Use of OncoScint scan to assess resectability of hepatic metastases from colorectal cancer.

Patients with limited hepatic metastases from colorectal cancer can potentially be cured by resection. A number of patients deemed resectable by standard imaging procedures are found to have extrahepatic disease at laparotomy and are thus unresectable. A test capable of identifying these patients would assist in better patient selection. OncoScint (Cytogen Corp, Princeton, NJ) scan targets colorectal cancer by interacting with a tumor-associated glycoprotein. Can OncoScint scan be used to reliably identify patients with extrahepatic disease preoperatively? Between February 1996 and August 1998 eight patients with colorectal metastases to the liver were enrolled prospectively. All patients received preoperative OncoScint scan (indium-111) and underwent laparotomy. The laparotomy findings were correlated with the results of OncoScint scan. In four of eight patients (50%) OncoScint scan showed no extrahepatic disease. This was confirmed at laparotomy. All of these patients underwent hepatic resection. One of eight patients (12.5%) had OncoScint findings suggestive of extrahepatic disease pathologically confirmed during laparotomy. Three of eight patients (37.5%) had OncoScint findings of extrahepatic disease not confirmed by laparotomy. All three patients underwent hepatic resection. One of the three patients is still disease free for more than 48 months after hepatic resection. If OncoScint scan had been used to determine resectability this patient with false positive scan would have been denied a potentially curative operation. Because of the unacceptably high false positive rate the study was terminated after eight patients. Because of its high false positive rate (37.5%) OncoScint scan is not a reliable test for the assessment of extrahepatic disease. Other tests need to be developed to accurately stage extrahepatic disease with an acceptably low false positive rate to prevent exclusion of patients who can potentially be cured.

Radiation after mastectomy in high-risk patients: is it necessary?

A retrospective review of cases from 1988 through 1992 was performed examining high-risk breast cancer patients treated with modified radical mastectomy without postoperative radiation at a single institution. Locoregional recurrence, distant metastases, overall survival, and number of lymph nodes removed were examined. This was compared with recent Danish and Canadian studies. Thirty-three premenopausal node-positive breast cancer patients had a 9 per cent locoregional recurrence rate. In the Danish and Canadian studies the locoregional recurrence rates were 32 and 21 per cent. These were reduced to 9 per cent and 10 per cent respectively in the radiated arms. Our locoregional recurrence in nonradiated patients was similar to that in the radiated arms of the studies and improved when compared with recurrence in their non-radiated controls. The adequacy of the axillary lymph node dissection was examined. In the current study a median of 18 lymph nodes were recovered with only 3 per cent containing less than 12 nodes. In the Danish study a median of seven lymph nodes were removed. Similarly in the Canadian trial a median of 11 nodes were removed. With complete axillary dissection results equivalent to those of postoperative adjuvant radiation is achieved. Further randomized controlled studies with standard axillary dissections are needed before the recommendation of routine postoperative radiotherapy.

Retroperitoneal soft-tissue sarcomas: prognosis and treatment of primary and recurrent disease.

Soft-tissue sarcomas of the retroperitoneum constitute a heterogeneous group of tumors with varying histology, potential for complete resection, and propensity for recurrent disease-making the development of effective treatment difficult and challenging. A retrospective review of 23 patients with retroperitoneal sarcomas from 1985 through 1998 was performed to assess the biological behavior and clinical outcomes and to identify factors that may influence prognosis and optimize treatment strategy. Liposarcomas were the most common pathology (61%); 79 per cent of these were of low grade. Leiomyosarcomas were the next most common pathology (30%); 43 per cent of these were of low grade. Low-grade sarcomas overall accounted for 62 per cent of the total group. Low-grade tumors independent of histologic type exhibited good prognosis for long-term survival with a median survival of 44 months. In contrast, intermediate- or high-grade tumors were associated with a median survival of only 9 months (P < 0.02). On the other hand, tumor histologic type independent of grade did not have a significant survival difference. Complete tumor resection was possible in 21 of 23 patients, which gives an overall resectability rate of 91 per cent. Eight patients (36%) remain disease-free after initial surgical treatment. However, local recurrence was common; this occurred in 11 of 22 patients (50%). Local recurrence, however, did not preclude long-term survival. Surgical resection of recurrent disease was done in nine patients with a median survival of 91 months (range 24-150 months). Three patients had as many as three operations for recurrent disease. With subsequent recurrences there was a decrease in interval from approximately 4 years to 2 years, and 33 per cent of these patients developed tumor dedifferentiation to high grade. An aggressive surgical approach with reoperation can produce prolonged survival in patients with low-grade retroperitoneal sarcoma.

Selective lymphoscintigraphy: a necessary adjunct to dye-directed sentinel node biopsy for breast cancer?

BACKGROUND: Dye-directed sentinel node biopsy (SNB) for breast cancer provides accurate staging with low morbidity, but for tumors distant from the axilla, its use has been questioned. HYPOTHESIS: Can preoperative breast lymphoscintigraphy (BL) applied selectively to medial hemisphere tumors predict a subset of patients who may not require surgical staging of the axilla? DESIGN: Prospective cohort study. SETTING: Tertiary, multidisciplinary breast center. PATIENTS: Thirty-two women with breast tumors located in the medial hemisphere (30) or inframammary crease (2). INTERVENTION: Peritumoral injection of 500 microCi of technetium Tc 99m sulfur colloid and biplanar imaging. Nonpalpable lesions were localized with ultrasound or mammography. At the time of definitive breast surgery, isosulfan blue dye-directed SNB was performed. Axillary dissection was performed when the SN contained a tumor or could not be identified. MAIN OUTCOME MEASURES: Regional nodal basins identified by BL; success rate of SNB. RESULTS: Preoperative BL demonstrated axillary drainage in 28 patients (88%); 2 patients (6%) had isolated internal mammary radionuclide uptake, and 2 patients had no nodal uptake. Dye-directed axillary SNB succeeded in 27 (87%) of 31 patients, including both patients with failed BL. Breast lymphoscintigraphy had predicted isolated internal mammary drainage in 2 of 4 patients whose SNs could not be identified. Metastases were found in 5 patients (16%). CONCLUSIONS: Axillary radionuclide uptake predicts but does not augment dye-directed SN identification. In those few patients with isolated internal mammary drainage, BL may obviate the need for surgical staging of the axilla.

Unresectable pancreatic carcinoma: correlating length of survival with choice of palliative bypass.

The preferred method of biliary bypass and the need for prophylactic gastroenterostomy in unresectable pancreatic carcinoma are dependent on the length of survival of the patient. From 1980 through 1996, 60 patients with biopsy-proven pancreatic cancer were found to be unresectable at exploration. The reasons for unresectability included major vascular involvement in 21 patients (35%), liver metastases in 16 (26.7%), celiac or portal lymph node metastases in 13 (21.7%), carcinomatosis in 5 (8.3%), and advanced age and/or comorbid medical condition in 4 patients (6.7%). One patient refused pancreaticoduodenectomy. Nine patients (15%) underwent Roux-en-Y choledochojejunostomy, and 51 (85%) underwent choledochoduodenostomy. Prophylactic gastroenterostomy was not performed routinely; however, in 9 patients (15%), gastrojejunostomy was performed for impending duodenal obstruction. Late biliary obstruction did not occur. Late gastric obstruction occurred in 6 of 51 patients (11.7%), at a median of 13.5 months after initial operation (range, 5-26 months). However, late gastric obstruction primarily occurred in 5 of 31 patients (16%) with locally advanced disease (major vessel involvement or lymph node metastases). The median survival was 12.0 months (range, 3.5-62 months) for patients with major vessel involvement, 11.5 months (range, 3-42 months) for patients with lymph node metastases, 4.5 months (range 0.5-24 months) for patients with liver metastases, 5.0 months (range, 4-7 months) for patients with carcinomatosis, and 9.0 months (range 2-27 months) for patients with significant comorbid medical illness and/or advanced age. Patients with liver metastases and carcinomatosis do not survive long enough to develop late obstruction. On the other hand, patients with locally advanced pancreatic carcinoma have a longer median survival and could be considered for prophylactic gastroenterostomy to avoid late gastric obstruction. Choledochoduodenostomy offers effective palliation for biliary obstruction.

Factors affecting early postoperative feeding following elective open colon resection.

HYPOTHESIS: If factors accounting for the inability to tolerate early postoperative feeding after elective open colon resection can be identified, then perhaps these factors can be modified to decrease future failures. DESIGN: Consecutive case series. SETTING: Tertiary referral center. PATIENTS: From 1993 to 1998, 200 consecutive patients undergoing elective open colon resection. INTERVENTION: Early postoperative feeding protocol consisting of clear liquids on the evening of postoperative day 2, regular diet on postoperative day 3, and discharged home as tolerated. A subgroup of patients was treated with metoclopramide. MAIN OUTCOME MEASURES: The ability to tolerate early feeding. Postoperative complications. Length of hospitalization. RESULTS: Twenty-seven (13.5%) of the 200 patients failed to tolerate early feeding. 16 patients (8.0%) were immediately unable to tolerate oral intake, whereas 11 patients (5.5%) initially tolerated early postoperative feeding but required hospital readmission due to emesis. There were no abdominal abscesses or anastomotic leaks. In patients who failed early feeding, no significant differences were noted for age, comorbid medical illness, operative time, or additional surgical procedures, when compared with patients who tolerated early oral intake. However, 18 (20.9%) of the 86 men failed early feeding, compared with 5 (6.8%) of the 73 women (P=.01). Additionally, patients undergoing total abdominal colectomy or total proctocolectomy (n = 11) failed 45.5% of the time, compared with 12.2% of the patients undergoing other types of colectomy (n = 189) (P = .01). The addition of metoclopramide therapy did not significantly improve the ability to tolerate early feeding. CONCLUSIONS: In patients undergoing elective open colon resection, early postoperative feeding is safe and effective, and produces a brief hospital stay compared with patients fed by traditional means. However, men and patients undergoing total abdominal colectomy are more likely to be intolerant of early postoperative feeding.

Safety and durability of single-layer, stentless, biliary-enteric anastomosis.

Biliary-enteric anastomosis has long been associated with significant complications of early bile leak, cholangitis, and late stricture formation, and controversy exists regarding which operative technique best prevents these problems. Biliary-enteric anastomosis was performed using a single-layer running 4-0 polyglactin (Vicryl) suture, without a transanastomotic stent, in 97 patients by a single surgeon over a 17-year period. Indications for operation included malignant obstruction (84.5%), benign stricture, choledocholithiasis, and choledochal cyst. The most common operation performed was a choledochoduodenostomy; the remaining operations were either Roux-en-Y choledochojejunostomy, hepaticoduodenostomy, or Roux-en-Y hepaticojejunostomy. Complications occurred in 14.1 per cent of patients; there was one perioperative death. There was only one case of anastomotic leak (1%), which resolved spontaneously within 1 week. Mean hospital stay was 8.7 days. The mean follow-up was 13.1 months in all patients. Among patients with benign disorders of the biliary tract, the mean follow-up was 21 months, during which time no patient developed an anastomotic stricture. One patient experienced postoperative cholangitis, although not as a result of anastomotic stricture. Biliary-enteric anastomosis for both benign and malignant disorders can be safely performed using a running, absorbable suture without a stent.

Prognosis and treatment of bile duct carcinoma.

Bile duct carcinomas present a therapeutic challenge because of different histologies, tumor locations, and resectabilities. The goal of our study was to identify prognostic factors to better delineate therapeutic options. Forty patients (30 males and 10 females) diagnosed with bile duct cancer, treated between 1985 and 1996, at Kaiser Permanente Medical Center, Los Angeles were retrospectively reviewed. Three prognostically significant variables were identified: tumor histology, tumor location, and resection. Papillary histology was the most significant determinant of long-term survival. Of six patients with papillary adenocarcinoma, four patients (67%) underwent resection, with all four achieving long-term survival. Lower-third lesions also demonstrated a survival advantage. Four out of 12 (33%) lower-third tumors were resected, with a median survival of 36 months. Irrespective of tumor histology or tumor location, tumor resection always afforded longer survival times than did palliative treatments. A prognostic classification system based on weighted values of significant variables is presented that accurately predicted long-term survival. In conclusion, bile duct tumors in general are incurable, except perhaps for a small subset of patients with papillary adenocarcinoma. Papillary histology is the most significant determinant of ultimate survival and cure. A multifunctional prognostic classification system can be helpful for this perplexing disease.

Choledochoduodenostomy for palliation in unresectable pancreatic cancer.

OBJECTIVE: To determine whether choledochoduodenostomy provides adequate long-term palliation of obstructive jaundice in unresectable pancreatic cancer. DESIGN: Consecutive case series. SETTING: Tertiary referral center. PATIENTS: From 1980 to 1997, 79 consecutive patients (45 men, 34 women; mean age, 67.8 years) with biopsy-proved pancreatic cancer found to be unresectable at operation. INTERVENTION: All patients had resectable disease by preoperative criteria. At exploratory laparotomy, unresectability was determined by the presence of liver or peritoneal metastases, encasement of major vascular structures by tumor, and/or celiac lymph node involvement. Choledochoduodenostomy for biliary bypass was performed in 71 (90%) of 79 patients; Roux-en-Y choledochojejunostomy was performed in the remaining 8 patients. MAIN OUTCOME MEASURES: Resolution of jaundice, duration of hospital stay, mean survival, postoperative complications, and evidence of recurrent biliary obstruction. RESULTS: All patients experienced rapid resolution of jaundice. Average hospital stay was 8.3 days. Mean survival after operation was 13.1 months (range, 2 weeks to 62 months). Postoperative mortality was 3%. There were no biliary or duodenal leaks. Four patients (6%) required hospitalization for gastrointestinal hemorrhage; however, only 1 (1%) was from peptic ulceration. No patient developed recurrent biliary obstruction. CONCLUSIONS: Choledochoduodenostomy provides rapid, long-lasting relief of jaundice, with little morbidity and a low rate of duodenal ulceration, and is the palliative operation of choice when patients are found to have unresectable disease at operation or when stenting procedures fail.

Segmental mastectomy and tamoxifen alone provide adequate locoregional control of breast cancer in elderly women.

We wished to determine whether tamoxifen and local excision without breast radiation or axillary lymph node dissection provides adequate local and regional control of breast cancer in elderly women. The records of 36 women with breast cancer who were more than 70 years old and were treated only with tamoxifen and local excision from January 1985 to July 1996 were retrospectively reviewed. These patients had refused, or were considered too ill for, standard therapy. The mean follow-up was 44.1 months. Twenty-two (61%) were alive without disease, and six (17%) died of unrelated causes, without recurrence. Two (6%) were alive with metastasis, and five (14%) died with metastasis. One patient developed a breast recurrence, which was reexcised. A second patient developed metastasis and axillary recurrence, which was treated with modified radical mastectomy. Pathologic grade, tumor size, and estrogen receptor and margin status were not predictive of recurrence. In conclusion, despite the omission of breast radiation and axillary dissection, there were only two locoregional recurrences, and both were easily treated surgically. In this select group of patients, local excision and tamoxifen provided adequate locoregional control of breast cancer in elderly women.

Surgical management of massive splenomegaly.

Recent reports document an increased mortality and morbidity associated with splenectomy for massive splenomegaly (<1000 g), with a morbidity that is 2 to 10-fold higher than that seen for splenectomy for normal size spleens. Preoperative angiographic embolization of the splenic artery has been advocated as a means to decrease this morbidity and mortality. In a retrospective review of 100 splenectomies performed at Kaiser Permanente (Los Angeles, CA), 20 were performed for splenomegaly, average weight 1811 g (1050-3700 g), and 80 were normal sized spleens. Mortality for normal sized spleens is 1.25 per cent, and for those performed for splenomegaly is zero. Likewise, the morbidity for splenectomy of normal sized spleens was 21.25 per cent, but for splenomegaly, a 20 per cent morbidity rate was observed. Average blood loss with splenomegaly was 696 mL, slightly higher than the 600 mL blood loss for normal sized spleens. Sixty-seven per cent of patients with splenomegaly required no transfusion, and none required more than two units. Of patients with normal size spleens, 75 per cent required no transfusion, and 96 per cent required two units or less. Splenectomy for splenomegaly is possible without an increase in morbidity or mortality. In this series, preoperative embolization was not performed; however, the morbidity and mortality rates compare favorably with series in which it was performed. Preoperative embolization of the splenic artery may be unnecessary and may expose the patient to additional expense, risk, and discomfort.

Safe and effective early postoperative feeding and hospital discharge after open colon resection.

Recent articles have stressed early postoperative feeding and hospital discharge as major benefits of laparoscopic colon surgery. From March 1993 to December 1994, an early feeding protocol after open colon resection consisting of clear liquid diet on postoperative Day (POD) 2, then advancing to regular diet on POD 3, and discharging home as tolerated was applied to 41 patients (Group A). We reviewed the charts of 41 consecutive patients from January 1992 to February 1993 who were operated immediately before the protocol and whose diet was started by traditional methods (Group B). Both groups were similar in age and types of procedures performed. Clear liquid diet was started earlier in Group A than Group B (all patients on POD 2 versus average POD 4.9 (range, 4-7 days), but it was tolerated by a similar number of patients in both groups (90% versus 85%). The mortality and morbidity in both Groups were similar. In Group B, four patients (9.8%) did not tolerate diet and needed nasogastric tube, whereas none required nasogastric tube in Group A. The average hospital stay was 4.2 days (range, 3-8 days) in group A versus 6.7 days (range, 5-34 days) in Group B. In Group A, 67 per cent were discharged home by POD 4 versus none in Group B. Neither group had readmission within 2 weeks for recurrent nausea or vomiting. The early postoperative feeding and hospital discharge are safe and effective after open colon surgery.

Breast cancer in pregnancy and lactation.

Although breast cancer is the most common malignancy in pregnancy, its overall incidence remains low. It appears that pregnancy and breast cancer are merely coincidental and that pregnancy does not directly contribute to the development or accelerated progression of breast cancer. The majority of studies have documented a significant delay in diagnosis secondary to physiologic changes of the breast during pregnancy and have reasoned that this is the likely explanation for the advanced stage of disease upon initial presentation. Although pregnant patients present at a later stage of breast cancer, survival stage for stage is the same when pregnant patients are compared with young nonpregnant patients with breast cancer. A suspicious breast mass in a pregnant patient should be biopsied and appropriately treated, without need for extensive preoperative staging. Therapeutic abortion should be performed only on an individual basis, namely in patients in whom necessary radiation or chemotherapy would be detrimental to the developing fetus and in whom a significant delay of this treatment would be harmful. In patients with early-stage disease, it is recommended to wait 2 years after treatment of breast cancer for subsequent pregnancy; however, in women with advanced disease, subsequent pregnancy should be discouraged.

Methods to decrease the morbidity of abdominoperineal resection.

Recent reports stress that abdominoperineal resections (APR) are associated with many complications, including hemorrhage, long hospitalization, and delayed closure of an open perineal wound. Thirty-five patients underwent an abdominoperineal resection for cancer at Kaiser Permanente Medical Center, Los Angeles, from January 1989 to December 1993. All patients, except two, had their perineal wound closed after closure of the peritoneum and insertion of closed system suction catheters. Ninety-one per cent of patients achieved successful primary healing. Three patients (8.5%) suffered perineal wound dehiscence. Overall morbidity was 55 per cent, with urinary retention being the most common, occurring in 23 per cent of patients. This was managed successfully by early in-and-out self catheterization. There was no incidence of urinary tract infections. There were no operative deaths. Length of stay averaged 8.6 days, with a median of 7 days. Five patients had previous radiation therapy. Of those, two (40%) had perineal wound dehiscence, compared to only one of 33 (3.3%) patients without previous radiation. APRs can be done with minimal mortality, although with an increased morbidity in irradiated patients. Primary closure and drainage of the perineal wound significantly lowers the complication rate, as does early Foley removal and self in-and-out catheterization.

Role of prophylactic gastroenterostomy for unresectable pancreatic carcinoma.

Methods of palliation and the use of prophylactic gastroenterostomy in the treatment of unresectable pancreatic carcinoma remain controversial. Gastroenterostomy has been linked with various complications. We conducted a 10-year (1982-1992) retrospective review of patients who had unresectable pancreatic carcinoma and underwent biliary decompression without prophylactic gastroenterostomy. 50 patients were studied. Only four patients (8%) developed duodenal obstruction and required reoperation for therapeutic gastroenterostomy. The mean time to obstruction was 15.75 months, whereas the mean overall survival was 12.99 months. The mean survival of patients who underwent therapeutic gastroenterostomy was 32.25 months, with an average palliation of 16.5 months after the second operation. We conclude that pancreatic carcinoma has a rapid natural progression, and most patients do not survive long enough to obstruct. The ones who do obstruct are unique in that they survive for a long period of time. We recommend that routine prophylactic gastroenterostomy is unnecessary, and selective use of gastroenterostomy should be exercised in case of present or impending duodenal obstruction.

Gastrointestinal complications associated with cardiopulmonary bypass procedures.

Gastrointestinal complications after cardiopulmonary bypass (CPB) procedures are rare, but when they do occur, they carry a significant incidence of morbidity and mortality. Over a 5-year period spanning 1988-1992, 4923 CPB procedures were performed and 64 patients were identified who suffered a GI complication, giving an incidence of 1.3 per cent. The most frequent complications were GI bleeding (40%) and pancreatitis (34%). Other complications included acute cholecystitis (11%), perforated duodenal ulcer (8%), ischemic bowel (5%), and diverticulitis (2%). Complications occurred most frequently in patients undergoing procedures with longer pump and cross-clamp times, such as valvular and combination (CABG/valve) procedures. Redo procedures and the use of an intra-aortic balloon pump increased the risk of developing a GI complication 2.5 and 12 times, respectively. Patients were treated aggressively both medically and surgically, but suffered a higher mortality (16%) as compared to those not suffering a GI complication (3%). We conclude that GI complications after CPB procedures are infrequent but lethal. Clinical features are often subtle, and a high index of suspicion is needed for early diagnosis and aggressive treatment.

Sacral resection. Operative technique and outcome.

Tumors involving the sacrum are difficult to treat. These include both primary tumors of the sacrum and locally invasive colorectal carcinomas. Sacral resection is often the only effective alternative for meaningful palliation or cure of sacral tumors. A review of 20 cases of sacral resections for primary sacral tumors (8) and locally invasive anorectal cancers (12) is presented. The mortality (0%) and morbidity (35% urinary complications, 25% wound disruptions, 1600-mL median blood loss) compare favorably with reports in the literature. Long-term survival was achieved with primary tumors of the sacrum. Local control of disease was achieved in the majority of patients with rectal cancer, with good palliation of preoperative pain. Long-term survival, however, is rare in this group. Surgical resection of sacral tumors can be undertaken with acceptable morbidity and mortality in selected patients.

Treatment of locally recurrent breast carcinoma.

Treatment of locally recurrent breast carcinoma remains a significant problem. The records of 106 patients with local chest wall recurrence were reviewed. Fifty-five percent eventually developed metastatic disease, while 45% remained free of systemic disease. Size of primary tumor (greater than 2 cm), number of recurrences (multiple), and disease-free interval from primary surgery (less than 2 years) were all highly significant for the development of metastatic disease. Negative estrogen receptors also predicted bad prognosis. Both irradiation and surgery used alone had high local failure rates of 83% and 62%, respectively, but combination radiation-surgery treatment failed only in 25%. Combination radiation-surgery treatment should be considered in patients with local recurrence, but further prospective trials with more patients will be needed to prove its effectiveness. Patients with unfavorable prognostic factors should be considered for adjuvant chemotherapy.

Prevention of complications in permanent central venous catheters.

As more patients are requiring permanent central venous catheters (PCC) for long term venous access, several associated complications have become evident, including: 1, sepsis; 2, thrombophlebitis; 3, insertion complications, such as unsuccessful placement, bleeding and pneumothorax, and 4, PCC transection with tip embolization. At our institution, 162 PCC were placed by way of cutdown or percutaneously. Sepsis occurred in 20 per cent (0.13 septic episodes per 100 catheter days), nearly always involving immunocompromised patients. Twenty-five per cent resolved with use of antibiotics and without removal of PCC. Two patients presented with clinical thrombophlebitis; both were treated with removal of PCC and anticoagulant medication. Failure of insertion was highest with the cephalic cutdown approach, and pneumothorax was highest with the subclavian approach. Transection of PCC is associated with the percutaneous subclavian approach and is heralded by intermittent catheter function and a "pinch-off" sign on roentgenogram. Methods of preventing these complications are emphasized herein.

The surgeon's role in treating acquired immunodeficiency syndrome.

The ever-increasing number of patients with acquired immunodeficiency syndrome (AIDS) will involve more surgeons in their diagnosis and treatment. The surgeon should be aware of the cause of AIDS, mode of transmission, method of diagnosis, usual cutaneous and abdominal manifestations, complications needing operative procedures, and precautions needed during surgery and the postoperative period. The gravity of AIDS requires the surgeon to be aware of the potential risks to other surgical patients by contaminated blood transfusions. From 110 cases of AIDS, we analyzed the indications, types of surgical procedures, and effect on final outcome in patients with AIDS.