Marginal Structural Models for Periodontal Treatments effects on A1C in Diabetes

描述（由申请人提供）：患有牙周病的糖尿病（DM）患者的糖化血红蛋白水平较高，因此假设慢性牙周炎治疗可以降低糖尿病患者的糖化血红蛋白水平。一项临床试验的荟萃分析显示，牙周病治疗在三个月的随访中使糖化血红蛋白水平降低0.4%，然而，由于研究是异质性的，样本量小，治疗并不总是随机的，因此这些结果并不确定。此外，牙周治疗对糖化血红蛋白的长期影响尚不清楚。因此，牙周治疗目前不是控制糖尿病患者糖化血红蛋白的推荐策略。美国牙周病学会（AAP）关于有效治疗和控制慢性牙周炎的指南包括初始治疗，然后每3至6个月进行牙周维持检查。据我们所知，基于AAP指南治疗和控制牙周炎的长期策略尚未与糖尿病患者的血糖控制进行评估。我们建议在前瞻性研究设计中，使用退伍军人管理局计算机数据库的数据，评估AAP推荐的慢性牙周炎治疗对糖尿病患者A1C的长期效果（超过2年）。为了克服时间依赖性混杂（牙周维护访问决定后续牙周治疗）的限制，我们建议使用Robins开发的边际结构模型（MSM）。据我们所知，这种方法以前没有被用于评估口腔系统疾病的关系。因此，这些分析将加强现代统计方法在口腔健康研究中的应用。这些分析的结果将接近一项随机对照试验的结果，该试验评估了长期牙周护理对糖尿病患者糖化血红蛋白的总体影响。控制T2DM患者的糖化血红蛋白已被证明可以预防并发症，但只有44%的美国糖尿病患者的糖化血红蛋白在推荐水平内。在T2DM患者中，糖化血红蛋白降低1%与任何糖尿病相关终点降低21%相关。如果慢性牙周炎治疗能适度降低糖尿病患者的糖化血红蛋白，这将对糖尿病和牙周炎患者的治疗具有重要意义。

DESCRIPTION (provided by applicant): A1C is higher among individuals with diabetes mellitus (DM) when periodontal disease is present, and therefore treatment of chronic periodontitis is hypothesized to reduce A1C levels among individuals with DM. A meta-analysis of clinical trials showed that periodontal treatment reduced A1C by 0.4% at three months follow-up, however, these results are not definitive because the studies were heterogeneous, they had small sample sizes, and treatment was not always randomized. Moreover, the long term impact of periodontal treatment on A1C is unknown. Consequently periodontal treatment is currently not a recommended strategy to control A1C in DM. The American Academy of Periodontology (AAP) guidelines for effective treatment and control of chronic periodontitis consists of initial treatment followed by periodontal maintenance visits every 3 to 6 months. To the best of our knowledge the long-term strategy to treat and control periodontitis based on AAP guidelines has not been evaluated in relation to glycemic control among individuals with DM. We propose to evaluate the long-term effect (over 2 years) of AAP recommended treatment for chronic periodontitis on A1C among individuals with DM using data from the Veterans Administration computerized database in a prospective study design. To overcome the limitation of time-dependent confounding (periodontal maintenance visit determines subsequent periodontal treatment) we propose to use Marginal Structural Models (MSM) developed by Robins. This approach has, to our knowledge, not been previously used to assess oral-systemic disease relations. These analyses will therefore enhance the application of modern statistical methods to oral health research. The results from these analyses will approach those from a randomized controlled trial evaluating overall effect of long term periodontal care on A1C among individuals with DM. Controlling hemoglobin A1C in T2DM has been shown to prevent complications, yet just 44% of all Americans with DM have A1C within the recommended levels. A 1% reduction in A1C is associated with a 21% reduction for any diabetes-related endpoint among individuals with T2DM. If treatment of chronic periodontitis even modestly reduces A1C in individuals with DM, this will have important implications for the management of individuals with DM and periodontitis.