Table 4 reports the estimated regression coefficients on the indicator variable for living in the high pollution area. Even after controlling for age, sex, smoking status, pack-years smoked, and occupational exposure, statistically significant elevated scores of the histopathologic parameters were observed for those individuals who had lived in the high pollution areas. The associations for inflammation and wall thickness with high pollution exposure are quite strong. Not only are they highly statistically significant, but nearly identical associations were observed even when the other factors were not controlled for in the models. Hypersecretion and gland/wall ratio were also elevated for those from the high pollution area, but the associations were smaller and only marginally statistically significant. Similar effect estimates were obtained from both the OLS and Tobit regression models for all of the parameters except hypersecretion. Over half of the scores for hypersecretion equaled zero, suggesting the need for the Tobit model that accounts for left censoring at zero.
The largest association with pollution exposure was with anthracosis. However, because the difference in anthracosis scores between the high and low exposure groups was relatively small compared with variability within groups, this difference was only marginally statistically significant. While anthracosis may be treated as a histopathologic health end point, as a measure of deposition of carbon pigment, it may also be considered a bioindicator of exposure to fossil fuel combustion-related air pollution. In Figures 2 and 3, individual scores for inflammation and wall thickness are plotted over individual scores for anthracosis. Although substantial stochastic variability in the scores is evidenced by the wide scatter observed in these figures, positive associations for inflammation and wall thickness with anthracosis can clearly be observed. To more fully explore these associations, regression models for inflammation, wall thickness, hypersecretion, and gland/wall ratio were estimated with the mean scores for anthracosis included in the models replacing the indicator variable for high pollution air exposure. The estimated regression coefficients on the anthracosis mean scores are reported in Table 5. Highly statistically significant associations with anthracosis were observed for inflammation and wall thickness but not for hypersecretion and gland/wall ratio.
Table 4—Estimated Regression Coefficients (and SEs) on Indicator Variable for Living in High Pollution Area Controlling for Smoking, Age, Sex, and Occupational Exposure
|OLS Model||Tobit Model|
|Anthracosis||0.95 (0.51 )f||1.03 (0.50)s|
|Wall thickness||0.47 (0.16)’||0.49 (0.16)’|
|Hypersecretion||0.13 (0.07)’||0.31 (0.16)’|
|Gland/wall ratio||0.03 (0.02)’||0.03 (0.02)|
Table 5—Estimated Regression Coefficients (and SEs) on the Anthracosis Score, Controlling for Smoking, Age, Sex, and Occupational Exposure
|OLS Model||Tobit Model|
|Inflammation||0.18 (0.03)’||0.18 (0.03)’|
|Wall thickness||0.18 (0.03)f||0.18 (0.03)f|
|Hypersecretion||0.02 (0.02)||0.06 (0.04)|
|Gland/wall ratio||0.00 (0.00)||0.00 (0.00)|
Figure 2. Scatter plot for inflammation mean scores over anthracosis mean scores.
Figure 3. Scatter plot for wall thickness mean scores over anthracosis mean scores.
Blog invites submissions of review articles, reports on clinical techniques, case reports, conference summaries, and articles of opinion pertinent to the control of pain and anxiety in dentistry.