This letter is in regard to the November JADA study by Dr. Greg Huang and Ms. Tessa Rue, "Third-Molar Extraction As a Risk Factor for Temporomandibular Disorder." While we would like to commend the authors for addressing an important problem in dentistry, we have some concerns about the conclusions put forth in the article.
Our first concern is in regard to the estimate of the risk of developing TMD following third-molar (M3) extraction. The authors report that patients who have their M3s removed are 60 percent more likely to experience TMD than those who don’t have their M3s removed, after controlling for sex and dental care utilization.
The overall frequency of TMD in the sample was 1.1 percent. Using the adjusted relative risk, this translates into adjusted frequencies of 0.4 percent and 0.7 percent in the non-M3 extraction and M3-extraction groups, respectively, a risk increase of 0.3 percent. This risk increase is of the same magnitude as those reported for inferior alveolar and lingual nerve injuries, which are known complications for which the benefit of M3 removal exceeds the risk.
In addition to the discordance between the authors’ conclusions and the clinically relevant magnitude of the reported effect, there is the question of whether the choice of analysis was appropriate. Generally, rigorous survival analysis is appropriate for "time to event" cohort data. One might argue against using a survival analysis in this study, as the "time to event" is not necessarily something akin to the development of a disease that has a latent period (for example, cancer).
The mechanism proposed for relating TMD to M3 removal may be trauma associated with the extraction or maintaining an open mouth position for the duration of the procedure. It is difficult to make a convincing argument that the study population is continually at risk of developing TMD at time points significantly after the extraction. If the proposed mechanism is injury associated with the extraction, the risk for developing TMD should be inversely related to duration of time after extraction. For example, patients are more likely to develop TMD within a month after the extraction, rather than 24 months after the extraction.
Another dimension to consider is, during the "at risk" period, some of the subjects had other dental procedures completed, which could have caused TMD in a similar manner as proposed for M3 extraction, or possibly exacerbated a condition that initially was caused by the M3 extraction to the extent that treatment was subsequently required. While the authors controlled for "other dental procedures" in their analyses, this may not adequately adjust for procedures associated with extended operating time. Thus, perhaps the exposure should be "duration of mouth opening for a procedure," rather than M3 extraction.
If the onset of TMJ symptoms is related to prolonged mouth opening rather than the trauma of extraction itself, then, generally speaking, any evaluation between M3 extraction and TMD will be overestimated without controlling for the true intermediate variable "prolonged mouth opening."
Finally, there is the issue of selection bias. The authors reported only 13 percent of the subjects met the inclusion criteria for the study. Given this small proportion of patients who were included, nothing can be said about the remaining 87 percent. Though the sample size is large, the large number of patients lost to follow-up weakens the conclusions of the study.
