ORIGINAL ARTICLE. Confirmations and Surprises in the Association of Tobacco Use With Sinusitis

Confirmations and Surprises in the Association of Tobacco Use With Sinusitis Judith E. C. Lieu, MD; Alvan R. Feinstein, MD ORIGINAL ARTICLE Objective: To generate estimates of sinusitis prevalence for adults in association with the use of tobacco or passive smoke exposure. Design: Analysis of data from the Third National Health and Nutrition Examination Survey, 1988-1994. Setting: Sample of the noninstitutionalized civilian population of the United States. Participants: A total of 20050 adults aged 17 years or older. Main Outcome Measure: Presence of self-reported sinusitis or sinus problems. Results: In the United States, 66 million adults, constituting 35% of the adult population, reported having sinusitis or sinus problems at least once during the previous 12 months. Female sex, non-hispanic white or non- Hispanic black race, higher income levels, and progressively higher educational levels were associated with increased prevalence of sinusitis. The prevalence of both acute and recurrent or chronic sinusitis increased with direct cigarette and other tobacco use but did not rise with passive exposure to cigarette smoke. Conclusions: Consistent with data for other respiratory ailments, the direct use of tobacco confers a small increased risk of developing sinusitis in the adult population, but contrary to expectation, passive smoke does not. The demographic variables of sex, race/ethnicity, and educational level demonstrated unexpectedly strong associations with the prevalence of sinusitis and should be analyzed and controlled for in future studies of sinusitis. Arch Otolaryngol Head Neck Surg. 2000;126:940-946 From the Robert Wood Johnson Clinical Scholars Program, Yale University School of Medicine, New Haven, Conn. ALTHOUGH SINUSITIS was the most frequently reported chronic condition in 1995, affecting 14.1% of the US population, 1 the impact of any simple etiologic factor has been difficult to define for individual patients. Tobacco smoke was suggested as a risk factor for the aggravation and prolongation of sinusitis in the 1964 Surgeon General s report on smoking, 2 but no clinical studies have documented this risk. Several lines of evidence link sinusitis to tobacco exposure through its effect on nasal mucociliary function. Cigarette smoke 3 and the inhalation of nasal snuff 4 have a profound effect on nasal airway mucociliary function, demonstrated by in vivo nasal mucociliary clearance tests. Agius et al 5 showed that cotinine, a metabolite of nicotine, caused a significant decrease in the ciliary beat frequency of nasal mucosal cells in vitro. Environmental tobacco smoke has been shown to alter nasal mucociliary clearance in nonsmokers. 6 People with allergic rhinitis and chronic sinusitis have been found to have nasal mucociliary clearance dysfunction, 7 but a clear clinical association with severity of symptoms has not been shown. Although pathophysiologically plausible, changes in nasal mucociliary function from tobacco do not necessarily translate into clinical sinusitis. This study was undertaken to examine whether the use of tobacco increases the risk of sinusitis in the US population and whether passive smoke exposure is also associated with an increased prevalence of sinus problems. RESULTS EFFECT OF DEMOGRAPHIC VARIABLES Table 1 shows that 35% of the adult population of the United States reported having sinusitis or sinus problems at least once during the past 12 months. Female 940

MATERIALS AND METHODS DATA SOURCE We used data from the Third National Health and Nutrition Examination Survey (NHANES III), a national crosssectional health survey, performed in 1988-1994, of 33994 persons aged 2 months and older, representing the noninstitutionalized civilian population of the United States. Details of the complex sampling design, data collection, and weighting approach have been described elsewhere. 8 Briefly, NHANES III used a stratified, multistage probability sample design, with oversampling of young children ( 5 years old), older persons ( 59 years old), black persons, and Mexican Americans. To obtain a distribution of participants that would be similar to the US population as a whole, sampling weights can be used during the analysis to incorporate the differential probabilities of selection and to include adjustments for noncoverage and nonresponse. To generate estimates of national prevalence, the present analysis is limited to the 20050 persons who comprised the adult population (age 17 years). DEFINITION OF VARIABLES In the NHANES III interviews on demographic, tobacco use, and health characteristics, race/ethnicity was defined by the participant as non-hispanic white, non-hispanic black, or Mexican American. Any person who did not choose those categories was considered to be other. Poverty income ratio (PIR) is the ratio of reported household annual income to the poverty threshold defined by the US Census Bureau, with adjustment for family size. This ratio was coded as a continuous variable in the NHANES III database and was demarcated for our analysis into 4 ordinal categories: less than 1.00, 1.00 through 1.99, 2.00 through 2.99, or 3.00 or greater. Educational level completed was coded in the NHANES III database as never attended or kindergarten only or in individual levels of 1 through 17 years, with 17 entered for those with 17 or more years of education. From these codes we categorized educational level as no schooling, 1 through 6 years (elementary school), 7 through 9 years (middle school), 10 and 11 years (some high school), 12 years (graduated high school), and 13 through 17 years (at least some college). The dependent variable sinusitis was self-reported for the following question: During the past 12 months, have you had sinusitis or sinus problems? A positive response was followed by a question about the number of episodes, which could be answered with an integer number or continuously, constantly. Anyone who responded positively was considered to have any sinusitis. For the present analysis, we defined acute sinusitis as having 1 to 3 episodes, and recurrent or chronic sinusitis as having at least 4 episodes or continuous sinus problems during the past 12 months. Tobacco exposure variables were the direct use of cigarette, cigar, pipe, and smokeless forms of tobacco; passive exposure to cigarette smoke; and the quantification of cigarette exposure, both direct and passive. Ever used tobacco was defined in the survey as having smoked at least 100 cigarettes, 20 cigars, 20 pipefuls of tobacco, or ever having used chewing tobacco or snuff. Those who had ever used cigarettes, cigars, pipes, chewing tobacco, or snuff were asked about current use of those forms of tobacco. We defined former smokers as those who had ever used cigarettes but did not currently smoke them. Direct cigarette exposure was quantified as the number of cigarettes currently smoked per day and was also dichotomized as 10 or fewer cigarettes per day vs more than 10 cigarettes per day. Passive cigarette smoke exposure in the survey was determined by asking about the existence and number of persons who smoked cigarettes in the household. We coded this passive exposure as yes or no for the presence of smokers in the household and quantified it by summing the number of persons who smoked in the home and the number of cigarettes those persons smoked per day. To quantify the dose of tobacco exposure in another way and to capture other passive smoke exposure outside the home, we also examined the NHANES results for serum cotinine, a metabolite of nicotine. Since they had been performed only on study subjects who participated during 1988-1991, serum cotinine measurements were available for only 60% of the total group. Serum cotinine concentrations were determined in a 2-step process. The enzyme immunoassay method was used as a screening method for differentiating low ( 142 nmol/l [ 25 ng/ml]) and high ( 142 nmol/l [ 25 ng/ml]) cotinine concentrations. Confirmatory analysis was then performed with a liquid chromatography mass spectrometric method in batches of 50 specimens, according to their low or high concentrations on enzyme immunoassay. Serum cotinine concentrations less than 28.4 nmol/l (5 ng/ml) generally indicate nonsmokers, levels of 28.4 to 85.2 nmol/l (5-15 ng/ ml) may indicate recent passive smoke exposure, and levels greater than 85.2 nmol/l (15 ng/ml) generally indicate active smokers. 9 STATISTICAL ANALYSIS Our analyses used SUDAAN software (Research Triangle Institute, Research Triangle Park, NC) to incorporate sampling weights consistent with the complex design of the NHANES III survey. 8 Bivariate comparisons of prevalence of sinusitis were assessed using 2 and risk ratios (RRs) with 95% confidence intervals (CIs) for dichotomous independent variables and logistic regression with odds ratios (ORs) and 95% CIs for the categories of polytomous independent variables, which were converted to dummy binary variables. For the continuous serum cotinine level variable, means, t tests, medians, and interquartile ranges were used. We adjusted for the demographic variables of sex, race/ethnicity, PIR, and educational level using multivariable logistic regression. For the multivariable logistic regression models, the group with the smallest prevalence was chosen as the reference level for the nominal variables sex and race/ethnicity. Adjusted ORs were converted into adjusted RRs, since the prevalences of any, acute, or recurrent or chronic sinusitis in this population were all found to be more than 10%. 10 The more common the outcome of interest, the more the OR overestimates the RR if it is greater than 1 or underestimates it if it is less than 1. Multiple stratification tables were used to look for possible interactions between the independent variables in affecting the prevalence of any sinusitis. 11 941

Table 1. Effect of Demographic Variables on the Prevalence of Any Sinusitis* Variable NHANES Weighted Unadjusted RR Total 5848/20 011 29.2 35.3... Sex Male 2315/9381 24.7 30.8 1.00 Female 3533/10 630 33.2 39.5 1.19 (1.15-1.24) Race/ethnicity Non-Hispanic white 3054/8459 36.1 38.3 1.88 (1.68-2.08) Non-Hispanic black 1604/5479 29.3 29.7 1.45 (1.29-1.62) Mexican American 1044/5299 19.7 20.4 1.00 Other 146/774 18.9 23.5 1.15 (0.95-1.37) Poverty income ratio 1.00 1015/4277 23.7 32.0 1.00 1.00-1.99 1367/5041 27.1 31.9 0.99 (0.87-1.13) 2.00-2.99 1089/3306 32.9 35.7 1.12 (0.96-1.28) 3.00 1860/5183 35.9 38.6 1.20 (1.04-1.38) Unknown 517/2204 23.5 31.5 0.99 (0.82-1.17) Educational level, y No schooling 69/519 13.3 16.6 1.00 1-6 400/2346 17.1 20.4 1.23 (0.79-1.83) 7-9 740/2988 24.8 29.3 1.77 (1.19-2.49) 10-11 795/2648 30.0 34.6 2.09 (1.43-2.86) 12 1950/5946 32.8 37.2 2.24 (1.58-3.00) 13-17 1857/5349 34.7 37.8 2.28 (1.56-3.11) Unknown 37/215 17.2 19.9 1.20 (0.66-2.02) *NHANES indicates National Health and Nutrition Examination Survey; RR, risk ratio (calculated from the weighted prevalence); and CI, confidence interval. ratio is the actual number of people in the survey who reported that they had sinusitis divided by the number of people in the category. prevalence is calculated from the actual ratio. Table 2. Effect of Cigarette Smoking Status on Prevalence of Any Sinusitis* Cigarette Smoking Status Prevalence, % NHANES Weighted Prevalence, % Adjusted RR Never 2854/10 223 27.9 33.8 1.00 Former 1466/4797 30.6 35.6 1.06 (0.98-1.14) Current 1523/4981 30.6 37.6 1.14 (1.06-1.22) Total 5843/20 001 29.2 35.3... *NHANES indicates National Health and Nutrition Examination Survey; RR, risk ratio (calculated from weighted prevalence); and CI, confidence interval. ratio is the actual number of people in the survey who reported that they had sinusitis divided by the number of people in the category. prevalence is calculated from the actual ratio. NHANES weighted prevalence incorporates the sampling weights to estimate prevalence for the US population. sex, non-hispanic white or non-hispanic black race, and higher income status (PIR 3.00) were all associated with an increased prevalence of any (including both acute and recurrent or chronic) self-reported sinusitis in bivariate analysis. This prevalence also progressively increased as educational level increased. In multiple stratification tables, all race/ethic groups had a higher prevalence with female sex and progressively increasing prevalence with increasing education. However, PIR did not affect the prevalence of any sinusitis in all race/ethnicity groups in the same way; although non-hispanic black and Mexican Americans did show increases with increasing PIR, non-hispanic white and others did not. EFFECT OF CIGARETTE SMOKING As shown in Table 2, current cigarette smokers had a higher prevalence of any sinusitis than former and never smokers. Table 3 shows that when stratified by the number of cigarettes smoked per day, those who smoked 11 or more cigarettes per day had a significantly higher prevalence of any sinusitis than those who smoked fewer than 11 cigarettes per day or did not smoke (adjusted RR, 1.16). Median serum cotinine concentrations increased with the numbers of cigarettes smoked per day, demonstrating a dose relationship, except for persons who smoked more than 40 cigarettes per day. When adjusted for the demographic variables of sex, race/ethnicity, educational level, and PIR, the risks remained essentially the same. However, the increase in prevalence is more consistent with a threshold effect than with a dose-response effect. In a threshold effect, the outcome response suddenly jumps after previously being unchanged at increasing levels of exposure. In contrast, in a dose-response effect, the outcome response rises or falls constantly (or monotonically ) with higher levels of exposure. Quantifying dose by increasing serum cotinine concentrations (Table 4) also did not show consistent increases in the prevalence of sinusitis. Table 5 divides the study population into those with acute sinusitis and those with recurrent or chronic sinusitis; it reveals that about 12% of adults reported recurrent or chronic problems with their sinuses, whereas about 21% had acute problems. (The sum of prevalences for acute sinusitis and recurrent or chronic sinusitis does not equal 35%, because 742 persons in the survey did not respond to the query about the number of 942

Table 3. Prevalence of Any Sinusitis by Number of Cigarettes Currently Smoked per Day* No. of Cigarettes Smoked per Day Median Serum Cotinine Level, NHANES Weighted Adjusted RR Total 3.35 (0.59) 5848/20 011 29.2 35.3... None 1.31 (0.23) 4320/15 020 28.8 34.4 1.00 1 40.1 (7.04) 78/295 26.4 31.1 0.95 (0.74-1.19) 1-4 263.4 (46.38) 170/668 24.4 36.2 1.17 (0.97-1.39) 5-9 682.39 (120.14) 184/679 27.1 30.7 0.95 (0.80-1.13) 10 1014.0 (178.53) 219/800 27.4 33.9 1.04 (0.84-1.26) 11-19 1235.3 (217.49) 156/490 31.8 43.4 1.24 (1.03-1.45) 20 1531.6 (269.65) 389/1230 31.6 38.0 1.12 (1.01-1.24) 21-39 1652.1 (290.86) 164/411 39.9 40.0 1.15 (0.98-1.32) 40 1735.1 (305.48) 120/300 40.0 41.0 1.22 (1.04-1.40) 40 1486.2 (261.66) 48/118 40.7 39.7 1.20 (0.83-1.62) 0-10 3.35 (0.59) 4971/17 462 28.5 34.3 1.00 10 1606.6 (282.85) 877/2549 34.4 39.8 1.16 (1.08-1.25) *NHANES indicates National Health and Nutrition Examination Survey; RR, risk ratio (calculated from weighted prevalence); and CI, confidence interval. ratio is the actual number of people in the survey who reported that they had sinusitis divided by the number of people in the category. prevalence is calculated from the actual ratio. Table 4. Prevalence of Self-Reported Sinusitis by Serum Cotinine Concentrations* Serum Cotinine Level, NHANES Weighted Adjusted RR Unknown 2313/8024 28.8 34.6 0.96 (0.77-1.18) 0-28.4 (0-5) 1525/5389 28.3 33.5 1.00 28.4-85.2 ( 5-15) 47/178 26.4 35.1 1.17 (0.79-1.61) 85.2-284 ( 15-50) 62/261 23.8 30.4 0.94 (0.70-1.21) 284-568 ( 50-100) 89/285 31.2 37.3 1.21 (1.00-1.44) 568-852 ( 100-150) 78/285 27.4 41.4 1.25 (0.86-1.67) 852-1136 ( 150-200) 91/291 31.3 34.0 1.09 (0.91-1.30) 1136-1420 ( 200-250) 92/300 30.7 33.2 1.04 (0.81-1.30) 1420-1704 ( 250-300) 80/275 29.1 36.6 1.12 (0.89-1.38) 1704-1988 ( 300-350) 79/234 33.8 41.4 1.27 (0.96-1.60) 1988 ( 350) 170/526 32.3 37.6 1.20 (0.99-1.42) *NHANES indicates National Health and Nutrition Examination Survey; RR, risk ratio (calculated from weighted prevalence); and CI, confidence interval. ratio is the actual number of people in the survey who reported that they had sinusitis divided by the number of people in the category. prevalence is calculated from the actual ratio. Table 5. Prevalence of Acute and Recurrent or Chronic Sinusitis by Cigarette Use Status* Acute Sinusitis Recurrent or Chronic Sinusitis Cigarette Use Weighted Prevalence, % Adjusted RR Weighted Prevalence, % Adjusted RR Total 20.7... 11.8... Smoking status Never 20.3 1.00 11.1 1.00 Former 19.8 0.99 (0.89-1.11) 12.3 1.12 (0.94-1.33) Current 22.2 1.18 (1.05-1.30) 12.7 1.22 (1.05-1.39) No. of cigarettes smoked per day 0-10 20.2 1.00 11.4 1.00 10 23.1 1.21 (1.07-1.36) 13.7 1.23 (1.07-1.42) *RR indicates risk ratio (calculated from the weighted prevalence); CI, confidence interval. Weighted prevalence incorporates the sampling weights to estimate prevalence of the US population. episodes during the previous 12 months.) Current smokers report an increased prevalence of both acute (adjusted RR, 1.18) and recurrent or chronic (adjusted RR, 1.22) sinusitis. Although the RRs are small, they represent a proportional increase of 14% in recurrent or chronic sinusitis between never smokers and current smokers. Examined as the number needed for one effect (calculated as the reciprocal of the absolute difference in preva- 943

Table 6. Effect of Type of Tobacco Currently Used on Prevalence of Any Sinusitis* Type of Tobacco Currently Used Median Serum Cotinine Level, Prevalence, % NHANES Weighted RR Never used any 1.08 (0.19) 2598/9376 27.7 33.5 1.00 Cigarettes 1297.5 (228.44) 1523/4981 30.6 37.6 1.18 (1.09-1.27) Chewing tobacco 1043.9 (183.78) 103/328 31.4 37.3 1.11 (0.81-1.53) Snuff 1721.3 (303.04) 84/229 36.7 40.6 1.21 (0.96-1.53) Pipe 667.17 (117.46) 44/146 30.1 40.0 1.19 (0.86-1.66) Cigar 554.8 (97.67) 88/296 29.7 37.1 1.11 (0.89-1.37) Noncigarette tobacco users 1195.0 (210.39) 227/708 32.1 38.1 1.37 (1.21-1.54) *NHANES indicates National Health and Nutrition Examination Survey; RR, risk ratio (calculated from the weighted prevalence); and CI, confidence interval. ratio is the actual number of people in the survey who reported that they had sinusitis divided by the number of people in the category. prevalence is calculated from the actual ratio. Risk ratios are unadjusted except for the cigarettes and noncigarette tobacco users categories. lence between current smokers and never smokers), 62 people must become current smokers to result in one excess case of recurrent or chronic sinusitis. When dichotomized by the number of cigarettes smoked per day, those who smoke more than 10 cigarettes per day had an increased risk of developing both acute and recurrent or chronic sinusitis. EFFECT OF NONCIGARETTE TOBACCO USE Table 6 shows that the prevalence of any sinusitis was higher for each form of tobacco used when compared with never users of any form of tobacco. The RRs for chewing tobacco, snuff, pipe, and cigar use were not adjusted to avoid problems with too few events per variable. 11,12 Because the 95% CIs for the RRs include 1.0, increased risk cannot be claimed for each separate form of tobacco used. When users of noncigarette forms of tobacco (ie, pipe, cigar, chewing tobacco, and snuff) were examined as a combined group, however, they had a statistically significant increase in the prevalence of any sinusitis (adjusted RR, 1.37) when compared with never users. INTERACTION OF DEMOGRAPHIC VARIABLES AND TOBACCO USE In data not tabulated herein, when race/ethnicity was analyzed with tobacco use variables in multiple stratification tables, non-hispanic whites and Mexican Americans showed a higher prevalence of any sinusitis in users compared with nonusers of tobacco. Non-Hispanic blacks tended to have a lower prevalence of any sinusitis with tobacco use, whereas other race/ethnicity groups tended to have a higher prevalence. In multiple logistic regression models that included tobacco use variables, female sex, 10 years or more of education, and non-hispanic white and non-hispanic black race/ethnicity groups were independent predictors of increased prevalence for sinusitis, whereas PIR was not. EFFECT OF PASSIVE SMOKE EXPOSURE Table 7 shows that among the never smokers exposure to passive smoke at home did not result in a higher prevalence of any sinusitis. Median serum cotinine concentrations were slightly higher for those exposed to passive smoke in the home, confirming their exposure. Quantifying the dose by the number of smokers, the daily total number of cigarettes smoked in the home, or serum cotinine concentrations also did not lead to a higher prevalence with increasing dose. When divided into acute and recurrent or chronic sinusitis groups, exposure to passive smoke did not increase the prevalence in either group, with adjusted RRs of 1.23 (95% CI, 0.97-1.54) and 0.96 (95% CI, 0.73-1.27) for acute and recurrent or chronic sinusitis, respectively. COMMENT The results of this study demonstrate that the direct use of tobacco, but not household passive tobacco smoke exposure, can be linked to an increased prevalence of sinusitis. This study, however, does not allow conclusions about whether the use of tobacco or exposure to its smoke affects the severity of symptoms, chronicity of disease, or refractoriness to treatment. The prevalence of any self-reported sinusitis shows variation with the numbers of cigarettes directly smoked but not a true dose response. Since sinusitis prevalence is also most likely increased by smokeless tobacco, direct tobacco smoke is probably not the only cause for mucociliary dysfunction. Nasal mucociliary dysfunction from nicotine is not likely to be the only pathophysiologic explanation for the increased prevalence of sinusitis from tobacco use, however, because serum cotinine concentrations did not consistently affect the prevalence of sinusitis. Other components in tobacco or tobacco smoke may cause changes to the cilia or nasal secretions that are responsible for the observed differences. In addition, individuals susceptibilities play an important role in their response to tobacco smoke. Bascom et al 6 found that in healthy nonsmoking adults, 6 of 12 subjects showed more rapid nasal mucociliary clearance with sidestream tobacco smoke than with just air, whereas 3 of 12 subjects had substantial decreases in clearance. Defining passive smoke exposure is limited by the quantification of household exposures only. Workplace exposures are unaccounted for in the survey. However, 944

Table 7. Effect of Passive Smoke Exposure on Prevalence of Any Sinusitis Among Never Users of Tobacco* Passive Smoke Exposure Median Serum Cotinine Level, NHANES Weighted Adjusted RR Total 1.08 (0.19) 2585/9330 27.7 33.5... Smokers in present house Yes 4.99 (0.88) 446/1709 26.1 34.4 1.15 (0.98-1.32) No 0.85 (0.15) 2139/7621 28.1 33.3 1.00 No. of persons who smoke in household None 0.85 (0.15) 2139/7621 28.1 33.3 1.00 1 4.43 (0.78) 347/1276 27.2 36.6 1.21 (1.02-1.41) 2 7.16 (1.26) 79/327 24.1 27.4 0.92 (0.67-1.23) 3 7.89 (1.39) 20/106 18.9 30.6 1.13 (0.64-1.72) No. of cigarettes per day of passive exposure None 0.85 (0.15) 2139/7621 28.1 33.3 1.00 1-10 3.64 (0.64) 267/1071 24.9 34.4 1.16 (0.96-1.37) 11-20 6.70 (1.18) 103/386 26.7 32.1 1.03 (0.79-1.31) 21-40 8.86 (1.56) 55/176 31.2 38.0 1.28 (0.96-1.62) 40 8.75 (1.54) 12/46 26.1 36.8 1.26 (0.61-2.02) *NHANES indicates National Health and Nutrition Examination Survey; RR, risk ratio (calculated from the weighted prevalence); and CI, confidence interval. ratio is the actual number of people in the survey who reported that they had sinusitis divided by the number of people in the category. prevalence is calculated from the actual ratio. serum cotinine levels are not significantly different between the exposed with sinusitis and without sinusitis. Given that more than 10 cigarettes had to be smoked before the prevalence among direct smokers exceeded that of nonsmokers in this study, the dose from passive tobacco smoke is probably insufficient to increase the risk of sinusitis in tobacco nonusers. The main limitation of this study is that sinusitis is self-reported in a questionnaire rather than documented by a physician using strict diagnostic criteria. Colds, flu, allergies, facial pain or pressure, headache, and nasal obstruction without sinusitis may all be included in self-reported sinusitis. Nevertheless, this method of self-report is similarly used for estimating prevalence in the United States, according to the National Health Interview Survey, an annual multistage probability sample survey conducted by the US Census Bureau for the National Center for Health Statistics. In that survey, during 1990-1992, the estimated US prevalence of chronic sinusitis was 13.6%, affecting 33.7 million people. 13 This result can be compared with the 11.8% of adults in the present analysis for an overlapping period. The estimates may differ, because the National Health Interview Survey defines chronic as duration of problem (3 months, vs our categorization of recurrent or chronic as more than 3 episodes or continuous problems). The estimates produced by our analysis seem just as valid as those given by the National Center for Health Statistics, which concluded that chronic sinusitis is now the most common chronic condition affecting the US adult population. 1 It is unclear why the prevalence of self-reported sinusitis varies with income level and educational level. Differences in access to medical care would produce a detection bias, leading to more diagnoses for symptom clusters in patients with more access. A different explanation is that those with more access to medical care may have higher expectations for health and may not tolerate symptoms that others with more barriers to access may consider not important enough to pursue with a physician. Differences in prevalence for race/ethnicity may be due to true differences in the baseline racial susceptibility to sinusitis. Race/ethnicity correlates with some differences in PIR and educational level, but these socioeconomic factors do not explain all the variation in prevalence with race/ethnicity. These differences in prevalence for sex, white and black race, and income level were also noted in the National Center for Health Statistics estimates for chronic sinusitis. 13 Because these variations with race/ethnicity, educational level, and sex seem robust, these demographic variables will be important to analyze and control for in future studies of sinusitis. CONCLUSIONS Consistent with data for other respiratory ailments, the direct use of tobacco confers a small increased risk of developing sinusitis in the adult population, but contrary to expectation, passive smoke does not. Although risk factors should be identified and altered when possible, it is also important to avoid burdening patients with recommendations for lifestyle alterations that may not affect their disease. There are enough other healthrelated reasons not to smoke or use tobacco without chiding people for causing their own or someone else s sinus problems. The demographic variables of sex, race/ ethnicity, and educational level demonstrate unexpectedly robust associations with the prevalence of sinusitis and should be analyzed and controlled for in future studies of sinusitis. Although the multifactorial risk factors for sinusitis make it difficult to quantitate the impact of any one factor, only by doing work like this can we take measures to prevent and treat this common problem. 945

Accepted for publication February 22, 2000. Corresponding author: Judith E. C. Lieu, MD, Department of Otolaryngology Head and Neck Surgery, Washington University School of Medicine, One Children s Place, Room 3S35, St Louis, MO 63110. Reprints: Alvan R. Feinstein, MD, Robert Wood Johnson Clinical Scholars Program, Yale University School of Medicine, 333 Cedar St, IE-456 SHM, PO Box 208025, New Haven, CT 06520-8025. REFERENCES 1. Benson V, Marano MA. Current estimates from the National Health Interview Survey, 1995. Vital Health Stat 10. 1998;No. 199:5, 90. 2. US Department of Health, Education and Welfare, US Public Health Service. Smoking and Health: Report of the Advisory Committee to the Surgeon General of the Public Health Service. Washington, DC: US Dept of Health, Education and Welfare, US Public Health Service; 1964. 3. Ogino S, Nose M, Irifune M, Kikumori H, Igarashi T. Nasal mucociliary in patients with upper and lower respiratory diseases. ORL J Otorhinolaryngol Relat Spec. 1993;55:352-355. 4. Chetan S. Nasal mucociliary clearance in snuff users. J Laryngol Otol. 1993;107: 24-26. 5. Agius AM, Wake M, Pahor AL, Smallman LA. Smoking and middle ear ciliary beat frequency in otitis media with effusion. Acta Otolaryngol (Stockh). 1995;115: 44-49. 6. Bascom R, Kesavanathan J, Fitzgerald TK, Cheng KH, Swift DL. Sidestream tobacco smoke exposure acutely alters human nasal mucociliary clearance. Environ Health Perspect. 1995;103:1026-1030. 7. Mahakit P, Pumhirun P. A preliminary study of nasal mucociliary clearance in smokers, sinusitis and allergic rhinitis patients. Asian Pac J Allergy Immunol. 1995;13:119-121. 8. National Center for Health Statistics. Plan and Operation of the Third National Health and Nutrition Examination Survey, 1988-1994. Hyattsville, Md: US Dept of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Center for Health Statistics; 1994. 9. Gunter EW, Lewis BG, Koncikowski SM. Laboratory Procedures Used for the Third National Health and Nutrition Examination Survey (NHANES III), 1988-1994. Hyattsville, Md: Centers for Disease Control and Prevention, National Center for Health Statistics; 1996. 10. Zhang J, Yu KF. What s the relative risk? a method of correcting the odds ratio in cohort studies of common outcomes. JAMA. 1998;280:1690-1691. 11. Feinstein AR. Multivariable Analysis: An Introduction. New Haven, Conn: Yale University Press; 1996. 12. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol. 1996;49:1373-1379. 13. Collins JG. Prevalence of selected chronic conditions: United States, 1990-1992. Vital Health Stat 10. 1997;No. 194:1-7. 946