| Sign In to gain access to subscriptions and/or personal tools. |
TO THE EDITOR:1 Department of Oral Health Policy and Epidemiology Harvard School of Dental Medicine and Department of Epidemiology Harvard School of Puplic Health Boston, MA 02115 Lopez et al. (2002) prospectively examined the association between maternal periodontitis and adverse pregnancy outcomes among poor Chilean women. Out of 881 women screened, 263 had periodontitis and 618 had gingivitis or mild periodontitis. Those in the latter (comparison) group were offered treatment to make a sharper contrast; 159 refused treatment. There were two main weaknesses of this study that cast doubts on its validity. Bias In this study, the comparison group was treated and the exposed (periodontitis) group was not, violating the counterfactual principle (Rothman and Greenland, 1998). Moreover, 26% (159/618) of the women in the comparison group dropped out vs. none in the exposed group, causing selection bias. Selection bias could be in either direction, depending upon the characteristics of the women who dropped out. The biases could have been avoided by not offering any intervention and comparing women with periodontitis with those without. Alternatively, women with periodontitis consenting to treatment could have been randomized to treated or untreated groups. Either comparison would have been between groups of women who differed only by exposure, and therefore unbiased. Residual Confounding
Smoking during pregnancy adversely affects birth outcomes (CDC, 2001). It is not clear here whether lifetime smoking or smoking during pregnancy was evaluated. Including non-smokers and smokers of REFERENCES
THE AUTHORS REPLY:2 Department of Conservative Dentistry Section of Periodontics University of Chile School of Dentistry Comment One The number of women excluded from the analysis was 83. This is 11.5% of the 722 women enrolled in the study, and not 26% as Pitiphat and Merchant stated in their letter. Of the 83 women excluded, 53 were in the control group (11.5% of 459 women) and 30 were in the exposed group (11.2% of 269 women). The percentage of women excluded was very similar in both groups. Therefore, there was not selection bias as a result of exclusion. Drs. Pitiphat and Merchant’s suggestion to avoid the biases by not offering any intervention and comparing women with periodontal disease with those without periodontal disease was not possible in the population of women we studied, because all the women had some type of periodontal infection, either gingivitis or marginal periodontitis. Since all the women were exposed to periodontal infection, the counterfactual ideal was impossible to achieve, because there were no periodontally healthy women. The only way to obtain a reference cohort of individuals not exposed to periodontal infection was giving treatment to women with gingivitis or mild periodontitis. Furthermore, according to Rothman and Greenland (1998), in a cohort study, "the theoretically ideal comparison or reference cohort for a cohort of exposed individuals would be the experience of the same individuals had they not been exposed, and this ideal is sometimes achievable in a crossover study." It is not possible to use a crossover study to determine the association between preterm birth and periodontal disease. We also did a randomized controlled clinical trial to determine the association between periodontal disease and preterm low birth weight. The results of the randomized study were similar to those of the follow-up current study, and the manuscript was accepted to be published shortly. Comment Two A moderate dose-response relationship between smoking and birth weight has been found, with birth-weight impairments increasing with number of cigarettes smoked per day (Kramer, 1987; Berkowitz and Papernik, 1993). Another study (Cnattingius et al., 1999) found a modest increase in risk for a pretern delivery among women who smoked from 1 to 9 cigarettes per day (odds ratio 1.4) and those who smoked 10 or more cigarettes per day (odds ratio 1.6), as compared with non-smokers. Of the 639 women in our study, 162 smoked 1 or more cigarettes per day. Of these women, 111 smoked more than 5 cigarettes per day. Since the percentage of women who smoked fewer than 5 cigarettes a day was very low (N = 51, 7.9%), and considering that the risk of preterm delivery has been found to increase with an increasing number of cigarettes smoked per day, we decided to use a case definition of non-smoker as one who smoked fewer than 5 cigarettes a day. Even though, in the present study, the proportion of smokers was significantly higher in the group exposed to periodontal disease (21.4% vs. 15%, P = 0.038), smoking did not show association with preterm delivery or with low birth weight. Additionally, we recorded tobacco use at the time the women entered the study, but we could not verify with reasonable accuracy if smokers kept smoking during pregnancy. Prenatal care given to the women of our study involved a strong anti-smoking counseling. Furthermore, "women are more likely to stop smoking during pregnancy than at other times in their lives" (Centers for Disease Control and Prevention, 2001), and as a consequence of the anti-smoking warning, it is highly probable that a large percentage of women quit smoking during pregnancy. Therefore, the residual confounding smoking effect in our study is unlikely. REFERENCES
Journal of Dental Research, Vol. 81, No. 8,
516-517 (2002)
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
5 cigarettes/day in one group increased misclassification and consequent confounding. Since more women with periodontitis smoked than those in the comparison group (21.4% vs. 15.0%, p = 0.038), smoking should have been included in every model, whether or not it was statistically significant. Residual confounding from smoking would probably overestimate the results.