Letters

Letters

The Editors welcome submissions for possible publication in the Letters section. Authors of letters should: • Include no more than 300 words of text, three authors, and five references • Type with double-spacing • Send three copies of the letter, a transfer-of-copyright form (see Table of Contents for location) signed by all authors, and a covering letter describing any conflicts of interest related to the contents of the letter • 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 Unpublished letters cannot be returned.

Neil L. Kao, MD University of Illinois College of Medicine Rockford, IL 61107-1897

Predisposition to Cytomegalovirus Infection of the Gastrointestinal Tract

G. Wendall Richmond, MD Rush Presbyterian-St. Luke's Medical Center Chicago, IL 60612-3833

To the Editor: Goodgame's interesting review of cytomegalovirus gastrointestinal disease (1) overlooked several studies of its treatment in patients with the acquired immunodeficiency syndrome (AIDS). In a placebo-controlled study, ganciclovir resulted in greater treatment success for cytomegalovirus colitis (63% compared with 37%), improved endoscopic scores, and decreased dissemination of cytomegalovirus disease (2). Foscarnet was shown to induce remission in 67% of patients who were unresponsive to ganciclovir (3). The concurrent use of these drugs may lead to response in 90% of patients who fail both drugs when they are used alone (4). The mortality rate for untreated cytomegalovirus gastrointestinal disease is high. Switching to foscarnet after failure of ganciclovir therapy prolongs life a median of 5 additional months (3), whereas combination therapy after failure of both agents when they are used alone may increase survival by an additional 6 to 7 months (4). Foscarnet used in a twice-daily dosing schedule (90 mg/kg) decreases adverse effects (5), which allows home administration and thus decreased hospital costs. Michael A. Poles, MD Edward A. Lew, MD Douglas T. Dieterich, MD New York University Medical Center New York, NY 10016 References 1. Goodgame RW. Gastrointestinal cytomegalovirus disease. Ann Intern Med. 1993;119:924-35. 2. Dieterich DT, Kotler DP, Busch DF, Crumpacker C, DuMond C, Dearmand B, et al. Ganciclovir treatment of cytomegalovirus colitis in AIDS: a randomized, double-blind, placebo-controlled multicenter study. J Infect Dis. 1993;167:278-82. 3. Dieterich DT, Poles MA, Dicker M, Tepper R, Lew E. Foscarnet treatment of cytomegalovirus gastrointestinal infections in AIDS patients who have failed ganciclovir induction. Am J Gastroenterol. 1993;88:542-8. 4. Dieterich DT, Poles MA, Lew EA, Mendez PE, Murphy R, Addessi A, et al. Concurrent use of ganciclovir and foscarnet to treat cytomegalovirus infection in AIDS patients. J Infect Dis. 1993;167:1184-8. 5. Dieterich DT, Faust M, Simpson K, Farrel K, Johnson J, Martin-

810

Munley S. Treatment of gastrointestinal cytomegalovirus infection with twice daily foscarnet: a pilot study with pharmacokinetics in patients with HIV [Abstract]. Ninth International Conference on AIDS. Berlin; 1993.

To the Editor: We read with interest the review by Goodgame (1) because one of our patients with systemic lupus erythematosus experienced cytomegalovirus infection complicated by a colonic perforation. I do not believe that therapy with corticosteroids alone encourages the development of widespread cytomegalovirus infection (2). However, the addition of cytotoxic drugs such as cyclophosphamide, azathioprine, or cyclosporine may reactivate latent infection, as in our patient (3, 4). Appreciating this association may help clinicians recognize and make an early diagnosis of cytomegalovirus infection in predisposed hosts.

References 1. Goodgame RW. Gastrointestinal cytomegalovirus disease. Ann Intern Med. 1993;119:924-35. 2. Rubin RH. Infection in the renal transplant patient. In: Rubin RH, Young LS, eds. Clinical Approach to Infection in the Compromised Host. New York: Plenum; 1981:553-605. 3. Dowling JN, Saslow AR, Armstrong JA, Ho M. Cytomegalovirus infection in patients receiving immunosuppressive therapy for rheumatologic disorders. J Infect Dis. 1976;133:399-408. 4. Dummer JS, Hardy A, Poorsattar A, Ho M. Early infections in kidney, heart, and liver transplant recipients on cyclosporine. Transplantation. 1983;36:259-67.

To the Editor: In a recent, extensive review of gastrointestinal cytomegalovirus disease, Goodgame (1) indicated that this condition can occur in patients with healthy immune systems, although it is more common in immunocompromised hosts. Most reports of this infection in normal hosts lack a thorough immunologic evaluation. We reported a case of cytomegalovirus-associated gastric ulcers in a patient with no detectable immunologic abnormalities (2). Our patient was a 42-year-old woman who presented with nausea, vomiting, and abdominal pain. Endoscopic examination of the upper gastrointestinal tract showed three gastric ulcers. Biopsy specimens showed cytomegalovirus inclusion bodies. Serologic test results for cytomegalovirus antibody were negative in serum stored 17 months earlier but showed IgM antibodies at the time of her illness. A thorough work-up for immunologic abnormalities (including antibody testing for the human immunodeficiency virus) was negative, as was a similar repeated evaluation 6 months later. The patient recovered and remains in good health almost 5 years later. David O. Arnar, MD Gunnar Gudmundsson, MD University of Iowa Hospitals and Clinics Iowa City, IA 52242 References 1. Goodgame, RW. Gastrointestinal cytomegalovirus disease. Ann Intern Med. 1993;119:924-35. 2. Arnar DO, Gudmundsson G, Theodors A, Valtysson G, Sigfusson A,

1 May 1994 • Annals of Internal Medicine • Volume 120 • Number 9

Jonasson JG. Primary cytomegalovirus infection and gastric ulcers in normal host. Dig Dis Sci. 1991;33:108-11.

In response: The important work by Dieterich and colleagues on the treatment of gastrointestinal cytomegalovirus infection was not overlooked but was published after submission of my review. Dr. Poles and coworkers have done readers a service by summarizing their results and providing a reference list. I do believe that corticosteroids predispose patients to gastrointestinal cytomegalovirus infection based on evidence from case reports quoted in my review (1-3) and from personal experience (see Figure 2 of the Annals review). However, many (but by no means all) such patients are elderly or have had cancer or another debilitating illness, making the role of corticosteroids in the activation of cytomegalovirus unclear. I certainly agree with Drs. Kao and Richmond's main point that the more potent immunosuppressive therapies are associated with a higher frequency of cytomegalovirus disease. I regret that my review did not include the important case report of Drs. Arnar and Gudmundsson showing that the disease can occur in normal hosts. Their 5-year follow-up, along with the quoted articles and more recent case reports (4, 5), will help clinicians to appreciate this. Richard Goodgame, MD Baylor College of Medicine Houston, TX 77030 References 1. Orloff J, Saito R, Lasky S, Dave H. Toxic megacolon in cytomegalovirus colitis. Am J Gastroenterol. 1989;84:794-7. 2. Wong T, Warner N. Cytomegalic inclusion disease in adults: report of 14 cases with review of literature. Arch Pathol. 1962;74:403-22. 3. Rosen P, Hajdu S. Cytomegalovirus inclusion disease at autopsy of patients with cancer. Am J Clin Path. 1971;55:749-56. 4. Maignan M, Wahl D, Thiaucourt D, Bach D, De Korwin JD, Vaillant G, et al. Self-limited primary cytomegalovirus colitis in an immunocompetent individual. J Intern Med. 1992;232:357-9. 5. Blair S, Forbes A, Parkins R. CMV colitis in an immunocompetent adult. J R Soc Med. 1992;85:238-9.

Prevalence of HIV-1 Syncytium-inducing Phenotype To the Editor: Koot and colleagues (1) reported that patients infected with human immunodeficiency virus (HIV) type 1 syncytium-inducing phenotype strains were more likely to progress to the acquired immunodeficiency syndrome (AIDS) than were patients with non-syncytium-inducing strains. Also, emergence of syncytium-inducing variants was associated with an accelerated CD4 cell decline, but it was not clear whether the relation was causal. We agreed that it was important to know the prevalence of the syncytium-inducing phenotype, especially in patients enrolled in clinical trials, to guard against potential confounding by biased allocation of syncytium- and non-syncytium-inducing phenorypes. Consequently, we did a cross-sectional analysis of HIV-infected patients at Stanford's Center for AIDS Research to determine the prevalence of the syncytium-inducing phenotype in our West Coast referral population and its correlation with CD4+ T-cell levels. Using an MT-2 cell assay, we tested viral isolates from 197 patients enrolled in study protocols at Stanford (2). The stage of disease varied from asymptomatic HIV infection to AIDS. The percentages of patients with the syncytium-inducing phenotype at various CD4 cell counts are shown in Table 1. Overall, 31% of patients harbored syncytium-inducing strains, and the prevalence of these strains was higher in the lower CD4 cell strata. These results concur with Koot and colleagues' finding that syncytium-inducing variants seem to emerge in patients with a CD4 count lower than 500 cells//xL. This finding is important because most clinical trials evaluating drug efficacy involve patients in this CD4 cell range. Most of our patients had significant declines in their CD4 cell counts in association with only the non-syncytium-inducing phenotype; this finding suggests that although a syncytium-inducing phenotype may be a cofactor in an accelerated decline in CD4 cells, it is not a prereq-

Table 1. Syne Determiiaations (MT-2 Tropism) in 197 Patients Ennriled in Protocols at Stailford University* CD4 Count, cells/iiL 0 to 100 101 to 200 201 to 300 301 to 400 401 to 500 501 to 600 >600 Total

Patients

Syncytiuminducing

n 20 51 45 42 24 8 7 197

10 (50) 18 (35) 11 (24) 16 (38) 6(25) 1(12) 0(0) 62 (31)

Non-syncytiuminducing n(%) 10 (50) 33 (65) 34 (75) 26 (62) 18 (75) 7(88) 7 (100) 135 (69)

* Patients had the acquired immunodeficiem cy syndrome (AIDS) or AIDS-related coimplex or were asymptomatic. ]>Jo patient had an active opportunistic inf ection at the 1 ime of phenotyp>e determinations.

uisite for significant immunologic deterioration in most patients. Michael J. Kozal, MD Rani V. Ramachandran, MD Robert W. Shafer, MD Stanford University Medical Center Stanford, CA 94305 References 1. Koot M, Keet IP, Vos AH, de Goede RY, Roos MT, Coutinho RA, et al. Prognostic value of HIV-1 syncytium-inducing phenotype for rate of CD4 + cell depletion and progression to AIDS. Ann Intern Med. 1993;118:681-8. 2. Koot M, Vos AH, Keet RP, de Goede RE, Dercksen MW, Terpstra FG, et al. HIV-1 biological phenotype in long-term infected individuals, evaluated with an MT-2 cocultivation assay. AIDS. 1992;6:4954.

Effects of Lovastatin and Pravastatin on Coronary Artery Disease To the Editor: In comparing the Monitored Atherosclerosis Regression Study (MARS) (1) with another recently published coronary prevention trial that also tested reductase inhibitors against placebo, it appears that a decreased coronary event rate may be related to the particular reductase inhibitor. The MARS trial was a randomized, double-blind, placebocontrolled, multicenter angiographic trial testing the effects of lovastatin on the progression of coronary artery disease in patients with angiographically proven coronary disease. Most patients had multiple risk factors for the disease. Treatment with diet and lovastatin (80 mg/d) for 2.2 years had a favorable effect on lipoprotein levels and angiographically proven coronary disease progression compared with diet and placebo. However, no statistical differences were noted between the placebo and lovastatin groups in regard to coronary events (myocardial infarction, percutaneous transluminal coronary angioplasty, coronary artery surgery, coronary death, and unstable angina). Another randomized, double-blind, placebo-controlled multicenter trial tested pravastatin in persons with cholesterol levels of 5.2 to 7.8 mmol/L (200 to 300 mg/dL) and additional coronary risk factors (2). Most had known coronary disease manifested by angina or previous myocardial infarction. Pravastatin had a favorable effect on lipoprotein levels during the 26-week trial period and significantly reduced cardiovascular event rates compared with placebo. Despite the expected increased reduction in low-density lipoprotein cholesterol levels in patients receiving lovastatin (80 mg/d) compared with pravastatin (20 mg/d) (comparable dosages in terms of potency), Jungnickel and colleagues (3) found that pravastatin, not lovastatin, favorably affected coronary event rates in high-risk persons. The reasons may relate to


811

chemical differences in the reductase inhibitors; lovastatin is lipophilic and pravastatin is hydrophilic (3). Only further, welldesigned comparison studies will resolve this issue. Mark R. Goldstein, MD Crozer-Chester Medical Center Upland, PA 19013 References 1. Blankenhorn DH, Azen SP, Kramsch DM, Mack WJ, Cashin-Hemphill L, Hodis HN, et al. Coronary angiographic changes with lovastatin therapy. The Monitored Atherosclerosis Regression Study (MARS). Ann Intern Med. 1993;119:969-76. 2. Effects of pravastatin in patients with serum total cholesterol levels from 5.2 to 7.8 mmol/liter (200 to 300 mg/dL) plus two additional atherosclerotic risk factors. The Pravastatin Multinational Study Group for Cardiac Risk Patients. Am J Cardiol. 1993;72:1031-7. 3. Jungnickel PW, Cantral KA, Maloley PA. Pravastatin: a new drug for the treatment of hypercholesterolemia. Clin Pharm. 1992;11:677-89.

In response: We thank Dr. Goldstein for his careful review of our work and for bringing attention to several important issues. However, we do not agree that patients in the MARS trial (1) and the Pravastatin Multinational Study trial (2) were "comparable." These studies report results in two distinctly different populations. Compared with patients in the MARS trial, those in the Pravastatin Multinational Study had greater low-density lipoprotein (LDL) cholesterol levels (181 compared with 157 mg/dL) and a lower mean age. The latter study also included more women, more patients with hypertension (uncontrolled hypertension was an exclusion criterion in the MARS trial), a greater number of current cigarette smokers, and fewer persons with a previous myocardial infarction. The only similar factor, angina pectoris, was significantly greater in patients receiving placebo in the Pravastatin Multinational Study. The only angiographic trials to show significant event reduction with lipid-lowering therapies enrolled patients with higher baseline LDL cholesterol levels (180 mg/dL in the Program on the Surgical Control of Hyperlipidemic Study [POSCH], 190 mg/dL in the Familial Atherosclerosis Treatment Study [FATS], and 190 mg/dL in the St. Thomas' Atherosclerosis Regression Study [STARS]. Selection criteria for the two studies differed. The MARS trial used angiographically defined coronary artery disease criteria and excluded patients with left-main disease and those requiring coronary artery bypass surgery (1). Patients in the Pravastatin Multinational Study were selected for an elevated cholesterol level and two additional coronary disease risk factors. Therefore, underlying status of the coronary arteries, including the extent and severity of disease, was probably dissimilar in these two patient groups. The Pravastatin Multinational Study, designed for efficacy and safety, included a mixed population of patients with and without coronary disease. It is therefore not valid to compare events between these two studies. Further, we do not agree with Dr. Goldstein that no favorable effect on events was seen in the MARS trial. Consistent with other coronary angiographic trials, this study showed a 29% (P = 0.22) reduction in coronary events with a decrease in LDL cholesterol levels over a 2-year period (1). Because neither MARS nor the Pravastatin Multinational Study was designed to show a reduction in events, caution should be exercised until large-scale trials designed to do so are completed. We have shown highly significant reductions in coronary events in long-term follow-up of the Cholesterol Lowering Atherosclerosis Study, which examined a cohort similar to that of the MARS trial (3). Therefore, long-term follow-up of coronary angiographic trials is indicated as an alternative to largescale trials. We agree with Dr. Goldstein that different agents may differentially affect end points and clinical events. We have shown that after LDL cholesterol is reduced, other factors (such as triglyceride-rich lipoproteins) assume control of atherogenesis (4, 5). Therefore, it is not unreasonable to assume that agents that beneficially affect other atherogenic fac812

tors may lead to further reduction of coronary artery disease progression as well as coronary events. Howard N. Hodis, MD Linda Cashin-Hemphill, MD Wendy J. Mack, PhD University of Southern California School of Medicine Los Angeles, CA 90033 References 1. Blankenhorn DH, Azen SP, Kramsch DM, Mack WJ, Cashin-Hemphill L, Hodis HN, et al. Coronary angiographic changes with lovastatin therapy. The Monitored Atherosclerosis Regression (MARS). Ann Intern Med. 1993;119:969-76. 2. The Pravastatin Multinational Study Group for Cardiac Risk Patients. Effects of pravastatin in patients with serum total cholesterol levels from 5.2 to 7.8 mmol/liter (220 to 300 mg/dL) plus two additional atherosclerotic risk factors. Am J Cardiol. 1993;72:1031-7. 3. Cashin-Hemphill L, Mack W, LaBree L, Hodis HN, Shircore A, Selzer RH, et al. Coronary progression predicts future cardiac events. Circulation. 1993;88:I363. 4. Hodis HN, Mack WJ, Pogoda JM, Alaupovic P, Hemphill LC, Blankenhorn DH. Effect of triglyceride-rich lipoproteins on progression of early atherosclerotic lesions as determined by coronary angiography in the Mevinolin (lovastatin) Atherosclerosis Regression Study (MARS) [Abstract]. J Am Coll Cardiol. 1993;21:71A. 5. Blankenhorn DH, Alaupovic P, Wickham E, Chin HP, Azen SP. Prediction of angiographic change in native human coronary arteries and aortocoronary bypass grafts. Lipid and non-lipid factors. Circulation. 1990;81:470-6. Distinguishing Central Nervous System Lymphoma from Toxoplasma Encephalitis To the Editor: In their review of primary central nervous system lymphoma, Fine and Mayer (1) recommend an empiric treatment trial for toxoplasmosis in all patients with the acquired immunodeficiency syndrome (AIDS) and a computed tomographic (CT) scan of the cranium suggestive of toxoplasmosis or lymphoma. An alternative approach would be to reserve a toxoplasmosis treatment trial for patients with the highest likelihood of toxoplasmosis, such as those with detectable serum anti-Toxoplasma IgG, those with multiple mass lesions on cranial magnetic resonance imaging (MRI), or those receiving Pneumocystis prophylaxis with agents other than trimethoprim-sulfamethoxazole. The likelihood of Toxoplasma encephalitis is low in human immunodeficiency virus (HlV)-infected patients with nondetectable serum mti-Toxoplasma IgG (2). In addition, a single lesion viewed by cranial MRI is more likely to be caused by primary central nervous system lymphoma than to toxoplasmosis (3). The presence of Epstein-Barr virus DNA in cerebrospinal fluid has been shown to be a sensitive and specific way to distinguish primary central nervous system lymphoma from Toxoplasma encephalitis (4). Preliminary data also suggest that thallium single-photon emission computed tomography (SPECT) may also be able to make this distinction accurately (5). Rather than undergoing a low-yield treatment trial, patients who are unlikely to have Toxoplasma encephalitis and in whom primary central nervous system lymphoma is suspected could be further evaluated with such tests as cerebrospinal fluid studies for Epstein-Barr virus DNA, thallium SPECT, or brain biopsy. Christina Marra, MD University of Washington School of Medicine Seattle, WA 98104-2499 References 1. Fine HA, Mayer RJ. Primary central nervous system lymphoma. Ann Intern Med. 1993;119:1093-104. 2. Luft BJ, Brooks RG, Conley FK, McCabe RE, Remington JS. Toxoplasmic encephalitis in patients with acquired immune deficiency syndrome. JAMA. 1984;252:913-7. 3. Ciricillo SF, Rosenblum ML. Imaging of solitary lesions in AIDS [Letter]. J Neurosurg. 1991 ;74:1029. 4. Cinque P, Brytting M, Vago L, Castagna A, Parravicini C, Zanchetta N, et al. Epstein-Barr virus DNA in cerebrospinal fluid from patients


with AIDS-related primary lymphoma of the central nervous system. Lancet. 1993;342:398-401. 5. Ruiz A, Ganz WI, Donnovan Post MJ. Use of thallium-201 brain SPECT to differentiate cerebral lymphoma from Toxoplasma encephalitis in AIDS patients. 31st Annual Meeting of the American Society of Neuroradiology, 16 to 20 May 1993, Vancouver, British Columbia, Canada.

In response: We agree with Dr. Marra that in a patient with AIDS who has a single intracerebral lesion and a negative serum anti-toxoplasmosis IgG, the likelihood of toxoplasmosis encephalitis is small. In such a case, proceeding directly to biopsy is reasonable. Thallium-SPECT scanning and assessment of Epstein-Barr virus DNA in cerebrospinal fluid are promising new diagnostic modalities; however, their value in differentiating between cerebral toxoplasmosis and lymphoma has not been tested in large-scale, prospective trials. In general, if some chance exists that newly diagnosed intracerebral lesions are the result of toxoplasmosis and the patient is neurologically stable, we believe a 10-day course of antitoxoplasmosis therapy is preferable to an invasive procedure. As Dr. Marra's letter points out, any proposed diagnostic algorithm merely represents a framework for physicians to make educated and appropriate decisions; however, each patient must be considered as an individual. Howard A. Fine, MD Robert J. Mayer, MD Dana-Farber Cancer Institute Harvard Medical School Boston, MA 02115

Charlton Wilson, MD United States Public Health Service Indian Hospital Mescalero, NM 88340 Brian Hjelle, MD Steven Jenison, MD University of New Mexico School of Medicine Albuquerque, NM 87131 Disclaimer: The opinion expressed is that of the authors and does not necessarily reflect the opinion of the U.S. Public Health Service. References 1. Outbreak of acute illness—Southwestern United States, 1993. MMWR Morb Mortal Weekly Rep. 1993;42:421-4. 2. Nichol ST, Spiropoulou CF, Morzunov S, Rollin PE, Ksiazek TG, Feldmann G, et al. Genetic identification of a hantavirus associated with an outbreak of acute respiratory illness. Science. 1993;262: 914-7. 3. Hjelle B, Jenison S, Torrez-Martinez N, Yamada T, Nolte K, Zumwalt R, et al. A novel hantavirus associated with an outbreak of fatal respiratory disease in the southwestern United States: evolutionary relationships to known hantaviruses. J Virol. 1994;68:592-6. 4. Hjelle B, Jenison S, Torrez-Martinez N, Damrow TA, Centers for Disease Control and Prevention. Progress in development of hantavirus diagnostic assays—United States. MMWR Morb Mortal Weekly Rep. 1993;42:770-2. 5. Jenison S, Yamada T, Morris C, Anderson B, Torrez-Martinez N, Keller N, et al. Characterization of human antibody responses to Four Corners hantavirus infections among patients with hantavirus pulmonary syndrome. J Virol. 1994; [In press].

5-Fluorouracil Dermatitis Prophylaxis with a Nicotine Patch

Probable Hantavirus Pulmonary Syndrome That Occurred in New Mexico in 1975 To the Editor: We describe a case of acute pulmonary disease consistent with the hantavirus pulmonary syndrome (1-3). In March 1975, fever and severe myalgias developed in a 37-year-old Native American ranch hand from New Mexico. He was hospitalized 3 days later with dyspnea and hypoxemia. Laboratory findings were as follows: arterial blood pH, 7.48; PC0 2 , 22 mm Hg; P 0 2 , 41 mm Hg; hematocrit, 0.59; leukocyte count, 31 900 x 106/L; and platelet count, 44 000 x 106/L. Peripheral blood contained immature neutrophils, including promyelocytes. Serum protein was 49 g/L, albumin was 28 g/L, lactate dehydrogenase level was 8.10 ^tkat/L, and aspartate aminotransferase was 1.33 ju,kat/L. A chest radiograph showed bilateral lower-lobe infiltrates. Bacterial cultures of sputum, blood, bone marrow, and cerebrospinal fluid were negative. Acid-fast bacilli cultures of sputum and bone marrow were negative. Cold agglutinins were not found. On the fifth day of illness, a chest radiograph showed pulmonary infiltrates suggestive of the adult respiratory distress syndrome. Respiratory insufficiency necessitated mechanical ventilation. The patient's arterial blood pH was 7.23, his P 0 2 was 72 mm Hg, and his P C 0 2 was 54 mm Hg with a delivered oxygen concentration of 1.0. His platelet count was 20 000 x 106/L. His respiratory status improved, and he was extubated after 5 days. His platelet count was normal. The discharge diagnosis was viral pneumonia. Since then, he has been generally healthy. During a November 1993 clinic visit, it was noted that the 1975 illness resembled the hantavirus pulmonary syndrome. No samples from 1975 were available. A serum sample obtained on 16 November 1993 was tested for antibodies to Four Corners hantavirus nucleocapsid and glycoprotein-1 recombinant proteins by Western immunoblot assay (3-5). The serum contained IgG antibodies to both Four Corners hantavirus nucleocapsid and glycoprotein-1 proteins; no IgM reactivities were detected. Nucleocapsid and glycoprotein-1 IgG reactivities were mapped to the same antigen locations as antibodies from patients with acute hantavirus pulmonary syndrome (5). Because the clinical features of the 1975 illness are highly suggestive of the syndrome, we believe it represents remote Four Corners hantavirus infection as confirmed by serologic test results.

To the Editor: Dermatologic toxicity (manifested by erythema, photosensitivity, hyperpigmentation, desquamation, xerosis or the "hand and foot syndrome") occurs in up to 20% of patients with cancer who receive 5-fluorouracil. Rarely, alopecia occurs as well. A 65-year-old woman with metastatic colorectal carcinoma started receiving chemotherapy consisting of leucovorin and 5-fluorouracil, given simultaneously as a continuous infusion over 24 hours once a week. Within a few months, the patient noticed hyperpigmentation of the hands, excluding the palms. A few months later, she developed 2 + desquamating erythema of the palms and 2 + hyperpigmentation of the rest of the hands and feet. No noticeable skin changes involved the soles. She also developed 1 + pitting of the nails, which were also brittle. A nicotine patch (7.0 mg) was applied to the skin 1 hour before the 5-fluorouracil infusion was begun and was removed 1 hour after completion of each 24-hour 5-fluorouracil infusion. Since then, the desquamation, erythema, and hyperpigmentation have completely resolved, despite continued chemotherapy. The improvement of the skin changes may have been caused by the vasoconstricting properties of the nicotine patch, thereby decreasing the delivery of 5-fluorouracil to the skin. Oncologists have been taking advantage of the beneficial effects of ice application to the scalp and to the oral mucosa in reducing the severity of chemotherapy-associated alopecia and mucositis, respectively. Presumably, the reduction of drug delivery to these areas by cooling-induced vasoconstriction is the mechanism of action. If so, chewing a stick of nicotine gum before bolus chemotherapy may help to reduce other dermatologic toxicities, including alopecia. Edwin C. Kingsley, MD Southwest Cancer Clinic Las Vegas, NV 89119 Pharmaceutical Desktop References To the Editor: Several compendia can be added to the six pharmaceutical desktop references described by Dr. Day (1). Widely available and distributed for free through drug company mailing lists is the Physician's Compendium of Drug

1 May 1994 • Annals of Internal Medicine

• Volume 120 • Number 9

813

Therapy (Compendium Publications Group, Secaucus, New Jersey). Its organization and general usefulness are most similar to Facts and Comparisons, but the drug descriptions are short. For physicians who frequently examine and treat travellers from abroad, two books besides Martindale: The Extra Pharmacopoeia are useful. The Index Nomina is also known as the International Drug Directory (Medpharm Publishers, Stutgart, Germany, 1990, available through Drug Intelligence, Inc., Hamilton, Illinois). Multiple synonymous international pharmaceutical brand names refer the reader to the generic name, which is accompanied by a short description of indications and use. Similar but more readily available and less expensive is Drugs Available Abroad (Gale Research, Detroit, Michigan). It may be found in community public libraries. Consulates and even liaison organizations such as Chambers of Commerce may have drug guides from their countries, but proficiency in the relevant language is necessary. It will probably be several years before these reference guides are electronically consolidated—informationally and linguistically—and available in the physician's office at a reasonable price. Mark A. LaPorta, MD Gateway Healthcare Miami Beach, F L 33141

To the Editor: Dr. Day's evaluation of tertiary drug reference texts (1) contains some inaccurate and incomplete information. Day states that the Physicians9 Desk Reference has four separate indexes, making location of generic names difficult. In fact, one index is for both generic and chemical names, and a quick scan of this index revealed only generic names. He also says that USP Drug Information and AHFS Drug Information organize the approved labeling more effectively than the Physicians9 Desk Reference. These monographs are based on primary references, not the package insert. Although neither contains bibliographic references, the editors of AHFS Drug Information may be contacted for literature citations. Although the product illustrations in USP Drug Information are more useful than those in the Physicians9 Desk Reference, neither is complete. Facts and Comparisons does not contain an illustrative section. The appendix in AHFS Drug Information to which Dr. Day refers is not an index of investigational drugs for the acquired immunodeficiency syndrome but rather a small and incomplete listing of treatment investigational new drugs. Most of the cited deficiencies of each reference reflect the delay between authorship and publication or differences in focus. For example, the Food and Drug Administration approved lisinopril for congestive heart failure in June 1993, but the 1993 edition of the Physicians9 Desk Reference was published in the fall of 1992. Because AHFS Drug Information primarily focuses on the needs of the institutional practitioner, many medications commonly used in the outpatient setting are not included. We strongly disagree with high commendation given to Martindale: The Extra Pharmacopoeia, editions of which are avail-

1 May 1994 • Annals of Internal Medicine

Arleen E. Petrucco, PharmD Hahnemann University Hospital Philadelphia, PA 19102 Sarah C. Erush, PharmD Hospital of the University of Pennsylvania Philadelphia, PA 19104 Angela A. Allerman, PharmD Thomas Jefferson University Hospital Philadelphia, PA 19107 Reference 1. Day RP. Desktop drug information. Ann Intern Med. 1993;119:936-8.

Reference 1. Day RP. Desktop drug information. Ann Intern Med. 1993;119:936-8.

814

able only every 3 years. This book contains information on many medications not available in the United States but lacks comprehensiveness regarding U.S. medications. In addition, its reference list for each monograph is often incomplete and outdated. No tertiary reference source is completely comprehensive, and the cost of a full reference collection is fairly prohibitive. Most physicians would be best served using university hospital and drug information centers staffed by specially trained pharmacists who can provide comprehensive drug information from a wide range of resources.

In response: I appreciate Dr. LaPorta's additions. The Physician's Compendium of Drug Therapy has a brand-name orientation, incomplete package insert information, and advertisements. Another work of this type is Physician GenRX, which gives the verbatim Food and Drug Administration (FDA) package inserts listed in alphabetical order by generic name. Petrucco and colleagues are correct that AHFS Drug Information lacks a bibliography and includes in its appendix several types of investigational drugs. However, they are incorrect in stating that the USP Drug Information has no references or that Facts and Comparisons has no illustration section. Regarding the latter discrepancy, I suspect they were referring to the loose-leaf version used by pharmacists, not the annual hardbound edition. Regarding the Physicians9 Desk Reference and its four indexes, my intention was that a person not familiar with a given drug name could not necessarily know in which index to look; editorial revision of this sentence obscured this. The other specific points raised by Petrucco and colleagues need further explanation. I refer to USP Drug Information and AHFS Drug Information as "reorganized FDA package insert information" because I was struck by a pattern of common phraseology. I cannot comment on the intent of the editors, only the result. In citing examples, I described in as much detail as space allowed the data each reference provided. I chose lisinopril to test whether a text was willing to go beyond current FDA indications. Although Martindale may only be available every 3 years, its inclusion of drugs available in Europe and not in the United States makes it less likely to be dated and is useful for European readers. Richard P. Day, MD Madison Internal Medicine Associates Madison, WI 53715

• Volume 120 • Number 9