Restoration Variables and Postoperative Hypersensitivity in Class I Restorations: PEARL Network Findings. Part 2

PEARL Network Practitioner Investigators (P-Is) requested that the PEARL Executive Committee investigate postoperative hypersensitivity (POH), as they considered it to be of major concern. The study described below was conducted in response to this request and to determine whether POH was related to materials and techniques employed in restorations. In this report, POH is considered to be pain that is present at 1 week or more post-treatment to the tooth having undergone restoration that is associated with mastication, clenching, or sensitivity to cold, hot, and sweet foods or beverages.⁴ Pain during clenching only usually indicates a restoration in hyper occlusion, while pain during chewing is considered a form of POH related to fluid flow associated with RBC restoration dentin debonding and tooth deformation during mastication.^5,6

While reports of POH following posterior RBC restorations vary widely, in general, there is some level of transient response among a proportion of patients. The majority of studies are small and associated with evaluation of a particular bonding agent or RBC formulation. The authors summarized the relevant studies and the range of POH reported in our companion manuscript.⁴

As POH is often a secondary focus of studies primarily examining restorative material and placement techniques, measurement approaches have not been uniform. The authors have adopted the approach of having patients, who have been given baseline training, report their POH experience by self-report at intervals following restoration. This approach is an attempt to ameliorate bias associated with the dentist-patient relationship. This PEARL protocol investigated a number of variables related to P-I choices in restorative materials and techniques possibly linked with POH that have not previously been combined in a single study to elucidate their relationships.

Following training and calibration, the staff in 45 dental offices enrolled eligible study subjects on an ongoing basis until 613 subjects had been identified with Class I carious lesions judged by the P-Is as requiring operative intervention and suitable for treatment with a resin-based composite restoration (RBC). The study protocol was approved by the New York University School of Medicine Institutional Review Board. PEARL P-Is recorded specific treatment data,¹ including the following: patient-reported presence of preoperative hypersensitivity using an 11-point (0-10) Numeric Pain Assessment Scale (NPAS) questionnaire for the following: sensitivity to cold, hot, sweets, clenching, or chewing; ranking of dentin caries activity after removal of enamel; and measured preparation depth, width, and length to the nearest 1 mm with a marked periodontal probe. In addition, data were collected by class and brand name for cavity disinfectant (antimicrobial), desensitizing agent and liner if used, dentin bonding agent (DBA), RBC applied, and manner of placement of the RBC—either bulk fill or layer fill.

Patients were surveyed using questionnaires comprising three distinct parts: the above analog scales for sensitivity; one for analgesic use; and finally, an Oral Health Impact Profile (OHIP-14) measuring quality of life (QOL).^7-10 Survey data were collected from patients in person preoperatively to familiarize them with the questionnaires, and were collected anonymously to the dentist and staff either via mail or electronically at 1 week, 4 weeks, and 13 weeks post-treatment.

Included were patients with fully erupted second molars who were not more than 60 years of age. The tooth to be restored had to be in occlusion with no existing restorations. Details of inclusion and exclusion criteria have previously been provided.³

Appreciable hypersensitivity (AH) is defined as a score of at least 3 on the 0 to 10 NPAS to any of the following: cold, hot, sweet, chewing, or clenching. The change in AH is categorized into four groups: did not develop AH, AH resolved, AH not resolved, and developed AH. The category “did not develop AH” is the referent group for all analyses. Cross-tabulations of restoration characteristics with change in AH were summarized using frequencies and percentages. Polychotomous logistic regression was used to evaluate the univariate relationships between the four levels of change in AH and the restoration characteristics. In the event of sparse data, Fisher’s exact test was used. Greatest sensitivity is defined as the highest of the five NPAS scores for the stimuli mentioned above. The change in greatest sensitivity is defined as the post-treatment score minus the value at baseline. Therefore, the theoretical range for change in greatest sensitivity is from -10 to 10. A positive change indicates a worsening of sensitivity, a negative value indicates an improvement, and a value of 0 means no change in sensitivity. To avoid distributional assumptions, the non-parametric Wilcoxon test was used to evaluate the univariate association between change in greatest sensitivity and the restoration characteristics. In cases of sparse data, an exact Wilcoxon was used.

Complete one-week post-treatment sensitivity information was available for 510 study teeth; and at the 4-week recall, there were 491 with complete data for all variables on 485 teeth. Gender, race, ethnicity, and age distributions have been covered in previous publications. Herein the authors consider techniques and materials employed.

The general distribution of restorative procedures is shown in Figure 1 as a combined pie chart. An antimicrobial was used in 17% of restorations (Figure 1). In the majority (96%) of instances, P-Is applied a chlorhexidine-containing product. A chlorhexidine digluconate–based product (Consepsis^® Scrub, Ultradent Products, Inc., www.ultradent.com) was specified in 39% of instances, and 18% indicated that they used a generic chlorhexidine. Other chlorhexidine-containing products employed were Peridex™ (3M ESPE, www.3MESPE.com) (22%), Cavity Cleanser™ (BISCO, Inc., www.bisco.com) (7%), Maxiclens (Henry Schein Dental, www.henryschein.com) (7%), and other (5%). Tubulicid Red (Global Dental Products, www.gdpdental.com) was applied in 2% of instances.

In 2% of preparations, a desensitizer containing glutaraldehyde (GLUMA) as the primary active ingredient was the only product employed. For an additional 6% of study teeth, the P-I indicated a desensitizer was used; however, the product was not an actual desensitizing agent. In subsequent analyses, only the former are considered as having used a desensitizing agent.

P-Is were queried as to whether they employed a cavity liner, and overall they reported use of a liner in 46% or restorations. This question created some misunderstanding, as P-Is considered use of a flowable composite over a DBA to be use of a liner. If flowable composite is removed from the liner category, it can be said that a liner was used in 33% of preparations (Figure 1). The majority (57%) applied glass ionomer or resin-modified glass ionomer; the authors have chosen to collapse this category into a single category, hereafter referred to as RMGI. In 43% of instances, a calcium hydroxide (CaOH)-based product was applied. In only one instance, both a CaOH and RMGI liner were used in combination in these generally small- to moderate-sized preparations.

In all but 1% of restorations, use of a DBA was reported (Figure 1). The use of a total-etch bonding agent and a self-etch DBA were almost equally distributed—53% and 47%, respectively. The choice of RBC material employed fell into the category of hybrid type (92%), with flowable in the remaining 8%. In the majority of instances, the RBC was placed and cured in layers (65%) and the rest with bulk curing.

As presented previously,⁴ 30% of patients had AH at baseline, and at 4 weeks restoration resulted in elimination of AH (but not necessarily all sensitivity) in 63% of instances. In contrast, among patients without AH, restoration led to development of AH, which remained at 4 weeks in 10% of these teeth. These changes in AH can be seen globally as changes in overall greatest sensitivity (Table 2).

Using polychotomous logistic regression and Fisher’s exact test when necessary, there were neither significant associations between materials and techniques employed nor any change in AH at 4 weeks (Table 1). In general, this lack of any significant relationship was true for change in greatest sensitivity (Table 2). Note, there is no difference between total-etch and self-etch DBAs when considering changes in AH. However, when considering change in greatest amount of sensitivity, there was a significant reduction (P = 0.004) in sensitivity with use of a self-etch as compared to a total-etch bonding agent, suggesting only subtle differences in outcomes between these systems in this application.

One unexpected finding was a significant relationship between change in AH at 1 week and use of an antimicrobial (P = 0.006). For study teeth with no AH at baseline, the odds of developing AH 1 week post-treatment increased with use of an antimicrobial agent (odds ratio = 2.83, 95% confidence interval: 1.40 to 5.73). This relationship appeared to be driven by use a chlorhexidine digluconate-containing product (Consepsis Scrub) on the dentin, which was significantly associated with 1-week post-restoration AH (P < 0.001). However, this relationship for overall use of an antimicrobial was not present at 4 weeks (P = 0.602), but there was a significant impact on AH with use of the chlorhexidine digluconate product (P = 0.010) that carried over from 1 week.

The average number of enrolled teeth per practice site in this study was 15 (range 1-53) across the 45 enrolling practices. P-Is used a variety of materials and material brands as well as technique. The percentage of study teeth with AH at 4 weeks varied across practices and ranged from 0% to 57% when considering sites that had outcome data on at least five study teeth.

The use of an antimicrobial solution on prepared dentin has appeal for many dentists, as caries is a microbial-based disease process, and it is known that preparation leaves viable bacteria in the dentinal tubules.^11,12 However, it is equally as well demonstrated that a preparation where soft carious dentin remains, if sealed by a restoration, results in a significant reduction in viable bacteria and hardening of the underlying dentin.^13-16 If POH is related to residual bacteria remaining and active after preparation and restoration, then application of a microbial might be justified. In this effectiveness study, the application of a chlorhexidine-containing product to the dentin prior to restoration was limited to 17% of restorations and was not related to change in AH nor to change in greatest sensitivity at 4 weeks. Rather, the application to teeth of chlorhexidine in the form of chlorhexidine digluconate was found to have a significant effect on AH at both 1 and 4 weeks postoperatively. Specifically, using that antimicrobial product more than doubles the odds of developing AH at 4 weeks post-restoration in teeth with no AH at baseline. Chlorhexidine is known to have a strong cytotoxic effect on cells in culture^17-20 and can cause inflammation if in sufficient concentration. It can also negatively affect nerve cells.²¹ Perhaps the observed postoperative AH was related to chlorhexidine diffusion that couples with the pulpal insult that occurs during complete caries removal. Note that complete caries removal was almost universally employed by P-Is (Figure 1).

Use of the desensitizer GLUMA was not found to have an appreciable effect on AH outcomes nor on change in greatest sensitivity. This goes counter to several studies,^22,23 where it has been suggested that the active ingredient glutaraldehyde (GA) can cause protein precipitation in the dentinal tubules, limiting fluid flow and reducing or eliminating sensitivity. Most studies where GA efficacy was demonstrated were conducted on sensitive noncarious cervical lesions.²⁴ The extension of these findings to occlusal restorations may not be justified. Two P-Is used both an antimicrobial and a desensitizer in one of their occlusal restoration procedures.

P-Is’ consideration of flowable RBC as a liner was unanticipated in this protocol. In order to consider the possible efficacy of the flowable as a liner, the authors analyzed the sensitivity data with inclusion of flowable RBC as a liner (see category Use of Any Liner in Table 1 and Table 2) and with use of the flowable liner excluded (see category Use of True Liner in Table 1 and Table 2). In neither category was there an effect on change in AH nor greatest sensitivity related to the liner placed at 4 weeks post-treatment. This finding is in line with a recent 7-year retrospective study where use of a flowable liner had no influence on clinical outcomes in Class II restorations.²⁵

Both glass ionomer and resin-modified glass ionomer have been suggested for some time as useful in reducing sensitivity in RBC restoration, particularly in the posterior dentition.^26,27 The authors have conducted subcategory analysis within “true liners” and find no influence of liner type, CaOH, and RMGI on development or change in AH; but CaOH was significantly related to greatest change in sensitivity at 4 weeks (P = 0.018). The clinical significance of this latter finding is questionable. The lack of efficacy of CaOH for reduction in AH in these moderate to shallow lesions is not unanticipated.^28-30 However, postoperative sensitivity may²⁹ or may not be associated with preparation depth in posterior RBC restorations.³⁰ Studies on dental patients receiving posterior RBC restorations with and without an RMGI liner (no use of CaOH) did not find a difference in these treatment groups.^31,32 The PEARL Network has just completed a randomized clinical effectiveness trial for Class I and II restorations placed with and without an RMGI liner to address this issue. The study closed with data collected on 341 patients, and the results are being analyzed.

PEARL P-Is were almost equally split in their use of self-etch as opposed to total-etch DBAs, 312 versus 353 treated teeth, respectively. In regard to dentin bonding agents, many clinicians feel that use of a self-etch bonding agent will result in less POH,³³ but this is not the finding in several clinical studies.^34-36 There was a statistically significant trend for a reduction in greatest sensitivity with use of self-etch bonding agents, but no difference with regard to change in AH. Any results are clouded by the wide range of DBAs employed by P-Is (more than 30 brands indicated). Also, in this effectiveness study, the authors were not able to determine to any extent whether the DBAs were placed following the manufacturers’ directions, but assume this to be the case. They did review the dentin bonding agents used by the P-Is that are advertised to be employed in either self-etch or total-etch mode for clarification. This resulted in identifying three P-Is using self-etch DBAs in total-etch mode where they at first had assigned these teeth to the self-etch category. However, the reassignment did not change the sensitivity findings.

The preferred type of RBC placed by P-Is was hybrid (92%), with the remainder using flowable, which is not unexpected in these shallow-to-moderate lesions. The authors investigated whether flowable RBC was applied in restorations of smaller volume. This used data upon which they are currently reporting related to preparation variables.⁴ They found that 55% of the flowable RBC applications were in instances of smaller restorations (volume < 25mm³). Seventy-six percent of P-Is used one type of RBC for all their restorations.

When it came to RBC placement technique, the majority of P-Is (65%) employed the layering approach, and the remainder bulk curing. The authors again anticipated that the bulk curing technique would be predominately employed in smaller restorations. Their findings were that use of the bulk technique was significantly associated with these small lesions (P < 0.001).

In reviewing their findings, the authors feel strongly that patient-reported outcome such as anonymously collecting their postoperative levels of sensitivity to various stimuli is a methodology that should be employed on a more extensive scale.

In summary, we find that in this Network study of 45 P-Is patients reporting their POH at 4 weeks, there is no “smoking gun” related to materials and techniques, whether examining a resolution of AH, development of AH, or any change in the greatest sensitivity. This suggests that the individual practitioner’s technique is more important than the materials employed, as AH at 4 weeks was reported by patients in some practices, while other practices using similar materials had almost no AH reported. The authors were quite surprised by the level of baseline AH found (30%) and that only 63% of these resolved by restoration. Another 10% of patients with no baseline AH had it at 4 weeks. Also, there was no relationship found for any of these changes or lack of change to cavity depth or volume (Berkowitz4). We, therefore, hypothesize that there may be a subset of patients who will have dentinal sensitivity related to caries and/or instrumentation of dentin regardless of the extent of the restoration and independent of the materials and techniques employed.

This study was supported by grant U01-DE016755 awarded to New York University College of Dentistry by the National Institute of Dental and Craniofacial Research.

Peter Blanchard, DDS
PEARL Network Practitioner-Investigator
The PEARL Network
New York, New York

Ying Wong, DDS PEARL
Network Practitioner-Investigator
The PEARL Network
New York, New York

Abigail G. Matthews, PhD
Biostatistician
The EMMES Corp.
Rockville, Maryland

Donald Vena, BS
Statistician
The EMMES Corp.
Rockville, Maryland

Ronald G. Craig, DMD, PhD
PEARL Network Director of Information Dissemination
Associate Professor
Basic Sciences and Craniofacial Biology
Department of Periodontology and Implant Dentistry
New York University College of Dentistry
New York, New York

Frederick A. Curro, DMD, PhD
PEARL Network Director of Recruitment
Retention and Operations; Clinical Professor Oral & Maxillofacial Pathology
Radiology and Medicine
New York University College of Dentistry
New York, New York

Van P. Thompson, DDS, PhD
PEARL Network Director of Protocol Development and Training
Professor
Chairperson
Biomaterials and Biomimetics
New York University College of Dentistry
New York, New York