Int J Oral Maxillofac Implants. 2021 Mar-Apr;36(2):388-394. doi: 10.11607/jomi.8983.
The objective of this study was to quantify the probability of implant failure over time from intrinsic patient factors using a population of patients from multiple private practices.
The records for this retrospective, multicenter cohort study were randomly selected from eight private practices. The primary outcome variable was time to event (implant failure or last known follow-up). The included independent variables were age, sex, diabetes status, smoking status, and arch location. Analyses were performed with Cox proportional hazards on three models: univariate, full multivariate, and systemic factor multivariate. The probability of implant survival at 1, 5, and 10 years was calculated using univariate time-to-event modeling on log-normal distribution with 95% CIs and Cox proportional hazard tests for significance. The Kaplan-Meier survival curve was calculated for patients < 71 years of age.
Eight hundred thirty-five implant-level records from 378 patients were collected for analyses. The mean patient age was 60 years, and 48% were men, 15% reported a history of smoking, and 16% reported having diabetes. The follow-up time was as long as 17 years, with a mean of 23.1 months, and a median of 7 months. The hazard ratio (HR) for implant failure due to sex (HR = 1.18; 95% CI: 0.52 to 2.66), smoking (HR = 1.30; 95% CI: 0.49 to 3.46), diabetes (HR = 1.17; 95% CI: 0.35 to 3.86), and arch location (HR = 2.13 to 3.39) failed to reach the threshold within any Cox proportional hazards model (P > .05).
The implant survival probability for patients ≤ 70 years of age is 86.4% at 10 years in the course of routine private practice by experienced providers. Cautious interpretation of these results is critical, as the effects of known systemic risk factors are likely tempered by effective modifications in clinical decisions and protocols with short- and long-term follow-up maintenance.
本研究旨在通过来自多个私人诊所的患者群体，从内在患者因素方面定量分析随时间推移种植体失败的概率。
本回顾性、多中心队列研究的记录是从 8 家私人诊所中随机选择的。主要结局变量是时间事件（种植体失败或最后一次已知随访）。纳入的独立变量包括年龄、性别、糖尿病状态、吸烟状况和牙弓位置。使用 Cox 比例风险分析在三个模型中进行分析：单变量、全多变量和系统性因素多变量。使用对数正态分布的单变量时间事件建模计算 1、5 和 10 年的种植体生存率，并进行 Cox 比例风险检验以确定显著性。对于<71 岁的患者，计算 Kaplan-Meier 生存曲线。
共收集了 378 名患者的 835 个种植体水平记录进行分析。患者平均年龄为 60 岁，48%为男性，15%有吸烟史，16%有糖尿病史。随访时间最长达 17 年，平均随访时间为 23.1 个月，中位数为 7 个月。由于性别（HR=1.18；95%CI：0.52 至 2.66）、吸烟（HR=1.30；95%CI：0.49 至 3.46）、糖尿病（HR=1.17；95%CI：0.35 至 3.86）和牙弓位置（HR=2.13 至 3.39）的种植体失败风险比在任何 Cox 比例风险模型中都未达到阈值（P>.05）。
在经验丰富的医生常规私人执业过程中，≤70 岁患者的种植体 10 年生存率为 86.4%。谨慎解释这些结果至关重要，因为在短期和长期随访维持下，临床决策和方案的有效调整可能会缓和已知系统性危险因素的影响。
