To determine whether continuing medical education affects the quality of clinical care, we randomly allocated 16 Ontario family physicians to receive or not receive continuing-education packages covering clinical problems commonly confronted in general practice. Over 4500 episodes of care, provided before and after study physicians received continuing education, were compared with preset clinical criteria and classified according to quality. Although objective tests confirmed that the study physicians learned from the packages, there was little effect on the overall quality of care. When the topics were of relatively great interest to the physicians, the control group (who did not receive the packages) showed as much improvement as did the study group. When the topics were not preferred, however, the documented quality of care provided by study physicians rose (P less than 0.05) and differed from that provided by control physicians (P = 0.01). Finally, there was no spillover effect on clinical problems not directly covered by the program. In view of the trend toward mandatory continuing education and the resources expended, it is time to reconsider whether it works.

In a prospective, controlled study, 50 children with frequently relapsing nephrotic syndrome who had steroid toxicity were treated for eight weeks with either cyclophosphamide (2 mg per kilogram of body weight per day) or chlorambucil (0.15 mg per kilogram per day), in combination with prednisone in tapering doses. Of the 34 children shows relapses had tended to occur after the prednisone dosage had been reduced or immediately after the drug was discontinued (the steroid-dependent group), 22 also had early relapses after cytotoxic-drug treatment. In contrast, cytotoxic drug treatment produced long-lasting remissions in 12 of the 16 children whose relapses usually occurred after prednisone treatment had been interrupted for more than 14 days (the non-steroid-dependent group). The difference in response between the two groups was highly significant (P less than 0.001). We conclude that patients with frequent relapses without steroid dependence can be treated successfully with an eight-week course of cytotoxic drugs, whereas those with steroid-dependent nephrotic syndrome do not profit from cytotoxic drugs in the low doses used.

Acute graft-versus-host disease is a major problem in allogeneic bone-marrow transplantation. We performed a randomized study to compare the effectiveness of two regimens in the prevention of acute graft-versus-host disease. Thirty-five patients received methotrexate alone, and 32 received methotrexate, antithymocyte globulin, and prednisone. Of the patients who received methotrexate alone, 48 per cent had acute graft-versus-host disease, as compared with 21 per cent of those who received methotrexate, antithymocyte globulin, and prednisone (P = 0.01). The age of the recipient was a significant factor in the development of acute graft-versus-host disease: Older patients had a higher incidence of the disease (P = 0.001). We conclude that the combination of methotrexate, antithymocyte globulin, and prednisone significantly decreased the incidence of acute graft-versus-host disease and should be used to prevent this disorder in patients receiving allogeneic marrow transplants.

We evaluated D-penicillamine in the treatment of primary biliary cirrhosis. In a prospective double-blind trial, 26 patients received D-penicillamine (250 mg four times a day), and 26 received an identical placebo. Although the desired urinary excretion of copper was achieved in patients taking D-penicillamine, there was no improvement in survival or symptoms after 28 months. Serum bilirubin and alkaline phosphatase increased equally in both groups. Alanine and aspartate aminotransferases were lower in the D-penicillamine group, but serum albumin was also lower in this group. Liver histology worsened equally in both groups. Major side effects, some appearing more than 24 months after the start of treatment, occurred in 31 per cent of the patients receiving D-penicillamine. Less serious side effects occurred in an additional 46 per cent. We conclude that D-penicillamine at the dosage we used is not effective in the treatment of primary biliary cirrhosis and is associated with a high incidence of serious side effects.

Previously, we compared fixed low doses of heparin with adjusted doses of warfarin for the long-term treatment of venous thrombosis; in that study low-dose heparin was ineffective in preventing recurrence in patients with proximal-vein thrombosis. We have now completed a randomized trial comparing adjusted doses of heparin and of warfarin for prevention of recurrent venous thromboembolism in patients with proximal-vein thrombosis. One hundred six consecutive patients with acute proximal-vein thrombosis confirmed by venography were treated with intravenous heparin and then randomized to secondary prophylaxis. Two of 53 patients receiving heparin, as compared with one of 53 receiving warfarin, had new episodes of objectively documented venous thromboembolism. Nine patients taking warfarin had bleeding complications (which were major in three patients), as compared with one patient taking heparin (P = 0.008). Our data indicate that adjusted-dose subcutaneous heparin therapy provides an effective alternative to warfarin sodium and is associated with a lower risk of bleeding.

It has been suggested that because propranolol decreases portal venous pressure, it may prevent gastrointestinal bleeding associated with portal hypertension. We randomly assigned 74 patients with cirrhosis, who were admitted because of gastrointestinal bleeding, to either oral propranolol given in doses that reduced the heart rate by 25 per cent (38 patients) or to a placebo (36 patients). The proportion of patients free of recurrent gastrointestinal bleeding one year after inclusion in this study was 96 per cent in the propranolol group and 50 per cent in the placebo group (P less than 0.0001). We conclude that continuous administration of propranolol by mouth is effective in preventing recurrent gastrointestinal bleeding in patients with cirrhosis.

We performed a double-blind study in 101 preterm infants who weighed less than or equal to 1500 g at birth, who had respiratory distress, and who survived for at least four weeks, to evaluate the efficacy of oral vitamin E in preventing the development of retrolental fibroplasia. Weekly indirect ophthalmologic examinations begun when the infants were three weeks old revealed a significant decrease in the incidence of retrolental fibroplasia greater than or equal to Grade III (P less than 0.03) and greater than or equal to Grade II (P less than 0.05) (McCormick classification) in the 50 infants given 100 mg of vitamin E per kilogram of body weight per day as compared with 51 given 5 mg per kilogram per day (controls). When multivariate analysis was applied to the controls, five risk factors were identified: gestational age, level and duration of administration oxygen, intraventricular hemorrhage, sepsis, and birth weight. When multivariate analysis was applied to both control and treatment groups, the severity of retrolental fibroplasia was found to be significantly reduced in infants given 100 mg of vitamin E (P = 0.012).

In a study of the effectiveness of high intravenous doses of metoclopramide as an antiemetic, 41 patients with advanced cancer who were being treated with cisplatin were entered into two double-blind trials. In the first trial patients were randomly assigned to receive either metoclopramide or placebo, and in the second trial they received either metoclopramide or prochlorperazine. Patients given metoclopramide had significantly fewer episodes of emesis than patients given placebo (medians, 1.0 vs. 10.5; P = 0.001) or prochlorperazine (medians, 1.5 vs. 12.0; P = 0.005). Metoclopramide was superior to placebo and to prochlorperazine in reducing the volume of emesis (P = 0.001 and P = 0.022, respectively) and was more effective than placebo in shortening the duration of nausea (P = 0.042) and vomiting (P = 0.028). Side effects from metoclopramide were minor, with mild sedation frequently observed; one patient had a brief extrapyramidal reaction. We conclude that metoclopramide in high intravenous doses has greater antiemetic activity than placebo or prochlorperazine in patients receiving cisplatin chemotherapy.

We concluded a prospective study of the value of maintenance therapy in disseminated testicular cancer after chemotherapy-induced complete remission or chemotherapy cytoreduction followed by surgical resection of residual teratoma. A group of 171 patients were randomized to treatment consisting of cisplatin, vinblastine, and bleomycin, or these drugs plus doxorubicin. There was no apparent difference between these two induction regimens. Complete remission was achieved in 113 patients (66 per cent), and 19 (11 per cent) were free of disease by surgical resection of residual tumor. Of the 171 who started, 113 were eligible to receive either maintenance doses of vinblastine (58 patients) or no further therapy (55 patients) after remission-induction therapy. There was a 9 per cent relapse rate during maintenance with vinblastine and a 7 per cent relapse rate with no maintenance therapy; the overall relapse rate was 8 per cent (nine of 113). Our data indicate that maintenance therapy is unnecessary in disseminated testicular cancer.

To evaluate the role of prophylactic granulocyte transfusions during remission-induction chemotherapy for acute myelogenous leukemia (AML) we randomized 102 infected patients either to receive daily granulocyte transfusions when blood granulocytes fell below 0.5 x 10(9) per liter (54 patients) or not to receive them (48). Although the percentage of patients acquiring any infection was similar in the transfusion and control groups (46 and 42 per cent, respectively), granulocyte transfusions decreased the proportion of patients with bacterial septicemia (9 per cent of those with transfusions vs. 27 per cent of the controls; P = 0.01). Granulocyte transfusions did not reduce the incidence of other infections or improve bone-marrow recovery, remission rate and duration, or survival. Seventy-two per cent of the patients given transfusions had transfusion reactions. Pulmonary infiltrates were more common in the transfusion group than in the control group (57 per cent vs. 27 per cent; P = 0.002). Thirty-five per cent of the patients with pulmonary filtrates died, as compared with 5 per cent of those without filtrates. We conclude that prophylactic granulocyte transfusions should not be used during remission-induction chemotherapy in AML because the risks outweigh the benefits.

We randomized 96 postmenopausal women with metastatic breast carcinoma to receive surgical adrenalectomy or medical therapy with an adrenal inhibitor, aminoglutethimide (AG), plus replacement hydrocortisone. Before randomization, women were stratified according to disease-free interval, site of dominant disease, and estrogen-receptor status. Of 40 evaluable women treated with AG and hydrocortisone, 53 per cent had objective responses, as compared with 45 per cent of 29 women undergoing surgical adrenalectomy (P value not significant). Responses lasted a mean of 17.2 months in the medical group and greater than 17.1 months in the surgical group (not significant). Estrogen levels fell similarly in response to either treatment, whereas AG and hydrocortisone preserved androgen production. A null hypothesis tested the single question asked by this study: "Is surgical adrenalectomy superior to treatment with AG and hydrocortisone?" Rejection at significance levels of P = 0.01 and P = 0.07 for differences of 20 per cent and 10 per cent, respectively, suggested that medical therapy with AG and hydrocortisone may be logically chosen in place of surgical adrenalectomy.

Within the past 15 years, regional neonatal-intensive-care programs have been introduced and have expanded rapidly. The efficacy of some of the individual interventions that constitute neonatal intensive care has been validated in randomized, controlled clinical trials. It is therefore generally assumed that neonatal-intensive-care programs that incorporate these maneuvers are effective in reducing death and disability. However, the overall effectiveness of these programs has not been tested experimentally. Moreover, much of the non-experimental evidence supporting their value is based on the experience of referral units and does not measure the impact on the populations they serve. A definitive economic evaluation of neonatal intensive care has not yet been reported, despite the high cost of such programs. We conclude that neonatal-intensive care programs require further evaluation with rigorous scientific methods.

Prolonged bed rest after myocardial infarction is thought to result in deconditioning, manifested by increased heart-rate and blood-pressure responses to exercise and decreased functional capacity. We studied the effects of early, supervised exercises in preventing deconditioning after acute myocardial infarction. Eighty-four patients were randomized to a control group and 174 to an exercise group. Enrollment in the exercise program occurred an average of 4.5 days after admission (range, one to nine). Discharge from the hospital occurred an average of 10.3 days after admission in the control group and 10.4 days in the exercise group. Most patients had a low-level treadmill test on the day before hospital discharge. There were no differences between the two groups in the clinical, hemodynamic, or electrocardiographic responses to the treadmill test. Incidences of complications and deaths (one death in each group) during hospitalization were not significantly different in the two groups, although six patients (3 per cent, all in the exercise group) required cardiac surgery--four because of recurrent chest pain and two because of rupture of heart muscle. Thus, we were unable to demonstrate any significant beneficial or deleterious effects of an early, in-hospital exercise program.

From 1973 to 1980, we carried out a controlled study at the National Cancer Institute in Milan to consider the value of a conservative procedure in patients with breast cancer of small size. We randomized 701 patients with breast cancer measuring less than 2 cm in diameter and with no palpable axillary lymph nodes to Halsted radical mastectomy or to "quadrantectomy" with axillary dissection and radiotherapy to the ipsilateral residual breast tissue. We treated 349 patients with Halsted mastectomy and 352 with quadrantectomy. The two groups were comparable in age distribution, size and site of primary tumor, menopausal status, and frequency of axillary metastases. There were three local recurrences in the Halsted group and one in the quadrantectomy group. Actuarial curves showed no difference between the two groups in disease-free or overall survival. From these results, mastectomy appears to involve unnecessary mutilation in patients with breast cancer of less than 2 cm and no palpable axillary nodes.

Fifty-three profoundly granulocytopenic patients with relapsed acute leukemia who were undergoing reinduction chemotherapy were prospectively randomized to receive either trimethoprim-sulfamethoxazole plus nystatin or gentamicin plus nystatin for prevention of infections. The acquisition of new organisms per patient during the total study period was similar in both groups. Thirty-five symptomatic infections (five of which were bacteremias) occurred in patients receiving trimethoprim-sulfamethoxazole plus nystatin, whereas 31 infections (eight bacteremias) occurred in patients receiving gentamicin plus nystatin. Four deaths related to infection occurred in patients taking trimethoprim-sulfamethoxazole, and eight occurred in patients taking gentamicin. We conclude that trimethoprim-sulfamethoxazole plus nystatin was approximately as effective as gentamicin plus nystatin for prophylaxis against infection in relapsed acute leukemia. Furthermore, side effects were fewer and compliance was better with trimethoprim-sulfamethoxazole plus nystatin.

To determine whether routine early endoscopy is beneficial to patients with upper-gastrointestinal-tract bleeding that ceases during hospitalization, we randomly assigned 206 patients to routine endoscopy (100 patients) or no routine endoscopy (106). Patients in the latter group underwent endoscopy only if recurrent bleeding occurred during hospitalization or if x-ray films disclosed gastric ulcer or suggested neoplasia. All patients were initially treated with an empiric antacid regimen. When the two groups were compared (experimental versus control), there were no significant differences in overall hospital deaths (11 versus eight), recurrence of bleeding (33 versus 32), number of transfusions required to treat recurrent bleeding (mean +/- S.E.M., 7.4 +/- 1.2 versus 6.3 +/- 0.7 units), deaths after recurrent bleeding (eight versus five), or duration of hospital stay. During the 12 months after discharge, there were also no significant differences in frequency of readmission to the hospital, incidence of further gastrointestinal bleeding, number of hemorrhage-related deaths, or frequency of gastrointestinal surgery. We conclude that endoscopy should not be a routine procedure in patients with upper-gastrointestinal-tract bleeding that ceases during treatment.

To compare the erythropoietic effects of nandrolone decanoate, testosterone enanthate, oxymetholone, and fluoxymesterone, we performed a randomized clinical trial in patients with anemia who were receiving maintenance hemodialysis (the women were not given testosterone enanthate). After a control period of at least two months, patients received one of the drugs for six months and then returned to control status; a second and third drug were administered in a similar fashion. Seventy-seven patients completed the first drug period, 56 the second, and 35 the third. The response to nandrolone and testosterone enanthate, the two drugs given by injection, was clearly superior to the response to oxymetholone or fluoxymesterone, given by mouth, in terms of the percentage of patients responding and the mean rise in hematocrit. Approximately half the patients had an increase of at least 5 percentage points in hematocrit after an injectable androgen was given; more than half the women responded. Patients who required transfusions regularly and those who had bilateral nephrectomies did not respond.

To assess the efficacy and safety of verapamil in variant angina pectoris, we entered 16 patients in a double-blind, randomized trial of nine months, duration. During treatment with verapamil, the frequency of angina fell substantially (12.6 +/- 25.9 chest pains per week with placebo, 1.7 +/- 2.8 pains per week with verapamil, mean +/- S.D.; P less than 0.01), as did the use of nitroglycerin tablets (14.4 +/- 34.4 tablets per week with placebo, 2.1 +/- 3.3 tablets per week with verapamil; P less than 0.05). The number of hospitalizations for clinical instability was significantly lower with verapamil (P less than 0.01). The number of episodes of transient ST-segment deviation during treatment with verapamil was reduced (33.1 +/- 39.3 ST-segment deviations per week with placebo, 7.7 +/- 11.7 deviations per week with verapamil; P less than 0.01). Verapamil caused no side effects forcing a reduction in dosage or a discontinuation. We conclude that verapamil is safe and effective in the therapy of variant angina pectoris.

A multicenter double-blind randomized study was carried out to compare the effect of timolol (10 mg twice daily) with that of placebo in patients surviving acute myocardial infarction. Treatment was started seven to 28 days after infarction in 1884 patients (945 taking timolol, and 939 placebo), who represented 52 per cent of those evaluated for entry; the patients were followed for 12 to 33 months (mean, 17). There were 152 deaths in the placebo group and 98 in the timolol group. When deaths that occurred during treatment or within 28 days of withdrawal were considered, the cumulated sudden-death rate over 33 months was 13.9 per cent in the placebo group and 7.7 per cent in the timolol group--a reduction of 44.6 per cent (P = 0.0001). The cumulated reinfarction rate was 20.1 per cent in the placebo group and 14.4 per cent in the timolol group (P = 0.0006). We conclude that long-term treatment with timolol in patients surviving acute myocardial infarction reduces mortality and the rate of reinfarction.