Review correlations of in vivo clinical performance with in vitro laboratory tests of restorative dental materials involving polymer-based matrices. Identify those factors interfering with the process.
Materials and methods
An evidence-based dentistry approach was used to identify clinical trials, critical reviews, and meta-analyses involving correlations. Factors impacting meaningful correlations were reviewed. The limited bona fide correlations were reviewed.
In vitro tests include physical, chemical, mechanical, and biological properties. Clinical research measurements routinely include 10–15 categories of clinical observations of performance such as color match, caries resistance, marginal integrity, surface texture, and others, but do not correspond well with laboratory properties. Clinical trials of restorative dental materials represent a small fraction of the total research in this arena (typically <10% of dental materials research over many years). Trials are generally short-term (2–5 years) and are designed primarily to test product “safety and efficacy.” A large number of risk factors (operator, design, material, intraoral location, patient) affect clinical outcomes and are not simulated well in laboratories. Little long-term information exists for clinical performance other than on composite wear. Very few meaningful correlations of laboratory tests and clinical results are demonstrated.
New studies should be focused on recovering restorations from service and characterizing them with the same tests as normally conducted in the laboratory. Many more long-term clinical trials that involve 10–20 years of observations are needed. Those trials should include planned restoration recovery to assess changes in laboratory properties of interest.
Correlation of key clinical performance behaviors with laboratory tests would allow much more rapid development of new materials. There is a keen interest in developing long-lasting resin-based materials for direct restorations to replace dental amalgam. However, there has been much less investment in clinical research to develop potential correlations.
Recent discussions of challenges for correlation of laboratory and clinical properties of dental materials have been reported . The conclusions generally are that there have been no strong correlations between a single laboratory property and cluster of properties with either short-term (2–5 years) or long-term (10–20 years) clinical performance of restorative materials. Most clinical trials report on one particular product. Collecting the clinical trials into a useful summary is difficult because of the scarcity of them.
There are key reasons why a correlation might not yet exist. Laboratory properties tend to be “static tests” of single properties representing physical, chemical, mechanical, or biological properties. Laboratory tests generally report bulk properties, even though surface properties can be evaluated. In contrast, clinical properties involve “observations” of the surfaces or margins or restorations. Information about pulpal changes and/or absence or presence of occlusal or interproximal contacts might also be included. Restorative materials change with time. Laboratory specimens in storage also change, but do not necessarily reflect the same service conditions or the same time scale.
During the past two decades, laboratory testing has tried to bridge the gap by simulating some of the major clinical conditions. However, the simulation protocols are crude and tested correlations are heuristic and not mechanistic ones. Many reports from clinical practice on types and patterns of failure, such as caries as the dominant reason for restoration failure, are not borne out by careful evidence collection from clinical practices or from carefully controlled trials. These problems confound the directions of development for simulations.
Map of laboratory/clinical correlations
Many different types of correlations are possible when considering laboratory and clinical information about dental materials. Fig. 1 presents a map for the properties, experimental variables, and potential correlations. At the left-hand side are all possible laboratory tests. Most are single-values tests (physical, chemical, mechanical, or biological). However, these may be combined sequentially (circles connected by a line) or simultaneously (circles within a circle) to create a simulation for clinic conditions. On the right-hand side of this figure are those circumstances generating clinical information. Most clinical trials of restorations or appliances are evaluated as direct intra-oral observations of events associated or described by USPHS categories (e.g., wear, caries) that will be expanded later. These observations may also be combined by clustering or weighting because of presumed relationships (e.g., marginal integrity, marginal staining, occurrence of dental caries). Indirect evaluations may be collected from intraoral events by using replica techniques to create laboratory casts for further rating, digital scanning, or microscopic evaluations (light, SEM) to collect more data. Finally, some tests of materials intraorally are conducted in removable appliances so that they can be retrieved for periodic examination.
Consider the laboratory tests (single categories or simulations) and clinical observations (direct USPHS, clustered or weighted groups, indirect data, or pseudo-clinical tests) in terms of the number of correlations that are possible. Correlations (or relationships, R ) are possible within laboratory experimental models (e.g., R 1, R 2) and within clinical studies (e.g., R 3) but these do not bridge the gap of interest. A very early meta-analysis of short-term clinical wear results demonstrated that concordance of study results was possible. Correlations between laboratory and clinical data (e.g., C 1, C 2, C 3, C 4) can occur in a number of combinations and permutations. Any single one of these correlations provides only part of the story. Ideally, all the key clinical events would be connected to a laboratory test or simulation.
Prevalence of clinical trials involving dental materials
Most of the restorative materials research occurs in the laboratory and not in clinical trials. Very little change has occurred in the prevalence of clinical research activity in 30 years . The total number of abstracts increases but the percentage of all abstracts stays about the same in the range of 5–10%. Clinical research abstracts among dental materials ones at IADR in 1990 was 7.4% and in 2000 was 5.3% . Fig. 2 summarizes the distribution of dental materials IADR abstracts presented over 30 years that are categorized as laboratory, tissue, and clinical research ones that are collected into applied and basic research types.