Clinical Research Mineral Trioxide Aggregate or Calcium Hydroxide Direct Pulp Capping: An Analysis of the Clinical Treatment Outcome Johannes Mente, DMD,* Beate Geletneky, DMD,* Marc Ohle,* Martin Jean Koch, MD, DMD, PhD,† Paul Georg Friedrich Ding, DMD,† Diana Wolff, DMD,† Jens Dreyhaupt, DSc,‡ Nicolas Martin, BDS, PhD, FDS,§ Hans Joerg Staehle, MD, DMD, PhD,† and Thorsten Pfefferle, DMD* Abstract Introduction: The use of mineral trioxide aggregate (MTA) might improve the prognosis of teeth after pulp exposure. The treatment outcome of teeth after direct pulp capping, either with mineral trioxide aggregate (MTA) or calcium hydroxide (controls), was investigated, taking into account possible confounding factors. Methods: One hundred forty-nine patients treated between 2001 and 2006 who received direct pulp capping treatment in 167 teeth met the inclusion criteria. Treatment was performed by supervised undergraduate students (72%) and dentists (28%). Assessment of clinical and radiographic outcomes was performed by calibrated examiners 12–80 months after treatment (median, 27 months). Results: One hundred eight patients (122 treated teeth) were available for follow-up (72.5% recall rate). A successful outcome was recorded for 78% of teeth (54 of 69) in the MTA group and for 60% of teeth (32 of 53) in the the calcium hydroxide group. The univariate analysis (generalized estimation equations model [GEE model] showed a significant difference in the success rate (odds ratio [OR], 2.36; 95% confidence interval [CI], 1.05–5.32; P = .04). In the multiple analysis (GEE model), the OR is marginally inside the nonsignificant range (OR, 0.43; 95% CI, 0.19–1.02; P = .05) when conspicuous confounding factors are stabilized (univariate analysis). Multiple analysis showed that teeth that were permanently restored $2 days after capping had a significantly worse prognosis in both groups (OR, 0.24; 95% CI, 0.09–0.66; P = .01). Conclusions: MTA appears to be more effective than calcium hydroxide for maintaining long-term pulp vitality after direct pulp capping. The immediate and definitive restoration of teeth after direct pulp capping should always be aimed for. (J Endod 2010;36:806–813) Key Words Calcium hydroxide, direct pulp capping, dental pulp exposure, humans, mineral trioxide aggregate, MTA, treatment outcome I n direct pulp capping the exposed pulp is dressed with a medicament or dental material, with the specific aim of maintaining pulpal vitality and health (1, 2). Direct pulp capping has been practiced for more than 200 years. In 1756 Phillip Pfaff covered exposed vital pulp with gold plate (3). Mixtures containing calcium hydroxide Ca(OH)2 for treating exposed pulp were first described about 100 years later (4). However, the use of Ca(OH)2 for direct pulp capping gained importance only after publication of investigations of Hermann (5) in 1930. Since then Ca(OH)2 has remained the material of choice, against which all other materials suggested for pulp capping are judged (1). Clinical success rates after direct pulp capping with Ca(OH)2 or with Ca(OH)2 compounds have been evaluated in different studies (6–12), and today this material is regarded as the gold standard. The spectrum of success rates ranges from 13% (11) to 96% (7). The difference in these rates is attributed to different potential prognostic factors that can influence the outcome of direct pulp capping such as length of follow-up (11), type of pulp exposure (carious or mechanical) (7, 12), presence of an extrapulpal blood clot between the pulp and the capping material (13), the area of pulp to which the capping material was applied (coronally or cervically) (12), time elapsed to placement of a definitive restoration of the pulp-capped tooth (11), type of Ca(OH)2 used (14), presence or absence of infection (as a result of bacteria still present or exposure to new bacteria from leakage) (12, 15, 16), as well as the age of the patients (10, 17). In addition, different definitions of success and failure must be considered when comparing and evaluating data in clinical studies. In recent years a new cement (mineral trioxide aggregate [MTA]), developed in the 1990s by Torabinejad and his coworkers at Loma Linda University (California), has become available as a root canal repair material and for direct pulp capping. During the setting process, MTA has an initial pH of 10.2, which increases to up to 12.5 during the first few hours (18). This is comparable with the pH range achieved by Ca(OH)2 preparations after application on the exposure area (19). In spite of this, there appear to be differences in pulpal tissue reaction to MTA compared with Ca(OH)2 in direct pulp caps (20). Dentin bridge formation with MTA seems to be more homogenous (fewer tunnel defects) and more localized than that formed with Ca(OH)2 (20–24). From the *Department of Conservative Dentistry Division of Endodontics, †Department of Conservative Dentistry, and ‡Institute of Medical Biometry and Informatics, Ruprecht-Karls-University of Heidelberg, Heidelberg, Germany; and §Department of Adult Dental Care, University of Sheffield, Sheffield, United Kingdom. Address requests for reprints to Dr Johannes Mente, Head, Division of Endodontics, Department of Conservative Dentistry, University Clinic Heidelberg, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany. E-mail address: johannes.mente@med.uni-heidelberg.de. 0099-2399/$0 - see front matter Copyright ª 2010 American Association of Endodontists. doi:10.1016/j.joen.2010.02.024 806 Mente et al. JOE — Volume 36, Number 5, May 2010 Clinical Research Various in vitro studies have shown MTA to be biocompatible (25) and to have good sealing properties (26). Initial clinical studies that have evaluated the use of MTA as a direct pulp-capping material have shown promising results (27–31). Positive results have also been obtained when using MTA after a pulpotomy (2, 32–34) and partial pulpotomy (35, 36). A drawback of these studies is that the number of cases is rather small and the observation periods are relatively short. A review of the current literature does not highlight any clinical studies that assess the long-term success (>1 year) of direct pulp capping with MTA compared with Ca(OH)2. The aim of this retrospective, single-center case-control study was to determine the clinical long-term success of direct pulp capping, undertaken under comparable conditions, by using MTA (ProRoot MTA; Dentsply-Maillefer, Ballaigues, Switzerland) or a non-setting Ca(OH)2 paste (Hypocal SN; Merz Dental, Lütjenburg, Germany). Material and Methods All the subjects in this study were retrieved from the records of patients who attended for routine conservative treatment at the Department of Conservative Dentistry at the University Hospital of Heidelberg between 2001 and 2006. Those patients were selected who had received direct pulp capping treatment. The inclusion criteria for this retrospective study were all teeth where the pulp was capped either with MTA or Ca(OH)2 and the interval between pulp capping and the last follow-up examination was at least 1 year. The following subjects were excluded from this study: those with compromised immune status, who were pregnant at the time of follow-up, who declined to participate in the study, or whose pretreatment and intratreatment records were incomplete. The study protocol was approved by the Ethics Committee of the University of Heidelberg (Ref. 132/2006). Recruitment of Patients Subjects who met the inclusion criteria (149 patients) were contacted by letter and subsequently by phone and were invited to attend the follow-up examinations. Participants who agreed to attend were mailed detailed information about the study. At the follow-up examination the patients were again given a detailed explanatory information sheet and were asked to sign a declaration of informed consent to participation in the study. Clinical and radiographic follow-up examinations were undertaken after written informed consent had been given. Calibration Two examiners (B.G., T.P.) were designated to carry out the clinical follow-up examinations of all subjects. They were calibrated by independently examining 24 patients on the same day and by recording the following clinical parameters: response to cold test (CO2) and percussion, probing depth, attachment loss, tooth mobility, type and quality of coronal restoration (visual inspection with mirror and explorer). A tolerance range of 1 mm was defined for probing depth and attachment loss. The recorded data were analyzed for interexaminer reliability by using absolute and relative frequencies of disagreement. Both examiners (B.G., T.P.) were also designated to carry out all the radiographic interpretations of intraoral periapical views. Before evaluating the study radiographs, one of these examiners (B.G.) was calibrated with the periapical index (PAI) calibration kit of 100 periapical radiographs (37). Intraexaminer reliability and interexaminer agreement with the calibration kit gold standard were assessed by using Cohen kappa statistic. JOE — Volume 36, Number 5, May 2010 Treatment Intervention All treatment procedures had been completed before the study was designed. Supervised undergraduate students completed 70% (48 teeth) of the direct pulp capping treatments with MTA and 76% (40 teeth) of the direct pulp capping treatments with Ca(OH)2. All other pulp cappings with MTA (21 teeth, 30%) and Ca(OH)2 (13 teeth, 25%) were performed by experienced dentists in the Department of Conservative Dentistry at the University Hospital of Heidelberg. Teeth had been treated in accordance with the current techniques for restoration of teeth. In every case teeth were isolated with rubber dam before all treatment procedures. Caries was removed by means of mechanical excavation with a slow-speed rose head bur. Both dentists and supervised students ensured that the peripheral caries was removed before the caries was excavated from the cavity walls near to the pulp, except for one carious spot, the removal of which resulted in exposure of the pulp. The cavities were routinely disinfected with 0.12% chlorhexidine solution (Glaxo Smith Kline GmbH, Bühl, Germany). If the caries had already extensively penetrated the pulp chamber or the tooth showed signs and symptoms of irreversible pulpitis, the pulp was not capped directly, but vital extirpation was performed; these teeth were not included in this study. Operative protocols for the endodontic management of exposed pulps are routinely followed by all clinicians in the department including students. The departmental operative protocol states the following clinical guidelines: if the pulp was exposed, a sterile cotton pellet soaked in 0.12% chlorhexidine solution was placed on the vital pulp. The exposure site was checked by the dentist (or the supervisor in case of treatment in the undergraduate clinic) to ensure that all carious hard tissues had been removed before pulp capping. The caries-free condition was verified by visual inspection with a mirror and explorer. Resolution of bleeding from the exposed pulp in less than 5 minutes was considered to be indicative of reversible inflammation and dressed with a direct pulp capping agent, MTA or non-setting Ca(OH)2. If bleeding from the exposed pulp persisted for 5 minutes or longer, irreversible inflammation of the pulp tissue was assumed, and a vital extirpation was performed. The pulp-capping material was placed exclusively directly over the exposed pulp. MTA cement was applied to the area of pulp exposure in small portions by using an MTA gun (Dentsply-Maillefer) and then carefully compressed into the pulp wound by using a rounded plugger (PFI 117; HU-Friedy, Chicago, IL). White ProRoot MTA (Dentsply-Maillefer) was used for the anterior teeth and gray for the posterior teeth. The non-setting Ca(OH)2 paste (Hypocal SN) was applied to the exposed pulp by using a mini ball burnisher (PLG 30/H 34; HUFriedy). When the capping material (ProRoot MTA or Hypocal SN) had been applied, it was overlaid with a thin protective layer of resinmodified, glass ionomer cement (Vitrebond; 3M Espe, St Paul, MN) to ensure that the capping material was not partially or wholly removed during subsequent treatment of the tooth (eg, cauterization and rinsing). The aim was to fill the cavities of the capped teeth immediately with composite fillings; however, in exceptional cases the teeth were treated with temporary filling material for a few days (IRM; Dentsply, Konstanz, Germany). This was replaced with composite as soon as possible. The final restoration consisted of an adhesive direct-placement composite restoration or a full-coverage crown. Preoperative and Intraoperative Data Pertinent preoperative and operative clinical information was gathered from the patients’ records and entered into a specifically MTA Compared with Ca(OH)2 for Direct Pulp Capping 807 Clinical Research designed database spreadsheet. Preoperative data included gender, age, tooth type, tooth location, clinical signs and symptoms, and studyrelevant information regarding medical history (eg, compromised immune status). Operative data included type of capping material (MTA or Ca(OH)2), date of direct pulp capping, type of pulp exposure (carious or mechanical), site of exposure (cervical or occlusal), response to cold test, tooth mobility, probing pocket depths and attachment loss, type and class of restoration placed immediately after pulp capping, time interval between pulp capping and placement of a permanent restoration, and treatment providers. Follow-up Examination The study design meant that the follow-up examinations were carried out at different time intervals ranging from 12–80 months after treatment, with a median follow-up period of 27 months. The presence of clinical signs and symptoms (sensitivity to percussion or bite test, any other pain or discomfort relating to the capped tooth), response to cold test, tooth mobility, type and quality of restoration, probing of pocket depths and attachment loss, furcation involvement, presence of a sinus tract, radiographic findings of root canal treatment, or extraction of the tooth were recorded and entered in a structured recall form specially designed for this study. The quality of the coronal restoration was assessed both clinically (visual inspection with mirror and explorer) and radiographically by looking for marginal gaps and radiolucencies consistent with and indicative of the presence of caries and then rated as acceptable or unacceptable (leakage probable). Further radiographic evaluation was carried out during the follow-up examination. Outcome Assessment Outcome was assessed on the basis of clinical and radiographic findings and was classified as success when there was a clear positive response to cold test (CO2), absence of clinical signs and symptoms, no indication of apical periodontitis (PAI score = 1), absence of internal root resorption, and no loss of function. Outcome was classified as failure when 1 or more of the following findings was observed at the follow-up examination with regard to the capped tooth: clinical signs and symptoms (including sensitivity to percussion or bite test, pain or discomfort related to the capped tooth), negative response to cold test, radiologic signs of apical periodontitis (PAI score $2), condensing apical periodontitis, presence of a sinus tract, root canal treatment or extraction of the pulp-capped tooth, internal root resorption, or loss of function (eg, grade III tooth mobility). Multirooted teeth were assessed according to the highest scored root on the PAI score. Follow-up radiographic assessment was undertaken by using periapical radiographs of each case. These were coded, stored, and assessed by a PAI-calibrated examiner (B.G.). They were then evaluated independently in a random sequence by the 2 examiners (B.G., T.P.). The radiographic examination sought to determine the presence or absence of any pathologic changes adjacent to the pulp-capped teeth (eg, internal root resorption or condensing apical periodontitis). Where they did not agree, the examiners met to discuss the radiographic findings and come to a consensus. All radiographs were evaluated in a darkened room by using an illuminated viewer box (Kentzler-Kaschner Dental GmbH, Ellwangen, Germany) with 2 magnification. Statistical Analysis Median, first and third quartile, minimum and maximum, as well as relative and absolute frequencies were calculated for descriptive analysis. In addition, the 95% confidence interval (CI) for the overall 808 Mente et al. success rate and for the success rates of the Ca(OH)2 and MTA groups was calculated. Generalized estimation equations model (GEE model) was used to investigate the relation between outcome and potential predictor variables (preoperative, intraoperative, and postoperative variables). GEE models allow an evaluation of dependent observations while accounting for the treatment of more than 1 tooth in individual subjects. The dependent variable in all analyses was the dichotomous outcome, success versus failure. The data were processed by using the SAS statistical package (Version 9.1; SAS Institute Inc, Cary, NC). Because of the exploratory nature of the study, no adjustment was made for multiple testing. All tests were performed at a significance level of .05. All investigated factors are listed in Table 1. Results Calibration Process In a blind procedure the 2 examiners assessed 24 patients (681 teeth), and the double examination of each patient by the 2 observers was performed on the same day. There was no discrepancy in the results of the cold or percussion tests and the evaluation of the tooth mobility recorded by both examiners for the 24 patients examined. A high level of consensus was also achieved with regard to probing depth (99.0%), attachment loss (86.9%), type of restoration (92.9%), and quality of restoration (99.7%). Intraexaminer reliability for the PAI calibration results was k = 0.96, and interexaminer agreement (examiner scores versus the calibration kit authorized scores) was k = 0.84. Study Group One hundred forty nine patients (167 teeth) were initially identified for potential inclusion. One hundred eight of these patients (122 teeth) took part in the follow-up examination (patient recall rate, 72.5%). At the time of the pulp exposure operative intervention, the ages of the patients ranged from 8–78 years, with a median age of 40.1 years (first quartile, 28.4; third quartile, 51.4). The group of patients treated with MTA (n = 62) ranged from age 8–75 years, with a median age of 37.1 years (first quartile, 26.4; third quartile, 49.1), and the group of patients treated with Ca(OH)2 (n = 46) ranged from age 13–78 years, with a median age of 44.4 years (first quartile, 33.8; third quartile, 57.8). The reasons for dropout were as follows: 15 patients (10%) could not be contacted in spite of repeated letter writing or telephone calls, 11 patients (7%) had moved away, and 19 patients (13%) refused to participate in the recall. Patients were classified according to treatment with MTA (n = 62, 69 teeth) and Ca(OH)2 (n = 46, 53 teeth). Table 1 shows the demographic and outcome distribution of teeth across preoperative, intraoperative, and postoperative variables in both treatment groups. Success Rate Of 122 treated teeth, the outcome was deemed successful in 86 teeth (70%; 95% CI, 62–78) and a failure in 36 teeth (30%). Twenty-two of the 36 failed teeth (61%) received postoperative root canal treatment, 10 teeth in the MTA group and 12 teeth in the Ca(OH)2 group. Five teeth (14%), all in the Ca(OH)2 group, were extracted by the patient’s dentist in private practice, or the patient’s records showed that the capped tooth had been extracted after direct pulp-capping treatment, 5 teeth (14%) (2 teeth in the MTA group and 3 teeth in the Ca(OH)2 group) showed evidence of pulp necrosis, JOE — Volume 36, Number 5, May 2010 Clinical Research TABLE 1. Outcome Distribution across Preoperative, Intraoperative, and Postoperative Variables for Both Treatment Groups of Teeth (MTA and Ca(OH)2) with Direct Pulp Capping MTA, total no. of teeth Variable Age <25 y 25–50 y >50 y Gender Female Male Tooth location Maxilla Mandibula Tooth type Anterior Posterior Type of restoration Temporary Permanent Unknown (tooth extracted) Time span before placement of permanent restoration <2 days $2 days Site of exposure Cervical Occlusal Type of pulp exposure Carious Mechanical Size of restoration Small Large Quality of coronal restoration at follow-up Acceptable Unacceptable Unknown (tooth extracted) Treatment providers Supervised undergraduate students Dentists Recall time 1–2 y 2–3 y >3 y MTA, pulp capping success* Ca(OH)2, total no. of teeth Ca(OH)2, pulp capping success* n % n % n % n % 17 40 12 25 58 17 15 30 9 88 75 75 7 28 18 13 53 34 2 21 9 29 75 50 44 25 64 36 35 19 80 76 23 30 43 57 13 19 57 63 44 25 64 36 36 18 82 72 30 23 57 43 21 11 70 48 16 53 23 77 15 39 94 74 14 39 26 74 7 25 50 64 2 67 0 3 97 0 1 53 – 50 79 – 1 47 5 2 89 9 0 32 – 0 68 – 60 9 87 13 49 5 82 56 42 11 79 21 29 3 69 27 11 58 16 84 10 44 91 76 10 43 19 81 6 26 60 61 59 10 86 15 47 7 80 70 47 6 89 11 29 3 62 50 12 57 17 83 8 46 67 81 6 47 11 89 5 27 83 57 64 5 0 93 7 0 52 2 – 81 40 – 44 4 5 83 8 9 31 1 – 71 25 – 48 21 70 30 38 16 79 76 40 13 75 25 27 5 68 38 29 26 14 42 38 20 24 18 12 83 69 86 20 20 13 38 38 25 15 11 6 75 55 46 MTA, mineral trioxide aggregate. *Based on radiographic and clinical assessment. and 4 teeth (11%) presented with periapical radiolucency at follow-up (2 teeth in the MTA group and 2 teeth in the Ca(OH)2 group). When the results were analyzed separately for both treatment groups, 54 teeth (78%; 95% CI, 67–87) in the MTA group were classified as successful and 15 teeth (22%) as failure. In the control group treated with Ca(OH)2, 32 teeth (60%; 95% CI, 46–74) were classified as successful and 21 teeth (40%) as failure (Fig. 1). The univariate GEE model showed a significantly higher success rate for teeth capped with MTA compared with Ca(OH)2-capped teeth (odds ratio [OR], 2.36; 95% CI, 1.05–5.32; P = .04). The tendency for the success rate in the teeth capped with Ca(OH)2 to drop after a follow-up period of 2–3 years and more than 3 years is striking; in the MTA group the success rate is relatively constant (Fig. 2, Table 1). Influence of Potential Prognostic Factors A breakdown of the sample across the different potential prognostic factors examined is shown in Table 1. To assess the influence JOE — Volume 36, Number 5, May 2010 of potential prognostic factors on the success rate, a univariate analysis (GEE model) was first performed for every treatment group (MTA group and Ca(OH)2 group), including the following variables discussed in the literature as possible influential factors: age, gender, tooth location, type of tooth, type of restoration, time span before placement of restoration, site of exposure, type of pulp exposure, size of restoration, quality of coronal restoration at follow-up, treatment providers, and recall time. Only 2 of these potential prognostic factors were conspicuous in the Ca(OH)2 group: time span before placement of a permanent restoration after pulp capping (OR, 0.17; 95% CI, 0.04–0.72; P = .02) and treatment provider (OR, 0.30; 95% CI, 0.09–1.02; P = .05). Interestingly, no conspicuous factors could be identified in the MTA group (Table 2). In a final multiple analysis (GEE model) the success rate of direct pulp capping with MTA or Ca(OH)2 was adjusted for the 3 conspicuous factors identified in the univariate analysis (P # .05): pulp-capping material, time span before placement of a permanent restoration after pulp capping, and treatment providers (Table 3). The OR for the MTA Compared with Ca(OH)2 for Direct Pulp Capping 809 Clinical Research TABLE 2. Univariate Analysis of Associations between Potential Prognostic Factors and the Success Rates for Both Treatment Groups (P values of the GEE model) Potential prognostic factor Figure 1. Overall and success rates of Ca(OH)2 and MTA groups expressed as a percentage. n = number of capped teeth in each group. Numbers in brackets denote the lower and upper limits of the 95% CI of the success rates. capping material adjusted for the factors of time span before placement of a permanent restoration and treatment providers is on the borderline between significant and not significant (OR, 0.43; 95% CI, 0.19–1.02; P = .05). All results of the multiple analyses are shown in Table 3. Discussion This retrospective study investigated the outcome of 108 patients with 122 pulp exposures, of which 69 teeth were directly capped with MTA (ProRoot MTA) and 53 with a non-setting Ca(OH)2 paste (Hypocal SN). The standardized study protocol was established in advance of the follow-up examinations, and it incorporated clinical and radiographic parameters derived from previous studies (9–12, 30, 37, 38). Because of the explorative nature of the study, no sample size calculation was made, and all eligible patients who had received direct pulp capping with MTA or Ca(OH)2 during the defined study period were considered. Because MTA was introduced to the University Hospital of Heidelberg in December 2000, only patients with teeth that had been treated with a direct pulp cap starting in 2001 could be included in the study. To avoid distortion between the 2 groups (MTA group and Ca(OH)2 group), only those patients in the Ca(OH)2 group whose treatment started after January 2001 were included in the assessment. The median follow-up interval of 27 months was sufficient to record a stable treatment outcome (39). The recall rate of 72.5% was comparable to that in many follow-up studies (39), but it fell short of the 80% required for high level of evidence (39). Independent calibrated examiners recorded the radiographic and clinical follow-up data. Intraexaminer reliability for the PAI calibration results was k = 0.96, and interexaminer agreement was k = 0.84; both kappa scores indicate almost perfect agreement (40). The interexa- Figure 2. Influence of follow-up period on the success rates of Ca(OH)2 and MTA groups expressed as a percentage. n = number of capped teeth in each group. Numbers in brackets denote the lower and the upper limits of the 95% CI of the success rates. 810 Mente et al. Age <25 vs >50 y 25–50 vs >50 y Gender Tooth location Tooth type Type of restoration Time span before placement of a permanent restoration Site of exposure Type of pulp exposure Size of restoration Quality of coronal restoration Treatment providers Recall time 1–2 vs >3 y 2–3 vs >3 y MTA group Ca(OH)2 group .35 1.00 .75 .35 .12 .36 .09 .26 .09 .63 .12 .33 .29 .50 .24 .06 .78 .80 .28 * ‡ .02† .98 .52 .25 .05† .09 .63 GEE, generalized estimation equations; MTA, mineral trioxide aggregate. *No convergence of the GEE model. † Conspicuous factors (P # .05). ‡ Not calculated; 5 missing values because 5 teeth extracted before follow-up. miner reliability with regard to clinical parameters, which was checked in the calibration process with independent, blind assessment of 24 patients by 2 people, also showed a very high level of consensus between the 2 examiners (see results of calibration process). The methodologic limitations of this study concern the allocation of the subjects, which was not randomized because of the retrospective nature of this study, the different qualifications of the treatment providers (which was somewhat compensated for by supervision of treatment in the student courses by experienced treatment providers), and the sample size of the study population. In addition, compromises with regard to standardization of clinical decision-making (eg, the decision to use one capping agent instead of another) are unavoidable in view of the retrospective nature of this study. The strengths of the study lie in the a priori calibration of the treatment providers in both clinical and radiographic assessment and in the advanced statistical method. The apical status at follow-up was assessed by the PAI introduced by Ørstavik et al (37). The PAI has been validated as a reproducible, unbiased method for interpretation of periapical radiographs (39). In the present study a successful outcome was strictly defined by completely normal radiographic and clinical findings with regard to the pulp-capped tooth (41). Therefore, to minimize misinterpretation, only teeth that had been scored as PAI 1 were accepted as healthy, and all other PAI scores were labeled as apical periodontitis. In some studies, directly capped teeth were classified at clinical follow-up into ‘‘uncertain’’ or ‘‘questionable’’ categories (6, 11). Although the study protocol of the present study did not include an uncertain category to avoid misinterpretations, there were in fact no teeth that could have been included in this category. Clinical symptoms such as sensitivity to percussion or bite test or pain or discomfort related to the capped tooth were seen in patients in the present study, but always in combination with other unequivocal findings that clearly classified them in the failure category. A detailed presentation of how many teeth in each group (MTA and Ca(OH)2) were classified in the failure category as a result of unequivocal findings (eg, root canal treatment, extraction, or periapical radiolucency) is shown in the Results section. JOE — Volume 36, Number 5, May 2010 Clinical Research TABLE 3. Multiple Analysis of Associations between Selected Factors and the Success Rate after Direct Pulp Capping by Using GEE Model Prognostic factor* OR for success 95% CI, lower limit 95% CI, upper limit P value* Treatment provider: supervised undergraduate student Reference: dentist Time span before placement of the permanent restoration: $2 days Reference: <2 days Material: Ca(OH)2 Reference: MTA 0.68 0.29 1.59 .37 0.24 0.09 0.66 .01 0.43 0.19 1.02 .05 GEE, generalized estimation equations; OR, odds ratio; CI, confidence interval; MTA, mineral trioxide aggregate. *Conspicuous factors (P # .05) identified by univariate analysis. In the present study the GEE model established by Zeger and Liang (42) was used to investigate the effect of potential outcome predictors. The GEE model takes into account that from some of the patients several teeth were included in the study and permits single and multiple analysis for a direct comparison of the success of both materials used. In contrast to classic regression models, a GEE model can solve the problem that the observations are not independent in a database. Thus the study units are the individual teeth rather than the test subjects. The statistical analysis was performed in 2 stages; first it was performed separately for each potential outcome predictor. For the final model only those factors were chosen that had been conspicuous in the previous analysis. The ages of the patients whose teeth had been capped with MTA ranged from 8–75 years (median, 37.1 years; first quartile, 26.4; third quartile, 49.1), and those of the patients capped with Ca(OH)2 ranged from 13–78 years, with a median of 44.4 years (first quartile, 33.8; third quartile, 57.8). Thus the age distribution of patients in both groups was almost identical. The influence of age itself was examined separately for both groups in the statistical analysis (GEE model), but it was shown not to be conspicuous in this study (Table 2). The influence of the patient’s age is controversial; a few studies point to the influence of age (10, 17), but most were unable to establish any effect on the outcome of direct pulp capping (8, 9, 11, 12). In total, the outcome in 54 of the 69 teeth in the MTA group (78%; 95% CI, 67–87) and in 32 of the 53 teeth in the Ca(OH)2 group (60%; 95% CI, 46–74) was successful (Fig. 1), which might indicate a clinically relevant difference between both groups. The failure rate in the Ca(OH)2 group (40%) was nearly double that in the MTA group (22%). In the univariate analysis (GEE model) this difference was significant (OR, 2.36; 95% CI, 1.05–5.32; P = .04). However, in the multiple analysis (taking into account all conspicuous potential outcome predictors as shown in Table 2), the difference in the success rates in the MTA group compared with the Ca(OH)2 group was borderline significant (OR, 0.43; 95% CI, 0.19–1.02; P = 0.05, GEE model). This might indicate an underpowered analysis as a result of the small sample size. On the basis of the results of this study, a post hoc power analysis was performed by using an ordinary multiple logistic regression model. The outcome was defined as success (on the basis of clinical and radiologic normalcy), and all conspicuous prognostic factors found in this study were included as covariates. The following assumptions were made:  proportion of supervised undergraduate student treatments, 72%; proportion of dentist treatments, 28%  proportion of teeth with time span before placement of restoration <2 days: 84%; proportion of teeth with time span before replacement of permanent restoration $2 days, 16%  material: proportion of MTA-treated teeth, 57%; proportion of Ca(OH)2-treated teeth, 43%. The overall probability for success was assumed as 70%. To show an OR of 0.43 with a power of 80% at a significance level of 5%, this JOE — Volume 36, Number 5, May 2010 scenario requires 211 patients (assuming that each patient contributes only 1 tooth), without consideration of possible dropout. Because the present study project will continue to include new patients prospectively, this prognostic factor will be reevaluated in the future with a larger sample size. Nine of 69 restorations in the MTA group (13%) and 11 of 53 restorations in the Ca(OH)2 group (21%) were permanently restored $2 days after direct pulp capping. This resulted in a reduction in the success rate in these teeth of 26% in the MTA group and of 42% in the Ca(OH)2 group (Table 1). The relatively low reduction in success rate associated with a late placement of the permanent restoration in the MTA group might be due to the effective sealing properties of the MTA cement against bacterial leakage (26) as well as to the protection afforded by the temporary cement (zinc oxide–eugenol, IRM). This might explain why in the univariate analysis (Table 2) the presence of the temporary cement was seen to have a significant influence only in the Ca(OH)2 group (OR, 0.17; 95% CI, 0.04–0.72; P = .02). In the multiple analysis (Table 3) the potential influencing factor of time span before placement of a permanent restoration after pulp capping ($2 days) also proved to be significant (OR, 0.24; 95% CI, 0.09–0.66; P = .01). In an animal study with follow-up periods of 1 and 2 years, Cox et al (43) showed that bacterial contamination after pulp capping can negatively affect the success of direct pulp capping long-term. The reduction in clinical success if a direct pulp capping is not followed immediately with permanent restoration has been shown in other clinical studies (11, 12). This might be because a permanent restoration protects the tooth structures exposed during the cavity preparation more effectively from microleakage than a temporary restoration. It might also be because bacterial contamination and mechanical irritation can occur when the temporary restoration is exchanged for a permanent one. Because only very few patients came for follow-up with a temporary restoration (Table 1: MTA group, n = 2 and Ca(OH)2 group, n = 1), this potential influential factor could not be meaningfully assessed in the present study. Some clinical studies following up direct pulp-capped teeth showed reduced clinical success rates in carious exposures compared with mechanical exposures (9, 12). This was not shown by the results of the present study (Tables 1 and 2); the differences in the success rates regarding this potential outcome predictor (as a percentage) were not statistically significant (Table 2). The parameter of treatment provider (student versus dentist), which showed a striking trend in the univariate analysis in the Ca(OH)2 group as a potential influential factor on the prognosis (OR, 0.30; 95% CI, 0.09–1.02; P = .05), proved not to be significant in the multiple analysis (GEE model), after adjusting all conspicuous influencing variables (OR, 0.68; 95% CI, 0.29–1.59; P = .37). Thus the results of the present study do not confirm the findings of Baume and Holz (9) that the success rate for direct pulp capping in the skilled hands of dentists is superior to that of students. MTA Compared with Ca(OH)2 for Direct Pulp Capping 811 Clinical Research In the present study, the longer the follow-up period, the more evident the trend became to a decline in the success rate of the teeth in the Ca(OH)2 group compared with the MTA group (Fig. 2, Table 1). However, this was not significant (Table 2), probably as a result of the small number of cases in the study, and should be reevaluated in the future with a larger sample size. The decrease in success rate with increased follow-up time when Ca(OH)2 was used for capping has been observed in many clinical studies (6, 8, 11). Interestingly, no time-dependent decline in the success rate was observed in the present study when MTA was used as capping material. This agrees with the results reported by Bogen et al (31) in a retrospective clinical study on the long-term success (up to 9 years of follow-up) of direct MTA capping. Other brands of MTA besides ProRoot MTA have become available in recent years. Some studies have identified differences between the different brands, eg, regarding composition, particle size, pH value after mixing, radiopacity, and the shape of these materials (44). In the only clinical study to date in which the MTA cements ProRoot MTA and MTA Angelus (Angelus Prod. Odont. Ltda, Londrina, PR, Brazil) were compared for direct pulp capping in human teeth (45), no significant histologic difference was established; however, the time period up to histologic evaluation was only 30 or 60 days, respectively. No conclusions can be drawn from these results regarding long-term clinical success, so that it is not possible to say whether long-term clinical success of direct pulp capping with other brands of MTA would be comparable. The results of the present study point to a difference in the success rates of MTA compared with Ca(OH)2 as a pulp-capping agent, which might be clinically relevant. In univariate analysis this difference was also significant, but when the previously identified potential outcome predictors were included, the difference was only borderline significant. This indicates the important differentiation between clinical relevance and statistical significance. It might be assumed that future clinical studies with larger sample sizes will confirm the superiority of MTA for direct pulp capping, backed up statistically by multiple models. A permanent restoration should always be placed immediately after direct pulp capping, regardless of the capping material used. Acknowledgments The authors would like to thank Mrs Joanna Voerste and Mrs Kirsten Stoik for their assistance in preparation of this manuscript and Mrs Ingrid Mente for her valuable help with this study. References 1. Camp JH, Fuks AB. Pediatrics: endodontic treatment for the primary and young permanent dentition. In: Cohen S, Hargreaves KM, eds. Pathways of the pulp. 9th ed. St Louis, MO: Mosby Elsevier; 2006:822–82. 2. Fuks AB. Vital pulp therapy with new materials for primary teeth: new directions and treatment perspectives. J Endod 2008;34:S18–24. 3. Pfaff P, quoted by Glass RL, Zander HA. Pulp healing. J Dent Res 1949;28:97–107. 4. Underwood. Über die Behandlung der bloßliegenden und erkrankten Zahnpulpa. Der Zahnarzt 1859;14:9. 5. Hermann BW. Dentinobliterationen der Wurzelkanäle nach Behandlung mit Calcium. Zahnärztl Rundschau 1930;39:888. 6. Ahrens G, Reuver J. [Follow-up studies of direct pulp capping in the daily dental practice]. Dtsch Zahnarztl Z 1973;28:862–5. 7. Cvek M. A clinical report on partial pulpotomy and capping with calcium hydroxide in permanent incisors with complicated crown fracture. J Endod 1978;4:232–7. 8. Haskell EW, Stanley HR, Chellemi J, Stringfellow H. Direct pulp capping treatment: a long-term follow-up. J Am Dent Assoc 1978;97:607–12. 9. Baume LJ, Holz J. Long term clinical assessment of direct pulp capping. Int Dent J 1981;31:251–60. 812 Mente et al. 10. Hørsted P, Søndergaard B, Thylstrup A, El Attar K, Fejerskov O. A retrospective study of direct pulp capping with calcium hydroxide compounds. Endod Dent Traumatol 1985;1:29–34. 11. Barthel CR, Rosenkranz B, Leuenberg A, Roulet JF. Pulp capping of carious exposures: treatment outcome after 5 and 10 years: a retrospective study. J Endod 2000; 26:525–8. 12. Al-Hiyasat AS, Barrieshi-Nusair KM, Al-Omari MA. The radiographic outcomes of direct pulp-capping procedures performed by dental students: a retrospective study. J Am Dent Assoc 2006;137:1699–705. 13. Schröder U. Effect of an extra-pulpal blood clot on healing following experimental pulpotomy and capping with calcium hydroxide. Odontol Revy 1973; 24:257–68. 14. Staehle HJ. ed. Calciumhydroxid in der Zahnheilkunde. München: Carl Hanser; 1990. 15. Nyborg H. Healing processes in the pulp on capping: a morphologic study—experiments on surgical lesions of the pulp in dog and man. Acta Odontol Scand 1955;13: 1–130. 16. Kakehashi S, Stanley HR, Fitzgerald RJ. The effects of surgical exposures of dental pulps in germ-free and conventional laboratory rats. Oral Surg Oral Med Oral Pathol 1965;20:340–9. 17. Auschill TM, Arweiler NB, Hellwig E, Zamani-Alaei A, Sculean A. [Success rate of direct pulp capping with calcium hydroxide]. Schweiz Monatsschr Zahnmed 2003;113:946–52. 18. Torabinejad M, Hong CU, McDonald F, Pitt Ford TR. Physical and chemical properties of a new root-end filling material. J Endod 1995;21:349–53. 19. Ida K, Maseki T, Yamasaki M, Hirano S, Nakamura H. pH values of pulp-capping agents. J Endod 1989;15:365–8. 20. Tecles O, Laurent P, Aubut V, About I. Human tooth culture: a study model for reparative dentinogenesis and direct pulp capping materials biocompatibility. J Biomed Mater Res B Appl Biomater 2008;85:180–7. 21. Witherspoon DE. Vital pulp therapy with new materials: new directions and treatment perspectives—permanent teeth. J Endod 2008;34:S25–8. 22. Pitt Ford TR, Torabinejad M, Abedi HR, Bakland LK, Kariyawasam SP. Using mineral trioxide aggregate as a pulp-capping material. J Am Dent Assoc 1996;127:1491–4. 23. Chacko V, Kurikose S. Human pulpal response to mineral trioxide aggregate (MTA): a histologic study. J Clin Pediatr Dent 2006;30:203–9. 24. Sawicki L, Pameijer CH, Emerich K, Adamowicz-Klepalska B. Histological evaluation of mineral trioxide aggregate and calcium hydroxide in direct pulp capping of human immature permanent teeth. Am J Dent 2008;21:262–6. 25. Camilleri J, Pitt Ford TR. Mineral trioxide aggregate: a review of the constituents and biological properties of the material. Int Endod J 2006;39:747–54. 26. Torabinejad M, Rastegar AF, Kettering JD, Pitt Ford TR. Bacterial leakage of mineral trioxide aggregate as a root-end filling material. J Endod 1995;21: 109–12. 27. Aeinehchi M, Eslami B, Ghanbariha M, Saffar AS. Mineral trioxide aggregate (MTA) and calcium hydroxide as pulp-capping agents in human teeth: a preliminary report. Int Endod J 2003;36:225–31. 28. Nair PN, Duncan HF, Pitt Ford TR, Luder HU. Histological, ultrastructural and quantitative investigations on the response of healthy human pulps to experimental capping with mineral trioxide aggregate: a randomized controlled trial. Int Endod J 2008;41:128–50. 29. Accorinte Mde L, Holland R, Reis A, et al. Evaluation of mineral trioxide aggregate and calcium hydroxide cement as pulp-capping agents in human teeth. J Endod 2008;34:1–6. 30. Farsi N, Alamoudi N, Balto K, Al Mushayt A. Clinical assessment of mineral trioxide aggregate (MTA) as direct pulp capping in young permanent teeth. J Clin Pediatr Dent 2006;31:72–6. 31. Bogen G, Kim JS, Bakland LK. Direct pulp capping with mineral trioxide aggregate: an observational study. J Am Dent Assoc 2008;139:305–15. 32. Naik S, Hegde AM. Mineral trioxide aggregate as a pulpotomy agent in primary molars: an in vivo study. J Indian Soc Pedod Prev Dent 2005;23:13–6. 33. Witherspoon DE, Small JC, Harris GZ. Mineral trioxide aggregate pulpotomies: a case series outcomes assessment. J Am Dent Assoc 2006;137:610–8. 34. Aeinehchi M, Dadvand S, Fayazi S, Bayat-Movahed S. Randomized controlled trial of mineral trioxide aggregate and formocresol for pulpotomy in primary molar teeth. Int Endod J 2007;40:261–7. 35. Qudeimat MA, Barrieshi-Nusair KM, Owais AI. Calcium hydroxide vs mineral trioxide aggregates for partial pulpotomy of permanent molars with deep caries. Eur Arch Paediatr Dent 2007;8:99–104. 36. Barrieshi-Nusair KM, Qudeimat MA. A prospective clinical study of mineral trioxide aggregate for partial pulpotomy in cariously exposed permanent teeth. J Endod 2006;32:731–5. 37. Ørstavik D, Kerekes K, Eriksen HM. The periapical index: a scoring system for radiographic assessment of apical periodontitis. Endod Dent Traumatol 1986;2: 20–34. JOE — Volume 36, Number 5, May 2010 Clinical Research 38. Hørsted-Bindslev P, Løvschall H. Treatment outcome of vital pulp treatment. Endodontic Topics 2002;2:24–34. 39. Friedman S. Expected outcomes in the prevention and treatment of apical periodontitis. In: Orstavik D, Pitt Ford T, eds. Essential endodontology: prevention and treatment of apical periodontitis. 2nd ed. Copenhagen: Blackwell Munksgaard Ltd; 2008: 408–69. 40. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159–74. 41. Ørstavik D, Qvist V, Stoltze K. A multivariate analysis of the outcome of endodontic treatment. Eur J Oral Sci 2004;112:224–30. JOE — Volume 36, Number 5, May 2010 42. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986;42:121–30. 43. Cox CF, Bergenholtz G, Heys DR, Syed SA, Fitzgerald M, Heys RJ. Pulp capping of dental pulp mechanically exposed to oral microflora: a 1-2 year observation of wound healing in the monkey. J Oral Pathol 1985;14:156–68. 44. Parirokh M, Torabinejad M. Mineral trioxide aggregate: a comprehensive literature review—part I: chemical, physical, and antibacterial properties. J Endod 2010;36: 16–27. 45. Accorinte ML, Loguercio AD, Reis A, et al. Evaluation of two mineral trioxide aggregate compounds as pulp-capping agents in human teeth. Int Endod J 2009;42:122–8. MTA Compared with Ca(OH)2 for Direct Pulp Capping 813