To the Editor.— Dr Howard and colleagues[1] report a greater rate of progression of atherosclerosis in people exposed to tobacco smoke who had diabetes compared with those who did not. However, some of the participants placed in the "without diabetes" category may have diabetes that has not been diagnosed. According to the study, participants were defined as having diabetes if they reported having diabetes, if they were taking blood glucose-lowering medications, or if they had an 8-hour fasting glucose level of at least 11.1 mmol/L (200 mg/dL). Although the chosen value of 11.1 mmol/L for an 8-hour fasting glucose level is specific, it is not sensitive enough to accurately classify participants in the "with diabetes" or "without diabetes" categories that were used clinically. Current clinical practice recommendations call for a diagnostic value of 7.0 mmol/L (126 mg/dL) for an 8-hour fasting glucose level.[2] The accepted value during the course of the study was 7.8 mmol/L (140 mg/dL) for an 8-hour fasting glucose level with 11.1 mmol/L reserved for random samples only.
The relationship of cigarette smoke to the progression of atherosclerosis in the population with diabetes is of utmost importance, especially when considering the added burden of an already compromised vascular endothelium. The authors should provide revised estimates based on these 2 additional diagnostic values to permit optimal public health efforts aimed at protecting those individuals in our society who are most vulnerable to cigarette smoke.
Michael R. Duenas, OD
Chattahoochee, FlaReferences
1. Howard G, Wagenknecht LE, Burke GL, et al. Cigarette smoking and progression of atherosclerosis: the Atherosclerosis Risk in Communities (ARIC) study. JAMA. 1998;279:119-124.
2. American Diabetes Association. Management of dyslipidemia in adults with diabetes. Diabetes Care. 1998;21(suppl I) 179-182.
(JAMA. 1998;280:32)
To the Editor.—Dr Howard et al[1] conclude, "Atherosclerosis progression appears to be largely related to the pack-years of cigarette exposure and not to present smoking status" and "the impact of exposure to ETS [environmental tobacco smoke] was 34% (5.9/17.1) as great as the impact of active smoking on the progression of atherosclerosis." The first conclusion is in stark contradiction with 2 huge bodies of literature, the mechanisms underlying the atherosclerotic process[2] and measured alterations in atherogenic and atherothrombotic markers of risk seen in current smokers but not in ex-smokers. Many studies in current smokers, but not ex-smokers, have reported elevated low-density lipoprotein cholesterol levels and lower high-density lipoprotein cholesterol levels; increased fibrinogen levels; acute platelet aggregation and peripheral white blood cell counts; transient increases in blood pressure, heart rate, and vasoconstriction; increased blood viscosity via elevations in red blood cells, hematocrit, and hemoglobin; and increases in oxidative stress. In summary, the growth rate of plaque in ex-smokers would be expected to be less than the plaque growth rate found in smokers currently experiencing the large number of atherogenic stimuli listed above. If the normalization of these changes in ex-smokers has no impact on the progression of atherosclerosis, then a conceptual paradigm shift regarding the biochemical and hemodynamic stimuli of atherosclerotic plaque growth is in order.
The second conclusion implies that ETS is a special case in which dose-response relationships do not apply. Dose response is absent among subjects reportedly exposed to ETS and is also absent across the vast range of doses represented by ETS exposure and active smoking. The relative atherogenic response of ETS exposure was 34% of that seen with active smoking. This response is disproportionately large compared with the tobacco smoke dose (conservatively estimated at <0.3% of the active smoking dose). As stated in a widely used toxicology text, [3] "This relationship [dose-response] is the most fundamental and pervasive concept in toxicology." Again, a conceptual paradigm shift is called forth.
What could account for these unexpected results? Given that the method of determining exposure dose, both for mainstream and secondhand smoke, was self-report without biomarker confirmation, the well-known bias of smoking status misclassification should be suspected. Howard et al[1] do not mention this bias, although 2 of the article's authors have shown that self-report underestimates the true prevalence of cigarette smoking by about 4%.[4] In addition, reported past smokers tend to misrepresent their current smoking status at a rate more than twice that of those who report having never smoked.[5] Proper correction for this known bias would be expected at least to reduce the calculated progression rate among those who never smoked.
Carr J. Smith, PhD, DABT
Michael W. Ogden, PhD
RJ Reynolds Tobacco Co
Winston-Salem, NCReferences
1. Howard G, Wagenknecht LE, Burke GL, et al. Cigarette smoking and progression of atherosclerosis: the Atherosclerosis Risk in Communities (ARIC) study. JAMA. 1998;279:119-124.
2. Brown G, Albers JJ, Fisher LD, et al. Regression of coronary artery disease as a result of intensive lipid-lowering therapy in men with high levels of apolipoprotein B. N Engl J Med. 1990;323:1289.
3. Eaton DL, Klaassen CD. Principles of toxicology. In: Klaasen CD, ed. Casarett & Doull's Toxicology. 5th ed. New York, NY: McGraw-Hill Book Co; 1996:13-33.
4. Wagenknecht LE, Burke GL, Perkins LL, Haley NJ, Friedman GD. Misclassification of smoking status in the CARDIA study. Am J Public Health. 1992;82:33-36.
5. Ogden MW, Morgan WT, Heavner DL, Davis RA, Steichen TJ. National incidence of smoking and misclassification among the US married female population. J Clin Epidemiol. 1997;3:253-263.
(JAMA. 1998;280:32-33)
To the Editor.—Dr Howard and colleagues[1] report that exposure merely to ETS increases carotid intimal-medial thickening (IMT) by 20% in 3 years, but the 3-year increase (7 microns) is only the width of a red blood corpuscle. The measuring techniques actually used could not detect changes 30 times as large, and the reported results are, in my opinion, a statistical fiction.
First, measurements with any medium (light, x-ray, or sound) are not accurate beyond a half wavelength. The shortest wavelength feasible for this study is 150 µm and more common use is 210 µm. Second, variability from one scan to the next, done a few days apart in the same patient, is commonly more than 100 µm.[2] Third, measurements are made after the technician places a cursor on the apparent margins of IMT. The cursor must be set on one of the screen's pixels, which have a finite number and a fixed position regardless of the image. Even if the technician could see the exact micrometer margin of the IMT, the cursor may lie only in the nearest pixel. The pixel measures 67 µm,[2] so that it always misses the true border by an average of 17 µm, and sometimes much more. Two pixels must be hand-marked for each measurement. Fourth, the Cardiovascular Health Study, in reviewing protocols for carotid ultrasound, concluded that a measured change of IMT must be at least 1400 µm to be certain of a significant difference in 1 individual. In large series, a difference of more than 200 µm could be significant.[3]
Howard et al[1] also calculated that active smoking increased IMT by 17 µm, less than the width of 3 red blood corpuscles, in 3 years. These calculated changes all are far below the threshold of ultrasound detection, but they nevertheless were compared so as to state that environmental smoke causes increases in IMT of 34% that of smokers, a figure that has been widely ballyhooed in the press. Ultrasound measurements are coarse, variable, and operator dependent, suitable at best for tenths of millimeters, not micrometers. The ultrasound data of Howard et al have margins of error far greater than the minute changes they calculated for IMT.
Everett Schultz, Jr, MD
Augusta, GaReferences
1. Howard G, Wagenknecht LE, Burke GL, et al. Cigarette smoking and progression of atherosclerosis: the Atherosclerosis Risk in Communities (ARIC) study. JAMA. 1998;279:119-124.
2. Bond MG, Barnes RW, Riley WA, et al. High-resolution B-mode ultrasound scanning methods in the Atherosclerosis Risk in Communities (ARIC) study. J Neuroimaging. 1991;1:68-73.
3. O'Leary DH, Polak JF, Wolfson SK Jr, et al. Use of sonography to evaluate carotid atherosclerosis in the elderly. Stroke. 1991;22:1155-1163.
(JAMA. 1998;280:33)
In Reply.—Our study showed that active smoking increases the rate of atherosclerotic progression by approximately 50%, past smoking by 25%, and exposure to ETS by 20%, and that the effects of all types of cigarette smoke appear amplified among persons with diabetes and hypertension. We agree with Dr Duenas that the impact of diabetes on the relationship between smoking and atherosclerosis requires further investigation. Similar concerns exist for the amplified harmful effects of smoking among people with hypertension. Furthermore, because of the high population prevalence of these diseases (12% of the population is diabetic, 10% is hypertensive), smoking in public places affects large numbers of highly susceptible people
It is not surprising that Dr Smith and Ogden, scientists representing the tobacco industry, have chosen to criticize this report. They discuss selected pathways for the harmful effects of smoking, but fail to provide a comprehensive account. There are numerous other mechanisms for the harmful effects of smoking, for example, damage to the endothelial layer of the arterial wall with effects that persist after smoking cessation and continue to underline progression of atherosclerosis. They are incorrect that we found no dose-response relationship for active smoking. In addition, they suggest that the absence of a dose-response relationship for ETS implies the absence of an association. There are many counterexamples to this position, and the reasons for the lack of this relationship are clearly discussed in our article. Smith and Ogden mention 1 class of measurement error (misclassification bias) that would potentially exaggerate the observed association, but fail to discuss other effects of measurement error such as regression-dilution bias that dampens the estimated association.
Dr Schultz is correct, but confuses the ability to describe change in an individual with the ability to reliably describe the change of the average value for a group. Individual variability was incorporated in our statistical analysis of the rate of change of the average value, and in the analysis of statistical significance of differences between groups. In addition, the harmful effects of smoking accrue for longer than the 3 years we studied. As the exposure period extends, we surmise that changes in average wall thickness will result in clinically significant incident cardiovascular disease.
George Howard, DrPH
Lynn E. Wagenknecht, DrPH
Wake Forest University School of Medicine
Winston-Salem, NCGregory L. Burke, MD, MS
F. Javier Nieto, MD, PhD
The Johns Hopkins School of Hygiene and Public Health
Baltimore, Md(JAMA. 1998; 280:33)
Comments on this posting?
Click here to post a public comment on the Trash Talk Bulletin Board.
Click here to send a private comment to the Junkman.
Copyright © 1998 Steven J. Milloy. All rights reserved on original material. Material copyrighted by others is used either with permission or under a claim of "fair use." Site developed and hosted by WestLake Solutions, Inc.
Material presented on this home page constitutes opinion of Steven J. Milloy.