Dental implants in irradiated patients: which factors influence implant survival?

Excerpt

We thank Korfage et al. very much for their review and complemented our article with an Erratum to consider the addressed items. However, in our opinion, these items do not influence the quality of our study as they are descriptive values. Concerning the jaw region, we wrote in the discussion: “Although a meta-analytic approach to examine the influence of the jaw region on implant survival was not feasible, some studies indicated a lower implant survival in the irradiated maxilla compared to the irradiated mandible [11–13].” Concerning time of implant placement, in our study is written: “The timing of implant placement in relation to the end of radiation therapy is another point that may contribute to the success or failure of osseointegration. In most of the studies investigated in this review, implants were inserted after radiation. There is no scientific evidence for the optimal timing of secondary implant placement, but the literature cumulatively supports implant placement between 6 and 12 months after irradiation [26, 28, 29, 32].” Therefore, errors in Table 1 had no influence on given conclusions. Concerning the six eligible studies published in 2007–2013, the following statements were made in these studies. Salinas et al. 2010 (15): “The cumulative survival rate was 91.9 %, and there was no difference in survival between implants placed in the fibula versus the native mandible or depending on whether the patient received radiation therapy.” Korfage et al. 2010 (16): “During the 5-year follow-up, total 14 implants were lost, 13 in irradiated bone (survival rate 89.4 %, dose 440 Gy) and one in non-irradiated bone (survival rate 98.6 %).” Klein et al. 2009 (17): “Univariate analysis revealed no significant impact of radiation therapy on implant survival.” Schoen et al. 2008 (18): “During the healing period prior to the abutment connection operation, two implants (2 patients) were lost in non-irradiated mandibular bone (implant survival rate 97 % in the 16 patients with functioning prostheses, n = 64). After loading no implants were lost. In the irradiated mandibular bone, two implants (1 patient) were lost after abutment connection but before fabrication of the prosthesis (implant survival rate of 97 % in the 19 patients with functioning prostheses, n = 76).” Linsen et al. 2012 (21): “There was no significantly lower implant survival for implants inserted into irradiated bone (p = .302).” Mancha de la Plata et al. 2012 (25): “Irradiated patients had a marginally significantly higher implant loss than non-irradiated patients (p = 0.063).” As the p value was above 0.05, this is a non-significant difference. Concerning the study of Linsen et al., the mean follow-up was 42 months. As data on the 60 months survival was given in this study and as long-term follow-up was up to 120 months, we included this study in the meta-analysis. If we exclude the study of Linsen et al., the meta-analysis will have the same result ([OR], 1.25; [CI], 0.56–2.81) and the same conclusion. We totally agree that study design can have a substantial impact on effect sizes of a meta-analysis. However, a comprehensive discussion of the strengths and limitations of all included studies is beyond the scope of our article. Therefore, we wrote in our study: “Generally, quality and level of evidence of the examined studies were low. Nearly all of the studies were retrospective analyses. Allocation concealment at high risk of bias, missing blind examiners to evaluate clinical outcomes, absence of reporting characteristics of dropouts, and lack of CONSORT adherence suggest to be careful with data interpretation and drawing general conclusions out of these studies.” In our opinion, this statement describes the problem of quality bias quite well, and we explicitly state that drawing definite conclusions out of our meta-analysis is not advisable. We performed this meta-analysis to the best of our knowledge. It is obvious that there are structural and methodological flaws in the quality of the included studies concerning this topic. However, we stated this risk of bias in our study. In our opinion, every meta-analysis has this risk of bias. We are very sorry for some descriptive flaws in our tables. However, we are in total agreement in our conclusions with already published systematic reviews [9–13] on this topic, and the addressed items have no influence on the conclusions of our study. Therefore, interpreting the literature on implants in irradiated bone, we still would draw the same conclusions. …