapy was associated with favorable outcomes despite the degree of immunosuppression. A mortality rate of 0 percent and a morbidity rate of 20 percent were noted. These rates contradicted the single-institution studies described previously (2, 6, 7). Based on this review of 12 patients, cholecystectomy was advocated for symptomatic relief and improvement in the quality of life (8).

Although prompt identification of biliary tract disease may be linked to outcomes, the use of standard diagnostic testing for biliary tract disease in HIV and AIDS has been controversial. The diagnostic evaluation of chronic, right upper quadrant abdominal pain in patients with HIV is assisted by using CT (9, 10) ultrasonography (11-13), and nuclear scintigraphy (14, 15). The diagnostic modality or combination best applied to this patient population remains unclear. In this study, although a large percentage of patients underwent a second or third diagnostic test (ie, nuclear scintigraphy or CT), no evidence indicated that this additional testing significantly altered the management plan. In several patients, a CT was performed after a definitive diagnosis of acute cholecystitis was obtained by ultrasonography, HIDA, or both. A delay from admission to operation was noted in patients undergoing such additional diagnostic evaluations.

The entity of AIDS cholangiopathy has been well described (5, 16). This term defines a spectrum of pathologic changes including ampullary papillitis, intrahepatic and extrahepatic biliary radical disease, sclerosing cholangitis, and cholecystitis (17, 18). Retrospective reviews have documented that papillary stenosis or sclerosing cholangitis with or without cholecystitis developed in 3 percent to 11 percent of patients with AIDS and systemic CMV infection (16). It seems that infection at any site in the biliary tree can be manifested alone or in combination with other sites. Interestingly, biliary dilatation has been identified in 8 percent of asymptomatic patients with AIDS who are receiving treatment for CMV-related nonbiliary tract infections (19). The prevalence of common bile duct disease may cause the persistence of symtoms after cholecysectomy.

The CMV and Cryptococcus organisms have been reported as two common pathogens of the biliary tree in AIDS (8, 20), although a multitude of others including Campylobacter, Candida, and Klebsiella organisms have been implicated (4, 2l, 22). In a substantial group of patients, no pathologic agent is found (5). The affinity of CMV for human bile duct epithelium may determine its pathogenicity (23), and CMV frequently colonizes the biliary tree along with the gastrointestinal tract. Disseminated CMV disease and viremia are associated risk factors for the development of cholangiopathy (16).

The pathologic finding in gallbladders infected with CMV is diffuse inflammation. The proposed mechanism of this intense gallbladder inflammation is a CMV-induced vasculitis (24). Despite this severe inflammation, progression to perforation has been rarely documented (6, 8). An alternative theory of AIDS cholangiopathy is that a vasculitis associated with HIV itself may cause ischemic injury of the biliary tract (25). Animal studies of simian immunodeficiency virus have identified biliary fibrosis and cholangitis associated with cryptosporidiosis, but have failed to identify evidence of vasculitis (26).

Evaluation of a large series of patients with HIV or AIDS undergoing cholecystectomy reveals a higher than expected frequency of common biliary disease. Although the predominance of unusual pathogens has been described in patients with AIDS (8, 20), almost 70 percent of the disease identified in this study was related to non-CMV cholelithiasis. Because unusual pathologic findings are often reported in the literature, the routine nature of some abdominal disease that affects patients with AIDS may be overlooked. Although others (2) have described unusual clinical manifestations of biliary disease, in this study the typical findings of right upper abdominal pain, nausea and vomiting, and abdominal tenderness were usually noted. Only a small percentage of patients complained of unusual gastrointestinalsymptoms.

Both laparoscopic and open cholecystectomy were well tolerated with no immediate mortality and a 5 percent wound infection rate. The procedure "conversion rate" was 8.2 percent, which is higher than expected for laparoscopic cholecystectomy. This higher rate did not seem related to operative pathologic findings, but rather to the year the procedure was performed. After 1994, no laparoscopic conversions were needed, and this may be related to a greater facility with the procedure. Only one open procedure was initiated after 1992. No common bile duct explorations were performed despite the abnormal elevation of several preoperative laboratory values (eg, lipase, amylase, aspartate aminotransferase, and gamma-glutamyltransferase). Despite this finding, no evidence of common bile duct obstruction was found during preoperative screening, and no evidence of retained common duct stones was found postoperatively. This suggests that the elevations in the preoperative laboratory tests may not have the same clinical significance as during non-HIV disease. Elevation of these serum enzymes alone should not be an indication to perform cholangiography. Indeed, biliary colonization without common bile duct obstruction in patients with AIDS cholangiography (including cholecystitis) has been associated with nonspecific elevations in these laboratory values (8, 20). Interestingly, no significant advantage in length of stay was found for the laparoscopic procedure, and in general, the postoperative length of stay was longer than expected. This prolonged recuperation period may be the result of associated disease and the generally debilitated state of many of these patients, or both.

The 30-day operative mortality rate of 7 percent was higher than expected for age-matched patients without AIDS. This outcome is similar to the results of one review (8), but contrasts with the poor outcomes noted by others (2, 6, 7). Different patient selection criteria and delays in intervention most likely explain the discrepancies in outcome. Cholecystectomy is a well-tolerated operation in patients with AIDS and should be considered reasonable therapy for biliary disease. Anecdotal evidence suggests that some surgeons are reluctant to operate on patients with end-stage AIDS because they believe that the expected survival from underlying immune disease does not warrant the risks of anesthesia and operation.

As expected, patients with the most advanced immune disease were less likely to be alive at long-term follow-up. Despite very low CD4 cell counts, survivorship was quite acceptable. A CD4 cell count of less than 50 is normally considered an indicator of likely morbidity; but, 2 years after cholecystectomy, almost 50 percent of patients with such a CD4 cell count are alive. Cholecystitis among patients in the terminal phases of AIDS does not seem to be a premorbid event. Furthermore, control of symptoms can be obtained with acceptable impact on the patient. No assessment of the effects of cholecystectomy or anesthesia on the progression of AIDS can be made without a control group with similar characteristics.

Although a wide spectrum of AIDS-related cholangiopathy has been described, this study demonstrates that cholecystitis caused by cholelithiasis is a more common entity in immunocompromised patients than previously considered. Cholecystectomy is a well-tolerated intervention with few complications and favorable outcomes despite the severity of preexisting immune disease. Only by critically evaluating specific operative procedures as well as their complications and outcomes can the role of operative treatment in the spectrum of AIDS be determined. Cholecystectomy for biliary disease has a clear role in the treatment of the patient with AIDS.

ACKNOWLEDGMENT
We thank Kathleen Ahern, PhD, RN, for statistical analysis, and the attending surgeons at St. Vincent's Hospital (Christopher Mills, MD, Jose Corvalon, MD, Vincent Scarpinato, MD, Albert Burchell, MD, Jack Yee, MD, Steven Wong, MD, Jack Slattery, MD, James Pacholka, MD, and Marc Wallack, MD) who managed the care of this complex group of patients.

REFERENCES

1. Barone, J. E., Gingold, B. S., Nealon, T. F., and Arvanitis, M. L. Abdominal pain in patients with acquired immune deficiency syndrome. Ann. Surg., 1986; 204: 619-623.
2. LaRaja, R. D., Rothenberg, R. E., Odom, J. W., and Mueller, S. C. The incidence of intra-abdominal surgery in acquired immunodeficiency syndrome: a statistical review of 904 patients Surgery, 1989, 105: 175-179.
3. Edwards, P., Wodak, A., Cooper, D. A., et al. The gastrointestinal manifestations of AIDS. Aust. N. Z. J. Med., 1990, 20: 141-148.
4. Cappell, M. S. Hepatobiliary manifestations of the acquired immune deficiency syndrome. Am. J. Gastroenterol., 1991, 86: 1-15.
5. Cello, J. P. Acquired immunodeficiency syndrome cholangiopathy: spectrum of disease. Am. J. Med., 1989, 86: 539-546.
6. Wind, P., Chevallier, J. M., Jones, D., et al. Cholecystectomy for cholecystitis in patients with acquired immune deficiency syndrome. Am. J. Surg., 1994, 168: 244-246.
7. Hinnant, K., Schwartz, A., Rotterdam, H., and Rudski, C. Cytomegaloviral and cryptosporidial cholecystitis in two patients with AIDS. Am. J. Surg. Pathol., 1989, 13: 57-60.
8. Adolph, M. D., Bass, S. N., Lee, S. K., et al. Cytomegaloviral acalculous cholecystitis in acquired immune deficiency syndrome patients. Am. Surg., 1993, 59: 679-684.
9. Dolmatch, B. L., Laing, F. C., and Federle, M. AIDS related cholangitis: radiographic findings in nine patients. Radiology, 1987, 163: 313-316.
10. Radin, D. R., Cohen, H., and Halls, J. M. Acalculous inflammatory disease of the biliary tree in acquired immunodeficiency syndrome: CT demonstration. J. Comput. Assist. Tomogr., 1987, 11: 775-778.
11. McCarthy, M., Choudhri, A. H., Helbert, M., and Crofton, M. E. Radiological features of AIDS related cholangitis. Clin. Radiol., 1989, 40: 582-585.
12. Rusin, J. A., Sivit, C. J., Rakusan, T. A., and Chandra, R. S. AIDS-related cholangitis in children: sonographic findings. Am. J. Roentgenol., 1992, 159: 626-627.
13. Defalque, D., Menu, Y., Girard, J. P., and Coulaud, J. P. Sonographic diagnosis of cholangitis in AIDS patients. Gastrointest. Radiol., 1989, 14: 143-147.
14. Brunetti, J. C., van Heertum, R. L., Kempf, J. S., et al. Tc-99m DISIDA hepatobiliary scintigraphy in AIDS cholangitis. Clin. Nucl. Med., 1994, 19: 36-42.
15. Quinn, D., Popock, N., Freund, J., et al. Radionuclide hepatobiliary scanning in patients with AIDS-related sclerosing cholangitis. Clin. Nucl. Med., 1993, 18: 417-422.
16. Jacobson, M. A., Cello, J. P., and Sande, M. A. Cholestasis and disseminated cytomegalovirus disease in patients with the acquired immunodeficiency syndrome. Am. J. Med., 1988, 84: 218-224.
17. Farman, J., Brunetti, J., Baer, J. W., et al. AIDS related cholangiopancreatographic changes. Abdom. Imaging, 1994, 19: 417-422.
18. Benhamou, Y., Caumes, E., Gerosa, Y., et al. AIDS-related cholangiopathy. Dig. Dis. Sci., 1993, 38: 1113-1118.
19. Csillag, A. J., McKinnon, A., Todhunter, L. P., and Doust, B. D. Biliary dilatation in asymptomatic AIDS patients. Australasian Radiology, 1994, 38: 254-255.
20. Teixidor, H. S., Godwin, T. A., and Ramirez, E. A. Cryptosporidiosis of the biliary tract in AIDS. Radiology, 1991, 180: 51-56.
21. Benator, D. A., French, A. L., Beaudet, L. M., et al. Isospora belli infection associated with acalculous cholecystitis in a patient with AIDS. Ann. Intern. Med., 1994, 121: 663-664.
22. Keshavjee, S. H., Magee, L. A., Mullen, B. J., et al. Acalculous cholecystitis associated with cytomegalovirus and sclerosing cholangitis in a patient with acquired immunodeficiency syndrome. Can. J. Surg., 1993, 36: 321-325.
23. Finegold, M. J., and Carpenter, R. J. Obliterative cholangitis due to cytomegalovirus disease in patients with the acquired immunodeficiency syndrome. Am. J. Med., 1988, 84: 218-224.
24. Burt, A. D., Scott, G., Schiach, C. R., and Isles, C. G. Acquired immunodeficiency syndrome in a patient with no known risk factors. A pathologic study. J. Clin. Pathol., 1984, 37: 471-474.
25. Gherardi, R., Lebargy, F., Gauland, G., et al. Necrotizing vasculitis and HIV replication in peripheral veins. N. Engl. J. Med., 1989, 321: 685-686.
26. Baskerville, A., Ramsay, A. D., Millward-Sadler, G. H., et al. Chronic pancreatitis and biliary fibrosis associated with cryptosporidiosis in simian AIDS. J. Comp. Pathol., 1991, 105: 415-421.

Caption for figure 1: Kaplan-Meier survival estimates for all patients with human immunodeficiency virus or acquired immunodeficiency syndrome who underwent cholecystectomy.

Caption for figure 2. Kaplan-Meier survival estimates for patients with human immunodeficiency virus or acquired immunodeficiency syndrome after cholecystectomy and the relationship to the preoperative CD4 cell count. Significant improvements in survival were noted for patients with CD4 cell counts greater than 200/mL3 at 6, 12, 24, 36, 48, and 60 months after the operation when compared with the survival of patients with CD4 cell counts less than 200/mL3.

__________________________

Journal of the American College of Surgeons
March 1997, Volume 184, Number 3