With interest, I read the article of Ö zeç et al. regarding the second molar distal caries and the relation with impacted wisdom teeth. Because of the controversies in the literature on this subject, well performed scientific research can give us valuable information. However, in this paper, some serious methodological inaccuracies are present which make the conclusions unreliable. The main shortcomings are listed below.
First, the aim of the study is to describe the prevalence of the affliction in a Turkish population. It is, however, not clear which population is meant by this statement. They indicated that the 485 patients evaluated in the study came from a clinic. What type of clinic was this? Further information on this group is missing. For example, were the patients referred to this clinic for removal of third molars or were they referred for treatment of caries or is this a primary dental health centre? It is very likely that there was a considerable selection bias in the patients included in this study, and therefore, the selected group is no true representation of the “Turkish” population.
Second, the authors describe a significant relation between age and the second molar distal caries. However, the statistics that were used to demonstrate this relation are dubious. Age is a continuous variable that should be analyzed (at least) by a t -test, to divide age in different categories and use a χ 2 -test is very questionable. Moreover, as the median is 25.2 years (according to the authors), the first age category (18–23) includes almost half the patients. This means that in each of the age category a different number of patients is represented, making this division in categories artificial and unreliable. Furthermore, age is almost certainly not equally distributed in the research population which makes the statistics even more undependable.
Third, this is not a randomized clinical trial and, therefore, the authors should take a considerable confounding variable into account. For example, if the group of 35 years and older includes only molars with an angulation of more than 31°, the significant difference in caries is not related to age but to angulation or vice versa. This should be statistically corrected (if possible) by multivariate regression techniques. There is most certainly a considerable number of confounders such as oral hygiene, diabetes, and dental care. The authors should have recognized this problem.
Fourth, there is no description of the controls. Therefore, we do not know anything about this group. For example, elements with an angulation of 31–70° showed caries in the second molar in 47% of cases. But how many cases were present in this group? As such, the percentage does not tell us anything, especially not if the absolute number is low. In every scientific paper, patients and controls should be properly reported.
In conclusion, although the subject of the research is most interesting, the outcome of this study is not reliable.