Nasal cytology in southwest metropolitan Mexico City inhabitants: a pilot intervention study.

Southwest metropolitan Mexico City (SWMMC) inhabitants have been exposed several hours per day for the last 6 years to photochemical smog, ozone being the most important oxidant pollutant. Subjects exposed to the SWMMC atmosphere develop several histopathological changes in their nasal mucosa: dysplasia is the most significant, affecting 78.72% of adult individuals within 60 or more days of residence in SWMMC. This study was originally designed to explore whether chemical intervention could modify nasal dysplasia, as determined by nasal cytology, in a defined adult population. In a placebo-controlled, randomized, double-blind trial, 177 healthy male subjects were divided into 5 groups to whom 5000 IU of vitamin A, 100 IU of vitamin E, a combination of vitamins A and E (5000 IU + 100 IU), 16 mg of beta-carotene, or placebo were administered daily for 4 months. Sixteen clinical and cytological variables were monitored. No effect on dysplasia was seen at the end of the 4-month trial; however, an apparent reversibility as well as progression of the dysplastic nasal lesions and high correlation coefficients between dysplasia and nasal cytology of polymorphonuclear leukocytes (PMNs; 0.85), squamous metaplasia (SM; 0.50), and nasal mucosa atrophy (NMA; 0.41) were found. A mathematical theoretical nasal dysplasia (tD) predictor equation for SWMMC adult male inhabitants is proposed (tD = 0.85 delta PMNs + 0.50 delta SM + 0.41 delta NMA + 0.98), in which PMNs are the best single dysplasia predictor, and all variables are independent.(ABSTRACT TRUNCATED AT 250 WORDS)

Retinoids are active in the promotional phase of carcinogenesis (6), and their chemopreventive actions are reversible upon withdrawal, a very important point for our SWMMC-exposed population, since theoretically it implies that the subjects have to continue with retinoids or their precursors for as long as they remain in the highly polluted urban environment. Because our final aim is to give the chemoprevention treatment for long periods of time to a large population, including women and children, the doses of vitamin A and f-carotene chosen are low compared to those used in other clinical trials. Vitamin E is a well-known antioxidant that acts in the promotional phase of carcinogenesis through its action on free radicals (2,2'). Absence of dietary vitamin E exacerbates lung injury from ozone inhalation in rats (8). Elsayed et al. (8) found that the magnitude of this protective effect does not increase proportionately with increased dietary vitamin E supplementation beyond a certain level.
The results of our 4-month pilot trial show no statistically significant difference between the treated and placebo groups on the chosen endpoint of nasal dysplasia, remission of established nasal dysplasias and/or inhibition of the development of new dysplasias. Furthermore, we present evidence of a high correlation coefficient between nasal dysplasia and polymorphonuclear leukocytes (PMNs) revealed by nasal cytology and also of a direct relationship between an increment in PMNs in the nasal turbinates and the total number of monthly exposure hours to ozone above 0.1 1 ppm. A theoretical nasal dysplasia predictor equation is proposed for SWMMC adult males.
We suggest that the nasal cytological changes in SWMMC inhabitants may constitute an adaptative response to the urban I !i -.* .. I --.i Iigo environmental pollutants. A long-term follow-up of these subjects will be necessary to establish the possible outcomes of the nasal abnormalities.

Methods
The project was approved by the Medical Ethical Committee of the Naval Medical Center, and informed consent was obtained from all subjects. We selected 206 healthy male subjects between 18 and 50 years of age (mean 30.4 ± 7.2 years) for the study. The participants were health workers, medical staff, and Marines working in SWMMC, where they spend . 8 hr (from 7:00 AM onward), 6-7 days per week. In addition, 49.75% of these subjects lived in the SWMMC urban area. We chose this population because of the many advantages it presented for an intervention project. The subjects were stationed in an SWMMC area for which we have atmospheric pollutant data 24 hr per day, all year long, for several years; their movements during the day (when the pollutant emissions are at their highest) were virtually restricted to a few city blocks; the participants had no known exposures to local sources of air pollutants; they could be readily reached to control the intake of medication; and, finally, they were very cooperative with researchers. Subjects were introduced into the study during November and December 1990, and the trial ended 4 months later.
Clinical data obtained included age, sex, place of residence within MMC, length of residence, occupation, personal history of allergic diseases, smoking and drinking habits, and otolaryngologic history (epistaxis, nasal dryness, nasal mucous, nasal obstruction, rhinorrea, and septal deviation). These latter six clinical variables form part of the clinico-cytological variables (CCV) and were scored as absent, minimal, moderate, or severe. Septal deviation was scored as absent or present. We excluded individuals with a history of surgical otolaryngologic procedures or in need of treatment, atopic rhinitis, asthma, upper respiratory infections, known job exposure to toxic substances, photocopying machine workers, cigarette smokers with a consumption of more than half a pack per week, subjects traveling to and from MMC, and those regularly taking vitamin preparations or undernourished.
The study was a double-blind randomized, placebo-controlled, 4-month trial, and the subjects were divided into 5 groups. All groups received daily doses of vitamin A (5000 IU), vitamin E (tocopheryl acetate; 100 IU), a combination of vitamins A and B (5000 LU + 100 LU), T carotene (16 mg), or placebo (vegetable oil). All participants were submitted to an ear-nose-throat examination by an otolaryngologist three times during the study period: days 0, 60, and 120. The ENT findings, upper respiratory symptomatology and medication side effects were recorded; the administration of the designated doses was closely monitored throughout.

Cytology Sampling
After the rhinoscopic examination, we scored the severity of the nasal mucosa findings by reference to five parameters (CCV); atrophy, hyperemia, hyaline mucous bridges, bleeding friable mucosa, and pale mucosa (1 = absent to 4 = severe). The nasal cytology specimen was taken by the same physician from behind the anterior curvature of the inferior turbinate of the nasal cavity with the best airflow passage. The second and third samples were obtained from the same side as the original one. A small piece of a wooden tongue depressor was used to take the cytology sample. We obtained two slides from each subject, and Papanicolaustained slides were prepared and examined by two pathologists separately,with no access to clinical information. The following cytological parameters were scored: ciliated respiratory-type cells, goblet cells, squamous metaplasia cells, inflammatory cells, and dysplasia (Figs. [1][2][3][4]. For the first three parameters, the scoring system was 1 = absent, 2 = 1-25%, 3 = 26-50%, 4 = 51-75%, and 5 = >76%. The last two parameters were scored absent, mild, moderate, and severe. In 86% of the cases, the scores of both observers coincided; in the .,   remaining cases, the samples were reexamined, and a joint determination of the appropriate score was reached. Only the results of the first and the third clinical and cytological evaluations are considered here.
In this study, the endpoint for assessing chemical intervention was dysplasia: remission of established nasal dysplasias and/or inhibition of the development of new dysplasias. The criteria of cytological dysplasia was made on the bases of the identification of dysplastic cells exhibiting mild to severe pleomorphism, both in the size and shape of cells and their nuclei, as well as alterations in the staining characteristics.

Statistical Analysis
We used Student's t-test, analysis of variance with contrast, Fisher's least significant difference test, Cochran-Mantel-Haenszel  chi-square test, and multivariate linear regression analysis for statistical analysis. For the development of the theoretical dysplasia predictor equation, multivariate linear analysis was used. Only p < 0.05 was considered significant.

Air Quality Data
Ozone (03) has been monitored continuously at the Atmospheric Science Center (located in SWMMC) since 1979 and not until 1981 were diurnal levels of 03 > 0.11 ppm recorded: 40 and 30 hr of 03 > 0.11 ppm were recorded in 1984 and 1985, respectively. The number of hours SWMMC inhabitants are exposed to 03 > 0.11 ppm has been steadily increasing since September 1986 and coincided with the marked decrease of tetraethyl lead in MMC gasoline. For 11 of the 12 months of 1990 and for every month in 1991, more than 85 hr of 03 > 0.11 ppm were recorded, for a total of 1403 hr in 1990 and 1561 hr in 1991 (Table 1). In fact, there were only 6 days in 1991 with diurnal levels of 03 < 0.11 ppm. A similar trend has been observed in the monthly average of maximal concentrations of 03 recorded in SWMMC ( Table 2): during 1990 and 1991 the maximum levels were mostly above 0.2 ppm, with the highest levels (> 0.5 ppm) recorded in October 1991.
Concentrations of sulfur dioxide in SWMMC for the years 1990-1991 have shown a diurnal cycle with maxima in the morning hours and in the winter months; however, the EPA Air Quality Standard of 30 ppb annual average for sulfur dioxide is not exceeded in SWMMC. Formaldehyde levels in SWMMC show a peak shortly before noon, and in samples taken between

Study Population
Of the initial 206 subjects, 177 were monitored throughout the 4-month trial; 29 subjects were excluded due to changes in work location. The participants had a mean length of residence in MMC of 165 ± 143 months (46% of them had < 72 months and 54% > 72 months residence in MMC); only 19 participants had less than 1 year of residence in MMC (5.94 ± 2.8 months). There were no statistical differences among groups at day 0 in regard to clinical, otolaryngological, and nasal cytology findings. No attempt was made in this study to determine the indoor/outdoor exposure schedule or the level of physical exercise of the participants, although the majority spent most of their working time indoors and were in the category of intermittent light and moderate physical exercise. No medication side effects were reported.
We grouped the subjects in terms of their cytological dysplasia and their chemo-prevention treatment (Table 3). Sixteen clinical and cytological variables were evaluated at days 0 and 120, and their score increments or decrements were determined at the end of the trial in this manner: ACCV= CCV2 -CCV1 , where ACCV represents the increment or decrement (> 0, < 0) between the two evaluations at the beginning and end of the trial. We then determined, by multivariated linear regression analysis, that 3 of the 16 CCVs studied had the highest correlation coefficients with cytological dysplasia: PMNs (0.76), SM (0.554), and NMA (0.515). These parameters were independent variables (correlation coefficients SM:PMNs, 0.162; SM:NMA, 0.085; PMNs; NMA 0.01660). The distribution of individual subjects by cytological dysplasia, chemoprevention treatment, and decrements or increments in the three important variables that correlate with dysplasia is available upon request. The sum of the increments or decrements in PMNs, SM, and NMA was designated "expected dysplasia" (eD). The response rate of the four treated groups versus the placebo was evaluated in terms of cytological dysplasia at the end of the trial, and no statistically significant difference was found; data in Table 3  Most subjects, regardless of treatment, showed an increased severity of squamous metaplasia (p < 0.01), nasal mucosa atrophy (p = 0.012), mucosal hyperemia, a decrease in the numbers of ciliated respiratory-type cells, and a decrease in the amount of nasal mucus (p = 0.0073). There were no changes in epistaxis and bleeding friable mucosa. A statistically significant difference in nasal symptomatology was present in the group receiving vitamin A in regard to nasal dryness (p = 0.01372) and nasal obstruction (p = 0.02155) and in the group receiving f3carotene in regard to nasal dryness (p = 0.0072). On the other hand, the placebo group showed a significant worsening in nasal dryness and nasal obstruction (p < 0.01). No differences were seen in terms of dysplasia or CCVs between nonsmokers and smokers (p = 0.267). The Gill  GIV  GV  GVI  patients  Gl  11  7  7  10  0  2  37  G2  5  7  9  11  0  2  34  G3  10  9  7  4  1  0  31  G4  5  4  13  11  1  2  36  G5   10  8  8  10  1  2  39  Residence  >1 year  40  32  39  38  2  7  158  <1 year  1  3  5  8  1  1  19  Total no.  41  35  44  46  3  8  177  of patients aGI, vitamin E; G2, vitamin A; G3, vitamins A and E; G4, 1-carotene; G5, placebo. bGl, subjects without dysplasias throughout the study; Gil, subjects with dysplasias that did not change throughout the study; Gill, subjects who went from no dysplasia to dysplasia at the end of the study; GIV, subjects who started with dysplasia, but had no dysplasia at the end of the study; GV, subjects who went from a low degree of dysplasia to a higher degree of dysplasia; GVI, subjects who went from a high degree of dysplasia to a lower degree. ed dysplasia (0.51). Furthermore, only 1 of the 19 newly arrived subjects showed no dysplasia during the 4-month observation period; the remaining 18 individuals occupied a place in each of the cytological dysplasia groups (Table 3).

Theoretical Nasal Dysplasia Predictor Equation
Based on the observation of high correlation coefficients between three CCVs and cytological dysplasia, the following equation was applied as a model: uD (unadjusted dysplasia) = 0.76 APMNs + 0.55 ASM + 0.52 ANMA (1) where APMNs, ASM, and ANMA represent the changes (increments or decrements) in these variables at the end of the 4-month observation period. Subsequently, the equation was written as follows: tD (adjusted, theoretical dysplasia) = N APMNs + K2 ASM + K3 ANMA + K4 (2) where K1, K2 and K3 are adjusted constants based on the original correlation coefficients between the pertinent CCV data, and K4 is a proportionality constant; this leads to a proposed final form: tD = 0.85 APMNs + 0.50 ASM + 0.41ANMA + 0.98.
Correlation coefficients for SM and NMA are lower after adjustment, while the value increased for PMNs. Figure 5 shows the tD scores versus increments or decrements over time of squamous metaplasia (ASM) and polymorphonuclear leukocytes (APMNs). We then analyzed eD and tD in relation to cytological dysplasia and determined that the majority of subjects without dysplasia throughout the study and subjects who started with dysplasia but who had no dysplasia at the end of the study correspond to negative or low positive score ranges (eD, -2-0 and tD, -0.28-1.50). Subjects with dysplasia that did not change throughout the study are in the middle of the scale, and subjects who went from no dysplasia to dysplasia at the end of the study correspond to the highest score ranges (eD, 3,4 and tD, >2.51). Theoretical dysplasia 0 < cu -J _. CL aJ <06 0% I _ J scores in relation to cytological dysplasia are shown in Table 4. Note that there are a certain number of subjects who display high tD scores in spite of the absence of cytological dysplasia (the 13 subjects in GI with tD values between 1.51 and 2.50), and the GII individuals with dysplasia that did not change throughout the study had tD values of 2.51-3.50 (n = 9). The implications of these findings are complex. On one hand, this relationship between cytological dysplasia and low or high tD scores could indicate the predictive value of the equation; namely, subjects without cytological dysplasia after two evaluations but with high tD scores are at risk of developing cytological dysplasia in the future, and the subjects with a steady dysplasia perhaps will go into a higher severity of dysplasia if their tD values remain high over a period of time. This prediction must be confirmed by periodical cytological exams in the same subjects. On the other hand, we have to take into account the inherent flaw in cytological procedures; that is, the presence of false positives or negatives. Longterm follow-up of these individuals will clarify these issues.
It is interesting to point out here that the correlation coefficient between resi-   Table 3 for a description of dysplasia groups.
Higher scores are seen in subjects with persistent dysplasia (GIl) and for subjects who went from no dysplasia to dysplasia at the end of the study (Gill).
dence in months and tD scores in the group of 19 subjects with less than 1 year of residence in MMC was 0.66. This relationship holds true for all subjects who had lived less than 20 months in MMC; it did not hold true, however, for individuals with more than 10 years in MMC (data not shown). Further, 65% of subjects with high tD scores had lived less than 72 months in MMC, whereas 61% of subjects with low scores had lived more than 72 months in MMC. The 72-month division is crucial because it marks the abrupt change in diurnal ozone values in SWMMC ( Table 2).
In an attempt to correlate PMN values for the population studied with actual 03 values during the study period, we established two points: P2 (x = 128.8, y = 48), where the value on the abscissa corresponds to the average number of hours with 03 > 0.11 ppm over the 5-month study period (November 1990 through March 1991; see data in Table 1) and the value on the ordinate corresponds to the percentage of patients at the end of the study with APMNs > 0 (48%) and PI (x = 113.8, y = 42.42), where 113.8 corresponds to the average number of hours with 03 > 0.11 ppm during the 5 months before the study (June through October 1990), and 42.41 represents the percentage of subjects living in SWMMC assumed to have developed an increment in PMNs, based on the P2 value. The straight line PI-P2 can be prolonged, and we established a presumptive 03 value when 50% of the SWMMC population has an increment of PMNs in their nasal turbinates; this value corresponds theoretically to 130 hr in a month with ozone values above 0.11 ppm (a rather common situation in SWMMC).

Discussion
This study is the first double-blind, placebo-controlled trial with low oral doses of vitamins A, E, and f-carotene administered to individuals chronically exposed to photochemical smog, of which ozone is the most important constituent. Nasal dysplasia was the chosen endpoint of the study.
As reviewed by Bertram et al. (2), vitamin A and retinoids allow normal growth and epithelial differentiation and have well-demonstrated effects on the inhibition of experimental and human carcinogenesis at several sites, including the oral cavity (3)(4)(5). In a population of betel quid chewers in India, a 6-month treatment with oral vitamin A (200,000 IU/week) induced remission of established oral leukoplakias in 57% of patients, preventing the formation of new leukoplakias within the trial period (4). Beta-carotene also has chemopreventive and antitumor activity (4-6).
Both vitamin A and f-carotene are active in the promotional phase of carcinogenesis; there is a delay in tumor onset with a slowdown in the progression to frank neoplasia (2). Vitamin E, on the other hand, is a lipid-phase antioxidant that inhibits cellular damage produced by oxygen free-radicals and is also active in the promotional phase of carcinogenesis.
The results of this study show no effect on dysplasia with any of the vitamins or the fl-carotene-neither remission of established nasal dysplasias nor inhibition of the development of new ones. Further, we found a worsening of squamous metaplasia in all patients, regardless of treatment, with only a few improved clinical parameters in treated subjects. The fact that no positive results were seen might be result of the short trial time and the lack of interim endpoints; positive effects on chemoprevention results will have to be studied in a larger number of subjects over a longer period. More importantly, however, this study allowed us to establish a relationship between nasal dysplasia and clinico-cytological variables that show increments in relation to high levels of ozone in the polluted SWMMC atmosphere.
Ozone injury involves the development of oxygen free-radical events, which are responsible for DNA strand breaks and the formation of base adducts such as 8hydroxyguanine (9). Borek et al. (10), using in vitro transformation, reported that 03 induced neoplastic transformation in primary hamster embryo cells and mouse fibroblast cultures, with enhanced levels of free-radical-mediated lipid peroxidation products. Borek at al. (10) concluded that "ozone acts as a direct carcinogen and co-carcinogen on susceptible cells, therefore having important consequences for public health." The nasal cavity absorbs 40-70% of inhaled ozone in dogs (11) and is a major target for environmental pollutants. Exposure of rats to 0.8 ppm 03 for 3 or 7 days results in hyperplasia of the nasal cuboidal transitional epithelium; the rat cuboidal transitional epithelium is the epithelial cell type most responsive to the effects of 03 exposure (12). Johnson et al. (12) encountered a variable response to ozone depending on concentration, duration of exposure, and cell type; these authors suggest that there is an immediate proliferative response to ozone (within 3 days) which then subsides and is followed by the development of hyperplastic lesions. Similar changes are described in the transitional and respiratory nasal epithelia in the Bonnet monkey after 6 days of exposure to 0.15 ppm 03 (13). Further, three 6-hr/ day exposures to 0.8 ppm 03 in rats triggered hyperplastic and metaplastic changes in the nasal nonciliated cuboidal epithelium (14). As reported earlier, basal cell hyperplasia and squamous metaplasia are seen in human nasal biopsies of subjects exposed for more than 2 months to levels of 03 > 0.11 ppm (1). In the nasal biopsy study we observed in some cases an acute inflammatory intraepithelial infiltrate as well as marked proliferation of capillaries in or against basement membranes, and submucosal proliferation of thin vascular channels. Hotchkiss et al. (15 ) reported large numbers of pavementing neutrophils within nasal turbinate blood vessels in rats exposed to 0.8 ppm 03 immediately and 18 and 66 hr after exposure, although they did not mention increased numbers of submucosal blood vessels. The same authors demonstrated that a single 6-hr exposure to 03 (0.12 and 0.8 ppm) induces an inflammatory response within the rat nasal cavity; furthermore, the number of PMNs recovered by nasal and bronchoalveolar lavage (NAL and BAL), accurately reflected the inflammatory response in the nasal cavity and lungs injured by the acute ozone toxicity. Human exposures to 0.4 ppm 03 for 2 hr elicit a 7.7-fold increase in PMN in NAL (immediately postexposure), and this increase is detectable 18 hr later (6fold), both in NAL and BAL (16).
In the present study, the percentage of patients with PMNs in their nasal cavity, regardless of treatment, was 60% at day 0 and 81% at day 120, and 48% of subjects showed an increment in PMNs at the end of the 4-month observation period. These results are in agreement with previous observations both in rats and humans (14)(15)(16) (14), we have the appropriate conditions for the development of hyperplastic, metaplastic, and dysplastic nasal lesions. The high correlation coefficient between nasal dysplasia and PMNs is an interesting but not surprising finding in view of the previous suggestion that release of neutrophil products is an inductive mechanism for secretory cell hyperplasia and metaplasia in 03exposed macaque nasal epithelium (17) and the considerable evidence that liberation or generation of oxygen free radicals contribute to carcinogenesis (9,(18)(19)(20)(21)(22)(23)(24)(25)(26).
Dysplastic nasal mucosa lesions in SWMMC inhabitants may well be a manifestation of a process of adaptation to prolonged exposure to environmental atmospheric pollutants. Farber and Rubin (27) make several points that could be relevant to this study: carcinogenic agents produce heritable changes in a high proportion of treated cells at both the initiation and promotion stages, although its obvious manifestations (e.g., focus-forming capacity, benign tumors) occur with low probability. Transformation represents an adaptative response to growth constraints and occurs only under specific physiological conditions; the optimal condition for transformation in vivo being growth stimulation.
Given the evidence of the rapid development of nasal dysplasia upon arrival in SWMMC (1), the question arises how these pathological changes might be elicited and what role ozone plays. Farber and Rubin (22) state that "the persistent perturbation of the concentration or localization in cells of proteases and other hydrolytic enzymes might result in a transmissible defect of growth regulation." In this regard, phagocytic cell activation and the resultant host of mediator substances (including proteases, oxidants, neuropeptides, etc.) are both markers and effectors of toxin-induced respiratory tract injury (15,16,22).
We propose that repeated, chronic exposures to 03 and other toxic and potential carcinogenic atmospheric pollutants in SWMMC trigger a series of adaptative responses in the nasal mucosa, characterized histologically by severe loss of normal mucociliary epithelium, basal cell hyperplasia, squamous metaplasia, focal atrophy of the mucosa, and atrophy of submucosal glands [with a consequent decrease in mucus and antioxidant protection (28)], submucosal vascular proliferation and telangiectatic blood vessels [the result, perhaps, of the activation of type C sensory nerves and the release of neuropeptides (22)], activation of monocytes and macrophages in the submucosa, and mild to severe dysplastic changes, with nasal dysplasia highly correlated with the increment of PMNs in the nasal cavity.
Subjects newly arrived in SWMMC rapidly display these pathological changes, and in less than 1 year they can be included in cytological dysplasia groups. These groups, far from being static, undergo relatively rapid changes, some toward a higher degree of atypia (two severe dysplasias were diagnosed during the 4-month observation period), others toward an apparent reversibility of the phenotypic changes. Although the cytological dysplasia groups seem to be well characterized, the application of the theoretical dysplasia equation reveals that some apparently phenotypically "nonatypical" nasal mucosas are at high risk of developing atypical, dysplastic lesions at a later time.
The theoretical dysplasia equation has been developed to encourage a systematic clinical and cytological approach for analysis and presentation of risk estimates in nasal samples taken from a well-characterized SWMMC population exposed to photochemical smog. Long-term follow-up of these populations will be necessary to establish the real value of the equation and several other elements ought to be added: exposures, doses, and lifetime individual risk of various indoor and outdoor air pollutants of a carcinogenic and toxic nature (29).
The possible outcomes of the cytological nasal findings described in this population remain unclear at this point. The spectrum of cytological and histological abnormalities in the nasal turbinates varies from benign reactive changes to lesions characterized by nuclear abnormalities such as enlargement, hyperchromasia, and disturbances of mitotic activity. We have already encountered severe dysplasias and in one case an squamous cell carcinoma in situ; however, the incidence of nasal squamous cell carcinoma remains low, at least in the reported cases to oncological hospitals in Mexico City. We suspect that this incidence could increase in the next few years if the problem of urban pollution continues to be severe.
Finally, although ozone is the focus of our investigation, other toxic and carcinogenic substances in the SWMMC may play a role in the nasal mucosa changes described here. In particular, formaldehyde can produce proliferative reactions such as hyperplastic-metaplastic lesions in xenotransplanted human nasal respiratory epithelium (30) and squamous cell carcinomas in the rat nasal cavity (31). Further work is urgently needed to determine the effect. of environmental pollutants in SWMMC inhabitants. Well-characterized groups of subjects should be monitored and have the benefit of a close follow-up of nasal mucosa changes.