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 with the letter 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 only if their letter is accepted. Unpublished letters cannot be returned. Cost Effectiveness of Treating Asymptomatic HIV Infection To the Editors: We read with interest the article by Schulman and colleagues (1). Our major concern about their method is the use of data on time to progression to the acquired immunodeficiency syndrome (AIDS) to make statements about survival. This is an extraordinarily important variable for which almost no data exist. They draw from the randomized AIDS Clinical Trial Group Protocol 019, which had only eight deaths, four in the placebo-treated group and four in the zidovudine-treated group (2). Zidovudine prolongs survival in human immunodeficiency virus (HlV)-infected patients once they progress to AIDS or to AIDS-related complex (3-5). It is not clear, however, if zidovudine will provide the same benefit for patients who are already receiving the drug when they progress. The small amount of data available from Protocol 019 suggests that the death rate may be higher for patients already taking zidovudine when they progress than for untreated patients who progress; these two groups of patients had the same number of deaths, but the latter group was followed longer (2). It is therefore possible that progression to AIDS or AIDS-related complex is delayed by early (as compared to later) treatment with zidovudine but that survival is either not affected or adversely affected. If this is the case, many of the conclusions of Dr. Schulman and colleagues would need to be modified. We believe that this issue is extremely important because of the already high cost of treatment for HIV infection. In addition, data on quality of life, cumulative toxicity of the drug, and emergence of viral resistance should be taken into account when analyzing the cost effectiveness of zidovudine therapy. John D. Hamilton, MD Michael S. Simberkoff, MD Pamela M. Hartigan, PhD The Veterans Affairs Cooperative Study Group on Treatment of AIDS Durham Veterans Affairs Medical Center Durham, NC 27705 References 1. Schulman KA, Lynn LA, Glick HA, Eisenberg JM. Cost effectiveness of low-dose zidovudine therapy for asymptomatic patients with human immunodeficiency virus (HIV) infection. Ann Intern Med. 1991; 114:798-802. 2. Volberding PA, Lagakos SW, Koch MA, et al. Zidovudine in asymptomatic human immunodeficiency virus infection: a controlled trial in
persons with fewer than 500 CD4-positive cells per cubic millimeter. N Engl J Med. 1990;322:941-9. 3. Fischl MA, Richman DD, Grieco MH, et al. The efficacy of azidothymidine (AZT) in the treatment of patients with AIDS and AIDSrelated complex. N Engl J Med. 1987;317:185-91. 4. Lemp GF, Payne SF, Neal D, et al. Survival trends for patients with AIDS. JAMA. 1990;263:402-6. 5. Moore RD, Hidalgo J, Sugland BW, Chaisson RE. Zidovudine and the natural history of the acquired immunodeficiency syndrome. N Engl J Med. 1991;324:1412-6. In response: The comments of Hamilton and colleagues about our paper (1) reflect the lack of available information on the ultimate survival benefit for patients with asymptomatic HIV infection who are treated with zidovudine. We examined the effects of changes in life expectancy after the development of AIDS in patients who received zidovudine while they were asymptomatic. Our model indicates that for the one-time effect, treated asymptomatic patients would need to lose 6 of their 16 months of survival after the development of AIDS or advanced AIDS-related complex for there to be no difference in discounted life expectancy between this group and patients receiving placebo. For the continuous effect model, even if asymptomatic, treated patients died immediately after developing AIDS or advanced AIDS-related complex, they still would live 2.76 discounted years longer than patients receiving a placebo. One reason why we did not address this issue in our original paper is that we do not share the authors' concerns about the mortality data presented in the report of the AIDS Clinical Trial Group Protocol 019 (2). Although there were equal numbers of deaths in placebo-treated and zidovudine-treated patients, only 428 patients were treated with placebo (approximately 502 person-years of follow-up) as compared with 910 patients treated with zidovudine (approximately 927 personyears of follow-up). Thus, there was an apparent—but not statistically significant (P > 0.2)—50% reduction in mortality rates in zidovudine-treated patients once progression to AIDS had occurred. The AIDS Clinical Trial Group study did not report this result, possibly because total mortality was not a primary study outcome (more than 4500 patients would have been needed in each arm of the trial to detect this mortality difference with an alpha of 0.05 and a beta of 0.8). We agree that in economic analyses of technologies early in their diffusion, there are large amounts of uncertain data that may affect their eventual cost-effectiveness ratios. Extensive sensitivity analyses are required to explore the effects of this uncertainty and to examine the robustness of the cost-effectiveness models. Given the wide range of cost-effectiveness ratios in our primary analyses and the robustness of these ratios to a large number of sensitivity analyses, we stand by our original conclusions. Kevin A. Schulman, MD Lorna A. Lynn, MD Henry A. Glick, MA Hospital of the University of Pennsylvania Philadelphia, PA 19104 References 1. Schulman KA, Lynn LA, Glick HA, Eisenberg JM. Cost effectiveness of low-dose zidovudine therapy for asymptomatic patients with human immunodeficiency virus (HIV) infection. Ann Intern Med. 1991; 114:798-802. 2. Volberding PA, Lagakos SW, Koch MA, et al. Zidovudine in asymptomatic human immunodeficiency virus infection: a controlled trial in persons with fewer than 500 CD4-positive cells per cubic millimeter:
15 October 1991 • Annals of Internal Medicine
Downloaded From: https://annals.org/ by a Univ of Utah User on 09/26/2017
• Volume 115 • Number 8
655
the AIDS Clinical Trials Group of the National Institute of Allergy and Infectious Diseases. N Engl J Med. 1990;322:941-9.
Fetal Hemoglobin Alters Hemoglobin A l c Measurements To the Editors: The letter of Holt and colleagues (1) emphasizes that diabetic subjects with hemoglobin S or C traits will have a falsely low hemoglobin A, c value if measured by certain methods. We also have observed this finding but would like to highlight another common condition which alters hemoglobin A l c values. Persons with the trait or who are homozygous for the hereditary persistence of fetal hemoglobin will have a falsely elevated hemoglobin A l c value (2) because hemoglobin F co-elutes with the hemoglobin A l c fraction. Some patients who were referred to our institution for presumed diabetes based on an elevated hemoglobin A l c were not judged to be glucose intolerant using other criteria (such as the presence of glycosuria and fasting or postprandial hyperglycemia). These patients were subsequently found to have elevated levels of hemoglobin F. Because the heterozygous and homozygous states of the hereditary persistence of fetal hemoglobin are generally asymptomatic (3), these persons may be first identified with discrepant findings between hemoglobin A l c and measures of glucose intolerance as in our cases. The prevalence of the heterozygous state for this trait is not precisely known because this condition is asymptomatic and several variants of hereditary persistence of fetal hemoglobin exist (3). Awareness of this condition is germane not only to the diagnosis of diabetes but also to its management because falsely elevated values may result in unrealistic patient goals and unnecessary hypoglycemia if the hemoglobin A l c value is used as a "gold standard" for glucose control. We have used the fructosamine assay rather than a more expensive method for the measurement of hemoglobin A l c for diabetic subjects with elevated levels of hemoglobin F, S, or C (4) because the fructosamine assay is independent of hemoglobin. Richard W. Bergstrom, MD James R. Kelley, MD W. Kenneth Ward, MD Portland Diabetes and Endocrinology Center and Good Samaritan Hospital Portland, OR 97210 References 1. Holt GS, Wofford JL, Velez R. Hemoglobinopathies affect hemoglobin A l c measurement [Letter]. Ann Intern Med. 1991;115:68-9. 2. Bunn HF. Evaluation of glycosylated hemoglobin in diabetic patients. Diabetes. 1981;30:613-7. 3. Weatherall DJ, Clegg JB. Hereditary persistence of fetal haemoglobin. Br J Haemat. 1975;29:191-8. 4. Baker JR, O'Connor JP, Metcalf PA, Lawson MR, Johnson RN. Clinical usefulness of estimation of serum fructosamine concentration as a screening test for diabetes mellitus. Br Med J. 1983;287: 863-7.
Barriers to Treating Obesity To the Editors: Dr. Bray's editorial (1) elucidates some of the important barriers to treating obesity. As stated, "the public perception that obesity is a disease resulting from a lack of will power" is a major barrier to the treatment of a disease with fundamental genetic, molecular determinants. "Limited research funding" is surely another barrier. However, I am concerned about Dr. Bray's wish to lower barriers to the use of appetite-suppressant drugs. I do not find "clear and convincing evidence that modern appetite-suppressant drugs are effective and safe." Fenfluramine, one of the drugs cited, has been shown to have potentially serious side effects. At least three studies have shown pulmonary hypertension occurring after treatment with fenfluramine (2-4). In one instance, the pulmonary hypertension did not resolve with cessation of treatment and was lethal (2). There is evidence that other anorectic agents, for example, 656
aminorex and chlorphentermine, may produce pulmonary hypertension more frequently than fenfluramine (4). If any currently available appetite suppressants were reliably associated with major persistent weight loss, then one might be willing to take the risks and use them, taking care not to overestimate their efficacy. A recent study of the use of dexfenfluramine in 822 obese patients in a multicenter international trial showed less than a 3-kg difference in weight loss when the drug-treated group was compared with the placebotreated group (5). Weight loss curves in this study show a slow upward creep after 8 months in both placebo-treated and drugtreated groups, suggesting that in the long run one might expect considerable weight regain despite continued treatment (5). I disagree with the editorial (1), which brands recent hearings on the "diet pill industry" as "legislative grand standing" which have "produced little else but publicity for the legislators." Congressman Wyden and others have become deeply interested in working with the National Institutes of Health (NIH) to widen the research base for obesity and to inform the public about the present inadequate status of obesity treatment. The greatest barrier to the treatment of obesity remains our incomplete understanding of the complex genetic, developmental, and psychosocial influences that lead to obesity. Jules Hirsch, MD The Rockefeller University New York, NY 10021 References 1. Bray GA. Barriers to the treatment of obesity [Editorial]. Ann Intern Med. 1991;115:152-3. 2. McMurray J, Bloom fie Id P, Miller HC. Irreversible pulmonary hypertension after treatment with fenfluramine. Br Med J. 1986;292: 239-40. 3. Douglas JG, Munro JF, Kitchin AH, Muir AL, Proudfoot AT. Pulmonary hypertension and fenfluramine. Br Med J. 1981;283:881-3. 4. Pouwels HM, Smeets JL, Cheriex EC, Wouters EF. Pulmonary hypertension and fenfluramine. Eur Respir J. 1990;3:606-7. 5. Guy-Grand B, Crepaldi G, Lefebvre P, Apfelbaum M, Gries A, Turner P. International trial of long-term dexfenfluramine in obesity. Lancet. 1989;1:1142-5.
To the Editors: Although we agree with Dr. Bray's (1) observations about attitudes toward obesity, we disagree with his assertion about the "convincing evidence" of appetite-suppressant drugs' safety and efficacy. As he noted, relatively few clinical trials have assessed these agents. Phenylpropanolamine, an over-the-counter medication, is as effective as other appetite-suppressant drugs (2). Despite its widespread availability, consumers still seek other therapies. Even when overthe-counter drugs, such as nonsteroidal anti-inflammatory drugs, are effective, companies actively investigate new "me-too" agents. Although long-term appetite suppressants may deserve further study, the effect of short-term use is reversed after therapy is discontinued. Craighead and colleagues (3) found that during treatment, fenfluramine plus behavior therapy resulted in greater weight loss than behavioral management alone. However, in the year after use, patients on the drug regained more weight, suggesting that its use compromised the development of permanent abilities to alter behavior. We concur that obesity cannot be "cured" and would welcome additional studies on its management. Some might argue that no obesity treatment has provided long-term efficacy (4). For the present, alterations in nutrition and physical activity, along with behavior modification, can be effective, especially when individuals develop personal strategies for weight loss (5). Although lifestyle alterations are difficult, we do not believe that appetite suppressants should replace patient education and use of nonpharmacologic therapies. Diane L. Elliot, MD Linn Goldberg, MD Oregon Health Sciences University Portland, OR 97201
15 October 1991 • Annals of Internal Medicine • Volume 115 • Number 8
Downloaded From: https://annals.org/ by a Univ of Utah User on 09/26/2017
References 1. Bray GA. Barriers to the treatment of obesity. Ann Intern Med. 1991;115:152-3. 2. Weintraub M, Bray GA. Drug treatment of obesity. Med Clin N Am. 1989;73:237. 3. Craighead LW, Stunkard AJ, O'Brien RM. Behavior therapy and pharmacotherapy for obesity. Arch Gen Psychiat. 1981;38:763-8. 4. Fitzgerald FT. The problem of obesity. Ann Rev Med. 1981;32:22131. 5. Kayman S, Bruvold W, Stern JS. Maintenance and relapse after weight loss in women: behavioral aspects. Am J Clin Nutr. 1990;52: 800-7.
In response: I am grateful for the letters from Drs. Eliot and Goldberg and from Dr. Hirsch because they provide an opportunity to re-emphasize my thesis (1) that barriers exist to the treatment of obesity. My central thesis is that both the public and health professionals view obesity as a character disorder resulting from a lack of will power rather than as a disease for which we currently have no cure. When drugs are discontinued, we expect the patient to be "cured." Regain in weight is used to condemn drugs rather than being viewed as the expected result of termination of treatment. A recent paper by Weintraub and colleagues (2) and an earlier review by Stunkard (3) support the contention that long-term use of appetite suppressants may temporarily "reset" the regulatory system for body weight. "Safety" and "effectiveness" are relative terms. For any overweight person, the initial treatment decision should be based on evaluation of the "risk" of obesity relative to the risks of any proposed treatment (4). Dr. Hirsch properly points out three reports of pulmonary hypertension and one death with fenfluramine among the many thousands who have been treated. If one compares this with the mortality of 0.5% to 5% associated with surgical treatment of obesity, one could argue that treatment with appetite suppressants might offer a "safer" alternative than surgery. Both experimental and clinical evidence show that appetite suppressants can produce long-term weight reduction (1, 2). Currently, the Food and Drug Administration requires an investigational new drug license to use appetite suppressants for long-term treatment of obese patients. Few physicians are likely to surmount this barrier. I am encouraged by the comments of Dr. Hirsch about the activity of Congressman Wyden and his committee. If the efforts of Dr. Hirsch and others are effective in increasing the support for obesity research by the NIH, we will have made substantial progress in breaking down the barriers to future advances in this field. George A. Bray, MD Pennington Biomedical Research Center Baton Rouge, LA 70808 References 1. Bray GA. Barriers to the treatment of obesity [Editorial]. Ann Intern Med. 1991;115:152-3. 2. Weintraub M, Fundaresan PR, Schuster B, Ginsberg G, Madan M, Stein EC, et al. National Heart, Lung and Blood Institute long-term weight control study: I-VII. Clin Pharmacol Ther. 1991; in press. 3. Stunkard AJ. Anorectic agents lower body weight set point. Life Sciences. 1982;39:2043-55. 4. Bray GA. Classification and evaluation of the obesities. Med Clin North Am. 1989;73:161-84.
Dementia Screening To the Editors: The review of dementia screening and evaluation by Siu (1) was well done. To be a useful clinical guide for the practicing physician, however, two points need to be emphasized. First, the use of a criterion-based definition of dementia is essential. The major source of error in diagnosing dementia is in properly assessing its presence or absence. O'Connor and colleagues (2) demonstrated in a large community survey with extensive clinical evaluation that 42% (87 of 208) of demented patients were not recognized as demented by their personal
physician. They also found that 22% (51 of 236) of nondemented persons were mistakenly diagnosed as demented. The misdiagnosis rate was 31% (138 of 444), with almost 40% of the errors being overdiagnoses, which are not generally the result of inadequate screening. Furthermore, Somerfield and colleagues (3) surveyed 127 physicians who had referred patients to a dementia clinic and found that fewer than 30% used either of the two sets of specific criteria to diagnose dementia. Using these criteria along with any of the screening tests would probably improve the diagnosis of dementia more than cerebral imaging, laboratory testing, or choice of screening instruments. Second, the educational bias of mental status tests was not stressed. My colleagues and I recently confirmed previous findings that scores on the Mini-Mental State Examination of Folstein and coworkers (4) are significantly affected by the educational level of those tested (5). Our findings agreed with Siu's conclusion that two cut-off points are needed for MiniMental State Examination scores. We found, however, the cut-off scores to be less than or equal to 17 for high likelihood, 18 to 23 for intermediate likelihood, and 24 to 30 for low likelihood of dementia in persons with less than a ninth-grade education. Additionally, a study on clock drawing also shows the effects of education on performance. The clinician must be aware of these effects when interpreting test results in poorly educated persons. In summary, the helpful suggestions by Siu can greatly improve the diagnosis of dementia, when combined with the use of specific criteria to define dementia and with altered cut-off points in test scores for poorly educated persons. Robert A. Murden, MD The Ohio State University Columbus, OH 43210 References 1. Siu AL. Screening for dementia and investigating its causes. Ann Intern Med. 1991;115:122-32. 2. O'Connor DW, Politt PA, Hyde JB, Brooks CPB, Reiss BB, Roth M. Do general practitioners miss dementia in elderly patients? Br Med J. 1988;297:1107-10. 3. Somerfield MR, Weisman CS, Ury W, Chase GA, Folstein MF. Physician practices in the diagnosis of dementing disorders. J Am Geriatr Soc. 1991;39:172-5. 4. Folstein MF, Folstein SE, McHugh PR. "Mini-Mental S t a t e ' : a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12:189-98. 5. Murden RA, McRae TD, Kaner S, Bucknam ME. Mini-Mental State Exam scores vary with education in blacks and whites. J Am Geriatr Soc. 1991;39:149-55.
To the Editors: Dr. Siu's recent review (1) fails to mention the growing role of emission tomography in screening for dementia. He is correct in pointing out the limits of conventional anatomic imaging modalities (computed tomography and magnetic resonance imaging) in the evaluation of these patients. Unlike stroke, dementia, particularly that seen in Alzheimer disease, does not result from a structural, gross anatomic defect. Dementia is probably a consequence of metabolic changes in the brain and is thus more likely to be demonstrated by functional imaging modalities. Many reports (2-5) argue strongly that positron-emission tomography and single photon emission computed tomography are of significant help in differentiating demented from nondemented patients. A report being prepared by the Consensus Panel on Dementia sponsored by the Brain Imaging Council of the Society of Nuclear Medicine supports this position. Ronald S. Tikofsky, PhD Medical College of Wisconsin Milwaukee, WI 53226 References 1. Siu AL. Screening for dementia and investigating its causes. Ann Intern Med. 1991;115:122-32. 2. Prohovnik I, Mayeux R, Sackeim HA, Smith G, Stern Y, Alderson PO. Cerebral perfusion as a diagnostic marker of early Alzheimer's disease. Neurol. 1988;38:931-7. 3. Jagust WJ, Budinger TF, Reed BR. The diagnosis of dementia with
15 October 1991 • Annals of Internal Medicine
Downloaded From: https://annals.org/ by a Univ of Utah User on 09/26/2017
• Volume 115 • Number 8
657
single photon emission computed tomography. Arch Neurol. 1987; 44:258-62. 4. Hellman RS, Tikofsky RS, Collier BD, et al. Alzheimer's disease: quantitative analysis of I-123-iodoamphetamine SPECT brain imaging. Radiology. 1989;172:183-8. 5. Johnson KA, Holman BL, Rosen TJ, Nagel JS, English RJ, Growdon JH. Iofetamine 1-123 single photon emission computed tomography is accurate in the diagnosis of Alzheimer's disease. Arch Intern Med. 1990;150:752-6.
In response: My article (1) addressed the methods by which primary care physicians can screen for dementia and investigate its various causes. I did not mention the role of emission tomography because the literature does not support its application in clinical practice at this time. All four articles (2-5) cited by Dr. Tikofsky investigate the use of emission tomography in distinguishing patients with Alzheimer disease (many with moderate to severe disease) from healthy subjects who were not suspected of having dementia. Yet it is almost inconceivable that a physician would have to use a $2340 test (the charge at a Los Angeles facility) to differentiate healthy older adults from patients with moderate to severe dementia. (Incidentally the radiologist's professional fee alone exceeds what many internists charge for an entire afternoon of office practice!) The biased spectrum of disease in these studies renders them virtually useless for clinical decision making. I generally agree with and thank Dr. Murden for his comments. Space limitations prevented me from elaborating on several related issues. I emphasized screening in my article because screening is clearly within the role of the primary care physician. When there is uncertainty about the diagnosis after screening, some internists may choose to use standardized criteria to make a diagnosis of dementia; others may wish to refer to appropriate specialists. Albert L. Siu, MD, MSPH University of California, Los Angeles Los Angeles, CA 90024 References 1. Siu AL. Screening for dementia and investigating its causes. Ann Intern Med. 1991;115:122-32. 2. Prohovnik I, Mayeux R, Sackeim HA, Smith G, Stern Y, Alderson PO. Cerebral perfusion as a diagnostic marker of early Alzheimer's disease. Neurol. 1988;38:931-937. 3. Jagust WJ, Budinger TF, Reed BR. The diagnosis of dementia with single photon emission computed tomography. Arch Neurol. 1987; 44:258-262. 4. Hellman RS, Tikofsky RS, Collier BD, Hoffmann RG, Palmer DW, Glatt SL, et al. Alzheimer's disease: quantitative analysis of 1-123iodoamphetamine SPECT brain imaging. Radiology. 1989;172:183188. 5. Johnson KA, Holman BL, Rosen TJ, Nagel JS, English RJ, Growdon JH. Iofetamine I 123 single photon emission computed tomography is accurate in the diagnosis of Alzheimer's disease. Arch Intern Med. 1990;150:752-756.
Histamine-2 Receptor Blockers and Delirium To the Editors: I read with interest the article by Cantu and Korek (1) in which they reviewed the literature on central nervous system reactions to histamine-2 receptor (H2) blockers. As the chief of a psychiatry consultation service in a university teaching hospital, I have seen at least five cases of delirium (as defined by the DSM-II-R [2]) associated with H2 blocker use. In all five cases, however, the medication was being given intravenously. Thus, I was surprised that the authors did not address the route of administration in their discussion of risk factors. A MEDLINE literature search yielded no publications exploring this phenomenon. Further, regarding differential "deliriogenic" potential of the various H2 blockers, in anecdotal and uncontrolled observations of hospitalized elderly patients, our service has observed apparent induction of delirium by intravenous cimetidine that resolved after switching the patient to intravenous ranitidine. I am curious whether other Annals readers have had similar observations. As a clinical aside, when presented with a case of delirium apparently secondary to administration of an H2 blocker, our 658
standard recommendation is to switch to sucralfate if the patient is able to take oral medications. P. V. Nickell, MD West Virginia University School of Medicine Morgantown, WV 26506 References 1. Cantu TG, Korek JS. Central nervous system reactions to histamine-2 receptor blockers. Ann Intern Med. 1991;114:1027-34. 2. Spitzer RL, Williams JB, eds. Diagnostic and Statistical Manual of Mental Disorders. 3d ed. Washington, D.C.: The American Psychiatric Association; 1987:103.
In response: Dr. Nickell raises two interesting points about central nervous system reactions to H2 blockers. The first is the intravenous route of administration as a risk factor for these reactions. We did not have adequate data in our study to reach any conclusions about relative risk based on the route of administration. In fact, only 26 of the 72 reported cases thought to be possibly or probably drug-related involved intravenous route of administration, and no comparative studies of these reactions have focused on the route of administration. Anecdotal experiences suggesting that intravenously administered H2 blockers are more likely to be associated with central nervous system reactions may result from patient selection. Patients receiving H2 blockers intravenously are more likely to be severely ill or to have metabolic dysfunction than patients receiving H2 blockers orally. These severely ill patients can have a high incidence of mental status changes independent of H2 blocker use (2). Although the intravenous route would provide more rapid and, perhaps, higher serum H2 blocker concentrations, there is no evidence that high serum concentrations are associated with a greater risk for these reactions (1). Rather, these reactions appear to be idiosyncratic and not dose related. As to the different "deliriogenic" potential of the H2 blockers, Dr. Nickell suggests (as do a handful of case reports) that some patients may "react" to one H2 blocker but tolerate another. In Dr. Nickell's experience, cimetidine-induced delirium resolved after the patient was switched to intravenous ranitidine. The reverse, however, has been noted by other authors (3, 4). We clearly have more to learn about the pharmacology and risk factors for these reactions. Thomas G. Cantu, PharmD The Johns Hopkins Hospital Baltimore, MD 21205 Joan S. Korek, PharmD University of Maryland at Baltimore Baltimore, MD 21201 References 1. Cantu TG, Korek JS. Central nervous system reactions to histamine-2 receptor blockers. Ann Intern Med. 1991;114:1027-34. 2. Cerra FB, Schentag JJ, McMillen M, Karawande SV, Fitzgerald GC, Leising M. Mental status, the intensive care unit and cimetidine. Ann Surg. 1982;196:565-70. 3. Hughes JD, Reed WD, Serjeant CS. Mental confusion associated with ranitidine [Letter]. Med J Aust. 1983;2:12-3. 4. Silverstone PH. Ranitidine and confusion [Letter]. Lancet. 1984; 1: 1071.
Levothyroxine-Cholestyramine Interaction Reemphasized To the Editors: Several clinically significant interactions have been reported with bile acid sequestrants, including digoxin, warfarin, thiazide diuretics, corticosteroids, and levothyroxine (1). To date, only one case report of cholestyramine interference with levothyroxine with subsequent clinical hypothyroidism has been reported (2). We report the case of a patient who received concomitant cholestyramine and levothyroxine and developed elevations in the thyroid-stimulating hormone (TSH) level. On the same dose of levothyroxine after discontinuation of the cholestyramine, the patient's TSH level returned to normal. A 51-year-old man with known coronary artery disease and hypothyroidism was noted on 5 July 1990 to have an elevated
15 October 1991 • Annals of Internal Medicine • Volume 115 • Number 8
Downloaded From: https://annals.org/ by a Univ of Utah User on 09/26/2017
low-density lipoprotein (LDL) cholesterol level of 4.32 mmol/L and a low serum TSH level of 0.13 mU/L. Medications at this time included levothyroxine, 0.15 mg/d (only Synthroid [Flint Laboratories, Inc., Deerfield, Illinois] was used; no generic substitutes were allowed), and aspirin, 325 mg/d. On 9 July, the levothyroxine dose was lowered from 0.15 mg/d to 0.125 mg/d because of the low TSH level. The patient was started on cholestyramine, 4 g/d, with the evening meal to achieve better control of his hyperlipidemia. The levothyroxine was taken routinely at 0430 hours and the cholestyramine at 1800 hours. On 27 September 1990, his serum TSH level was 0.17 mU/L and his LDL cholesterol level was 4.45 mmol/L. His levothyroxine dose was again lowered to 0.10 mg/d, whereas cholestyramine was increased to 8 g/d in divided doses (4 g taken at 0430 hours and 1800 hours). On 4 January 1991, the serum TSH and LDL cholesterol levels measured 20.65 mU/L and 3.75 mmol/L, respectively. Three days later the cholestyramine was withdrawn because of concern that it was interfering with levothyroxine absorption. The drug was replaced with lovastatin, 20 mg/d, with the evening meal. With no change in levothyroxine dose, the TSH level was 2.45 mU/L and the LDL cholesterol level was 2.22 mmol/L on 8 February 1991. The use of bile acid sequestrants alone or in combination with other lipid-lowering agents has become first-line therapy for hyperlipidemia. Because bile acid sequestrants have the ability to bind certain compounds, including bile acids and several medications, drug interactions involving reduced absorption are common (1). In-vitro work has shown that approximately 50 mg of cholestyramine resin is capable of irreversibly binding 3000 /xg of thyroxine (2). Studies in animals have shown significant decreases in the absorption of radioactive thyroxine when combined with diets containing cholestyramine (3). Approximately 5% to 30% of intravenously administered radioactive thyroxine appears in the stool as free thyroxine or conjugates (3-5). In patients with hyperlipidemia who are not receiving concomitant levothyroxine, no significant change in free thyroxine or TSH levels occurred as a result of therapy with bile acid sequestrants (4, 5). Northcutt and colleagues (2) evaluated the absorption of thyroxine 1-131 in two hypothyroid patients and five healthy volunteers. In both patients and healthy volunteers, when cholestyramine was administered within 4 to 5 hours of the levothyroxine dose, significant reduction (greater than 50%) in levothyroxine absorption occurred (2). Our patient developed significant increases in the TSH level when cholestyramine was given concomitantly with levothyroxine. With no dose adjustment of levothyroxine and discontinuation of cholestyramine, TSH values returned to normal. In patients receiving concomitant levothyroxine and bile acid sequestrants, the doses should be given at least 6 hours apart, or the bile acid sequestrants should be changed to alternate lipid-lowering medications. Susan M. Harmon, MD Presbyterian Professional Building Oklahoma City, OK 73104-5023 Charles F. Seifert, PharmD University of Oklahoma Health Sciences Center Oklahoma City, OK 73190-5040 References 1. Tatro DS, Olin BR, Hebel SK, et al., eds. Drug Interaction Facts. St. Louis, Missouri: Facts and Comparisons Division, J. B. Lippincott Company; 1991. 2. Northcutt RC, Stiel JN, Hollifield JW, Stant EG. The influence of cholestyramine on thyroxine absorption. JAMA. 1969;208:1857-61. 3. Bergman F, Heedman PA, van der Linden W. Influence of cholestyramine on absorption and excretion of thyroxine in Syrian hamster. Acta Endocrinol. 1966;53:256-63. 4. Witztum JL, Jacobs LS, Schonfeld G. Thyroid hormone and thyrotropin levels in patients placed on colestipol hydrochloride. J Clin Endocrinol Metab. 1978;46:838-40. 5. Cashin-Hemphill L, Spencer CA, Nicoloff JT, Blankenhorn DH, Nessim SA, Chin HP, et al. Alterations in serum thyroid hormonal indices with colestipol-niacin therapy. Ann Intern Med. 1987; 107: 324-9.
Cigarette Smoking and Facial Wrinkling To the Editors: Kadunce and colleagues (1) report that cigarette smoking is an independent risk factor for the development of facial wrinkles, although the mechanism is unclear. As noted by the authors, only a subgroup of subjects (27.5%) with a smoking history of more than 50 pack-years were severely wrinkled, a relation similar to that between smoking and chronic obstructive pulmonary disease, in which only 26% of heavy smokers developed airway obstruction. An inverse relation has been reported among vitamin A consumption, retinol levels, and the subsequent development of lung cancer in men (2). Smoking might contribute to the increased risk of facial wrinkling by decreasing vitamin A levels. Vitamin A may protect against oxygen radicals, thereby limiting damage to DNA and connective tissue (3). Damage to connective tissue is an important histologic feature of both chronic obstructive pulmonary disease and wrinkled skin (4), and vitamin A deficiency could contribute to this damage. Measuring vitamin A levels of the subjects might provide important new insight into how smoking damages facial skin. Ian Joffe, MD Albert Einstein Medical Center Philadelphia, PA 19141 References 1. Kadunce DP, Burr R, Gress R, Kanner R, Lyon JL, Zone JJ. Cigarette smoking: risk factor for premature facial wrinkling. Ann Intern Med. 1991;114:840-4. 2. Willett WC, MacMahon B. Diet and cancer: an overview. N Engl J Med. 1984;310:633-8. 3. Krinsky NI. Carotenoid protection against oxidation. Pure Appl Chem. 1979;51:649-60. 4. Chen VL, Fleischmajer R, Schwartz E, Palaia M, Timpl R. Immunochemistry of elastotic material in sun-damaged skin. J Invest Dermatol. 1986;87:334-7.
To the Editors: We read with interest the article by Kadunce and colleagues (1). This study parallels the study done in 1971 by Daniell (2). Both studies provide data about wrinkling and smoking without providing adequate controls for sunlight exposure. It is known that sun exposure, aging, and genetic makeup are the primary causes of wrinkling. To adequately control for the effect of sunlight exposure and wrinkle production, a population where the effects of sun are negated must be included. If there are no black patients in the study, then there are no adequate controls for sun exposure. All of Kadunce's study patients were white, and the pigmentation the authors noted may have developed after sun exposure, which in itself offers little or no protection from the effects of ultraviolet light. In 1973 my colleagues and I published a similar study of cigarette smoking and wrinkling that included black smokers (3). We did a histologic evaluation of the sun-exposed areas and the non-sun-exposed areas to see if there was any effect of cigarette smoking alone on the elastic tissue within the dermis in the non-sun-exposed areas. We concluded that cigarette smoking plays little or no role in wrinkle production. I agree with Kottke's editorial (4) that every effort should be used to help people stop smoking. But how do you answer your patients when they say, "well, doctor, my black colleagues smoke and they don't have wrinkles"? Bernett L. Johnson, MD Hospital of the University of Pennsylvania Philadelphia, PA 19104 References 1. Kadunce DP, Burr R, Gress R, Kanner R, Lyon JL, Zone JJ. Cigarette smoking: risk factor for premature facial wrinkling. Ann Intern Med. 1991;114:840-4. 2. Daniell HW. Smoker's wrinkles: a study in the epidemiology of "crow's feet." Ann Intern Med. 1971;75:873-80. 3. Allen HB, Johnson BL, Diamond SM. Smokers' wrinkles? JAMA. 1973;225:1067-9. 4. Kottke TE. Smoking and wrinkles: does it matter? [Editorial]. Ann Intern Med. 1991;114:908-9.
15 October 1991 • Annals of Internal Medicine
Downloaded From: https://annals.org/ by a Univ of Utah User on 09/26/2017
• Volume 115 • Number 8
659
In response: We thank Dr. Johnson for his comments on our paper (1) and for calling our attention to his report (2). We apologize for failing to cite this article. Our study dealt only with white patients because of known racial differences in the structure of the skin (3). Because "sun exposure, aging, and genetic makeup are the prime causes of wrinkling," inclusion of black patients would introduce another variable rather than serving as a control for sun exposure. The appropriate control is to use a single racial group and to control for such factors as skin pigmentation and sun exposure (4). These controls were used in our study, which conformed to the rigorous criteria established by other researchers (4, 5). In Dr. Johnson's article (2), however, the authors chose an arbitrary measure of sun exposure without validation. Furthermore, we were able to establish that the effects of smoking and sun exposure were not only independent of one another but were additive. In an earlier study (5), we were unable to confirm Dr. Johnson's findings that the changes seen by histology are sensitive enough to be a valid indicator of wrinkling or total sun exposure. Dr. Johnson suggests that we controlled for pigmentation based on our observations of each subject and that our classification of patients, therefore, was biased by tanning of the skin. Pigmentation is commonly controlled for in epidemiologic studies (as it was in our study) with a scoring system that takes into account the patient's physiologic response to sunlight, that is, the development of erythema, ancestry, and hair and eye coloration (4). This system is used to control for varying degrees of protection that pigmentation provides against ultraviolet damage regardless of race. We also used standard statistical methods to control for lifetime sun exposure and lifetime cigarette consumption. Our conclusions that among whites there is an association, independent of sun exposure, between cigarette smoking and facial wrinkling are based on the analysis of our data using accepted statistical methods. To the statement, "well, doctor, my black colleagues smoke and they don't have wrinkles," we would say that "it is known that sun exposure, aging, and genetic makeup are the prime causes of wrinkling," but that smoking is an additional contributor in white people. Donald P. Kadunce, MD Joseph L. Lyon, MD John J. Zone, MD University of Utah Health Sciences Center Salt Lake City, UT 84132 References 1. Kadunce DP, Burr R, Gress R, Kanner R, Lyon JL, Zone JJ. Cigarette smoking: risk factor for premature facial wrinkling. Ann Intern Med. 1991;114:840-4. 2. Allen HB, Johnson BL, Diamond SM. Smokers' wrinkles? JAMA. 1973;225:1067-9. 3. Montagna W, Carlisle K. The architecture of black and white facial skin. J Amer Acad Dermatol. 1991;24:929-37. 4. Scotto J, Fears T, Fraumeni J, Jr. Incidence of non-melanoma skin
660
cancer in the United States. Rockville, Maryland: United States Public Health Service; 1981; DHHS publication no (NIH)82-2433. 5. Blake JS, Kadunce DP, Piepkorn MW, Zone JJ. The effects of smoking on the skin. Clin Res. 1988;36:119A.
What's in a Title? To the Editors: Short declarative sentences are being used more frequently for titles of scientific articles. This practice should be discouraged because it is misleading. One example is the otherwise well-written article by Krishna and Kapoor (1) titled "Potassium Depletion Exacerbates Essential Hypertension." Do the authors really mean to imply that the issue is now and forever settled? I suggest three alternatives, beginning with the one I prefer: 1. The Effect of Potassium Depletion on Blood Pressure 2. Does Potassium Depletion Elevate Blood Pressure? 3. A Low Potassium Diet for Ten Days Was Associated With an Elevation of Blood Pressure in Each of Ten Hypertensive Patients Except One Whose Blood Pressure Was Actually Lower on the Potassium Restricted Diet. Mark S. McGowan, MD University of Southern California Los Angeles, CA 90007 Reference 1. Krishna GG, Kapoor SC. Potassium depletion exacerbates essential hypertension. Ann Intern Med. 1991;115:77-83.
Correction A Library for Internists VII A psychiatric textbook was incorrectly described on page 823 of the 1 May 1991 issue as having looseleaf format (1). The correct information is as follows: The American Psychiatric Press Textbook of Psychiatry. Edited by J.A. Talbott, R.E. Hales, and S.C. Yudofsky. American Psychiatric Press, Washington, DC, 1988. 1352 pages. $105.00. A comprehensive reference. The 38 chapters from 55 contributors cover theoretical foundations, assessment, psychiatric disorders, psychiatric treatments, and special topics. The emphasis is on the clinical relevance of psychiatry. Casebound book with study guide available separately. K. Loper American Psychiatric Press Washington, DC 20005 Reference 1. Ludmerer KM. A library for internists. Recommendations from the American College of Physicians. Ann Intern Med. 1991;114:811-32.
15 October 1991 • Annals of Internal Medicine • Volume 115 • Number 8
Downloaded From: https://annals.org/ by a Univ of Utah User on 09/26/2017
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 with the letter 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 only if their letter is accepted. Unpublished letters cannot be returned. Cost Effectiveness of Treating Asymptomatic HIV Infection To the Editors: We read with interest the article by Schulman and colleagues (1). Our major concern about their method is the use of data on time to progression to the acquired immunodeficiency syndrome (AIDS) to make statements about survival. This is an extraordinarily important variable for which almost no data exist. They draw from the randomized AIDS Clinical Trial Group Protocol 019, which had only eight deaths, four in the placebo-treated group and four in the zidovudine-treated group (2). Zidovudine prolongs survival in human immunodeficiency virus (HlV)-infected patients once they progress to AIDS or to AIDS-related complex (3-5). It is not clear, however, if zidovudine will provide the same benefit for patients who are already receiving the drug when they progress. The small amount of data available from Protocol 019 suggests that the death rate may be higher for patients already taking zidovudine when they progress than for untreated patients who progress; these two groups of patients had the same number of deaths, but the latter group was followed longer (2). It is therefore possible that progression to AIDS or AIDS-related complex is delayed by early (as compared to later) treatment with zidovudine but that survival is either not affected or adversely affected. If this is the case, many of the conclusions of Dr. Schulman and colleagues would need to be modified. We believe that this issue is extremely important because of the already high cost of treatment for HIV infection. In addition, data on quality of life, cumulative toxicity of the drug, and emergence of viral resistance should be taken into account when analyzing the cost effectiveness of zidovudine therapy. John D. Hamilton, MD Michael S. Simberkoff, MD Pamela M. Hartigan, PhD The Veterans Affairs Cooperative Study Group on Treatment of AIDS Durham Veterans Affairs Medical Center Durham, NC 27705 References 1. Schulman KA, Lynn LA, Glick HA, Eisenberg JM. Cost effectiveness of low-dose zidovudine therapy for asymptomatic patients with human immunodeficiency virus (HIV) infection. Ann Intern Med. 1991; 114:798-802. 2. Volberding PA, Lagakos SW, Koch MA, et al. Zidovudine in asymptomatic human immunodeficiency virus infection: a controlled trial in
persons with fewer than 500 CD4-positive cells per cubic millimeter. N Engl J Med. 1990;322:941-9. 3. Fischl MA, Richman DD, Grieco MH, et al. The efficacy of azidothymidine (AZT) in the treatment of patients with AIDS and AIDSrelated complex. N Engl J Med. 1987;317:185-91. 4. Lemp GF, Payne SF, Neal D, et al. Survival trends for patients with AIDS. JAMA. 1990;263:402-6. 5. Moore RD, Hidalgo J, Sugland BW, Chaisson RE. Zidovudine and the natural history of the acquired immunodeficiency syndrome. N Engl J Med. 1991;324:1412-6. In response: The comments of Hamilton and colleagues about our paper (1) reflect the lack of available information on the ultimate survival benefit for patients with asymptomatic HIV infection who are treated with zidovudine. We examined the effects of changes in life expectancy after the development of AIDS in patients who received zidovudine while they were asymptomatic. Our model indicates that for the one-time effect, treated asymptomatic patients would need to lose 6 of their 16 months of survival after the development of AIDS or advanced AIDS-related complex for there to be no difference in discounted life expectancy between this group and patients receiving placebo. For the continuous effect model, even if asymptomatic, treated patients died immediately after developing AIDS or advanced AIDS-related complex, they still would live 2.76 discounted years longer than patients receiving a placebo. One reason why we did not address this issue in our original paper is that we do not share the authors' concerns about the mortality data presented in the report of the AIDS Clinical Trial Group Protocol 019 (2). Although there were equal numbers of deaths in placebo-treated and zidovudine-treated patients, only 428 patients were treated with placebo (approximately 502 person-years of follow-up) as compared with 910 patients treated with zidovudine (approximately 927 personyears of follow-up). Thus, there was an apparent—but not statistically significant (P > 0.2)—50% reduction in mortality rates in zidovudine-treated patients once progression to AIDS had occurred. The AIDS Clinical Trial Group study did not report this result, possibly because total mortality was not a primary study outcome (more than 4500 patients would have been needed in each arm of the trial to detect this mortality difference with an alpha of 0.05 and a beta of 0.8). We agree that in economic analyses of technologies early in their diffusion, there are large amounts of uncertain data that may affect their eventual cost-effectiveness ratios. Extensive sensitivity analyses are required to explore the effects of this uncertainty and to examine the robustness of the cost-effectiveness models. Given the wide range of cost-effectiveness ratios in our primary analyses and the robustness of these ratios to a large number of sensitivity analyses, we stand by our original conclusions. Kevin A. Schulman, MD Lorna A. Lynn, MD Henry A. Glick, MA Hospital of the University of Pennsylvania Philadelphia, PA 19104 References 1. Schulman KA, Lynn LA, Glick HA, Eisenberg JM. Cost effectiveness of low-dose zidovudine therapy for asymptomatic patients with human immunodeficiency virus (HIV) infection. Ann Intern Med. 1991; 114:798-802. 2. Volberding PA, Lagakos SW, Koch MA, et al. Zidovudine in asymptomatic human immunodeficiency virus infection: a controlled trial in persons with fewer than 500 CD4-positive cells per cubic millimeter:
15 October 1991 • Annals of Internal Medicine
Downloaded From: https://annals.org/ by a Univ of Utah User on 09/26/2017
• Volume 115 • Number 8
655
the AIDS Clinical Trials Group of the National Institute of Allergy and Infectious Diseases. N Engl J Med. 1990;322:941-9.
Fetal Hemoglobin Alters Hemoglobin A l c Measurements To the Editors: The letter of Holt and colleagues (1) emphasizes that diabetic subjects with hemoglobin S or C traits will have a falsely low hemoglobin A, c value if measured by certain methods. We also have observed this finding but would like to highlight another common condition which alters hemoglobin A l c values. Persons with the trait or who are homozygous for the hereditary persistence of fetal hemoglobin will have a falsely elevated hemoglobin A l c value (2) because hemoglobin F co-elutes with the hemoglobin A l c fraction. Some patients who were referred to our institution for presumed diabetes based on an elevated hemoglobin A l c were not judged to be glucose intolerant using other criteria (such as the presence of glycosuria and fasting or postprandial hyperglycemia). These patients were subsequently found to have elevated levels of hemoglobin F. Because the heterozygous and homozygous states of the hereditary persistence of fetal hemoglobin are generally asymptomatic (3), these persons may be first identified with discrepant findings between hemoglobin A l c and measures of glucose intolerance as in our cases. The prevalence of the heterozygous state for this trait is not precisely known because this condition is asymptomatic and several variants of hereditary persistence of fetal hemoglobin exist (3). Awareness of this condition is germane not only to the diagnosis of diabetes but also to its management because falsely elevated values may result in unrealistic patient goals and unnecessary hypoglycemia if the hemoglobin A l c value is used as a "gold standard" for glucose control. We have used the fructosamine assay rather than a more expensive method for the measurement of hemoglobin A l c for diabetic subjects with elevated levels of hemoglobin F, S, or C (4) because the fructosamine assay is independent of hemoglobin. Richard W. Bergstrom, MD James R. Kelley, MD W. Kenneth Ward, MD Portland Diabetes and Endocrinology Center and Good Samaritan Hospital Portland, OR 97210 References 1. Holt GS, Wofford JL, Velez R. Hemoglobinopathies affect hemoglobin A l c measurement [Letter]. Ann Intern Med. 1991;115:68-9. 2. Bunn HF. Evaluation of glycosylated hemoglobin in diabetic patients. Diabetes. 1981;30:613-7. 3. Weatherall DJ, Clegg JB. Hereditary persistence of fetal haemoglobin. Br J Haemat. 1975;29:191-8. 4. Baker JR, O'Connor JP, Metcalf PA, Lawson MR, Johnson RN. Clinical usefulness of estimation of serum fructosamine concentration as a screening test for diabetes mellitus. Br Med J. 1983;287: 863-7.
Barriers to Treating Obesity To the Editors: Dr. Bray's editorial (1) elucidates some of the important barriers to treating obesity. As stated, "the public perception that obesity is a disease resulting from a lack of will power" is a major barrier to the treatment of a disease with fundamental genetic, molecular determinants. "Limited research funding" is surely another barrier. However, I am concerned about Dr. Bray's wish to lower barriers to the use of appetite-suppressant drugs. I do not find "clear and convincing evidence that modern appetite-suppressant drugs are effective and safe." Fenfluramine, one of the drugs cited, has been shown to have potentially serious side effects. At least three studies have shown pulmonary hypertension occurring after treatment with fenfluramine (2-4). In one instance, the pulmonary hypertension did not resolve with cessation of treatment and was lethal (2). There is evidence that other anorectic agents, for example, 656
aminorex and chlorphentermine, may produce pulmonary hypertension more frequently than fenfluramine (4). If any currently available appetite suppressants were reliably associated with major persistent weight loss, then one might be willing to take the risks and use them, taking care not to overestimate their efficacy. A recent study of the use of dexfenfluramine in 822 obese patients in a multicenter international trial showed less than a 3-kg difference in weight loss when the drug-treated group was compared with the placebotreated group (5). Weight loss curves in this study show a slow upward creep after 8 months in both placebo-treated and drugtreated groups, suggesting that in the long run one might expect considerable weight regain despite continued treatment (5). I disagree with the editorial (1), which brands recent hearings on the "diet pill industry" as "legislative grand standing" which have "produced little else but publicity for the legislators." Congressman Wyden and others have become deeply interested in working with the National Institutes of Health (NIH) to widen the research base for obesity and to inform the public about the present inadequate status of obesity treatment. The greatest barrier to the treatment of obesity remains our incomplete understanding of the complex genetic, developmental, and psychosocial influences that lead to obesity. Jules Hirsch, MD The Rockefeller University New York, NY 10021 References 1. Bray GA. Barriers to the treatment of obesity [Editorial]. Ann Intern Med. 1991;115:152-3. 2. McMurray J, Bloom fie Id P, Miller HC. Irreversible pulmonary hypertension after treatment with fenfluramine. Br Med J. 1986;292: 239-40. 3. Douglas JG, Munro JF, Kitchin AH, Muir AL, Proudfoot AT. Pulmonary hypertension and fenfluramine. Br Med J. 1981;283:881-3. 4. Pouwels HM, Smeets JL, Cheriex EC, Wouters EF. Pulmonary hypertension and fenfluramine. Eur Respir J. 1990;3:606-7. 5. Guy-Grand B, Crepaldi G, Lefebvre P, Apfelbaum M, Gries A, Turner P. International trial of long-term dexfenfluramine in obesity. Lancet. 1989;1:1142-5.
To the Editors: Although we agree with Dr. Bray's (1) observations about attitudes toward obesity, we disagree with his assertion about the "convincing evidence" of appetite-suppressant drugs' safety and efficacy. As he noted, relatively few clinical trials have assessed these agents. Phenylpropanolamine, an over-the-counter medication, is as effective as other appetite-suppressant drugs (2). Despite its widespread availability, consumers still seek other therapies. Even when overthe-counter drugs, such as nonsteroidal anti-inflammatory drugs, are effective, companies actively investigate new "me-too" agents. Although long-term appetite suppressants may deserve further study, the effect of short-term use is reversed after therapy is discontinued. Craighead and colleagues (3) found that during treatment, fenfluramine plus behavior therapy resulted in greater weight loss than behavioral management alone. However, in the year after use, patients on the drug regained more weight, suggesting that its use compromised the development of permanent abilities to alter behavior. We concur that obesity cannot be "cured" and would welcome additional studies on its management. Some might argue that no obesity treatment has provided long-term efficacy (4). For the present, alterations in nutrition and physical activity, along with behavior modification, can be effective, especially when individuals develop personal strategies for weight loss (5). Although lifestyle alterations are difficult, we do not believe that appetite suppressants should replace patient education and use of nonpharmacologic therapies. Diane L. Elliot, MD Linn Goldberg, MD Oregon Health Sciences University Portland, OR 97201
15 October 1991 • Annals of Internal Medicine • Volume 115 • Number 8
Downloaded From: https://annals.org/ by a Univ of Utah User on 09/26/2017
References 1. Bray GA. Barriers to the treatment of obesity. Ann Intern Med. 1991;115:152-3. 2. Weintraub M, Bray GA. Drug treatment of obesity. Med Clin N Am. 1989;73:237. 3. Craighead LW, Stunkard AJ, O'Brien RM. Behavior therapy and pharmacotherapy for obesity. Arch Gen Psychiat. 1981;38:763-8. 4. Fitzgerald FT. The problem of obesity. Ann Rev Med. 1981;32:22131. 5. Kayman S, Bruvold W, Stern JS. Maintenance and relapse after weight loss in women: behavioral aspects. Am J Clin Nutr. 1990;52: 800-7.
In response: I am grateful for the letters from Drs. Eliot and Goldberg and from Dr. Hirsch because they provide an opportunity to re-emphasize my thesis (1) that barriers exist to the treatment of obesity. My central thesis is that both the public and health professionals view obesity as a character disorder resulting from a lack of will power rather than as a disease for which we currently have no cure. When drugs are discontinued, we expect the patient to be "cured." Regain in weight is used to condemn drugs rather than being viewed as the expected result of termination of treatment. A recent paper by Weintraub and colleagues (2) and an earlier review by Stunkard (3) support the contention that long-term use of appetite suppressants may temporarily "reset" the regulatory system for body weight. "Safety" and "effectiveness" are relative terms. For any overweight person, the initial treatment decision should be based on evaluation of the "risk" of obesity relative to the risks of any proposed treatment (4). Dr. Hirsch properly points out three reports of pulmonary hypertension and one death with fenfluramine among the many thousands who have been treated. If one compares this with the mortality of 0.5% to 5% associated with surgical treatment of obesity, one could argue that treatment with appetite suppressants might offer a "safer" alternative than surgery. Both experimental and clinical evidence show that appetite suppressants can produce long-term weight reduction (1, 2). Currently, the Food and Drug Administration requires an investigational new drug license to use appetite suppressants for long-term treatment of obese patients. Few physicians are likely to surmount this barrier. I am encouraged by the comments of Dr. Hirsch about the activity of Congressman Wyden and his committee. If the efforts of Dr. Hirsch and others are effective in increasing the support for obesity research by the NIH, we will have made substantial progress in breaking down the barriers to future advances in this field. George A. Bray, MD Pennington Biomedical Research Center Baton Rouge, LA 70808 References 1. Bray GA. Barriers to the treatment of obesity [Editorial]. Ann Intern Med. 1991;115:152-3. 2. Weintraub M, Fundaresan PR, Schuster B, Ginsberg G, Madan M, Stein EC, et al. National Heart, Lung and Blood Institute long-term weight control study: I-VII. Clin Pharmacol Ther. 1991; in press. 3. Stunkard AJ. Anorectic agents lower body weight set point. Life Sciences. 1982;39:2043-55. 4. Bray GA. Classification and evaluation of the obesities. Med Clin North Am. 1989;73:161-84.
Dementia Screening To the Editors: The review of dementia screening and evaluation by Siu (1) was well done. To be a useful clinical guide for the practicing physician, however, two points need to be emphasized. First, the use of a criterion-based definition of dementia is essential. The major source of error in diagnosing dementia is in properly assessing its presence or absence. O'Connor and colleagues (2) demonstrated in a large community survey with extensive clinical evaluation that 42% (87 of 208) of demented patients were not recognized as demented by their personal
physician. They also found that 22% (51 of 236) of nondemented persons were mistakenly diagnosed as demented. The misdiagnosis rate was 31% (138 of 444), with almost 40% of the errors being overdiagnoses, which are not generally the result of inadequate screening. Furthermore, Somerfield and colleagues (3) surveyed 127 physicians who had referred patients to a dementia clinic and found that fewer than 30% used either of the two sets of specific criteria to diagnose dementia. Using these criteria along with any of the screening tests would probably improve the diagnosis of dementia more than cerebral imaging, laboratory testing, or choice of screening instruments. Second, the educational bias of mental status tests was not stressed. My colleagues and I recently confirmed previous findings that scores on the Mini-Mental State Examination of Folstein and coworkers (4) are significantly affected by the educational level of those tested (5). Our findings agreed with Siu's conclusion that two cut-off points are needed for MiniMental State Examination scores. We found, however, the cut-off scores to be less than or equal to 17 for high likelihood, 18 to 23 for intermediate likelihood, and 24 to 30 for low likelihood of dementia in persons with less than a ninth-grade education. Additionally, a study on clock drawing also shows the effects of education on performance. The clinician must be aware of these effects when interpreting test results in poorly educated persons. In summary, the helpful suggestions by Siu can greatly improve the diagnosis of dementia, when combined with the use of specific criteria to define dementia and with altered cut-off points in test scores for poorly educated persons. Robert A. Murden, MD The Ohio State University Columbus, OH 43210 References 1. Siu AL. Screening for dementia and investigating its causes. Ann Intern Med. 1991;115:122-32. 2. O'Connor DW, Politt PA, Hyde JB, Brooks CPB, Reiss BB, Roth M. Do general practitioners miss dementia in elderly patients? Br Med J. 1988;297:1107-10. 3. Somerfield MR, Weisman CS, Ury W, Chase GA, Folstein MF. Physician practices in the diagnosis of dementing disorders. J Am Geriatr Soc. 1991;39:172-5. 4. Folstein MF, Folstein SE, McHugh PR. "Mini-Mental S t a t e ' : a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12:189-98. 5. Murden RA, McRae TD, Kaner S, Bucknam ME. Mini-Mental State Exam scores vary with education in blacks and whites. J Am Geriatr Soc. 1991;39:149-55.
To the Editors: Dr. Siu's recent review (1) fails to mention the growing role of emission tomography in screening for dementia. He is correct in pointing out the limits of conventional anatomic imaging modalities (computed tomography and magnetic resonance imaging) in the evaluation of these patients. Unlike stroke, dementia, particularly that seen in Alzheimer disease, does not result from a structural, gross anatomic defect. Dementia is probably a consequence of metabolic changes in the brain and is thus more likely to be demonstrated by functional imaging modalities. Many reports (2-5) argue strongly that positron-emission tomography and single photon emission computed tomography are of significant help in differentiating demented from nondemented patients. A report being prepared by the Consensus Panel on Dementia sponsored by the Brain Imaging Council of the Society of Nuclear Medicine supports this position. Ronald S. Tikofsky, PhD Medical College of Wisconsin Milwaukee, WI 53226 References 1. Siu AL. Screening for dementia and investigating its causes. Ann Intern Med. 1991;115:122-32. 2. Prohovnik I, Mayeux R, Sackeim HA, Smith G, Stern Y, Alderson PO. Cerebral perfusion as a diagnostic marker of early Alzheimer's disease. Neurol. 1988;38:931-7. 3. Jagust WJ, Budinger TF, Reed BR. The diagnosis of dementia with
15 October 1991 • Annals of Internal Medicine
Downloaded From: https://annals.org/ by a Univ of Utah User on 09/26/2017
• Volume 115 • Number 8
657
single photon emission computed tomography. Arch Neurol. 1987; 44:258-62. 4. Hellman RS, Tikofsky RS, Collier BD, et al. Alzheimer's disease: quantitative analysis of I-123-iodoamphetamine SPECT brain imaging. Radiology. 1989;172:183-8. 5. Johnson KA, Holman BL, Rosen TJ, Nagel JS, English RJ, Growdon JH. Iofetamine 1-123 single photon emission computed tomography is accurate in the diagnosis of Alzheimer's disease. Arch Intern Med. 1990;150:752-6.
In response: My article (1) addressed the methods by which primary care physicians can screen for dementia and investigate its various causes. I did not mention the role of emission tomography because the literature does not support its application in clinical practice at this time. All four articles (2-5) cited by Dr. Tikofsky investigate the use of emission tomography in distinguishing patients with Alzheimer disease (many with moderate to severe disease) from healthy subjects who were not suspected of having dementia. Yet it is almost inconceivable that a physician would have to use a $2340 test (the charge at a Los Angeles facility) to differentiate healthy older adults from patients with moderate to severe dementia. (Incidentally the radiologist's professional fee alone exceeds what many internists charge for an entire afternoon of office practice!) The biased spectrum of disease in these studies renders them virtually useless for clinical decision making. I generally agree with and thank Dr. Murden for his comments. Space limitations prevented me from elaborating on several related issues. I emphasized screening in my article because screening is clearly within the role of the primary care physician. When there is uncertainty about the diagnosis after screening, some internists may choose to use standardized criteria to make a diagnosis of dementia; others may wish to refer to appropriate specialists. Albert L. Siu, MD, MSPH University of California, Los Angeles Los Angeles, CA 90024 References 1. Siu AL. Screening for dementia and investigating its causes. Ann Intern Med. 1991;115:122-32. 2. Prohovnik I, Mayeux R, Sackeim HA, Smith G, Stern Y, Alderson PO. Cerebral perfusion as a diagnostic marker of early Alzheimer's disease. Neurol. 1988;38:931-937. 3. Jagust WJ, Budinger TF, Reed BR. The diagnosis of dementia with single photon emission computed tomography. Arch Neurol. 1987; 44:258-262. 4. Hellman RS, Tikofsky RS, Collier BD, Hoffmann RG, Palmer DW, Glatt SL, et al. Alzheimer's disease: quantitative analysis of 1-123iodoamphetamine SPECT brain imaging. Radiology. 1989;172:183188. 5. Johnson KA, Holman BL, Rosen TJ, Nagel JS, English RJ, Growdon JH. Iofetamine I 123 single photon emission computed tomography is accurate in the diagnosis of Alzheimer's disease. Arch Intern Med. 1990;150:752-756.
Histamine-2 Receptor Blockers and Delirium To the Editors: I read with interest the article by Cantu and Korek (1) in which they reviewed the literature on central nervous system reactions to histamine-2 receptor (H2) blockers. As the chief of a psychiatry consultation service in a university teaching hospital, I have seen at least five cases of delirium (as defined by the DSM-II-R [2]) associated with H2 blocker use. In all five cases, however, the medication was being given intravenously. Thus, I was surprised that the authors did not address the route of administration in their discussion of risk factors. A MEDLINE literature search yielded no publications exploring this phenomenon. Further, regarding differential "deliriogenic" potential of the various H2 blockers, in anecdotal and uncontrolled observations of hospitalized elderly patients, our service has observed apparent induction of delirium by intravenous cimetidine that resolved after switching the patient to intravenous ranitidine. I am curious whether other Annals readers have had similar observations. As a clinical aside, when presented with a case of delirium apparently secondary to administration of an H2 blocker, our 658
standard recommendation is to switch to sucralfate if the patient is able to take oral medications. P. V. Nickell, MD West Virginia University School of Medicine Morgantown, WV 26506 References 1. Cantu TG, Korek JS. Central nervous system reactions to histamine-2 receptor blockers. Ann Intern Med. 1991;114:1027-34. 2. Spitzer RL, Williams JB, eds. Diagnostic and Statistical Manual of Mental Disorders. 3d ed. Washington, D.C.: The American Psychiatric Association; 1987:103.
In response: Dr. Nickell raises two interesting points about central nervous system reactions to H2 blockers. The first is the intravenous route of administration as a risk factor for these reactions. We did not have adequate data in our study to reach any conclusions about relative risk based on the route of administration. In fact, only 26 of the 72 reported cases thought to be possibly or probably drug-related involved intravenous route of administration, and no comparative studies of these reactions have focused on the route of administration. Anecdotal experiences suggesting that intravenously administered H2 blockers are more likely to be associated with central nervous system reactions may result from patient selection. Patients receiving H2 blockers intravenously are more likely to be severely ill or to have metabolic dysfunction than patients receiving H2 blockers orally. These severely ill patients can have a high incidence of mental status changes independent of H2 blocker use (2). Although the intravenous route would provide more rapid and, perhaps, higher serum H2 blocker concentrations, there is no evidence that high serum concentrations are associated with a greater risk for these reactions (1). Rather, these reactions appear to be idiosyncratic and not dose related. As to the different "deliriogenic" potential of the H2 blockers, Dr. Nickell suggests (as do a handful of case reports) that some patients may "react" to one H2 blocker but tolerate another. In Dr. Nickell's experience, cimetidine-induced delirium resolved after the patient was switched to intravenous ranitidine. The reverse, however, has been noted by other authors (3, 4). We clearly have more to learn about the pharmacology and risk factors for these reactions. Thomas G. Cantu, PharmD The Johns Hopkins Hospital Baltimore, MD 21205 Joan S. Korek, PharmD University of Maryland at Baltimore Baltimore, MD 21201 References 1. Cantu TG, Korek JS. Central nervous system reactions to histamine-2 receptor blockers. Ann Intern Med. 1991;114:1027-34. 2. Cerra FB, Schentag JJ, McMillen M, Karawande SV, Fitzgerald GC, Leising M. Mental status, the intensive care unit and cimetidine. Ann Surg. 1982;196:565-70. 3. Hughes JD, Reed WD, Serjeant CS. Mental confusion associated with ranitidine [Letter]. Med J Aust. 1983;2:12-3. 4. Silverstone PH. Ranitidine and confusion [Letter]. Lancet. 1984; 1: 1071.
Levothyroxine-Cholestyramine Interaction Reemphasized To the Editors: Several clinically significant interactions have been reported with bile acid sequestrants, including digoxin, warfarin, thiazide diuretics, corticosteroids, and levothyroxine (1). To date, only one case report of cholestyramine interference with levothyroxine with subsequent clinical hypothyroidism has been reported (2). We report the case of a patient who received concomitant cholestyramine and levothyroxine and developed elevations in the thyroid-stimulating hormone (TSH) level. On the same dose of levothyroxine after discontinuation of the cholestyramine, the patient's TSH level returned to normal. A 51-year-old man with known coronary artery disease and hypothyroidism was noted on 5 July 1990 to have an elevated
15 October 1991 • Annals of Internal Medicine • Volume 115 • Number 8
Downloaded From: https://annals.org/ by a Univ of Utah User on 09/26/2017
low-density lipoprotein (LDL) cholesterol level of 4.32 mmol/L and a low serum TSH level of 0.13 mU/L. Medications at this time included levothyroxine, 0.15 mg/d (only Synthroid [Flint Laboratories, Inc., Deerfield, Illinois] was used; no generic substitutes were allowed), and aspirin, 325 mg/d. On 9 July, the levothyroxine dose was lowered from 0.15 mg/d to 0.125 mg/d because of the low TSH level. The patient was started on cholestyramine, 4 g/d, with the evening meal to achieve better control of his hyperlipidemia. The levothyroxine was taken routinely at 0430 hours and the cholestyramine at 1800 hours. On 27 September 1990, his serum TSH level was 0.17 mU/L and his LDL cholesterol level was 4.45 mmol/L. His levothyroxine dose was again lowered to 0.10 mg/d, whereas cholestyramine was increased to 8 g/d in divided doses (4 g taken at 0430 hours and 1800 hours). On 4 January 1991, the serum TSH and LDL cholesterol levels measured 20.65 mU/L and 3.75 mmol/L, respectively. Three days later the cholestyramine was withdrawn because of concern that it was interfering with levothyroxine absorption. The drug was replaced with lovastatin, 20 mg/d, with the evening meal. With no change in levothyroxine dose, the TSH level was 2.45 mU/L and the LDL cholesterol level was 2.22 mmol/L on 8 February 1991. The use of bile acid sequestrants alone or in combination with other lipid-lowering agents has become first-line therapy for hyperlipidemia. Because bile acid sequestrants have the ability to bind certain compounds, including bile acids and several medications, drug interactions involving reduced absorption are common (1). In-vitro work has shown that approximately 50 mg of cholestyramine resin is capable of irreversibly binding 3000 /xg of thyroxine (2). Studies in animals have shown significant decreases in the absorption of radioactive thyroxine when combined with diets containing cholestyramine (3). Approximately 5% to 30% of intravenously administered radioactive thyroxine appears in the stool as free thyroxine or conjugates (3-5). In patients with hyperlipidemia who are not receiving concomitant levothyroxine, no significant change in free thyroxine or TSH levels occurred as a result of therapy with bile acid sequestrants (4, 5). Northcutt and colleagues (2) evaluated the absorption of thyroxine 1-131 in two hypothyroid patients and five healthy volunteers. In both patients and healthy volunteers, when cholestyramine was administered within 4 to 5 hours of the levothyroxine dose, significant reduction (greater than 50%) in levothyroxine absorption occurred (2). Our patient developed significant increases in the TSH level when cholestyramine was given concomitantly with levothyroxine. With no dose adjustment of levothyroxine and discontinuation of cholestyramine, TSH values returned to normal. In patients receiving concomitant levothyroxine and bile acid sequestrants, the doses should be given at least 6 hours apart, or the bile acid sequestrants should be changed to alternate lipid-lowering medications. Susan M. Harmon, MD Presbyterian Professional Building Oklahoma City, OK 73104-5023 Charles F. Seifert, PharmD University of Oklahoma Health Sciences Center Oklahoma City, OK 73190-5040 References 1. Tatro DS, Olin BR, Hebel SK, et al., eds. Drug Interaction Facts. St. Louis, Missouri: Facts and Comparisons Division, J. B. Lippincott Company; 1991. 2. Northcutt RC, Stiel JN, Hollifield JW, Stant EG. The influence of cholestyramine on thyroxine absorption. JAMA. 1969;208:1857-61. 3. Bergman F, Heedman PA, van der Linden W. Influence of cholestyramine on absorption and excretion of thyroxine in Syrian hamster. Acta Endocrinol. 1966;53:256-63. 4. Witztum JL, Jacobs LS, Schonfeld G. Thyroid hormone and thyrotropin levels in patients placed on colestipol hydrochloride. J Clin Endocrinol Metab. 1978;46:838-40. 5. Cashin-Hemphill L, Spencer CA, Nicoloff JT, Blankenhorn DH, Nessim SA, Chin HP, et al. Alterations in serum thyroid hormonal indices with colestipol-niacin therapy. Ann Intern Med. 1987; 107: 324-9.
Cigarette Smoking and Facial Wrinkling To the Editors: Kadunce and colleagues (1) report that cigarette smoking is an independent risk factor for the development of facial wrinkles, although the mechanism is unclear. As noted by the authors, only a subgroup of subjects (27.5%) with a smoking history of more than 50 pack-years were severely wrinkled, a relation similar to that between smoking and chronic obstructive pulmonary disease, in which only 26% of heavy smokers developed airway obstruction. An inverse relation has been reported among vitamin A consumption, retinol levels, and the subsequent development of lung cancer in men (2). Smoking might contribute to the increased risk of facial wrinkling by decreasing vitamin A levels. Vitamin A may protect against oxygen radicals, thereby limiting damage to DNA and connective tissue (3). Damage to connective tissue is an important histologic feature of both chronic obstructive pulmonary disease and wrinkled skin (4), and vitamin A deficiency could contribute to this damage. Measuring vitamin A levels of the subjects might provide important new insight into how smoking damages facial skin. Ian Joffe, MD Albert Einstein Medical Center Philadelphia, PA 19141 References 1. Kadunce DP, Burr R, Gress R, Kanner R, Lyon JL, Zone JJ. Cigarette smoking: risk factor for premature facial wrinkling. Ann Intern Med. 1991;114:840-4. 2. Willett WC, MacMahon B. Diet and cancer: an overview. N Engl J Med. 1984;310:633-8. 3. Krinsky NI. Carotenoid protection against oxidation. Pure Appl Chem. 1979;51:649-60. 4. Chen VL, Fleischmajer R, Schwartz E, Palaia M, Timpl R. Immunochemistry of elastotic material in sun-damaged skin. J Invest Dermatol. 1986;87:334-7.
To the Editors: We read with interest the article by Kadunce and colleagues (1). This study parallels the study done in 1971 by Daniell (2). Both studies provide data about wrinkling and smoking without providing adequate controls for sunlight exposure. It is known that sun exposure, aging, and genetic makeup are the primary causes of wrinkling. To adequately control for the effect of sunlight exposure and wrinkle production, a population where the effects of sun are negated must be included. If there are no black patients in the study, then there are no adequate controls for sun exposure. All of Kadunce's study patients were white, and the pigmentation the authors noted may have developed after sun exposure, which in itself offers little or no protection from the effects of ultraviolet light. In 1973 my colleagues and I published a similar study of cigarette smoking and wrinkling that included black smokers (3). We did a histologic evaluation of the sun-exposed areas and the non-sun-exposed areas to see if there was any effect of cigarette smoking alone on the elastic tissue within the dermis in the non-sun-exposed areas. We concluded that cigarette smoking plays little or no role in wrinkle production. I agree with Kottke's editorial (4) that every effort should be used to help people stop smoking. But how do you answer your patients when they say, "well, doctor, my black colleagues smoke and they don't have wrinkles"? Bernett L. Johnson, MD Hospital of the University of Pennsylvania Philadelphia, PA 19104 References 1. Kadunce DP, Burr R, Gress R, Kanner R, Lyon JL, Zone JJ. Cigarette smoking: risk factor for premature facial wrinkling. Ann Intern Med. 1991;114:840-4. 2. Daniell HW. Smoker's wrinkles: a study in the epidemiology of "crow's feet." Ann Intern Med. 1971;75:873-80. 3. Allen HB, Johnson BL, Diamond SM. Smokers' wrinkles? JAMA. 1973;225:1067-9. 4. Kottke TE. Smoking and wrinkles: does it matter? [Editorial]. Ann Intern Med. 1991;114:908-9.
15 October 1991 • Annals of Internal Medicine
Downloaded From: https://annals.org/ by a Univ of Utah User on 09/26/2017
• Volume 115 • Number 8
659
In response: We thank Dr. Johnson for his comments on our paper (1) and for calling our attention to his report (2). We apologize for failing to cite this article. Our study dealt only with white patients because of known racial differences in the structure of the skin (3). Because "sun exposure, aging, and genetic makeup are the prime causes of wrinkling," inclusion of black patients would introduce another variable rather than serving as a control for sun exposure. The appropriate control is to use a single racial group and to control for such factors as skin pigmentation and sun exposure (4). These controls were used in our study, which conformed to the rigorous criteria established by other researchers (4, 5). In Dr. Johnson's article (2), however, the authors chose an arbitrary measure of sun exposure without validation. Furthermore, we were able to establish that the effects of smoking and sun exposure were not only independent of one another but were additive. In an earlier study (5), we were unable to confirm Dr. Johnson's findings that the changes seen by histology are sensitive enough to be a valid indicator of wrinkling or total sun exposure. Dr. Johnson suggests that we controlled for pigmentation based on our observations of each subject and that our classification of patients, therefore, was biased by tanning of the skin. Pigmentation is commonly controlled for in epidemiologic studies (as it was in our study) with a scoring system that takes into account the patient's physiologic response to sunlight, that is, the development of erythema, ancestry, and hair and eye coloration (4). This system is used to control for varying degrees of protection that pigmentation provides against ultraviolet damage regardless of race. We also used standard statistical methods to control for lifetime sun exposure and lifetime cigarette consumption. Our conclusions that among whites there is an association, independent of sun exposure, between cigarette smoking and facial wrinkling are based on the analysis of our data using accepted statistical methods. To the statement, "well, doctor, my black colleagues smoke and they don't have wrinkles," we would say that "it is known that sun exposure, aging, and genetic makeup are the prime causes of wrinkling," but that smoking is an additional contributor in white people. Donald P. Kadunce, MD Joseph L. Lyon, MD John J. Zone, MD University of Utah Health Sciences Center Salt Lake City, UT 84132 References 1. Kadunce DP, Burr R, Gress R, Kanner R, Lyon JL, Zone JJ. Cigarette smoking: risk factor for premature facial wrinkling. Ann Intern Med. 1991;114:840-4. 2. Allen HB, Johnson BL, Diamond SM. Smokers' wrinkles? JAMA. 1973;225:1067-9. 3. Montagna W, Carlisle K. The architecture of black and white facial skin. J Amer Acad Dermatol. 1991;24:929-37. 4. Scotto J, Fears T, Fraumeni J, Jr. Incidence of non-melanoma skin
660
cancer in the United States. Rockville, Maryland: United States Public Health Service; 1981; DHHS publication no (NIH)82-2433. 5. Blake JS, Kadunce DP, Piepkorn MW, Zone JJ. The effects of smoking on the skin. Clin Res. 1988;36:119A.
What's in a Title? To the Editors: Short declarative sentences are being used more frequently for titles of scientific articles. This practice should be discouraged because it is misleading. One example is the otherwise well-written article by Krishna and Kapoor (1) titled "Potassium Depletion Exacerbates Essential Hypertension." Do the authors really mean to imply that the issue is now and forever settled? I suggest three alternatives, beginning with the one I prefer: 1. The Effect of Potassium Depletion on Blood Pressure 2. Does Potassium Depletion Elevate Blood Pressure? 3. A Low Potassium Diet for Ten Days Was Associated With an Elevation of Blood Pressure in Each of Ten Hypertensive Patients Except One Whose Blood Pressure Was Actually Lower on the Potassium Restricted Diet. Mark S. McGowan, MD University of Southern California Los Angeles, CA 90007 Reference 1. Krishna GG, Kapoor SC. Potassium depletion exacerbates essential hypertension. Ann Intern Med. 1991;115:77-83.
Correction A Library for Internists VII A psychiatric textbook was incorrectly described on page 823 of the 1 May 1991 issue as having looseleaf format (1). The correct information is as follows: The American Psychiatric Press Textbook of Psychiatry. Edited by J.A. Talbott, R.E. Hales, and S.C. Yudofsky. American Psychiatric Press, Washington, DC, 1988. 1352 pages. $105.00. A comprehensive reference. The 38 chapters from 55 contributors cover theoretical foundations, assessment, psychiatric disorders, psychiatric treatments, and special topics. The emphasis is on the clinical relevance of psychiatry. Casebound book with study guide available separately. K. Loper American Psychiatric Press Washington, DC 20005 Reference 1. Ludmerer KM. A library for internists. Recommendations from the American College of Physicians. Ann Intern Med. 1991;114:811-32.
15 October 1991 • Annals of Internal Medicine • Volume 115 • Number 8
Downloaded From: https://annals.org/ by a Univ of Utah User on 09/26/2017
