Letters
The Editors welcome submissions for possible publication in the Letters section. Authors of letters should: • Include no more than 400 words of text, three authors, and five references • Type with double-spacing • Send three copies of the letter and a transfer-of copyright form (see Table of Contents for location) signed by all authors • Provide a self-addressed envelope if they want to be notified that the letter was received Letters commenting on an Annals article will be considered if they are received within 6 weeks of the time the article was published. Only some of the letters received can be published. Published letters are edited and may be shortened; tables and figures are included only selectively. Authors will be notified that the letter has been received. If the letter is selected for publication, the author will be notified about 3 weeks before the publication date. Unpublished letters cannot be returned. Pharmacologic Stress Echocardiography To the Editors: Martin and colleagues (1) present interestingdata, which shows significant differences among agents currently used for pharmacologic stress echocardiography with regard to their sensitivity and specificity for coronary artery disease as well as to their induction of symptoms. Unfortunately, our replication of the chi-square tests, supposedly done to test statistical significance between groups (with Yates' correction [x2y] when appropriate), does not consistently confirm the reported statistical significance. (Contact author for calculations.) Statistical re-analysis of the data show that only 3 of 10 statistical contrasts attain their reported level of significance; 3 attain lesser degrees of statistical significance; and 4 are not statistically different (P > 0.05). Richard B. Devereux, MD New York Hospital-Cornell Medical Center New York, NY 10021 Reference 1. Martin TW, Seaworth JF, Johns JP, Pupa LE, Condos WR. Comparison of adenosine, dipyridamole, and dobutamine in stress echocardiography. Ann Intern Med. 1992;116:190-6. To the Editors: We congratulate Martin and colleagues on their well-designed study comparing adenosine, dipyridamole, and dobutamine in stress echocardiography (1). Although the authors conclude that their data are insufficient to determine the best pharmacologic agent for stress echocardiography, we believe they show that adenosine is superior to either dipyridamole or dobutamine at the doses used. Their conclusion is based, in part, on the diagnostic accuracies of the three different methods. Accuracy, however, is only a rough, global measure of a diagnostic test, highly dependent on disease prevalence. Positive predictive value, the proportion of patients with a positive test who have the disease, and negative predictive value, the proportion of patients with a negative test who do not have the disease, generally are more useful than is accuracy in evaluating diagnostic tests. By our calculations of positive and negative predictive values, adenosine was clearly better than dipyridamole and was at least as effective as dobutamine (Table 1). When we calculated 168
Table 1. Predictive Vainles and Accuracies fc>r Adenosine, Dipyridamole, and Dobutamin e in Stress; Echocardiography Variable
< Study population (disease prevalence = 63%) Adenosine Dipyridamole Dobutamine Typical patient population (disease prevalence = 30%) Adenosine Dipyridamole Dobutamine
>
%
91 74 76
48 48 60
60 60 70
71 42 45
78 78 85
77 64 65
these values using a lower disease prevalence (30%), which is more likely in typical clinical settings, adenosine performed better than either dipyridamole or dobutamine. Mark J. Eisenberg, MD, MPH University of California Medical Center San Francisco, CA 94143 Louise Pilote, MD, MPH Stanford University Palo Alto, CA 94304 Reference 1. Martin WM, Seaworth JF, Johns JP, Pupa LE, Condos WR. Comparison of adenosine, dipyridamole, and dobutamine in stress echocardiography. Ann Intern Med. 1992;116:190-6. In response: Dr. Devereux's chi-square analysis is correct. The reason for our error was a misapplication of our statistical software. We deeply regret this mistake, and we thank Dr. Devereux for bringing it to our attention. Our new statistical consultant believes the Fisher exact test to be more appropriate because some comparisons involve a small number of patients. Using the Fisher exact test, 8 of the 10 contrasts are statistically different (P < 0.05). For the other two, the 20% difference in sensitivity between dipyridamole and dobutamine (P = 0.12) and the 26% difference in specificity between dipyridamole and adenosine (P = 0.08) are not statistically different. Thus, we cannot confidently conclude that dipyridamole is less sensitive than dobutamine or less specific than adenosine, but our other conclusions are supported. Drs. Eisenberg and Pilote illustrate how disease prevalence affects the accuracy and predictive value of a test. We are not certain, however, that the predictive value is the most important measure of a test. The only prevalence-independent indicators are likelihood ratios, which were shown in our article (1). These ratios confirm that adenosine is most useful for ruling in coronary artery disease and dobutamine is most useful for ruling it out. The following practical concerns underlay our hesitancy to
15 July 1992 • Annals of Internal Medicine • Volume 117 • Number 2
Downloaded from https://annals.org by University of Otago user on 01/19/2019
Positive Negative Accuracy Predictive Predictive Value Value
declare adenosine "best." First, in populations with a high prevalence of disease, for example, patients with atypical chest pain and peripheral vascular disease, dobutamine testing may be more useful because of its greater sensitivity. Second, in situations such as testing of airline pilots, a more sensitive test may be preferred to minimize false-negative results. Third, clinicians already familiar with dipyridamole and dobutamine and their side effects may find them easier or safer to use. Fourth, dobutamine may be more appropriate in patients with reactive airway disease because of the potential for bronchospasm with adenosine or dipyridamole. Finally, more patients preferred dobutamine than adenosine or dipyridamole (1). Thus, the "best" drug for pharmacologic stress testing may vary depending on the clinical circumstances and the goals of the clinician. Timothy W. Martin, MD William Beaumont Army Medical Center El Paso, TX 79920-5001 Joseph P. Johns, MD Lawrence E. Pupa, MD Brooke Army Medical Center Fort Sam Houston, TX 78234-6200 Reference 1. Martin TW, Seaworth JF, Johns JP, Pupa LE, Condos WR. Comparison of adenosine, dipyridamole, and dobutamine in stress echocardiography. Ann Intern Med. 1992;116:190-6. Cerebral Toxoplasmosis and Prophylaxis for Pneumocystis carinii Pneumonia
2. Clotet B, Sirera G, Romeu J, Gimeno JM, Jou A, Condom MJ, et al. Twice-weekly dapsone-pyrimethamine for preventing PCP and cerebral toxoplasmosis [Letter]. AIDS. 1991;5:601-2.
In response: Drugs systemically administered against Toxoplasma gondii are also effective against Pneumocystis carinii and vice versa. Clotet and colleagues suggest that dapsone and pyrimethamine, given to prevent P. carinii pneumonia, also prevent toxoplasmosis. Their study was small and nonrandomized, but preliminary results from larger trials point in the same direction. At the Third European Conference on Clinical Aspects and Treatment of HIV Infection in Paris (March 1213, 1992), investigators from Switzerland and another group from France presented evidence suggesting that the combination of dapsone and pyrimethamine might protect against toxoplasmosis. Both studies compared the two oral drugs with inhaled pentamidine. The intention-to-treat analysis was clouded by the numerous crossovers: patients who were randomized to receive dapsone-pyrimethamine, who experienced side effects, and who were actually taking pentamidine. There was almost no incidence of toxoplasmosis, however, in those patients who continued taking dapsone and pyrimethamine. The same finding was true of trimethoprim-sulfamethoxazole in an American study also presented at the conference in Paris, which compared this drug combination to inhaled pentamidine. Bernard Hirschel, MD Hdpital Cantonal Universitaire de Geneve CH-1211 Geneva 4, Switzerland Gold Therapy for Rheumatoid Arthritis
References
To the Editors: In the Annals editorial on treatment of rheumatoid arthritis (1), various studies were reviewed which concluded that gold therapy was both ineffective and effective. I was introduced to gold therapy while on the Arthritis Service of the Mayo Clinic in the late 1940s and used it for over 40 years, until I retired last August. I developed a healthy respect for the drug when one of my patients in 1951 developed the only serious reaction that I saw, namely stomatitis, hepatitis, and a severe exfoliative dermatitis on the third injection. This may have been a true allergic reaction. Fortunately, the patient survived. One reason for the ineffectiveness of gold therapy may be differences among patients in the excretion rates for gold. Smith and colleagues (2) concluded that patients who responded poorly to gold therapy were hyperexcretors of gold and those who developed toxicity were usually hypoexcretors. These researchers found that individualized programs induced remissions in 82% of the patients. I found that patients who developed a skin rash, the most common side effect of gold therapy, could have the drug withdrawn until the rash cleared. Therapy could then be reinstituted at a lower dose than the usual 50-mg maintenance dose, and a good therapeutic response could be obtained without recurrence of the side effect. Later, I followed a woman for 2 years who had been receiving gold therapy for some time. She stayed in an excellent state of remission unless she skipped her weekly dose. She never had a side effect and died of an unrelated cause while still in remission. After that experience I became bolder and increased maintenance doses to 60 to 90 mg in those patients who responded poorly. Many patients who were on monthly doses for over 18 years remained functional or in remission as long as the dose was titrated as needed. Over the years I have been amazed that many rheumatologists with whom I discussed the use of gold therapy were unfamiliar with the work of Smith and colleagues (2). Although methotrexate is replacing gold therapy as first choice as a remission agent in rheumatoid arthritis, gold therapy is still a viable alternative and, if properly used, can produce excellent results.
1. Heald A, Flepp M, Chave JP, Malinverni R, Ruttiman S, Gabriel V, et al. Treatment for cerebral toxoplasmosis protects against Pneumocystis carinii pneumonia in patients with AIDS. Ann Intern Med. 1991;115:760-3.
John M. McMahon, MD Baptist Medical Center Princeton Birmingham, AL 35211
To the Editors: Recently, Heald and coworkers (1) focused on the decreased risk in acquired immunodeficiency syndrome (AIDS) patients for developing subsequent Pneumocystis carinii pneumonia after being treated for cerebral toxoplasmosis. According to these authors, "decreased risk is prcbably the result of chronic suppressive treatment with pyrimethamine and sulfonamides." We would like to point out that the reverse is also true for patients receiving primary or secondary prophylaxis for P. carinii pneumonia. We have reported that primary toxoplasmosis did not occur in 54 human immunodeficiency virus (HlV)-positive patients (26 had AIDS, and 28 belonged to stage IV-C2); these patients had immunoglobulin G antibodies to toxoplasma greater than 1/16, had a mean follow-up of 15 months (range, 8 to 18 months), and were receiving dapsone, 100 mg, and pyrimethamine, 25 mg, twice weekly (2). In fact, we have seen only a few cases of primary toxoplasmosis since 1990 when we started to prescribe dapsone and pyrimethamine as a prophylaxis for P. carinii pneumonia. We have registered 322 new AIDS cases during this period, and 26 of 28 new episodes of toxoplasma encephalitis have been observed among this group, either in patients who were not compliant with the former prophylaxis or in those receiving inhaled pentamidine with or without pyrimethamine as prophylaxis for P. carinii pneumonia. We agree with Heald and colleagues (1) about the efficacy of protection against P. carinii pneumonia in patients on secondary prophylaxis for cerebral toxoplasmosis. In 55 patients receiving sulfadiazine, 2000 mg, and pyrimethamine, 25 mg, twice weekly for prevention of secondary toxoplasmosis, we have not diagnosed, any new episodes of P. carinii pneumonia. B. Clotet, MD J. Romeu, MD G. Sirera, MD Hospital Universitari "Germans Trias i Pujol" 08916 Badalona, Barcelona, Catalonia, Spain
15 July 1992 • Annals of Internal Medicine • Volume 117 • Number 2 Downloaded from https://annals.org by University of Otago user on 01/19/2019
169
References 1. Pincus T, Wolfe F. Treatment of rheumatoid arthritis: challenges to traditional paradigms. Ann Intern Med. 1991;115:825-7. 2. Smith RT, Peak WP, Kron KM, Hermann IF, DelToro RA, Goldman M. Increasing the effectiveness of GOLD therapy in rheumatoid arthritis. JAMA. 1958;167:1197-1204. In response: The comments of Dr. McMahon echo positive experiences of many rheumatologists in the use of gold therapy. Analyses of study results in the use of gold and penicillamine (1) indicated that fewer than 2% of patients had sustained remissions for 3 years. Most of 75 patients with rheumatoid arthritis, who were monitored for 9 years, experienced significant functional declines, work disability, and increased mortality rates, regardless of use of gold or other drugs (2). Studies of patients observed for more than 3 years suggest that effective results in rheumatoid arthritis are generally not sustained (3). Nonetheless, 10% to 20% of patients experience meaningful clinical benefit with gold therapy over 5 years or longer, as noted in our editorial. We have suggested that documenting such clinical experience is now possible with effective databases and rigorous statistical procedures (4). The advance that facilitates this goal involves routine distribution to each patient at each visit of a brief, "patient-friendly," self-report questionnaire such as the health assessment questionnaire or its modified version (4). Data from patient questionnaires are correlated with data from traditional measures such as the joint count, radiograph, and erythrocyte sedimentation rate (5). Patient questionnaire data are effective in monitoring clinical status in clinical trials as well as in predicting long-term functional declines, work disability, and mortality in patients with rheumatoid arthritis (4). Such databases could facilitate translation of the clinical experience of careful observers such as Dr. McMahon into quantitative "scientific" data for rapid dissemination and improved patient care. Theodore Pincus, MD Vanderbilt University School of Medicine Nashville, TN 37232 Frederick Wolfe, MD University of Kansas School of Medicine Wichita, KS 67214 References 1. Wolfe F, Hawiey DJ. Remission in rheumatoid arthritis. J Rheumatol. 1985;12:245-52. 2. Pincus T, Callahan LF, Sale WG, Brooks AL, Payne LE, Vaughn WK. Severe functional declines, work disability, and increased mortality in seventy-five rheumatoid arthritis patients studied over nine years. Arthritis Rheum. 1984;27:864-72. 3. Hawiey DJ, Wolfe F. Are the results of controlled clinical trials and observational studies of second line therapy in rheumatoid arthritis valid and generalizable as measures of rheumatoid arthritis outcome: analysis of 122 studies. J Rheumatol. 1991;18:1008-14. 4. Wolfe F, Pincus T. Standard self-report questionnaires in routine clinical and research practice—an opportunity for patients and rheumatologists. J Rheumatol. 1991;18:643-6. 5. Pincus T, Callahan LF, Brooks RH, Fuchs HA, Olsen NJ, Kaye JJ. Self-report questionnaire scores in rheumatoid arthritis compared with traditional physical, radiographic, and laboratory measures. Ann Intern Med. 1989;110:259-66. Graft-versus-Host Disease and Liver Transplantation To the Editors: We offer another case in support of the supposition by Collins and colleagues (1) that graft-versus-host disease (GVHD) in liver transplantation is more common than previously thought. An 18-year-old white woman (A-positive) received an ABOmismatched orthotopic liver transplant from a previously healthy man (O-positive) for cryptogenic cirrhosis after her first transplant failed secondary to biliary obstruction. The procedure was uncomplicated. Her hemoglobin fell rapidly from 10.2 mg/dL to 6.6 mg/dL between days 5 and 9 after surgery without evidence of bleeding. High titers of anti-A antibodies were found in the patient's serum, believed to be 170
the result of donor B-cell "engraftment" or B-cell GVHD causing hemolysis. She received O-positive washed red blood cells without complication. After 14 days she developed a fever of 39 °C with no discernible focus; the fever was unresponsive to broad-spectrum antibiotics. She subsequently developed pancytopenia, diarrhea, and a skin rash involving the trunk and extremities, including the palms and soles. Skin biopsy was suggestive of GVHD. A bone-marrow biopsy specimen was 65% cellular with few mature granulocytes and megakaryocytic hyperplasia. The diagnosis of GVHD was made, and the patient was treated with hydrocortisone and antithymocyte globulin. Karyotyping on a bone-marrow aspirate showed a mosaic 46 XX/46 XY, confirming the diagnosis. (Karyotyping done on the patient 3 months earlier was 46 XX.) She died 49 days after surgery. This case exemplifies both the early hemolytic anemia reported in solid organ transplants (2) and the T-cell-mediated GVHD seen in liver transplant patients (1). Still classified as "rare," these cases seem to be recognized more frequently now that the entity has been described. Recently, a report (3) described 4 cases among 175 transplants. We feel it is prudent, given the severity of the disease, to emphasize prevention. One modality not mentioned might be treatment of the donor organ with a dose of radiation that impairs lymphocyte survival without compromising organ survival (4). Alex M. DePaoli, MD University of Chicago Hospitals Chicago, IL 60637 Jacob Bitran, MD Lutheran General Hospital Park Ridge, IL 60068 References 1. Collins RH, Cooper C, Nikaein A, Klintmalm G, Fay J. Graft versus host disease in a liver transplant recipient. Ann Intern Med. 1992; 116:391-2. 2. Ramsey G, Nusbacher J, Starzl TE, Lindsay GD. Isohemagglutinins of graft origin after ABO-unmatched liver transplantation. N Engl J Med. 1984;311:1167-70. 3. Roberts JP, Ascher NL, Lake J, Capper J, Purohit S, Garovoy M, et al. Graft vs. host disease after liver transplantation in humans: a report of four cases. Hepatology. 1991;14:274-81. 4. Von Fliedner V, Higby DJ, Kim U. Graft-versus-host reaction following blood product transfusion. Am J Med. 1982;72:951-61. In response: We appreciate Drs. DePaoli and Bitran's description of another case of graft-versus-host disease (GVHD) in a liver transplant recipient. We suspect the syndrome will be increasingly recognized as the index of suspicion increases (indeed, we have recently observed an additional case). Although irradiation of the liver might be an effective means of preventing GVHD, we would be concerned about possible damage to the liver from a large dose of radiation given over a short time. Careful study of this (or any other) manipulation of the graft is necessary before clinical application. Too little is currently known about the incidence, risk factors, and mechanisms of GVHD in liver transplant recipients to warrant prophylactic measures. Routine depletion of donor organ lymphocytes could possibly have adverse effects. Animal studies, for example, have suggested that limited donor engraftment, with the establishment of a state of low-level, mixed chimerism, may promote acceptance of the organ (1). If this were true in humans, then the depletion of donor lymphocytes could lead to an increased incidence of rejection. Thus, further study of the delicate interplay between donor and host immune systems is needed before attempting further manipulations aimed at reducing the immunologic complications— GVHD and rejection—of liver transplantation. Robert H. Collins, Jr., MD Goran Klintmalm, MD, PhD Joseph W. Fay, MD Baylor University Medical Center Dallas, TX 75246
15 July 1992 • Annals of Internal Medicine • Volume 117 • Number 2
Downloaded from https://annals.org by University of Otago user on 01/19/2019
Reference 1. Sachs DH. Specific transplantation tolerance. N Engl J Med. 1992; 325:1240-2. Chlamydia, Smoking, and Heart Disease
2.
3.
To the Editors: Saikku and colleagues (1) present evidence of an association between chronic Chlamydia pneumoniae infection and the development of coronary heart disease in the Helsinki Heart Study. This association, if confirmed, could have a profound effect on the prevention and management of coronary heart disease. The odds ratios associating evidence of C. pneumoniae infection and coronary heart disease were attenuated (from 2.7 to 2.3 for immunoglobulin A [IgA] titers, 2.1 to 1.8 for immune complexes, and 2.9 to 2.6 for the presence of both factors). There was no statistical difference for IgA titers and immune complexes after controlling for other coronary risk factors, including current smoking. Have the authors considered the possibility that smoking may be related to coronary heart disease through promotion of C. pneumoniae respiratory tract infection, facilitation of penetration of the organism into the blood stream, or both? If true, an association might be found between the interaction of smoking and C. pneumoniae infection and coronary heart disease in their data. David L. Hahn, MD Arcand Park Clinic Madison, WI 53704 Reference 1. Saikku P, Leinonen M, Tenkanen L, linnanmaki E, Ekman MR, Manninen V, et al. Chronic Chlamydia pneumoniae infection as a risk factor for coronary heart disease in the Helsinki Heart Study. Ann Intern Med. 1992;116:273-8. In response: Smoking is known to cause suppression of immune defense mechanisms (1) and, as we showed (2), is associated with markers of a possible chronic Chlamydia pneumoniae infection. The idea that smoking could promote chronic C. pneumoniae infection, as proposed by Dr. Hahn, is quite logical, but markers of chronic C. pneumoniae were found to be independent risk factors for coronary heart disease in both our earlier and current studies (2-4). The proposal that smoking could facilitate the pathogenic mechanisms by which G pneumoniae causes coronary heart disease, if true, would be reflected in the joint effect of smoking and markers of C. pneumoniae on the risk of coronary heart disease. The Helsinki Heart Study (5) instituted a thorough examination of these two risk factors (Leinonen M. and colleagues. Unpublished data.), and, as Dr. Hahn guessed, smoking did have a more than additive effect with markers of C. pneumoniae on the risk of coronary heart disease. Patterns of joint effects seen in the data, however, do not necessarily imply similar patterns at the biological level. The possibility that some symptoms commonly associated with smoking are partly caused by promotion of a chronic C. pneumoniae infection in the lungs, with all the destructive scarring and hypersensitivity typical in chronic chlamydial infections, remains to be studied. Pekka Saikku, MD, PhD University of Helsinki Helsinki, Finland
4.
5.
smoke and other atmospheric contaminants in man and experimental animals. Bacteriol Rev. 1977;41:205-16. Saikku P, Leinonen M, Tenkanen L, Linnenmaki E, Ekman MR, Manninen V, et al. Chronic Chlamydia pneumoniae infection as a risk factor for coronary heart disease in the Helsinki Heart Study. Ann Intern Med. 1992;116:273-8. Saikku P, Leinonen M, Mattila K, Ekman MR, Nieminen MS, Makela PH, et al. Serologic evidence of an association of a novel Chlamydia, TWAR, with chronic coronary heart disease and acute myocardial infarction. Lancet. 1988;2:983-5. Leinonen M, Linnanmaki E, Mattila K, Nieminen MS, Valtonen V, Leirisalo-Repo M, et al. Circulating immune complexes containing chlamydial lipopolysaccharide in acute myocardial infarction. Microb Pathog. 1990;9:67-73. Frick MH, Elo O, Haapa K, Heinonen OP, Heinsalmi P, Helo P, et al. Helsinki Heart Study: primary-prevention trial with gemfibrozil in middle-aged men with dyslipidemia. Safety of treatment, changes in risk factors, and incidence of coronary heart disease. N Engl J Med. 1987;317:1237-45.
Minimal Hepatitis C Infectivity in Semen To the Editors: Both hepatitis B virus and cytomegalovirus are known to be transmitted sexually. To assess the possibility that hepatitis C virus (HCV), a known blood-based infection, could be transmitted by sexual intercourse, samples of serum, semen, and urine were examined for HCV RNA using the polymerase chain reaction. Informed consent was obtained from six men with antibody to HCV (anti-HCV) positivity and chronic hepatitis (mean age, 46.8 ± 1 1 . 7 years). They denied homosexual behavior and intravenous drug abuse; half had a history of blood transfusion. Semen samples were obtained after urine samples were collected to prevent contamination of the specimens. No blood was found in the specimens, and microscopy of the urinary sediment showed erythrocytes at a maximum of 1 to 3 cells per high-power field. Anti-HCV titers (cl00-3 antigen) were measured by enzymelinked immunosorbent assay (ELISA). Hepatitis C virus RNA was purified from 100 /*L of each patient sample, using acid guanidinium thiocyanate-phenol-chloroform extraction (1). Using antisense primer (positions 229-248 of the Okamoto 5' end noncoding region, 5'-AACAT TACTC GGCTA GCAGT-3') and sense primer (positions 7-26 of the Okamoto, 5'-ACTCC ACCAT AGATC ACTCC-3') (2), 45 polymerase chain reaction cycles were done (3). Three of the six patients had spouses. To ascertain the infectivity of HCV among couples, serum antiHCV and HCV-RNA levels were also determined in the spouses by the same method. Hepatitis C virus RNA was detected in the serum of all six patients. However, HCV RNA was not detected in the semen or urine of any of the patients. No relation existed between their clinical data (transfusion history, disease duration, and biochemical data) and HCV-RNA positivity. Neither HCV RNA nor anti-HCV was detected in the serum of the three spouses. Esteban and colleagues (4) have reported that the positive rate of anti-HCV among couples was low. Although anti-HCV positivity does not always correspond to the existence of HCV, the infectivity through body secretions was also suggested to be low by Hsu and colleagues (5). The sensitivity of the polymerase chain reaction was 10 to 100 copies/100 /*L of sample. Transmission of HCV by semen, therefore, is not completely ruled out. Our results, however, suggest low infectivity of semen and urine except when contaminated by blood. Soichiro Terada, MD Koichi Kawanishi, MD Kawaga Medical School Kawaga, Japan
Maija Leinonen, PhD National Public Health Institute Helsinki, Finland
Kazuhiko Katayama, MD Biomedical Laboratory Kawagoe City, Japan
Leena Tenkanen, MSc University of Tampere Tampere, Finland References 1. Holt PG, Keast D. Environmentally induced changes in immunological function: acute and chronic effects of inhalation of tobacco
References 1. Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol chloroform extraction. Anal Biochem. 1987;162:156-9.
15 July 1992 • Annals of Internal Medicine • Volume 117 • Number 2 Downloaded from https://annals.org by University of Otago user on 01/19/2019
171
2. Okamoto H, Okada S, Sugiyama Y, Yotsumoto S, Tanaka T, Yoshizawa H, et al. The 5'-terminal sequence of the hepatitis C virus genome. Jpn J Exp Med. 1990;60:167-77. 3. Saiki RK, Scharf S, Faloona F, Mullis KB, Horn GT, Erlich HA, et al. Enzymic amplification of beta-globin genomic sequences and restriction site analysis for diagnosis of sickle cell anemia. Science. 1985; 230:1350-4. 4. Esteban JI, Esteban R, Viladomiu L, Lopez-Tolavera JC, Gonzalez A, Hernandez JM, et al. Hepatitis C virus antibodies among risk groups in Spain. Lancet. 1989;2:294-7. 5. Hsu HH, Wright TL, Luba D, Martin M, Feinstone SM, Garcia G, et al. Failure to detect hepatitis C virus genome in human secretions with the polymerase chain reaction. Hepatology. 1991;14:763-7. Gallbladder Disease and Angina Pectoris To the Editors: The article by Ryan and colleagues (1) calls to mind the occasional occurrence of high substernal chest pain, clearly related to exercise and relieved promptly by nitroglycerine in patients who also have gallstones. In some patients, removal of the gallstones is followed by disappearance of the angina. The late Dr. Chester Jones at Massachusetts General Hospital, who may have been the first to note this phenomenon, proposed that an overflow of autonomic nervous stimuli from the gallbladder raised the level of autonomic activity in the coronary vessels. Consequently, a smaller exercise stimulus would be needed to exceed the threshold level of coronary autonomic activity required to produce spasm and anginal pain. This "autonomic overflow," as Dr. Jones called it, may be analogous to the production of paralytic ileus by a hip fracture or kidney stones. This association has not been studied in a scientific manner, but it is not a rare observation. In my observations, the anginal pain has not been accompanied by electrocardiographic changes. Arthur B. French, MD Henry Ford Hospital Detroit, MI 48202 Reference 1. Ryan ET, Pak PH, DeSanctis RW. Myocardial infarction mimicked by acute cholecystitis. Ann Intern Med. 1992;116:218-20. In response: As Dr. French correctly points out, the apparent association between gallbladder disease and angina pectoris has been observed for many years. The late Dr. Chester Jones made many major contributions in the first half of this century to understanding the mechanisms of pain arising from abdominal organs, including the gallbladder and the common bile duct. The first reports of angina pectoris attributed to biliary disease, however, precede his time, as one can find in Miller's 1932 review (1) of this subject. The mechanism of this apparent association is still not clear, although the "autonomic overflow" hypothesis is an attractive one. The case we described (2) may be part of the spectrum of cardiac manifestations of primary gallbladder disease. Peter H. Pak, MD Edward T. Ryan, MD Roman W. DeSanctis, MD Massachusetts General Hospital Boston, MA 02114 References 1. Miller CH. The gall-bladder and cardiac pain. Lancet. 1932;1:767-72. 2. Ryan ET, Pak PH, DeSanctis RW. Myocardial infarction mimicked by acute cholecystitis. Ann Intern Med. 1992;116:218-20. Scientific Rigor of Economic Analyses To the Editors: The article by Udvarhelyi and colleagues (1) highlights the methodologic inadequacies of economic analyses in the medical literature. Failure to adhere to basic methodologic principles and inappropriate use of economic terminology are not limited to the medical literature (2). We assessed the methodologic soundness of 65 articles in sue pharmacy journals from 1985 to 1990, using a process similar 172
to that used by Udvarhelyi and colleagues. We used 10 criteria suggested by Drummond and colleagues (3) to assess whether appropriate methodologic techniques were used in each of the 65 articles. Additionally, we assessed the use of the term "cost effective," using criteria suggested by Doubilet and coworkers (4). Only 3 of the 10 methodologic criteria were fulfilled by 50% or more of the studies evaluated. Particular deficiencies included the identification of all relevant costs and consequences, the adjustment of costs and consequences for differential timing (discounting), and the performance of sensitivity and incremental analyses. We also found misinterpretation of the term "cost effective" in 57% of the studies evaluated. No differences were found when articles were subdivided by journal or by year of publication. We conclude that a multidisciplinary initiative is warranted to educate interested researchers, practitioners, and third-party payors in the necessary components of a thorough economic analysis. Jeffrey T, Lee, PharmD Glaxo Research Institute Research Triangle Park, NC 27709 Lisa A. Sanchez, PharmD Brigham and Women's Hospital Boston, MA 02115 References 1. Udvarhelyi S, Colditz GA, Rai A, Epstein AM. Cost-effectiveness and cost-benefit analyses in the medical literature: are the methods being used correctly? Ann Intern Med. 1992;116:238-44. 2. Lee JT, Sanchez LA. Interpretation of "cost-effective" and soundness of economic evaluations in the pharmacy literature. Am J Hosp Pharm. 1991;48:2622-7. 3. Drummond MF, Stoddart GL, Torrance GW. Methods for the economic evaluation of health care programmes. Oxford: Oxford University Press; 1986. 4. Doubilet P, Weinstein MC, McNeil BJ. Use and misuse of the term "cost effective" in medicine. N Engl J Med. 1986;314:253-6. Prince Dracula, Rabies, and the Vampire Legend To the Editors: Immortalized a century ago by Bram Stoker in his novel about Dracula, the legend of the human vampire is a popular fictional theme. In Eastern Europe, written records of vampirism date back to the thirteenth century (1). According to this lore, the vampire is a person whose soul stays in his or her body after death. During the day the vampire sleeps in a grave, leaving it at dusk to assault victims. To do this, he or she may transmogrify into an animal, for example, a bat. The victim usually survives, somewhat anemic, with only toothmarks as signs of what happened. Following a period of incubation, the victim's personality changes and becomes irrational and aggressive. The newly created vampire tries to bite other people, and those who are bitten either die or end up as vampires themselves. Garlic and crucifixes were ascribed some prophylactic value against vampirism, but the only "cure" was to drive a wooden stake through the vampire's heart, preferably before burial. If a vampire was suspected to be on the loose in a village, the churchyard would be dug up, and all corpses that had moved in their graves or were suspiciously well preserved were then treated with the stake (or decapitated as in Graffen zum Stein's 1732 account of an official Turkish autopsy of 13 Serbian vampires). Although we may still be fascinated by the vampire legend, we all now know that the human vampire never really existed. Or did he? A bite from an irrationally aggressive animal leads to aggressively psychotic behavior in the human victim. Doesn't it sound like rabies? In the agony of rabies, all affected mammals may display such a hyperexcitable phase. Even otherwise placid insectivorous bats have been reported to attack humans and other mammals (2). In human rabies a hyperexcitable psychotic phase is also seen. Although genuine biting behavior has rarely been reported, Lindtjorn mentions a 15-year-old rabid boy who bit off one of his mother's fingertips (3). Neither history nor folklore indicate Prince Dracula of
15 July 1992 • Annals of Internal Medicine • Volume 117 • Number 2
Downloaded from https://annals.org by University of Otago user on 01/19/2019
Wallachia (1431-76) to have been a vampire (4, 5), but he was responsible for the impalement of 40 000 to 100 000 people and for committing various cruelties against countless others. The Turkish sultan Mohammad II, after capturing Constantinople, is said to have been deterred from further conquests in Europe when, outside Dracula's capital Tirgoviste (in present Rumania), he saw a forest of 20 000 impaled bodies. The genuine vampire was hardly this cruel prince, although it could very well have been a poor, rabid peasant who ended up buried with a stake driven through his heart.
Correction: Preoperative Cardiac Risk Assessment A sentence in the conclusion of a review by Wong and colleagues (1) appeared incorrectly. The sentence should read: "Therefore, we suggest that preoperative coronary artery revascularization should not be used to protect patients from cardiac complications after vascular surgery unless the patient's cardiac disease warrants these procedures independently of the need for peripheral vascular surgery." Reference 1. Wong T, Detsky AS. Preoperative cardiac risk assessment for patients having peripheral vascular surgery. Ann Intern Med. 1992; 116: 743-53.
Alex Heick, MD, PhM Hvidovre Hospital Copenhagen 2650, Denmark References 1. Sturm D, Volker K. Von denen Vampiren oder Menschensaugern. Miinchen: Carl Hanser; 1968. 2. Constantine DG. Bat rabies in the southwestern United States. Public Health Rep. 1967;82:867-88. 3. Lindtj0rn B. Clinical features of rabies in man. Trop Doct. 1982;12: 9-12. 4. McNally RT, Florescu R. In Search of Dracula. Greenwich, Connecticut: New York Graphic Society; 1972. 5. Florescu R, McNally RT. Dracula: A Bibliography of Vlad the Impaler. London: Robert Hale: 1973.
Correction: Vagal Reflexes and the Aerodigestive Tract A review (1) incorrectly noted an author's degree. The correct degree for Bronwyn Jones is FRCR. Reference 1. Cunningham ET Jr, Ravich WJ, Jones B, Donner MW. Vagal reflexes referred from the upper aerodigestive tract: an infrequently recognized cause of common cardiorespiratory responses. Ann Intern Med. 1992;116:575-82.
15 July 1992 • Annals of Internal Medicine • Volume 117 • Number 2 Downloaded from https://annals.org by University of Otago user on 01/19/2019
173
The Editors welcome submissions for possible publication in the Letters section. Authors of letters should: • Include no more than 400 words of text, three authors, and five references • Type with double-spacing • Send three copies of the letter and a transfer-of copyright form (see Table of Contents for location) signed by all authors • Provide a self-addressed envelope if they want to be notified that the letter was received Letters commenting on an Annals article will be considered if they are received within 6 weeks of the time the article was published. Only some of the letters received can be published. Published letters are edited and may be shortened; tables and figures are included only selectively. Authors will be notified that the letter has been received. If the letter is selected for publication, the author will be notified about 3 weeks before the publication date. Unpublished letters cannot be returned. Pharmacologic Stress Echocardiography To the Editors: Martin and colleagues (1) present interestingdata, which shows significant differences among agents currently used for pharmacologic stress echocardiography with regard to their sensitivity and specificity for coronary artery disease as well as to their induction of symptoms. Unfortunately, our replication of the chi-square tests, supposedly done to test statistical significance between groups (with Yates' correction [x2y] when appropriate), does not consistently confirm the reported statistical significance. (Contact author for calculations.) Statistical re-analysis of the data show that only 3 of 10 statistical contrasts attain their reported level of significance; 3 attain lesser degrees of statistical significance; and 4 are not statistically different (P > 0.05). Richard B. Devereux, MD New York Hospital-Cornell Medical Center New York, NY 10021 Reference 1. Martin TW, Seaworth JF, Johns JP, Pupa LE, Condos WR. Comparison of adenosine, dipyridamole, and dobutamine in stress echocardiography. Ann Intern Med. 1992;116:190-6. To the Editors: We congratulate Martin and colleagues on their well-designed study comparing adenosine, dipyridamole, and dobutamine in stress echocardiography (1). Although the authors conclude that their data are insufficient to determine the best pharmacologic agent for stress echocardiography, we believe they show that adenosine is superior to either dipyridamole or dobutamine at the doses used. Their conclusion is based, in part, on the diagnostic accuracies of the three different methods. Accuracy, however, is only a rough, global measure of a diagnostic test, highly dependent on disease prevalence. Positive predictive value, the proportion of patients with a positive test who have the disease, and negative predictive value, the proportion of patients with a negative test who do not have the disease, generally are more useful than is accuracy in evaluating diagnostic tests. By our calculations of positive and negative predictive values, adenosine was clearly better than dipyridamole and was at least as effective as dobutamine (Table 1). When we calculated 168
Table 1. Predictive Vainles and Accuracies fc>r Adenosine, Dipyridamole, and Dobutamin e in Stress; Echocardiography Variable
< Study population (disease prevalence = 63%) Adenosine Dipyridamole Dobutamine Typical patient population (disease prevalence = 30%) Adenosine Dipyridamole Dobutamine
>
%
91 74 76
48 48 60
60 60 70
71 42 45
78 78 85
77 64 65
these values using a lower disease prevalence (30%), which is more likely in typical clinical settings, adenosine performed better than either dipyridamole or dobutamine. Mark J. Eisenberg, MD, MPH University of California Medical Center San Francisco, CA 94143 Louise Pilote, MD, MPH Stanford University Palo Alto, CA 94304 Reference 1. Martin WM, Seaworth JF, Johns JP, Pupa LE, Condos WR. Comparison of adenosine, dipyridamole, and dobutamine in stress echocardiography. Ann Intern Med. 1992;116:190-6. In response: Dr. Devereux's chi-square analysis is correct. The reason for our error was a misapplication of our statistical software. We deeply regret this mistake, and we thank Dr. Devereux for bringing it to our attention. Our new statistical consultant believes the Fisher exact test to be more appropriate because some comparisons involve a small number of patients. Using the Fisher exact test, 8 of the 10 contrasts are statistically different (P < 0.05). For the other two, the 20% difference in sensitivity between dipyridamole and dobutamine (P = 0.12) and the 26% difference in specificity between dipyridamole and adenosine (P = 0.08) are not statistically different. Thus, we cannot confidently conclude that dipyridamole is less sensitive than dobutamine or less specific than adenosine, but our other conclusions are supported. Drs. Eisenberg and Pilote illustrate how disease prevalence affects the accuracy and predictive value of a test. We are not certain, however, that the predictive value is the most important measure of a test. The only prevalence-independent indicators are likelihood ratios, which were shown in our article (1). These ratios confirm that adenosine is most useful for ruling in coronary artery disease and dobutamine is most useful for ruling it out. The following practical concerns underlay our hesitancy to
15 July 1992 • Annals of Internal Medicine • Volume 117 • Number 2
Downloaded from https://annals.org by University of Otago user on 01/19/2019
Positive Negative Accuracy Predictive Predictive Value Value
declare adenosine "best." First, in populations with a high prevalence of disease, for example, patients with atypical chest pain and peripheral vascular disease, dobutamine testing may be more useful because of its greater sensitivity. Second, in situations such as testing of airline pilots, a more sensitive test may be preferred to minimize false-negative results. Third, clinicians already familiar with dipyridamole and dobutamine and their side effects may find them easier or safer to use. Fourth, dobutamine may be more appropriate in patients with reactive airway disease because of the potential for bronchospasm with adenosine or dipyridamole. Finally, more patients preferred dobutamine than adenosine or dipyridamole (1). Thus, the "best" drug for pharmacologic stress testing may vary depending on the clinical circumstances and the goals of the clinician. Timothy W. Martin, MD William Beaumont Army Medical Center El Paso, TX 79920-5001 Joseph P. Johns, MD Lawrence E. Pupa, MD Brooke Army Medical Center Fort Sam Houston, TX 78234-6200 Reference 1. Martin TW, Seaworth JF, Johns JP, Pupa LE, Condos WR. Comparison of adenosine, dipyridamole, and dobutamine in stress echocardiography. Ann Intern Med. 1992;116:190-6. Cerebral Toxoplasmosis and Prophylaxis for Pneumocystis carinii Pneumonia
2. Clotet B, Sirera G, Romeu J, Gimeno JM, Jou A, Condom MJ, et al. Twice-weekly dapsone-pyrimethamine for preventing PCP and cerebral toxoplasmosis [Letter]. AIDS. 1991;5:601-2.
In response: Drugs systemically administered against Toxoplasma gondii are also effective against Pneumocystis carinii and vice versa. Clotet and colleagues suggest that dapsone and pyrimethamine, given to prevent P. carinii pneumonia, also prevent toxoplasmosis. Their study was small and nonrandomized, but preliminary results from larger trials point in the same direction. At the Third European Conference on Clinical Aspects and Treatment of HIV Infection in Paris (March 1213, 1992), investigators from Switzerland and another group from France presented evidence suggesting that the combination of dapsone and pyrimethamine might protect against toxoplasmosis. Both studies compared the two oral drugs with inhaled pentamidine. The intention-to-treat analysis was clouded by the numerous crossovers: patients who were randomized to receive dapsone-pyrimethamine, who experienced side effects, and who were actually taking pentamidine. There was almost no incidence of toxoplasmosis, however, in those patients who continued taking dapsone and pyrimethamine. The same finding was true of trimethoprim-sulfamethoxazole in an American study also presented at the conference in Paris, which compared this drug combination to inhaled pentamidine. Bernard Hirschel, MD Hdpital Cantonal Universitaire de Geneve CH-1211 Geneva 4, Switzerland Gold Therapy for Rheumatoid Arthritis
References
To the Editors: In the Annals editorial on treatment of rheumatoid arthritis (1), various studies were reviewed which concluded that gold therapy was both ineffective and effective. I was introduced to gold therapy while on the Arthritis Service of the Mayo Clinic in the late 1940s and used it for over 40 years, until I retired last August. I developed a healthy respect for the drug when one of my patients in 1951 developed the only serious reaction that I saw, namely stomatitis, hepatitis, and a severe exfoliative dermatitis on the third injection. This may have been a true allergic reaction. Fortunately, the patient survived. One reason for the ineffectiveness of gold therapy may be differences among patients in the excretion rates for gold. Smith and colleagues (2) concluded that patients who responded poorly to gold therapy were hyperexcretors of gold and those who developed toxicity were usually hypoexcretors. These researchers found that individualized programs induced remissions in 82% of the patients. I found that patients who developed a skin rash, the most common side effect of gold therapy, could have the drug withdrawn until the rash cleared. Therapy could then be reinstituted at a lower dose than the usual 50-mg maintenance dose, and a good therapeutic response could be obtained without recurrence of the side effect. Later, I followed a woman for 2 years who had been receiving gold therapy for some time. She stayed in an excellent state of remission unless she skipped her weekly dose. She never had a side effect and died of an unrelated cause while still in remission. After that experience I became bolder and increased maintenance doses to 60 to 90 mg in those patients who responded poorly. Many patients who were on monthly doses for over 18 years remained functional or in remission as long as the dose was titrated as needed. Over the years I have been amazed that many rheumatologists with whom I discussed the use of gold therapy were unfamiliar with the work of Smith and colleagues (2). Although methotrexate is replacing gold therapy as first choice as a remission agent in rheumatoid arthritis, gold therapy is still a viable alternative and, if properly used, can produce excellent results.
1. Heald A, Flepp M, Chave JP, Malinverni R, Ruttiman S, Gabriel V, et al. Treatment for cerebral toxoplasmosis protects against Pneumocystis carinii pneumonia in patients with AIDS. Ann Intern Med. 1991;115:760-3.
John M. McMahon, MD Baptist Medical Center Princeton Birmingham, AL 35211
To the Editors: Recently, Heald and coworkers (1) focused on the decreased risk in acquired immunodeficiency syndrome (AIDS) patients for developing subsequent Pneumocystis carinii pneumonia after being treated for cerebral toxoplasmosis. According to these authors, "decreased risk is prcbably the result of chronic suppressive treatment with pyrimethamine and sulfonamides." We would like to point out that the reverse is also true for patients receiving primary or secondary prophylaxis for P. carinii pneumonia. We have reported that primary toxoplasmosis did not occur in 54 human immunodeficiency virus (HlV)-positive patients (26 had AIDS, and 28 belonged to stage IV-C2); these patients had immunoglobulin G antibodies to toxoplasma greater than 1/16, had a mean follow-up of 15 months (range, 8 to 18 months), and were receiving dapsone, 100 mg, and pyrimethamine, 25 mg, twice weekly (2). In fact, we have seen only a few cases of primary toxoplasmosis since 1990 when we started to prescribe dapsone and pyrimethamine as a prophylaxis for P. carinii pneumonia. We have registered 322 new AIDS cases during this period, and 26 of 28 new episodes of toxoplasma encephalitis have been observed among this group, either in patients who were not compliant with the former prophylaxis or in those receiving inhaled pentamidine with or without pyrimethamine as prophylaxis for P. carinii pneumonia. We agree with Heald and colleagues (1) about the efficacy of protection against P. carinii pneumonia in patients on secondary prophylaxis for cerebral toxoplasmosis. In 55 patients receiving sulfadiazine, 2000 mg, and pyrimethamine, 25 mg, twice weekly for prevention of secondary toxoplasmosis, we have not diagnosed, any new episodes of P. carinii pneumonia. B. Clotet, MD J. Romeu, MD G. Sirera, MD Hospital Universitari "Germans Trias i Pujol" 08916 Badalona, Barcelona, Catalonia, Spain
15 July 1992 • Annals of Internal Medicine • Volume 117 • Number 2 Downloaded from https://annals.org by University of Otago user on 01/19/2019
169
References 1. Pincus T, Wolfe F. Treatment of rheumatoid arthritis: challenges to traditional paradigms. Ann Intern Med. 1991;115:825-7. 2. Smith RT, Peak WP, Kron KM, Hermann IF, DelToro RA, Goldman M. Increasing the effectiveness of GOLD therapy in rheumatoid arthritis. JAMA. 1958;167:1197-1204. In response: The comments of Dr. McMahon echo positive experiences of many rheumatologists in the use of gold therapy. Analyses of study results in the use of gold and penicillamine (1) indicated that fewer than 2% of patients had sustained remissions for 3 years. Most of 75 patients with rheumatoid arthritis, who were monitored for 9 years, experienced significant functional declines, work disability, and increased mortality rates, regardless of use of gold or other drugs (2). Studies of patients observed for more than 3 years suggest that effective results in rheumatoid arthritis are generally not sustained (3). Nonetheless, 10% to 20% of patients experience meaningful clinical benefit with gold therapy over 5 years or longer, as noted in our editorial. We have suggested that documenting such clinical experience is now possible with effective databases and rigorous statistical procedures (4). The advance that facilitates this goal involves routine distribution to each patient at each visit of a brief, "patient-friendly," self-report questionnaire such as the health assessment questionnaire or its modified version (4). Data from patient questionnaires are correlated with data from traditional measures such as the joint count, radiograph, and erythrocyte sedimentation rate (5). Patient questionnaire data are effective in monitoring clinical status in clinical trials as well as in predicting long-term functional declines, work disability, and mortality in patients with rheumatoid arthritis (4). Such databases could facilitate translation of the clinical experience of careful observers such as Dr. McMahon into quantitative "scientific" data for rapid dissemination and improved patient care. Theodore Pincus, MD Vanderbilt University School of Medicine Nashville, TN 37232 Frederick Wolfe, MD University of Kansas School of Medicine Wichita, KS 67214 References 1. Wolfe F, Hawiey DJ. Remission in rheumatoid arthritis. J Rheumatol. 1985;12:245-52. 2. Pincus T, Callahan LF, Sale WG, Brooks AL, Payne LE, Vaughn WK. Severe functional declines, work disability, and increased mortality in seventy-five rheumatoid arthritis patients studied over nine years. Arthritis Rheum. 1984;27:864-72. 3. Hawiey DJ, Wolfe F. Are the results of controlled clinical trials and observational studies of second line therapy in rheumatoid arthritis valid and generalizable as measures of rheumatoid arthritis outcome: analysis of 122 studies. J Rheumatol. 1991;18:1008-14. 4. Wolfe F, Pincus T. Standard self-report questionnaires in routine clinical and research practice—an opportunity for patients and rheumatologists. J Rheumatol. 1991;18:643-6. 5. Pincus T, Callahan LF, Brooks RH, Fuchs HA, Olsen NJ, Kaye JJ. Self-report questionnaire scores in rheumatoid arthritis compared with traditional physical, radiographic, and laboratory measures. Ann Intern Med. 1989;110:259-66. Graft-versus-Host Disease and Liver Transplantation To the Editors: We offer another case in support of the supposition by Collins and colleagues (1) that graft-versus-host disease (GVHD) in liver transplantation is more common than previously thought. An 18-year-old white woman (A-positive) received an ABOmismatched orthotopic liver transplant from a previously healthy man (O-positive) for cryptogenic cirrhosis after her first transplant failed secondary to biliary obstruction. The procedure was uncomplicated. Her hemoglobin fell rapidly from 10.2 mg/dL to 6.6 mg/dL between days 5 and 9 after surgery without evidence of bleeding. High titers of anti-A antibodies were found in the patient's serum, believed to be 170
the result of donor B-cell "engraftment" or B-cell GVHD causing hemolysis. She received O-positive washed red blood cells without complication. After 14 days she developed a fever of 39 °C with no discernible focus; the fever was unresponsive to broad-spectrum antibiotics. She subsequently developed pancytopenia, diarrhea, and a skin rash involving the trunk and extremities, including the palms and soles. Skin biopsy was suggestive of GVHD. A bone-marrow biopsy specimen was 65% cellular with few mature granulocytes and megakaryocytic hyperplasia. The diagnosis of GVHD was made, and the patient was treated with hydrocortisone and antithymocyte globulin. Karyotyping on a bone-marrow aspirate showed a mosaic 46 XX/46 XY, confirming the diagnosis. (Karyotyping done on the patient 3 months earlier was 46 XX.) She died 49 days after surgery. This case exemplifies both the early hemolytic anemia reported in solid organ transplants (2) and the T-cell-mediated GVHD seen in liver transplant patients (1). Still classified as "rare," these cases seem to be recognized more frequently now that the entity has been described. Recently, a report (3) described 4 cases among 175 transplants. We feel it is prudent, given the severity of the disease, to emphasize prevention. One modality not mentioned might be treatment of the donor organ with a dose of radiation that impairs lymphocyte survival without compromising organ survival (4). Alex M. DePaoli, MD University of Chicago Hospitals Chicago, IL 60637 Jacob Bitran, MD Lutheran General Hospital Park Ridge, IL 60068 References 1. Collins RH, Cooper C, Nikaein A, Klintmalm G, Fay J. Graft versus host disease in a liver transplant recipient. Ann Intern Med. 1992; 116:391-2. 2. Ramsey G, Nusbacher J, Starzl TE, Lindsay GD. Isohemagglutinins of graft origin after ABO-unmatched liver transplantation. N Engl J Med. 1984;311:1167-70. 3. Roberts JP, Ascher NL, Lake J, Capper J, Purohit S, Garovoy M, et al. Graft vs. host disease after liver transplantation in humans: a report of four cases. Hepatology. 1991;14:274-81. 4. Von Fliedner V, Higby DJ, Kim U. Graft-versus-host reaction following blood product transfusion. Am J Med. 1982;72:951-61. In response: We appreciate Drs. DePaoli and Bitran's description of another case of graft-versus-host disease (GVHD) in a liver transplant recipient. We suspect the syndrome will be increasingly recognized as the index of suspicion increases (indeed, we have recently observed an additional case). Although irradiation of the liver might be an effective means of preventing GVHD, we would be concerned about possible damage to the liver from a large dose of radiation given over a short time. Careful study of this (or any other) manipulation of the graft is necessary before clinical application. Too little is currently known about the incidence, risk factors, and mechanisms of GVHD in liver transplant recipients to warrant prophylactic measures. Routine depletion of donor organ lymphocytes could possibly have adverse effects. Animal studies, for example, have suggested that limited donor engraftment, with the establishment of a state of low-level, mixed chimerism, may promote acceptance of the organ (1). If this were true in humans, then the depletion of donor lymphocytes could lead to an increased incidence of rejection. Thus, further study of the delicate interplay between donor and host immune systems is needed before attempting further manipulations aimed at reducing the immunologic complications— GVHD and rejection—of liver transplantation. Robert H. Collins, Jr., MD Goran Klintmalm, MD, PhD Joseph W. Fay, MD Baylor University Medical Center Dallas, TX 75246
15 July 1992 • Annals of Internal Medicine • Volume 117 • Number 2
Downloaded from https://annals.org by University of Otago user on 01/19/2019
Reference 1. Sachs DH. Specific transplantation tolerance. N Engl J Med. 1992; 325:1240-2. Chlamydia, Smoking, and Heart Disease
2.
3.
To the Editors: Saikku and colleagues (1) present evidence of an association between chronic Chlamydia pneumoniae infection and the development of coronary heart disease in the Helsinki Heart Study. This association, if confirmed, could have a profound effect on the prevention and management of coronary heart disease. The odds ratios associating evidence of C. pneumoniae infection and coronary heart disease were attenuated (from 2.7 to 2.3 for immunoglobulin A [IgA] titers, 2.1 to 1.8 for immune complexes, and 2.9 to 2.6 for the presence of both factors). There was no statistical difference for IgA titers and immune complexes after controlling for other coronary risk factors, including current smoking. Have the authors considered the possibility that smoking may be related to coronary heart disease through promotion of C. pneumoniae respiratory tract infection, facilitation of penetration of the organism into the blood stream, or both? If true, an association might be found between the interaction of smoking and C. pneumoniae infection and coronary heart disease in their data. David L. Hahn, MD Arcand Park Clinic Madison, WI 53704 Reference 1. Saikku P, Leinonen M, Tenkanen L, linnanmaki E, Ekman MR, Manninen V, et al. Chronic Chlamydia pneumoniae infection as a risk factor for coronary heart disease in the Helsinki Heart Study. Ann Intern Med. 1992;116:273-8. In response: Smoking is known to cause suppression of immune defense mechanisms (1) and, as we showed (2), is associated with markers of a possible chronic Chlamydia pneumoniae infection. The idea that smoking could promote chronic C. pneumoniae infection, as proposed by Dr. Hahn, is quite logical, but markers of chronic C. pneumoniae were found to be independent risk factors for coronary heart disease in both our earlier and current studies (2-4). The proposal that smoking could facilitate the pathogenic mechanisms by which G pneumoniae causes coronary heart disease, if true, would be reflected in the joint effect of smoking and markers of C. pneumoniae on the risk of coronary heart disease. The Helsinki Heart Study (5) instituted a thorough examination of these two risk factors (Leinonen M. and colleagues. Unpublished data.), and, as Dr. Hahn guessed, smoking did have a more than additive effect with markers of C. pneumoniae on the risk of coronary heart disease. Patterns of joint effects seen in the data, however, do not necessarily imply similar patterns at the biological level. The possibility that some symptoms commonly associated with smoking are partly caused by promotion of a chronic C. pneumoniae infection in the lungs, with all the destructive scarring and hypersensitivity typical in chronic chlamydial infections, remains to be studied. Pekka Saikku, MD, PhD University of Helsinki Helsinki, Finland
4.
5.
smoke and other atmospheric contaminants in man and experimental animals. Bacteriol Rev. 1977;41:205-16. Saikku P, Leinonen M, Tenkanen L, Linnenmaki E, Ekman MR, Manninen V, et al. Chronic Chlamydia pneumoniae infection as a risk factor for coronary heart disease in the Helsinki Heart Study. Ann Intern Med. 1992;116:273-8. Saikku P, Leinonen M, Mattila K, Ekman MR, Nieminen MS, Makela PH, et al. Serologic evidence of an association of a novel Chlamydia, TWAR, with chronic coronary heart disease and acute myocardial infarction. Lancet. 1988;2:983-5. Leinonen M, Linnanmaki E, Mattila K, Nieminen MS, Valtonen V, Leirisalo-Repo M, et al. Circulating immune complexes containing chlamydial lipopolysaccharide in acute myocardial infarction. Microb Pathog. 1990;9:67-73. Frick MH, Elo O, Haapa K, Heinonen OP, Heinsalmi P, Helo P, et al. Helsinki Heart Study: primary-prevention trial with gemfibrozil in middle-aged men with dyslipidemia. Safety of treatment, changes in risk factors, and incidence of coronary heart disease. N Engl J Med. 1987;317:1237-45.
Minimal Hepatitis C Infectivity in Semen To the Editors: Both hepatitis B virus and cytomegalovirus are known to be transmitted sexually. To assess the possibility that hepatitis C virus (HCV), a known blood-based infection, could be transmitted by sexual intercourse, samples of serum, semen, and urine were examined for HCV RNA using the polymerase chain reaction. Informed consent was obtained from six men with antibody to HCV (anti-HCV) positivity and chronic hepatitis (mean age, 46.8 ± 1 1 . 7 years). They denied homosexual behavior and intravenous drug abuse; half had a history of blood transfusion. Semen samples were obtained after urine samples were collected to prevent contamination of the specimens. No blood was found in the specimens, and microscopy of the urinary sediment showed erythrocytes at a maximum of 1 to 3 cells per high-power field. Anti-HCV titers (cl00-3 antigen) were measured by enzymelinked immunosorbent assay (ELISA). Hepatitis C virus RNA was purified from 100 /*L of each patient sample, using acid guanidinium thiocyanate-phenol-chloroform extraction (1). Using antisense primer (positions 229-248 of the Okamoto 5' end noncoding region, 5'-AACAT TACTC GGCTA GCAGT-3') and sense primer (positions 7-26 of the Okamoto, 5'-ACTCC ACCAT AGATC ACTCC-3') (2), 45 polymerase chain reaction cycles were done (3). Three of the six patients had spouses. To ascertain the infectivity of HCV among couples, serum antiHCV and HCV-RNA levels were also determined in the spouses by the same method. Hepatitis C virus RNA was detected in the serum of all six patients. However, HCV RNA was not detected in the semen or urine of any of the patients. No relation existed between their clinical data (transfusion history, disease duration, and biochemical data) and HCV-RNA positivity. Neither HCV RNA nor anti-HCV was detected in the serum of the three spouses. Esteban and colleagues (4) have reported that the positive rate of anti-HCV among couples was low. Although anti-HCV positivity does not always correspond to the existence of HCV, the infectivity through body secretions was also suggested to be low by Hsu and colleagues (5). The sensitivity of the polymerase chain reaction was 10 to 100 copies/100 /*L of sample. Transmission of HCV by semen, therefore, is not completely ruled out. Our results, however, suggest low infectivity of semen and urine except when contaminated by blood. Soichiro Terada, MD Koichi Kawanishi, MD Kawaga Medical School Kawaga, Japan
Maija Leinonen, PhD National Public Health Institute Helsinki, Finland
Kazuhiko Katayama, MD Biomedical Laboratory Kawagoe City, Japan
Leena Tenkanen, MSc University of Tampere Tampere, Finland References 1. Holt PG, Keast D. Environmentally induced changes in immunological function: acute and chronic effects of inhalation of tobacco
References 1. Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol chloroform extraction. Anal Biochem. 1987;162:156-9.
15 July 1992 • Annals of Internal Medicine • Volume 117 • Number 2 Downloaded from https://annals.org by University of Otago user on 01/19/2019
171
2. Okamoto H, Okada S, Sugiyama Y, Yotsumoto S, Tanaka T, Yoshizawa H, et al. The 5'-terminal sequence of the hepatitis C virus genome. Jpn J Exp Med. 1990;60:167-77. 3. Saiki RK, Scharf S, Faloona F, Mullis KB, Horn GT, Erlich HA, et al. Enzymic amplification of beta-globin genomic sequences and restriction site analysis for diagnosis of sickle cell anemia. Science. 1985; 230:1350-4. 4. Esteban JI, Esteban R, Viladomiu L, Lopez-Tolavera JC, Gonzalez A, Hernandez JM, et al. Hepatitis C virus antibodies among risk groups in Spain. Lancet. 1989;2:294-7. 5. Hsu HH, Wright TL, Luba D, Martin M, Feinstone SM, Garcia G, et al. Failure to detect hepatitis C virus genome in human secretions with the polymerase chain reaction. Hepatology. 1991;14:763-7. Gallbladder Disease and Angina Pectoris To the Editors: The article by Ryan and colleagues (1) calls to mind the occasional occurrence of high substernal chest pain, clearly related to exercise and relieved promptly by nitroglycerine in patients who also have gallstones. In some patients, removal of the gallstones is followed by disappearance of the angina. The late Dr. Chester Jones at Massachusetts General Hospital, who may have been the first to note this phenomenon, proposed that an overflow of autonomic nervous stimuli from the gallbladder raised the level of autonomic activity in the coronary vessels. Consequently, a smaller exercise stimulus would be needed to exceed the threshold level of coronary autonomic activity required to produce spasm and anginal pain. This "autonomic overflow," as Dr. Jones called it, may be analogous to the production of paralytic ileus by a hip fracture or kidney stones. This association has not been studied in a scientific manner, but it is not a rare observation. In my observations, the anginal pain has not been accompanied by electrocardiographic changes. Arthur B. French, MD Henry Ford Hospital Detroit, MI 48202 Reference 1. Ryan ET, Pak PH, DeSanctis RW. Myocardial infarction mimicked by acute cholecystitis. Ann Intern Med. 1992;116:218-20. In response: As Dr. French correctly points out, the apparent association between gallbladder disease and angina pectoris has been observed for many years. The late Dr. Chester Jones made many major contributions in the first half of this century to understanding the mechanisms of pain arising from abdominal organs, including the gallbladder and the common bile duct. The first reports of angina pectoris attributed to biliary disease, however, precede his time, as one can find in Miller's 1932 review (1) of this subject. The mechanism of this apparent association is still not clear, although the "autonomic overflow" hypothesis is an attractive one. The case we described (2) may be part of the spectrum of cardiac manifestations of primary gallbladder disease. Peter H. Pak, MD Edward T. Ryan, MD Roman W. DeSanctis, MD Massachusetts General Hospital Boston, MA 02114 References 1. Miller CH. The gall-bladder and cardiac pain. Lancet. 1932;1:767-72. 2. Ryan ET, Pak PH, DeSanctis RW. Myocardial infarction mimicked by acute cholecystitis. Ann Intern Med. 1992;116:218-20. Scientific Rigor of Economic Analyses To the Editors: The article by Udvarhelyi and colleagues (1) highlights the methodologic inadequacies of economic analyses in the medical literature. Failure to adhere to basic methodologic principles and inappropriate use of economic terminology are not limited to the medical literature (2). We assessed the methodologic soundness of 65 articles in sue pharmacy journals from 1985 to 1990, using a process similar 172
to that used by Udvarhelyi and colleagues. We used 10 criteria suggested by Drummond and colleagues (3) to assess whether appropriate methodologic techniques were used in each of the 65 articles. Additionally, we assessed the use of the term "cost effective," using criteria suggested by Doubilet and coworkers (4). Only 3 of the 10 methodologic criteria were fulfilled by 50% or more of the studies evaluated. Particular deficiencies included the identification of all relevant costs and consequences, the adjustment of costs and consequences for differential timing (discounting), and the performance of sensitivity and incremental analyses. We also found misinterpretation of the term "cost effective" in 57% of the studies evaluated. No differences were found when articles were subdivided by journal or by year of publication. We conclude that a multidisciplinary initiative is warranted to educate interested researchers, practitioners, and third-party payors in the necessary components of a thorough economic analysis. Jeffrey T, Lee, PharmD Glaxo Research Institute Research Triangle Park, NC 27709 Lisa A. Sanchez, PharmD Brigham and Women's Hospital Boston, MA 02115 References 1. Udvarhelyi S, Colditz GA, Rai A, Epstein AM. Cost-effectiveness and cost-benefit analyses in the medical literature: are the methods being used correctly? Ann Intern Med. 1992;116:238-44. 2. Lee JT, Sanchez LA. Interpretation of "cost-effective" and soundness of economic evaluations in the pharmacy literature. Am J Hosp Pharm. 1991;48:2622-7. 3. Drummond MF, Stoddart GL, Torrance GW. Methods for the economic evaluation of health care programmes. Oxford: Oxford University Press; 1986. 4. Doubilet P, Weinstein MC, McNeil BJ. Use and misuse of the term "cost effective" in medicine. N Engl J Med. 1986;314:253-6. Prince Dracula, Rabies, and the Vampire Legend To the Editors: Immortalized a century ago by Bram Stoker in his novel about Dracula, the legend of the human vampire is a popular fictional theme. In Eastern Europe, written records of vampirism date back to the thirteenth century (1). According to this lore, the vampire is a person whose soul stays in his or her body after death. During the day the vampire sleeps in a grave, leaving it at dusk to assault victims. To do this, he or she may transmogrify into an animal, for example, a bat. The victim usually survives, somewhat anemic, with only toothmarks as signs of what happened. Following a period of incubation, the victim's personality changes and becomes irrational and aggressive. The newly created vampire tries to bite other people, and those who are bitten either die or end up as vampires themselves. Garlic and crucifixes were ascribed some prophylactic value against vampirism, but the only "cure" was to drive a wooden stake through the vampire's heart, preferably before burial. If a vampire was suspected to be on the loose in a village, the churchyard would be dug up, and all corpses that had moved in their graves or were suspiciously well preserved were then treated with the stake (or decapitated as in Graffen zum Stein's 1732 account of an official Turkish autopsy of 13 Serbian vampires). Although we may still be fascinated by the vampire legend, we all now know that the human vampire never really existed. Or did he? A bite from an irrationally aggressive animal leads to aggressively psychotic behavior in the human victim. Doesn't it sound like rabies? In the agony of rabies, all affected mammals may display such a hyperexcitable phase. Even otherwise placid insectivorous bats have been reported to attack humans and other mammals (2). In human rabies a hyperexcitable psychotic phase is also seen. Although genuine biting behavior has rarely been reported, Lindtjorn mentions a 15-year-old rabid boy who bit off one of his mother's fingertips (3). Neither history nor folklore indicate Prince Dracula of
15 July 1992 • Annals of Internal Medicine • Volume 117 • Number 2
Downloaded from https://annals.org by University of Otago user on 01/19/2019
Wallachia (1431-76) to have been a vampire (4, 5), but he was responsible for the impalement of 40 000 to 100 000 people and for committing various cruelties against countless others. The Turkish sultan Mohammad II, after capturing Constantinople, is said to have been deterred from further conquests in Europe when, outside Dracula's capital Tirgoviste (in present Rumania), he saw a forest of 20 000 impaled bodies. The genuine vampire was hardly this cruel prince, although it could very well have been a poor, rabid peasant who ended up buried with a stake driven through his heart.
Correction: Preoperative Cardiac Risk Assessment A sentence in the conclusion of a review by Wong and colleagues (1) appeared incorrectly. The sentence should read: "Therefore, we suggest that preoperative coronary artery revascularization should not be used to protect patients from cardiac complications after vascular surgery unless the patient's cardiac disease warrants these procedures independently of the need for peripheral vascular surgery." Reference 1. Wong T, Detsky AS. Preoperative cardiac risk assessment for patients having peripheral vascular surgery. Ann Intern Med. 1992; 116: 743-53.
Alex Heick, MD, PhM Hvidovre Hospital Copenhagen 2650, Denmark References 1. Sturm D, Volker K. Von denen Vampiren oder Menschensaugern. Miinchen: Carl Hanser; 1968. 2. Constantine DG. Bat rabies in the southwestern United States. Public Health Rep. 1967;82:867-88. 3. Lindtj0rn B. Clinical features of rabies in man. Trop Doct. 1982;12: 9-12. 4. McNally RT, Florescu R. In Search of Dracula. Greenwich, Connecticut: New York Graphic Society; 1972. 5. Florescu R, McNally RT. Dracula: A Bibliography of Vlad the Impaler. London: Robert Hale: 1973.
Correction: Vagal Reflexes and the Aerodigestive Tract A review (1) incorrectly noted an author's degree. The correct degree for Bronwyn Jones is FRCR. Reference 1. Cunningham ET Jr, Ravich WJ, Jones B, Donner MW. Vagal reflexes referred from the upper aerodigestive tract: an infrequently recognized cause of common cardiorespiratory responses. Ann Intern Med. 1992;116:575-82.
15 July 1992 • Annals of Internal Medicine • Volume 117 • Number 2 Downloaded from https://annals.org by University of Otago user on 01/19/2019
173
