| Sign In to gain access to subscriptions and/or personal tools. |
TO THE EDITOR:1 47-140 Heno Place Kaneohe, HI 96744rethman{at}hotmail.com In light of Hujoel et al.'s ("Pre-existing Cardiovascular Disease and Periodontitis: A Follow-up Study", J Dent Res 81:186-191, 2002) comparison of means (bounded by large and overlapping confidence intervals), a report that the presence of periodontitis apparently cuts the risk of recurrent coronary heart disease (CHD) events by 34% should come as no surprise. Such post hoc "statistical shopping" can often produce bizarre, albeit compelling, results. Indeed, several putative relationships highlighted in this paper are similarly odd. In the midst of trying to explain the many peculiar inferences they distilled from their analyses, the authors gamely admitted that the methods used to gather the NHANES I data may have been insensitive (high false-negative rate) to the true incidence of periodontal infections in the patient sample. However, the authors continued on, rhetorically, to pose a challenge to the reader, namely, How could the potentially insensitive but (what they suspect to have been) highly specific Russell Index "bias the relative risk from increased risk (of CHD events putatively the result of periodontal infections) to the protective relationships witnessed in this study"? Well, here are a couple of stabs:
What scientists and clinicians ought to recognize is that it appears that if any causal link truly exists between periodontitis and CHD events, the work of many investigators, including Dr. Hujoel and his collaborators, has repeatedly suggested that the presence of periodontitis will not have the same predictive capability as does the relationship between a failed parachute and a skydiver's subsequent health. Moreover, some JDR readers need reminding that there is ample clinical and basic science evidence suggesting that periodontitis MAY play a role in CHD events. But, most important, what readers need to realize is that until well-run intervention studies' results are reported, key questions will remain unanswered. And, in the context of this last observation, scarce NIDCR research dollars spent on additional re-hashes of the putative relationship of periodontitis to CHD events using the NHANES I database are ill-advised. Instead, research emphasis now needs to shift to the clinic. Fortunately, a multi-center randomized clinical trial has recently been funded. Until the results of this or other adequately powered and well-run clinical studies are reported, the minimal risks of what some might consider "periodontal overtreatment" (at least in terms of systemic benefits that may accrue) amounts to sound public policy. Or perhaps I miss the point. Maybe we should all start investigating methods to induce periodontitis in heart patients...as Dr. Hujoel et al.'s analyses also suggest.
THE AUTHORS REPLY:2 Department of Dental Public Health Sciences School of Dentistry Box 357475 University of Washington Seattle, WA 98195 hujoel{at}u.washington.edu We agree that the protective effect of periodontitis on secondary coronary heart disease was statistically non-significant and that the tendency toward a protective effect may have been a spurious finding. These points were the primary conclusions of our study and are consistent with the current overall evidence that periodontitis is either not or weakly related to coronary heart disease (Danesh, 1999; Joshipura et al., 2000). The suggestion that NHANES I and the Russell Index are systematically biased needs to be substantiated with evidence. Simply because findings disagree with beliefs is insufficient evidence. In fact, the oddball coincidences cited to suggest that "there is something amiss with the data" more likely suggest that there is something wrong with the hypothesis. The Russell Index has documented diagnostic properties (using radiographic bone loss as the gold standard) (Russell, 1956; Sheiham and Striffler, 1970), and it is not a data-generated criterion devised to fit prior beliefs of the periodontitis-CHD association (Mills, 1993). The purported "good" periodontitis measures of some of the positive studies should become similarly validated against some accepted gold standard of periodontitis. More importantly, periodontitis-systemic disease associations based on data-generated indices should be verified in a dataset other than the one that was used to create the index. Until then, criticisms of negative "Russell Index-based" studies appear self-serving, especially when some of the critics themselves embrace the Russell Index when positive associations can be identified. The suggestion that existing data not be analyzed is puzzling. The scientific value of existing databases is widely recognized, and NIH is planning to implement policies by January, 2003, that will further encourage data sharing ("Scope", 2002). The NHANES I Epidemiologic Follow-up Study (NHEFS) reflects an over 20-year research effort funded by 12 National Institutes, included comprehensive dental examinations, and has detailed information on a variety of systemic disease outcomes. The study represents one of the few available opportunities to investigate periodontitis-systemic disease associations. Indeed, the data from NHEFS, based on the Russell Index, have served as the basis for some of the claims made concerning periodontitis-systemic disease relationships (DeStefano et al., 1993; Wu et al., 1999, 2000). The key is whether those claims can withstand the scrutiny of additional analyses. The cost of such re-analyses is miniscule when compared with the cost of initiating a randomized controlled trial. It appears only logical that, to justify the investment in a randomized controlled trial, all available evidence should first be analyzed and synthesized. Finally, recommending intensive periodontal treatments, such as the repeated and routine use of antibiotics, with the aim of achieving hypothesized systemic benefits is not a sound public health policy. Public health policy should be based on "evidence, not hope" (Guyatt and Rennie, 1993). The history of medicine is filled with interventions that appeared beneficial based on basic science, yet turned out to be harmful. Even an intervention as benign as vitamin supplementation turned out to increase mortality risk (The Alpha-Tocopherol Beta Carotene Cancer Prevention Study Group, 1994; Omenn et al., 1996). Therefore, until it is shown that the hypothesized systemic benefits of intensive periodontal therapies outweigh harms such as antibiotic resistance (Vincent, 2000), the prevention of systemic diseases should not be used as the basis for recommending periodontal treatments. REFERENCES
Journal of Dental Research, Vol. 81, No. 6,
372-374 (2002) Related articles in Journal of Dental Research:
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
