|
|
| |||||||||||||||||||||||||||||||
Background Pneumocystis carinii pneumonia and toxoplasmic encephalitis are frequent life-threatening opportunistic infections in patients with human immunodeficiency virus (HIV) infection. Primary prophylaxis against P. carinii pneumonia is now common, but there are few data on regimens for primary prophylaxis against toxoplasmosis.
Methods We conducted a randomized trial that compared two prophylactic regimens: dapsone (50 mg per day) plus pyrimethamine (50 mg per week) was compared with aerosolized pentamidine (300 mg per month). The patients had symptomatic HIV infection, no history of P. carinii pneumonia or symptomatic toxoplasmosis, and CD4+ counts below 200 per cubic millimeter (0.2 x 109 per liter).
Results In an intention-to-treat analysis, after a median follow-up of 539 days P. carinii pneumonia developed in 10 patients in each group, whereas toxoplasmosis developed in 32 of 176 patients in the pentamidine group and 19 of 173 patients in the dapsone-pyrimethamine group. Those assigned to pentamidine had a risk of P. carinii pneumonia that was similar to the risk in those assigned to dapsone-pyrimethamine (adjusted relative risk, 1.13; 95 percent confidence interval, 0.44 to 2.92; P = 0.79), but a higher risk of toxoplasmosis (adjusted relative risk, 1.81; 95 percent confidence interval, 1.12 to 2.94; P = 0.02). Among the 262 patients with serologic evidence of past exposure to Toxoplasma gondii, the relative risk of symptomatic toxoplasmosis was 2.37 times higher in those assigned to pentamidine (95 percent confidence interval, 1.3 to 4.4; P = 0.006). More patients discontinued dapsone-pyrimethamine than pentamidine because of toxicity (42 vs. 3; P<0.001). Survival was similar in the two groups.
Conclusions For primary prevention of P. carinii pneumonia, dapsone-pyrimethamine is as effective, though not as well tolerated, as aerosolized pentamidine. Dapsone-pyrimethamine also prevents first episodes of toxoplasmosis.
In contrast to the declining incidence of P. carinii pneumonia, the incidence of toxoplasmic encephalitis, the most frequent presentation of opportunistic toxoplasmosis, is increasing markedly24,25. The development of opportunistic toxoplasmosis is mainly thought to be due to the reactivation of latent infection; this life-threatening disease is expected to develop in about one third of HIV-infected persons with antibodies to Toxoplasma gondii26. The highest incidence of toxoplasmic encephalitis in patients with the acquired immunodeficiency syndrome (AIDS) has been reported in Europe,25,27 where the seroprevalence of T. gondii in adults may exceed 70 percent28. In France, for example, it is the first AIDS-defining manifestation in 23 percent of patients25. Because aerosolized pentamidine has no activity against T. gondii and the efficacy of trimethoprim-sulfamethoxazole remains to be proved, there is a great need for an effective prophylaxis. The drug combinations used to treat toxoplasmosis (pyrimethamine-sulfadiazine and pyrimethamine-clindamycin) are poor candidates for primary prophylaxis because of their side effects29,30. Sulfone compounds inhibit the T. gondii dihydropteroate synthetase,31 and dapsone is synergistic with pyrimethamine both in vitro and in animal models32,33.
In a randomized trial, we compared aerosolized pentamidine with the dapsone-pyrimethamine combination for the primary prevention of P. carinii pneumonia and assessed the value of dapsone-pyrimethamine as primary prophylaxis against opportunistic toxoplasmosis.
Methods
Study Population
The study design was approved by the ethics committee of the Hopital Bichat-Claude Bernard, Paris, and all participants gave written informed consent before enrollment. The participants were adults with HIV infection classified as belonging to group IV according to the definition of the Centers for Disease Control and Prevention (CDC)34; they had no history of P. carinii pneumonia or toxoplasmic encephalitis, had not received prophylaxis against these diseases, and had CD4+ T-lymphocyte counts below 200 per cubic millimeter. Patients receiving zidovudine were eligible for the study. The criteria for eligibility also included a hemoglobin level of more than 10 g per deciliter, a total neutrophil count of more than 750 per cubic millimeter (0.75 x 109 per liter), and serum aspartate and alanine aminotransferase levels less than 10 times the upper limit of normal. The criteria for exclusion were pregnancy, asthma requiring long-term treatment, a history of adverse reactions to sulfonamides or sulfones, glucose-6-phosphate dehydrogenase deficiency, ischemic cardiomyopathy, and a serum creatinine concentration greater than 2.26 mg per deciliter (200 µmol per liter). In addition, patients who had a history of central nervous system disease or diseases requiring maintenance therapy with drugs with activity against P. carinii or T. gondii or who had contraindications to the use of aerosolized pentamidine (active pulmonary tuberculosis or pulmonary Kaposi's sarcoma) were not eligible. Other forms of prophylaxis against P. carinii pneumonia or toxoplasmosis were not allowed.
Study Design
This open, randomized, controlled study was conducted in 16 French centers. Block randomization using sealed envelopes was carried out at each center to ensure that three of every six consecutive patients would be assigned to treatment with pentamidine and three to treatment with dapsone-pyrimethamine. The randomization was stratified according to the patients' clinical status. Stratum A included patients with AIDS-related complex who had not received antiretroviral therapy or who had received it for less than six months. Stratum B included patients with AIDS, regardless of previous zidovudine therapy, and patients with AIDS-related complex who had received zidovudine for more than six months.
Treatment according to the protocol was started within a month of randomization and was planned to last for two years. Aerosolized pentamidine isethionate (Pentacarinat, Laboratoires Roger Bellon, Neuilly-sur-Seine, France) was delivered through a Respirgard II nebulizer (Marquest, Englewood, N.J.) at a dosage of 300 mg once a month (±3 days) under the supervision of a nurse. The administration of a bronchodilator before the inhalations was recommended.
Dapsone (50 mg once a day; Laboratoires Specia, Paris), pyrimethamine (50 mg once a week; Malocide, Laboratoires Specia), and folinic acid (25 mg once a week; Laboratoires Roger Bellon) were dispensed in blister packs.
Evaluation of the Patients
The participants underwent a thorough clinical examination every three months. Laboratory tests, including a differential blood count and CD4+ T-lymphocyte count, were performed at entry and every three months. Serologic tests for T. gondii were performed every six months when the results of such testing were negative before the patient's inclusion in the study. In addition, the reticulocyte count, methemoglobin level, and serum haptoglobin level were monitored every three months in the patients receiving dapsone-pyrimethamine.
Clinical End Points
Episodes of P. carinii pneumonia were defined by the onset of pneumonia clinically and radiologically consistent with acute P. carinii pneumonia, even in the absence of microscopical confirmation, if it responded to antipneumocystis therapy. Confirmed cases of P. carinii pneumonia were those in which P. carinii was detected in induced sputum, bronchoalveolar-lavage fluid, or lung tissue.
Toxoplasmic encephalitis was diagnosed if T. gondii tachyzoites or cysts were identified in brain tissue obtained by biopsy or autopsy or if a ring-enhancing lesion or lesions were found on computed tomography or magnetic resonance imaging. Less characteristic radiologic abnormalities were also considered diagnostic for toxoplasmic encephalitis if there was improvement during specific treatment (i.e., pyrimethamine plus either sulfadiazine or clindamycin). Histologic proof was not required, because brain biopsy is not routinely indicated in this setting. In the light of contemporary observations,35,36,37 the protocol was amended during the enrollment period to include extracerebral toxoplasmosis as a clinical end point. Extracerebral toxoplasmosis was diagnosed when there were characteristic funduscopic features responsive to specific therapy (retinal toxoplasmosis) or when tachyzoites or cysts were identified in bronchoalveolar fluid or lung tissue (pulmonary toxoplasmosis).
Deaths during short-term treatment of P. carinii pneumonia or toxoplasmosis were considered to be due to these diseases. All critical events involving the occurrence of P. carinii pneumonia, toxoplasmosis, or death were reviewed by a Critical Events Committee that was unaware of the patients' treatment assignments.
Statistical Analysis
The enrollment target of 280 patients was increased to 360 patients in September 1990 on the basis of recent data on the incidence of opportunistic infections in zidovudine-treated patients in France38.
The statistical analyses were based on the intention-to-treat principle. Only patients who never started the planned prophylactic regimen were excluded from this analysis. The results were analyzed by the chi-square test or Fisher's exact test in the case of categorical variables and by one-way and two-way analyses of variance and Student's t-test in the case of continuous variables. The time-to-event distributions were plotted with Kaplan-Meier curves and compared by a log-rank test. Cox proportional-hazards regression was used to estimate the difference between the two regimens after adjustment for the following base-line covariates39: study center, age, sex, risk factors for HIV infection, stratum of randomization, HIV disease stage according to the CDC classification, history of herpes zoster, and CD4+ T-lymphocyte count. In addition, the results of T. gondii serologic testing obtained before enrollment in the study were included in the model for comparison of the occurrence of toxoplasmosis in the two study groups. All P values were two-tailed, and P values of 0.05 or less were considered to indicate statistical significance.
Given the numerous critical events observed, the Data and Safety Monitoring Committee requested an interim analysis in October 1991. On the basis of the rules of O'Brien and Fleming,40 discontinuation of the trial on December 18, 1991, was recommended. The analysis presented here is based on data obtained up to December 19, 1991.
Results
Study Population
From July 1989 through November 1990, 362 patients were randomized. Thirteen of these patients (four in the pentamidine group and nine in the dapsone-pyrimethamine group) did not receive their assigned regimens within a month of randomization and were not included in the analysis. Among the 349 who could be evaluated, 176 were randomly assigned to pentamidine, and 173 to dapsone-pyrimethamine; 256 patients were included in stratum A, and 93 in stratum B. The two study groups were well matched with respect to base-line characteristics (Table 1), the only significant difference being one concerning a history of herpes zoster (8 percent in the pentamidine group vs. 16 percent in the dapsone-pyrimethamine group; P = 0.03).
|
The median follow-up was 539 days and did not differ significantly between the two groups. Twenty episodes of P. carinii pneumonia, 51 episodes of toxoplasmosis, and 86 deaths occurred during the study. One hundred thirty-eight patients completed the study -- i.e., they continued to receive the assigned therapy and did not reach an end point. Eight patients in the pentamidine group and four in the dapsone-pyrimethamine group were lost to follow-up at least six months before the end of the trial, and therefore data on these patients were censored at the last visit. One hundred thirty patients discontinued the assigned treatment because of drug intolerance, violations of protocol, or other reasons, but they were monitored nonetheless.
The mean (±SE) daily dosage of zidovudine tended to be lower in the dapsone-pyrimethamine group than in the pentamidine group: 413 ±25 and 475 ±21 mg per day, respectively (P = 0.06). Twenty-one patients in the dapsone-pyrimethamine group and 25 patients in the pentamidine group were switched from zidovudine to didanosine.
Occurrence of P. carinii Pneumonia
P. carinii pneumonia developed in 10 of the 176 patients in the pentamidine group (5.7 percent) and in 10 of the 173 patients in the dapsone-pyrimethamine group (5.8 percent), corresponding to rates of 0.32 and 0.33 cases, respectively, per 100 patient-months. The cumulative proportion of patients remaining free of P. carinii pneumonia is shown in Figure 1. One patient in the pentamidine group had an upper-lobe localization of the pneumonia. Seventeen of the 20 episodes of P. carinii pneumonia were confirmed microscopically. In the dapsone-pyrimethamine group, two patients in whom the pneumonia developed were receiving didanosine concomitantly. In the Cox proportional-hazards regression model, the risk of P. carinii pneumonia in the participants assigned to pentamidine was 1.13 times that in the participants assigned to dapsone-pyrimethamine (95 percent confidence interval, 0.44 to 2.92; P = 0.79). No cases of P. carinii pneumonia occurred after an episode of toxoplasmosis. One death in the dapsone-pyrimethamine group was due to P. carinii pneumonia.
|
Among the 51 patients with toxoplasmosis, the localization was cerebral in 46, ocular in 4, and pulmonary in 1. Toxoplasmosis developed in 32 of the 176 patients in the pentamidine group (18.2 percent) as compared with 19 of the 173 patients in the dapsone-pyrimethamine group (11.0 percent), corresponding to rates of 1.08 and 0.63 cases, respectively, per 100 patient-months (Figure 2). The adjusted risk of toxoplasmosis in the participants assigned to aerosolized pentamidine was 1.81 times that in the participants assigned to dapsone-pyrimethamine (95 percent confidence interval, 1.12 to 2.94; P = 0.02). All but one of the patients in whom toxoplasmosis developed had serologic evidence of past exposure to T. gondii at the time of inclusion in the study. The remaining patient had seroconverted by the time toxoplasmosis developed. In the subgroup of patients who were seropositive for T. gondii at enrollment, the adjusted relative risk of toxoplasmosis in the pentamidine group was 2.37 times that in the dapsone-pyrimethamine group (95 percent confidence interval, 1.3 to 4.4; P = 0.006).
|
In the pentamidine group, the independent predictive factors for toxoplasmosis were a positive serologic test for T. gondii (P<0.001), lower CD4+ T-lymphocyte count (P = 0.03), and lower daily dose of zidovudine during the trial (P = 0.02). Four patients in each study group died of toxoplasmosis. In each group, one case of toxoplasmosis occurred in a patient in whom P. carinii pneumonia had already developed.
Survival
Forty-one and 45 patients died in the pentamidine and dapsone-pyrimethamine groups, respectively (relative risk, 0.74; 95 percent confidence interval, 0.47 to 1.16; P = 0.20). The respective median survival times were 576 and 581 days. Among the 86 deaths, the primary cause of death was P. carinii pneumonia or toxoplasmosis in 9, another AIDS-related disease in 72 (including neoplasia in 18), and a disease not related to AIDS in 5.
Side Effects
Table 2 shows the observed adverse reactions and those causing interruptions of prophylaxis in the two study groups. The dose of dapsone-pyrimethamine was reduced by half or temporarily discontinued in 14 and 4 patients, respectively. Interruptions of prophylaxis because of drug-related side effects were significantly more frequent in the dapsone-pyrimethamine group (42 of 173) than in the pentamidine group (3 of 176) (P<0.001 by the log-rank test). The probability of discontinuing oral prophylaxis was highest during the first three months (Figure 3). Only the neutrophil count obtained before enrollment in the study was predictive of the discontinuation of dapsone-pyrimethamine. The relative risk of discontinuing this treatment in patients with a neutrophil count below 2000 per cubic millimeter (2 x 109 per liter) was 1.92 (95 percent confidence interval, 1.04 to 3.54; P = 0.04) as compared with patients who had a higher neutrophil count. No deaths related to the prophylaxis were reported.
|
|
Discussion
This open, randomized trial found aerosolized pentamidine and dapsone-pyrimethamine to be equally effective in preventing first episodes of P. carinii pneumonia. Dapsone-pyrimethamine drastically reduced the incidence of opportunistic toxoplasmosis, although it was less well tolerated than aerosolized pentamidine.
Prevention of P. carinii Pneumonia
From previous cohort data25,41 one can estimate a two-year incidence of P. carinii pneumonia in this population of about 40 percent in the absence of prophylaxis. The one-year incidence in the placebo group of a controlled trial of aerosolized pentamidine for primary prevention of P. carinii pneumonia was 27 percent11. The high efficacy of pentamidine in our trial contrasts with the relatively lower success rate found in a comparative trial with trimethoprim-sulfamethoxazole,13 possibly because that trial involved secondary prophylaxis in patients who were more severely immunosuppressed and in whom deposition of the drug in the lung may be altered. In other trials of pentamidine for primary prophylaxis, the one-year incidence of P. carinii pneumonia was 8.6 percent11 and 11 percent,12 as compared with 4.7 percent in our trial. Tolerance of aerosolized pentamidine and compliance with this treatment were both good in the present trial, as in others7,8,9,10,11,12,13.
The combination with pyrimethamine allowed the daily dosage of dapsone to be reduced; the incidence of side effects was thus lower than in a previous trial, in which 42 percent of the patients withdrew because of drug intolerance22. Discontinuation of prophylaxis was significantly more frequent in the dapsone-pyrimethamine group than in the pentamidine group. Side effects were mild and resolved after drug withdrawal. Even at the lower dosage of dapsone, however, we did observe anemia due at least in part to chronic hemolysis, a well-known effect of the drug. Agranulocytosis reported with dapsone alone42 or in combination with pyrimethamine43 did not occur. One patient discontinued the treatment because of asymptomatic methemoglobinemia.
As first-line primary prophylaxis against P. carinii pneumonia, the two test regimens did not differ in efficacy, but a number of drawbacks have been reported with aerosolized pentamidine, including treatment failures,7,8,9,10,11,12,13 localization of the pneumonia to the upper lungs,44 disseminated P. carinii pneumonia,45 and environmental drug contamination and dissemination of respiratory pathogens46. Despite a higher incidence of side effects with dapsone-pyrimethamine, this combination appears to be an attractive alternative. The risk of severe hemolysis in patients with glucose-6-phosphate dehydrogenase deficiency, although this condition is rare in non-Mediterranean whites, may require screening before dapsone is used. Pharmacologic interactions with rifampin47 have been reported, as well as reduced absorption of dapsone when it is given concomitantly with buffered drugs, such as didanosine48.
Trimethoprim-sulfamethoxazole, which appeared to be more effective than aerosolized pentamidine in two recent trials,12,13 has been officially recommended in the United States as first-line prophylaxis against P. carinii pneumonia49. A comparative trial with dapsone-pyrimethamine is needed to determine which regimen has the best risk-benefit ratio.
Prevention of Toxoplasmosis
Dapsone-pyrimethamine reduced the incidence of toxoplasmosis, for which primary prophylaxis has not been recommended in HIV-infected patients50. In the pentamidine group, the two-year actuarial incidence of toxoplasmosis (24.3 percent) was consistent with previous reports37.
The decision to discontinue the trial at the interim analysis was prompted by the significant protective effect of dapsone-pyrimethamine against toxoplasmosis (P = 0.018 by the log-rank test; adjusted relative risk, 2.33; 95 percent confidence interval, 1.19 to 4.55; P = 0.01). At the final analysis, this effect was confirmed by the low incidence of toxoplasmosis among the patients who continued to take this combination of drugs as prophylaxis.
Primary prophylaxis against opportunistic toxoplasmosis in HIV-infected patients is warranted because of the serious morbidity and sequelae of this life-threatening disease29. In the subgroup of patients seropositive for T. gondii and with CD4+ T-lymphocyte counts below 100 per cubic millimeter (0.1 x 109 per liter) in the pentamidine group, the probability of toxoplasmosis after 18 months was 40 percent, according to multivariate analysis (data not shown).
Recent trials and observational studies suggest that trimethoprim-sulfamethoxazole, primarily prescribed for prophylaxis against P. carinii pneumonia, may also reduce the risk of toxoplasmosis13,51. Further studies with this combination are warranted.
This trial demonstrates the feasibility of targeting more than one opportunistic pathogen in HIV-infected patients. Because the trial included two opportunistic infections as end points, potential bias due to the cross-reactivity of curative therapy must be taken into account. If only the first event (P. carinii pneumonia or toxoplasmosis) was considered, the conclusions were the same (data not shown). However, we cannot exclude the possibility in the intention-to-treat analysis that in the pentamidine group the risk of P. carinii pneumonia was reduced by the higher incidence of toxoplasmosis, which led to short-term and maintenance therapy with activity against P. carinii52.
Although this trial was not specifically designed to evaluate mortality, there was no advantage in terms of survival for the dapsone-pyrimethamine group in relation to the pentamidine group, despite the efficacy of the drug combination in preventing toxoplasmosis. Interestingly, previous trials of primary or secondary prophylaxis against opportunistic infections (P. carinii pneumonia and cryptococcal meningitis) in HIV-infected patients have failed to demonstrate a survival benefit despite remarkable preventive efficacy7,8,53.
The data presented here provide a basis for the primary prevention of two major life-threatening opportunistic infections in HIV-infected patients with low CD4+ T-lymphocyte counts. When the prevention of P. carinii pneumonia alone is required, aerosolized pentamidine is effective and well tolerated, but it costs more than oral regimens. The available oral regimens against P. carinii, including dapsone-pyrimethamine, are associated with a high rate of discontinuation because of side effects, and there is clearly a need for new, better-tolerated oral drugs. In addition to P. carinii pneumonia, toxoplasmosis must also be prevented in persons seropositive for T. gondii who have low CD4+ T-lymphocyte counts. Dapsone-pyrimethamine will be the preventive regimen of choice until better-tolerated regimens are proved effective.
Supported by a grant (ANRS 003) from the Agence Nationale de Recherches sur le SIDA and by grants from the Institut de Medecine et d'Epidemiologie Tropicale, Paris, and Laboratoires Roger Bellon, Neuilly-sur-Seine, with the participation of the Association de Recherche, de Communication, et d'Action pour le Traitement du SIDA and Le Cercle des Medecins, Paris.
We are indebted to Jean-Jacques Pocidalo, Pharm.D., Claude Bazin, M.D., Nadine Martin, M.D., Bernard Regnier, M.D., Bernard Dupuis, M.D., Charles Mayaud, M.D., Erik Oksenhendler, M.D., Muriel Vray, and Michel Wolff, M.D., for their expertise; to David Young for his assistance in the preparation of the manuscript; and to the patients who participated in the trial.
Source Information
From the Institut National de la Sante et de la Recherche Medicale, Unite 13, Paris (P.-M.G., R.L., Claude Gaudebout, P.J.); the Pharmacie (A.C.) and the Services des Maladies Infectieuses et Tropicales (R.O., A.G.S., Christiane Gaudebout, E.B., J.-P.C.), Hopital Bichat-Claude Bernard, Paris; the Service des Maladies Infectieuses et Tropicales, Hopital Saint Louis, Paris (T.L.); and the Service de Medecine Interne, Clamart (F.B.) -- all in France. The members of the PRIO (Prevention des Infections Opportunistes) Study Group are listed in the Appendix.
Address reprint requests to Dr. Girard at INSERM, Unite 13, 190 Blvd. MacDonald, 75019 Paris, France.
References
In addition to the study authors, the following persons and institutions participated in the PRIO Study Group.
Hopital Bichat-Claude Bernard, Paris: H. Balloul, S. Bartczak, H. Bideault, O. Bouchaud, R. Bousquet, F. Camus, C. Carbon, P. Coulaud, P. De Truchis, A. Dia, S. Fegueux, S. Halimi, S. Kernbaum, G. Kerouani, A. Lepretre, S. Mas, C. Maslo, S. Matheron, F. Mignot, G. Moreau, C. Picard, M.-H. Prevot, M. Ruquet, C. Ruggieri, C. Samson, Y. Tolila, F. Vachon, and C. Voyer; Hopital Broussais, Paris: J. Gilquin, M. Kazatchkine, and L. Weiss; Hopital Cochin, Paris: D. Sicard; Hopital Hotel Dieu, Paris: J.M. Brechot, P. Poubeau, and J. Rochemaure; Hopital Lariboisiere, Paris: J. Cervoni and C. Caulin; Hopital Pasteur, Paris: B. Dupont and G. Pialoux; Hopital Pitie-Salpetriere, Paris: E. Dohin, M. Gentilini, G. Guermonprez, and C. Katlama; Hopital Saint Louis, Paris: V. Averous, J.M. Decazes, G. Kouchner, J. Modai, and J.M. Molina; Hopital Antoine Beclere, Clamart: A. Blanc, J.F. Delfraissy, J. Dormont, and J.L. Martin; Hopital Louis Mourier, Colombes: M. Bloch, C. Michon, A.M. Simonpoli, and P. Vinceneux; Hopital Jean Verdier, Montfermeil: V. Jeantils, M.C. Schneider, and M. Thomas; Hopital Delafontaine, Saint Denis: P. Babinet, M.A. Khuong, D. Mechali, and G. Walckenaer; Hopital Henri Mondor, Creteil: J.C. Benaym, P. Boudes, P. De Vlaemink, and A. Sobel; Centre Hospitalier Universitaire de Dijon: P. Chavanet, J.P. Kisterman, H. Portier, and A. Waldner; Centre Hospitalier Universitaire de Lyon: P. Boibieux, D. Peyramond, and G. Tempelhoff; Centre Hospitalier Universitaire de Montpellier: N. Fegueux and J.-M. Navarro; Centre Hospitalier Universitaire de Nice: E. Rosenthal and J.P. Cassuto; Centre Hospitalier Universitaire de Villeneuve-Saint-Georges: J.E. Malkin and C. Lafaix; and Institut National de la Sante et de la Recherche Medicale, Unite 13: J.J. Pocidalo.
| |||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
