Magnitude of pulp space narrowing over time and contributing factors in teeth with vital pulp therapy: a retrospective cohort study

Study design and ethics approval

This retrospective cohort study was approved by the Human Experimentation Committee at the Faculty of Dentistry, Chiang Mai University, Chiang Mai, Thailand (registration number 39/2017). The study was conducted in accordance with the Declaration of Helsinki and reported following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [17].

Study participants

The study focused on the permanent mandibular molars of patients aged 6–18 years who underwent VPT, including a protective liner and base (LB), direct pulp capping (DPC), partial pulpotomy (PP), and coronal pulpotomy (CP), at the Pediatric Dentistry Clinic, Faculty of Dentistry, Chiang Mai University, Thailand from September 2012 to June 2022. The contralateral sound molars of the same patients served as controls.

For comprehensive treatment planning, preoperative bitewing radiographs on both sides were obtained for all patients; these served as the baseline for the control group. VPT was performed on the deep carious tooth within 1 to 2 weeks following the previously described VPT protocol in a published article from the same institution [18,19]. The immediate post-operative radiograph was taken and used as the baseline for the VPT group. Follow-up appointments were scheduled approximately every six months, during which bilateral bitewing radiographs were obtained, consistent with American Academy of Pediatric Dentistry recommendations for high-caries-risk patients [20]. The radiographs were positioned parallel to the crowns of the maxillary and mandibular teeth with the resultant images showing the contact areas as “opened” [21].

Sample size calculation

Sample size calculations were conducted using the program from http://powerandsamplesize.com [22], based on prior studies reporting incidences of pulp space narrowing following VPT: 35% in CP [4] and 5% in PP [1], thus at least 20 teeth was required for each treatment group. However, this study included all mandibular molars that met the inclusion criteria.

Inclusion and exclusion criteria

To be included in this study, the VPT-treated mandibular molars must have (1) a baseline and at least one follow-up bitewing radiograph of good quality; (2) no restorations or devices that interfere with the radiological evaluation of the pulp chamber; (3) all independent variables (predictors) from patient charts, including age (years, months), sex (male, female), follow-up period (recorded at one or more of the following intervals, as available: 6, 12, 18, 24, 30, 36, 48, 60, and 72 months); a preoperative pulpal diagnosis (normal pulp, reversible pulpitis, irreversible pulpitis) [23]; preoperative periapical lesion status (absence, presence); root formation status (mature, immature); type of VPT (LB, DPC, PP, CP); pulp-dressing material (ProRoot MTA, Dentsply Sirona, Tulsa, OK, USA; Biodentine, Septodont, Saint‑Maur‑des‑Fossés, France); type of irrigant (sodium hypochlorite, normal saline solution); site of exposure (mesial, distal); and VPT outcome (success, failure).

Regarding the VPT outcome, treatment was considered successful if the treated tooth demonstrated functional clinical and radiographic survival. Clinically, the tooth did not have pulpitis symptoms, abnormal mobility, or fistula. Radiographically, the tooth showed continued root formation or improved periapical conditions, the absence of prominent lesions, and no internal or external root resorption.

The contralateral sound molars from the same patients with VPT were screened and included as controls if they had (1) the baseline and at least one bitewing radiograph from routine check-ups, and (2) all required independent variables (age, sex, follow-up period, and root formation status).

Bitewing radiographs were excluded if the necessary reference points in Figure 1A, including the cementoenamel junction points on the distal and mesial external tooth surfaces, the highest point of the pulp chamber floor, and the highest point of the furcation of the tooth, were not identifiable by the primary investigator (AB).

Figure 1.

Reference points and lines on a bitewing radiograph used in this study. (A) Reference points: (1) blue points were the cementoenamel junction points on the distal and mesial external tooth surfaces; (2) a red point was the highest point of the pulp chamber floor; and (3) a yellow point was the highest point of the furcation. (B) Reference lines (RF line): the line connecting the mesial and distal cementoenamel junction points; B0 is a horizontal line/distance, paralleling to the RF line, between the mesial and the distal external root surfaces at the level of the fornix of furcation; C0 is the vertical and perpendicular line/distance between the RF line and the B0 line; A0 is the horizontal line/distance, paralleling to the RF line, between the mesial and the distal external root surfaces at the level of the occlusal one fourth of the C0 distance. (C) RF lines: A1 is the horizontal distance of pulp chamber from the mesial to the distal external pulp chamber surfaces at the level of the occlusal one fourth of the C0 distance; B1 is the horizontal distance of the mesial canal at the level of the fornix of furcation; B2 is the horizontal distance of the distal canal at the level of the fornix of furcation and C1 is the vertical distance between the floor of pulp chamber point and the fornix of furcation point.

Data collection and outcome

The bitewing radiographs that met the inclusion criteria were saved as JPEG files and then randomly placed into sets, each containing 25 radiographs. The blinded, calibrated examiners independently assessed each set of radiographs using the Chiang Mai University–Pulp Space Measurement (CMU–PSM) program on the same computer with the same settings in the same ambient environment. The evaluation duration was limited to 40 minutes per set.

To avoid discrepancies in radiographic image expansion or distortion, the semi-automated CMU–PSM program was specifically developed to linearly measure the pulp space using four pulp/tooth proportion values on bitewing radiographs. When using the CMU–PSM program, the evaluators manually selected the reference points in Figure 1A; the program then automatically created the reference lines in Figure 1B and C. The evaluator checked and manually adjusted the margin of the reference lines as appropriate. Finally, the CMU–PSM program automatically calculated four pulp/tooth proportion values, including pulp chamber width (A1/A0), pulp canal widths (mesial, B1/B0, and distal, B2/B0), and pulpal floor thickness (C1/C0). However, the A1/A0 value was not available for teeth treated with CP because their pulp chamber was completely occupied with pulp-dressing material. All pulp/tooth proportion values were exported to a Microsoft Excel table, and the change (∆) in these values from the baselines was computed as follows:

∆A1/A0 = A1/A0 (baseline) − A1/A0 (follow-up)

∆B1/B0 = B1/B0 (baseline) − B1/B0 (follow-up)

∆B2/B0 = B2/B0 (baseline) − B2/B0 (follow-up)

∆C1/C0 = C1/C0 (baseline) − C1/C0 (follow-up)

Sample size for reliability analyses was estimated with an online intraclass correlation coefficient (ICC) calculator (wnarifin.github.io/ssc/ssicc.html) [24] targeting an expected ICC of 0.80 with ± 0.10 precision at the 95% confidence level. This yielded required samples of 51 radiographs for intra-examiner reliability (two ratings per subject) and 29 radiographs for inter-examiner reliability. The radiographs were independently evaluated by four blinded evaluators (three faculty members and one master’s student in pediatric dentistry), with reassessment after a 2-week interval for intra-examiner reliability.

The primary investigator (AB) extracted all independent variables from patients’ charts. All values were derived from radiographic measurements performed by the calibrated independent evaluators. The dependent variables were the changes from baseline in four pulp/tooth proportion values (∆A1/A0, ∆B1/B0, ∆B2/B0, ∆C1/C0) assessed at each follow-up. In cases of total PCC, defined in this study as the complete radiographic absence of a discernible pulp space throughout the pulp chamber and root canal, quantitative measurements could not be performed, and these cases were recorded as total PCC.

The study addressed two main outcomes: (1) differences in changes of pulp space from baseline, by measuring the pulp/tooth proportion at different follow-ups between the VPT-treated and contralateral sound molars and (2) contributing factors associated with changes of the pulp space from baselines in the VPT-treated teeth.

Statistical methods

For intra-examiner reliability, the four pulp/tooth proportion values from the first and second measurements from each evaluator were analyzed using the ICC (two-way mixed model; type, single measurement; definition, absolute agreement). For inter-examiner reliability, the same four proportion values from all evaluators were analyzed using the ICC (two-way random model; type, single measure; definition, absolute agreement). ICCs were calculated using IBM SPSS ver. 19.0 (IBM Corp, Armonk, NY, USA), with an acceptable score set at 0.8 (good reliability), and scores above 0.9 considered excellent reliability [25].

Longitudinal missing data were handled using the last observation carried forward (LOCF) approach [26]. To evaluate the change in pulp/tooth proportion values from baseline, linear mixed-effects models were fitted, assuming a Gaussian distribution for the outcome. The models included fixed effects for treatment type (VPT vs. control), follow-up time (in months, modeled as a continuous variable with both linear and quadratic terms), and their interaction. Additional fixed covariates were included if they were potential confounders of the main determinants. To account for the hierarchical and longitudinal data structure, random effects were specified at the tooth level, with a random slope for time to model repeated measurements within each tooth. No additional patient-level random effects were included because each tooth represented the unit of analysis. Wald tests were applied to assess the statistical significance of associations between independent variables and each outcome at each follow-up time point. Stata software ver. 16.0 (StataCorp LLC, College Station, TX, USA) was used for analysis.

Demographic data

Of 635 teeth from 400 patients, a total of 134 VPT-treated teeth from 131 patients, with 382 bitewing radiographs, met the inclusion criteria. The follow-up period ranged from 6 to 84 months (mean, 27.12 ± 17.67 months). Missing data amounted to 58.2% of the total. From the same patients who received VPT, 20 contralateral sound molars with 60 bitewing radiographs at different recall periods were included in the control group.

Demographic data of patients with VPT-treated teeth and the control group are summarized in Tables 1 and 2, respectively. Table 3 shows the incidence of total PCC, which was observed in 14/134 teeth (10.4%), involving a higher number of teeth with PP (7/134, 5.2%) and CP (6/134, 4.5%). Only one VPT failure was found in this total PCC group (1/14, 7.1%). This failed tooth also exhibited a fracture in the composite resin restoration.

Table 1.

Demographic data of the study group with vital pulp therapy-treated teeth

Table 2.

Demographic data of the control group

Table 3.

Incidence of total pulp canal calcification and failure rate from different types of vital pulp therapy

Comparison of the changes from baseline in pulp/tooth proportion values between molars with different types of vital pulp therapy and controls

Figure 2 illustrates the change from baseline in pulp/tooth proportion between teeth treated with different types of VPT and controls. A general trend of progressive decreases in all four pulp/tooth proportion values over time was observed. Notably, for ∆A1/A0, the proportion changes were significantly greater in the PP group compared to the control group at 12, 18, 24, 30, 36, and 48 months in Figure 2A. For ∆B1/B0, the changes were significantly greater in the CP group than in the controls at all times, except for 72 months in Figure 2B. For ∆B2/B0, the CP group also showed significantly greater change compared to the controls at all follow-up points in Figure 2C. In contrast, the changes in ∆C1/C0 of each VPT group were not significantly different from those of the control group at 6–60 months; however, the LB group displayed significantly lower change values than the controls at 72 months in Figure 2D. The fixed- and random-effects models for changes from baseline in pulp/tooth proportion values between teeth treated with different types of VPT and controls, including their marginal predictions and pairwise contrast p-values, are presented in Supplementary Tables 1–12.

Figure 2.

Comparison of the changes from baseline in pulp/tooth proportion values (A–D) between teeth treated with different types of vital pulp therapy and controls. LB, protective liner and base; DPC, direct pulp capping; PP, partial pulpotomy; CP, coronal pulpotomy. ^*p < 0.05, statistically significant.

Differences in the changes from baseline of pulp/tooth proportion values between the subgroups

The factors that contributed to significant differences between subgroups in change from baseline of pulp/tooth proportion values were VPT type and preoperative pulpal diagnosis.

Type of vital pulp therapy

Figure 3 illustrates changes from baseline in pulp/tooth proportion values in the DPC, PP, and CP groups against those in the LB group. Notably, for the ∆A1/A0 value, changes were significantly greater in the PP group than in the LB group at 12, 18, 24, 30, 36, 48, 60, and 72 months in Figure 3A. In contrast, the changes in ∆B1/B0 of the CP, PP, and DPC groups did not significantly differ from those of the LB group at any follow-up point in Figure 3B. Furthermore, for ∆B2/B0, decreases were significantly greater in the CP group than in the LB group at all follow-up points, except at 6 months (Figure 3C). Finally, for ∆C1/C0, decreases were significantly greater in the PP group than in the LB group at 60 and 72 months in Figure 3D. Figure 4 illustrates different magnitudes of pulp space narrowing over time in different types of VPT cases. The fixed- and random-effects models for changes from baseline in pulp/tooth proportion values (ΔA1/A0) in the DPC, PP, and CP groups compared with those in the LB group, including their marginal predictions and pairwise contrast p-values, are presented in Supplementary Tables 13–24.

Figure 3.

Comparison of the changes from baseline in pulp/tooth proportion values (A–D) in the DPC, PP, and CP groups against those in the LB group. LB, protective liner and base; DPC, direct pulp capping; PP, partial pulpotomy; CP, coronal pulpotomy. ^*p < 0.05, statistically significant.

Figure 4.

Radiographs demonstrate pulp space narrowing over time following different types of vital pulp therapy, a greater magnitude of pulp space narrowing was observed in teeth treated with partial and coronal pulpotomy.

Preoperative pulpal diagnosis

Figure 5 illustrates changes from baseline of pulp/tooth proportion values in the different groups of preoperative pulpal diagnosis, against the normal pulp group. For the ∆A1/A0 value, changes were significantly greater in the irreversible pulpitis group compared to the normal pulp group at 18, 24, 30, and 36 months in Figure 5A. The changes in ∆B1/B0 for the irreversible and reversible pulpitis groups did not differ significantly from those in the normal pulp group at any follow-up point in Figure 5B. For ∆B2/B0, the change was significantly greater in the irreversible pulpitis group compared to the normal pulp group at 72 months in Figure 5C. In terms of ∆C1/C0, the decreases in proportion were significantly greater in the irreversible pulpitis group than in the normal pulp group at 48, 60, and 72 months in Figure 5D. The fixed- and random-effects models for changes from baseline in pulp/tooth proportion values (ΔA1/A0) across different preoperative pulpal diagnosis groups, compared with the normal pulp group, including their marginal predictions and pairwise contrast p-values, are presented in Supplementary Tables 25–36.

Figure 5.

Comparison of the changes from baseline of pulp/tooth proportion values (A–D) in the different groups of preoperative pulpal diagnosis, against the normal pulp group. ^*p < 0.05, statistically significant.