I have written about AAP’s centennial and the series of papers celebrating the “literature that shaped modern Periodontology” in several posts, see here and here. In this month’s contribution (in the Academy’s Journal of Periodontology), Steven Offenbacher and James Beck attracted my attention when writing about “Changing Paradigms in the Oral Disease–Systemic Disease Relationship.” Twenty-five years after the first report by Finnish researchers on a certain risk of poor oral health for cardiovascular disease, the incredibe surge of studies all over the world (“Floss or Die!”) may have ceased, paving the way for another overdue paradigm shift.
The paper disappoints. Instead of putting the initial and long-lasting exciting in at least some relation in view of results of recent very large intervention studies which were by and large not able to confirm a postulated benefica effect of treatment of periodontal disease on cardiovascular disease, low birth weight or diabetes, Offenbacher and Beck opened the bottom drwaer and pulled a couple of now questionable studies which, well, misled the public and thousands of researcher worldwide alike. For instance, Frank De Stefano’s paper of 1993 who, based on NHANES I follow-up data, reported a 25% increase of risk in patients with periodontitis. The data had been re-analyzed by Phillipe Hujoel in 2000 including much more careful adjustment of cofounders. They could not find an association. I have reported about the apalling letter to the editors of JAMA by Robert Genco and colleagues here.
And then, of course, the paper by Beck et al. (1996) from the VA Longitudinal Study.
“Our paper showed that periodontal disease was a significant independent risk factor for CVD events after adjusting for traditional risk factors and displayed a dose-dependent increase in risk with increasing periodontal disease severity. Importantly, this paper described a two-component mechanistic working model that linked periodontal disease to cardiovascular disease via systemic bacterial dissemination interacting with the vasculature and activation of the hepatic acute phase response as an inflammatory trigger for CVD. We also hypothesized that there may be an underlying hyperinflammatory trait that served to increase risk for both conditions, suggesting that bacterial exposure in those susceptible individuals would result in even more risk for CVD. In addition, by placing importance on inflammation, it explained how oral infection could influence multiple conditions such as diabetes. Over the next 10 years, many studies supported this basic model. The model is now more than 15 years old, and we would modify it only by adding specific details, such as the hyperinflammatory trait likely being attributable to genetic differences in the innate immune response.” (Emphasis added.)
In Robert Genco’s letter to the editor when questioning conclusions by Hujoel et al. (2000) [pdf], he referred to a “strong association” of periodontitis and cardiovascular disease found by Beck et al. (1996). Well, the adjusted odds ratio describing the association between total coronary heart disease (CHD) and periodontitis was just 1.5. What is more concerning is that this particular model was not adjusted for smoking. In another model for fatal CHD, adjusting for smoking led to considerable attenuation of the odds ratio. So, is it possible that authors selectively reported whenever there was “an association” but concealed when there was no? That they adjusted for smoking only as long as there was still an “association” left? If that was so, it must be considered ignoble. (Note that Bob Genco was “awarded” the 1996 IgNobel Prize for economics for reporting on an association between periodontitis and financial strain.)
At the end of their Centennial paper, Offenbacher and Beck won’t change their strong opinion in case of premature delivery either.
“As compared to cardiovascular disease studies [intervention studies on low birth weight], which require decades of follow-up to reach endpoints, randomized controlled trials have been conducted to examine the potential benefit of limited periodontal therapy on pregnancy outcomes. These studies have failed to show any reduction in pregnancy complications, a finding that is similar to what has occurred many times in clinical intervention trials in obstetrics. For example, bacterial vaginosis (BV) is a well-established risk factor for prematurity, but therapeutic trials have failed to show benefit. In the obstetric world, the exposure of interest (BV) is still considered to be a cause or risk factor—even if not readily modifiable—and research has continued. However, in dental research the field is now facing a growing opinion that failure to achieve treatment success with scaling and root planing is proof of noncausality. If the treatments had been successful in reducing the rate of prematurity, for example, the evidence for causality would have undeniably been strengthened, but the absence of a positive benefit from treatment does not prove a null effect of the exposure. […] This minority opinion of non-causality is perhaps a convenient concession to the fact that we do not as yet know how to best manage oral infections to prevent systemic sequelae.”
I do not want to quote out of context here. Offenbacher and Beck continue,
“Perhaps a clue to what may lie ahead is that we have shown in pregnant women that maternal periodontal disease is a risk for prematurity only when there is fetal exposure to maternal oral organisms, which can occur only if the mother is seronegative to the periodontal pathogens early in pregnancy. If the mother has a high titer to specific pathogens early in pregnancy, the fetus does not get exposed and is protected, irrespective of the maternal periodontal status. Just like the littermates in the animal experiments, placental exposure clearly is important, not changes in probing depth. Hopefully, just as we recognize that tooth extraction was not the ideal way to manage oral infections back in the time of the focal infection theory, we will exercise more insight this time and continue to delineate the underlying mechanisms and identify the appropriate treatments and endpoints before we rush to therapeutic trials. We may know how to manage the oral infection to prevent oral problems, but we are only beginning to understand how to manage oral infection to prevent systemic complications.”
This will, of course go on. As fact of the matter, periodontists (scientists) have not solved the most urgent questions of the profession (improving oral health by smoking cessation and in low socio-economic classes, development of affordable and reliable therapies), but outreach to pregancy outcomes and cardiovascular event prevention despite lack of any evidence. Future generations of dentists may describe the current era of periodontal medicine as interesting but, well, misleading and more or less useless.
30 May 2014 @ 9:00 am.
Last modified June 24, 2014.