REVIEW ANALYSIS AND EVALUATION ARTICLE TITLE AND BIOGRAPHICAL INFORMATION Systematic review of implant outcomes in treated periodontitis subjects. Ong CTT, Ivanoski S, Needleman IG, Retzepi M, Moles DR, Tonetti MS, et al. J Clin Periodontol 2008;35:438-62.
REVIEWER Clark Stanford, DDS, PhD
PURPOSE/QUESTION To determine the effect of a past history of periodontitis on survival and success of dental implants in partially edentulous patients.
SOURCE OF FUNDING Self-funded/university institutional support.
Dental Implant Outcomes may Vary in Patients With a History of Periodontal Disease SUMMARY Selection Criteria This systematic review used a set of defined inclusion criteria and identified 4448 publications. These were sorted and 10 were identified with sufficient strength to be considered in the systematic review. The authors used a series of inclusion criteria to sort though the identified systematic reviews on this topic. Criterion included the following: use of longitudinal studies (up to March 2006) or case series in which all subjects were included, studies addressing only partially edentulous subjects in which a variety of implant systems were used, and reviews that classified and ranked quality of the studies (eg, > 6-month recall, > 10 subjects). All studies were located through a search of MEDLINE and EMBASE.
Key Study Factor The primary prognostic factor in this study was to identify dental implant outcomes (survival or success as defined by the respective studies) and reports of peri-implantitis in subjects with a previous history of periodontitis.
Main Outcome Measure TYPE OF STUDY/DESIGN Systematic review.
The authors focused on 3 outcome measures: implant survival, implant success, and peri-implantitis. The authors’ inclusion criteria were generally all inclusive and did not limit studies based on definitions of these 3 outcomes.
LEVEL OF EVIDENCE Level 2: Limited-quality patientoriented evidence
STRENGTH OF RECOMMENDATION GRADE Grade B: Inconsistent or limitedquality patient-oriented evidence
J Evid Base Dent Pract 2010;10:46-48 1532-3382/$36.00 Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.jebdp.2009.11.016
Main Results The authors state that ‘‘overall, the non-periodontitis patients demonstrated better outcomes than treated periodontitis patients,’’ yet they admit that there is a ‘‘medium to high level of bias’’ with a lack of consistent outcome variables. They correctly state that owing to the heterogeneity of study designs (p. 459), they were not able to provide a meta-analysis. Unfortunately, the strength of the associations claimed are difficult to assess because there were no reports of odds ratios or confidence intervals. They also report significant inconsistencies between studies for the key outcome variables. For instance, only 3 of the 9 studies reviewed reported any form of peri-implant bone loss associated with inflammation (peri-implantitis). Bone loss was statistically observed in only 1 study,1 with a nonsignificant trend of greater bone loss in the subjects who had a history of periodontitis in the other studies reviewed.
Conclusions The authors suggest there is ‘‘some evidence’’ that patients treated for periodontitis may experience more implant loss and complications around implants than patients without periodontitis.
JOURNAL OF EVIDENCE-BASED DENTAL PRACTICE
COMMENTARY AND ANALYSIS For the reader to understand this complex issue, the systematic review approach taken by the authors will be discussed regarding the studies’ strengths, weaknesses, opportunities, and threats of the outcomes and conclusions.
Strengths Implant outcomes are a complex combination of patient-, procedure-, and device-related risk factors. A fundamental question is, ‘‘Does a previous history of periodontal bone loss act as a prognostic factor for outcomes of implant therapy?’’ There is a long history of device-related attributes used by manufacturers to promote certain advantages for implant designs and material features of specific medical devices. These may or may not affect outcomes in specific patient populations. In regard to patient-specific risk factors, one aspect of the chronic cycle of inflammation is the impact on outcomes of implant therapy. Adult chronic periodontitis is a complex disease affecting the supporting ligament and osseous attachment of natural teeth to bone. The development of implants and their use in patients with a reported history of periodontitis has led to the assumption that disease pathways leading toward periodontal tooth loss also apply to integrated dental implants. This systematic review was designed to address this important issue and sought to assimilate the range of current clinical trials on this topic. To this end, the authors performed a comprehensive literature search of comparative trials using the definitions of periodontitis, success, and survival used by the individual authors. The methodology is complete and similar to other recent systematic reviews.2,3 Detailed efforts were made to evaluate the studies cited, and efforts were made to weight bias in the cited studies. This approach is very appropriate and brings great value to the reader to understand the process of systematic review. The authors also correctly discuss the confounder of clustering of implant failures and the restriction this creates on combining data at the implant level. If this is done in a systematic review, it is possible to obtain large sample sizes; however, one loses the level of significance (patient level), and because the risk of implant failure is related to the patient, statistical independence is not ensured, violating many conventional parametric tests. The question of patient-level risk in outcomes of care is very important for the dental profession to understand, and this article approaches this topic in an appropriate and systematic manner.
Weaknesses (or challenges to interpretation) The challenges for consistency in reporting, extraction of data, and making conclusions are immense in this systematic review. The initial challenge is managing a range of Volume 10, Number 1
study designs including 2 cohort studies, 4 case series with controls or a retrospective control sample, and 3 studies that used a subgroup analysis to compare treated and nontreated periodontally involved subjects. The next challenge involves risk factor evaluation for etiology of failures needed to separate causative etiology from coincident association. Various authors suggest that bacterial periodontal pathogens can translate to implant sites from adjacent natural teeth. Although it is logical to measure and classify the bacteria present around implants, this does not mean this surrogate measure will affect the bone-implant interface in the same manner as natural teeth. As outlined, there was no uniform definition of periodontitis used among studies (some defined disease as mild, moderate, or severe and grouped these into the same category), and 2 studies included aggressive periodontitis as well as chronic adult periodontitis, all of which complicates outcomes, as the severity of disease (eg, frequency of inflammatory episodes, periods of bone loss) may be associated with specific genetic risk factors.4 This can be important in clarifying the role of clustering of implants with bone loss or implant losses within a subject, yet this cannot be determined from the studies or this article, and these ‘‘outliers’’ could lead to spurious conclusions. For instance, the authors outlined that of the 9 studies reviewed, only 1 suggested a difference in implant success,5 and bone loss was reported in only 5 of the 9 studies, of which only 11 suggested a difference in bone levels on implants in patients with a history of periodontitis. Sample size is important to consider in any of these studies. For instance, the study by Karoussis et al5 followed only 8 subjects with a history of periodontitis as compared to 45 without. In a follow-up study by Mengel et al6 (which was included in a follow-up systematic review by Renvert and Persson7) 5 subjects with rapidly progressive periodontitis relative to 5 periodontally healthy subjects were followed over a 10-year period and showed no difference in implant loss but suggested a greater range of bone loss in the periodontitis group. This outlines an important issue in the literature—that the comparison studies only report on a small number of subjects in which clustering of complications strongly influences the outcomes and conclusions. Second, reviews are often made with observations of risk but do not quantitatively report on the significance of the risk (eg, relative risk or odds ratios and 95% confidence intervals), making interpretation of the degree of risk difficult. This creates a threat to the validity of conclusions based on a lack of consistent observations between studies. This systematic review is also adversely impacted by aggregating studies for review that use inconsistent or poorly defined definitions, such as treated periodontitis, which could include scaling and root planing alone, chemotherapeutics, recall maintenance intervals, and so forth. In addition, there was no common definition for peri-implantitis used by the authors in the cited papers, creating another level of 47
JOURNAL OF EVIDENCE-BASED DENTAL PRACTICE
ambiguity. Given this, it is interesting to note that only 3 of the 9 studies reported incidences of peri-implantitis. Admitting that the heterogeneity of the studies did not allow for a systematic meta-analytic analysis, the authors proceeded with a more conventional narrative review of the cited studies. The problem with this approach is that the conclusions of individual studies can become the implied overriding conclusion of the review. A more minor issue is that the implant systems represented in the 9 cited studies are primarily off the market and consisted of either the turned surface Branemark system or the rough titanium plasma–sprayed (TPS) or hydroxyapatite (HA)-plasma spray-coated implant surfaces. The later surfaces have higher levels of bone loss in general when exposed to the oral environment.8
with different study designs add significance to the generalizability of the results. The challenge in this case is to understand the degree of risk of implant complications in patients with a history of periodontitis. Although it is very reasonable to discuss with specific patients who have been treated for periodontitis that they may be at higher risk for complications with dental implants, this may be because of the same genetic and environmental risk factors leading to an elevated risk for periodontitis, but that the risk of periodontitis by itself may be unrelated to the outcomes of implant-based tooth replacement therapy. It is an issue of cause and effect, and at this time the scientific literature does not support this. Without a quantization of risk, we do need to be careful in how we relate this information to patients by clearly articulating the limits of our knowledge to them.
Opportunity One of the key opportunities this systematic review provides is to outline the need for systematic reporting of data in publications. The conclusion section outlines the specific limitations of each issue raised (eg, lack of consecutive enrollment) and then summarizes why this is a threat to validity. The conclusions also outline that in an area of uncertainty, the potential risk should be explained to patients. The next issue, which is ongoing ‘‘supportive care,’’ is more unclear. The implied suggestion is that implants may need a different recall schedule or management by a periodontal specialist (a political and economic issue), which needs to be carefully evaluated in ongoing studies. As an opportunity, understanding chronic disease processes, which by definition have long-term, cyclic patterns often related to immune changes, must be appreciated to understand clinical risk factors in discussing implant therapy (or any therapy for that matter) with patients. The role of immunity and proinflammatory mediators is critical to many clinical issues, and implant therapy is but one of these. The opportunity allowed by this study is to outline the limitations of our knowledge on this topic, address areas of research, and propose field trial–appropriate definitions of operational parameters that can be used across the world.
Threats The value of systematic reviews is to sort though and identify what is supported across different clinical trials. Outcomes that are identified in different populations
48
REFERENCES 1. Hardt CR, Grondahl K, Lekholm U, Wennstrom JL. Outcomes of implant therapy in relation to experienced loss of periodontal bone support: a retrospective 5 year study. Clin Oral Impl Res 2002;13:488-94. 2. Karoussis IK, Kotsovilis S, Fourmousis I. A comprehensive and critical review of dental implant prognosis in periodontally compromised partially edentulous patients. Clin Oral Impl Res 2007;18:669-79. 3. Quirynen M, Abarca M, Van Assche M, Nevins M, van Steenbeghe D. Impact of supportive periodontal therapy and implant surface roughness on implant outcomes in patients with a history of periodontitis. J Clin Periodontol 2007;34:805-15. 4. Heitz-Mayfield LJA. Peri-implant disease: diagnosis and risk indicators. J Clin Periodontol 2008;35(Suppl 8):292-304. 5. Karoussis IK, Salvi GE, Heitz-Mayfield LJ, Bragger U, Hammerle CH, Lang NP. Long-term implant prognosis in patients with and without a history of chronic periodontitis: a 10-year prospective cohort study of the ITI Dental Implant System. Clin Oral Impl Res 2003;14:329-39. 6. Mengel R, Behle M, Flores-de-Jacoby L. Osseointegrated implants in subjects treated for generalized aggressive periodontitis: 10 year results of a prospective long-term cohort study. J Clin Periodontol 2007;78:2229-37. 7. Renvert S, Persson GR. Periodontists as a potential risk factor for peri-implantitis. J Clin Periodontol 2009;36(Suppl 10):9-14. 8. Stanford CM, Schneider GB. Functional behaviour of bone around dental implants. Gerodontology 2004;21(2):71-7.
REVIEWER Clark Stanford, DDS, PhD N419 Dental Science Building North, University of Iowa, Iowa City, Iowa 52242, (319) 335-7381
[email protected]
March 2010