Three‐Year Clinical and Economic Evaluation of Selective Caries Removal and Full Pulpotomy for Extensive Caries: An Exploratory Randomised Controlled Trial

ABSTRACT

Aim

The aim of this study was to compare pulp survival following selective caries removal (SCR) and full pulpotomy treatment in teeth with extensive caries over 3 years.

Methodology

This two‐arm, exploratory randomised controlled trial included vital mature permanent teeth with extensive primary or secondary caries diagnosed radiographically as being at least 75% or more into the thickness of dentine, without clinical signs of symptomatic irreversible pulpitis or radiographic evidence of a periapical lesion. Carious teeth were randomly allocated to receive either SCR or full pulpotomy. All teeth were reviewed clinically and radiographically at 6 months, 1 year, and 3 years post‐treatment. Log‐rank tests and Cox proportional hazards regressions were used to compare the outcomes of SCR and pulpotomy, adjusting for clustering using a robust variance estimator. Cost‐effectiveness analysis was carried out using the healthcare system and societal perspectives. Significance level was set at 5%.

Results

At 3 years, 44/58 teeth (75.9%) in the SCR group and 37/55 teeth (67.3%) in the full pulpotomy group were reviewed. A total of 13 teeth in the SCR group (29.5%) and 4 teeth in the pulpotomy group (10.8%) required further intervention with root canal treatment (RCT) or extraction (p = 0.039). Survival rates were 74% for SCR and 89% for pulpotomy (p = 0.041). Overall, 79.0% of teeth treated with either SCR or pulpotomy survived without requiring further intervention over the 3‐year period. In this study with more than two‐thirds of the enrolled teeth classified as extremely deep lesions, pulpotomy was more cost‐effective than SCR in the management of extensive caries above a willingness‐to‐pay threshold of $917 and $928 to avoid one tooth extraction or RCT from the healthcare system and societal perspective, respectively.

Conclusion

Full pulpotomy demonstrated greater effectiveness than SCR in avoiding further intervention over a 3‐year period in teeth with extensive caries, of which the majority were extremely deep lesions. Further long‐term studies and broader health economic evaluations are warranted to guide clinical decision‐making and policy development.

Trial Registration

ClinicalTrials.gov identifier: NCT04672070

1. Introduction

Traditionally, teeth diagnosed with extensive dentinal caries and irreversible pulpitis are managed with root canal treatments (RCT) (European Society of Endodontology ²⁰⁰⁶). However, RCT can weaken the tooth structure, increasing the risk of fracture (Kishen 2006; Lang et al. ²⁰⁰⁶; Reeh et al. ¹⁹⁸⁹; Zelic et al. ²⁰¹⁵) and can be costly for patients (Wigsten et al. 2024). Recently, the focus of various international professional bodies has shifted towards minimally invasive approaches that prioritise the preservation of pulp vitality, reduce the need for extensive procedures and potentially enhance long‐term oral health outcomes (Duncan et al. 2019). The International Caries Consensus Collaboration has advocated for techniques such as selective caries removal (SCR), stepwise removal, and atraumatic restorative treatment for managing deep lesions at risk of pulp exposure (Schwendicke et al. 2016). Similarly, the European Society of Endodontology (2006) recommends biologically based, minimally invasive therapies aimed at preserving pulpal health and preventing apical periodontitis (Duncan et al. 2019). These approaches are collectively referred to as vital pulp therapies. Among these, SCR and pulpotomy are two techniques that are commonly utilised and distinct approaches to the management of extensive caries. However, their relative success in preserving pulpal health in permanent teeth remains underexplored.

SCR has demonstrated high success rates, with a systematic review reporting 88% restoration survival and 96% pulp vitality at 2 years (Hoefler et al. 2016). However, longer term studies show a decline in the proportion of teeth maintaining vitality, from 82%–83% at 5 years to 63% at 10 years (Maltz et al. 2011, 2018). Since exposing the pulp can irreversibly damage odontoblasts (Bjørndal et al. 2019), SCR is recommended to reduce pulpal stress and minimise unnecessary tooth structure removal, helping to avoid the restorative cycle. The long‐term effectiveness of SCR is uncertain, as bacteria have been observed in both ‘leathery’ dentine as well as hard carious dentine (Langeland and Langeland 1968; Langeland ¹⁹⁸¹; Ricucci et al. ²⁰²⁰). These remnant bacteria may cause ongoing pulpal irritation, with potential pulp necrosis if immunity is compromised or the seal fails (Ricucci et al. 2019). Advances in understanding pulp repair and the development of improved bioactive materials (Nair et al. 2008) have led to pulpotomy being considered a viable alternative to root canal therapy for carious pulp exposures (Simon et al. 2013; Tan et al. ²⁰²⁰), even for teeth showing signs of irreversible pulpitis (Taha et al. 2017; Taha and Khazali ²⁰¹⁷; Careddu and Duncan ²⁰²¹). Success rates for pulpotomy in permanent teeth vary, with full pulpotomy achieving a 97.4% clinical success rate at 12 months and partial pulpotomy showing a 75% success rate in irreversible cases and 98% in reversible cases (Taha and Khazali 2017; Elmsmari et al. ²⁰¹⁹). While pulpotomy has shown comparable success rates to root canal treatment in mature molars with carious pulp exposures (Asgary et al. 2017; Galani et al. ²⁰¹⁷), evidence is limited and studies have failed to report factors that may determine treatment success, such as preoperative pulp status and caries depth (Cushley et al. 2019). Early pulpotomy failures are frequently attributed to misdiagnosed pulpitis severity, while late failures are linked to asepsis and restoration quality (Zanini et al. 2016; Bjørndal et al. ²⁰¹⁹; Tan et al. ²⁰²⁰). Most studies comparing minimally invasive treatments are short‐term. Previous findings from a 1‐year randomised controlled trial (Chua et al. 2023) suggest that pulpotomy may offer higher short‐term survival rates than SCR for teeth with extensive caries. While radiographic caries depth was not found to be a statistically significant prognostic factor potentially due to insufficient power, a high proportion of failures occurred in teeth with extremely deep caries (Chua et al. 2023), in line with previous findings that these lesions are associated with bacterial penetration into tertiary dentine and more extensive pulpal inflammation (Bjørndal et al. 2019; Demant et al. ²⁰²¹). However, questions remain about the relative longer term clinical outcomes of both treatments for deep and extremely deep caries, due to the lack of direct comparisons within a single trial setting. Larger restorations due to pulpotomy procedures may lead to a higher rate of restoration failures over time (Zanini et al. 2016; Bjørndal et al. ²⁰¹⁹; Tan et al. ²⁰²⁰), while SCR presents a greater risk of pulpal failure over time, increasing the likelihood of costly RCTs and crowns (Chua et al. 2023; Taha et al. ²⁰²⁴). Uncertainties about the long‐term outcomes of minimally invasive treatments have direct consequences for patient satisfaction, oral health–related quality of life (QoL), and cumulative treatment costs (Wigsten et al. 2021).

Few studies have addressed the cost‐effectiveness of vital pulp therapies. Economic evaluations (EEs) have become essential tools for decision‐making in health care, ensuring efficient resource allocation, especially under budget constraints (Gafni et al. 2003; Glied and Teutsch ²⁰¹⁶; Schwendicke and Herbst ²⁰²³). They allow clinicians and policymakers to compare the costs and benefits of different interventions, ensuring the selection of options that provide the best value for resources (Gafni et al. 2003; Vallejos et al. ²⁰¹⁴). This is particularly relevant in oral health care, where resources are limited (Peres et al. 2019). Chua et al. ²⁰²³; reported that pulpotomy generally incurs a higher initial cost and takes longer to complete than SCR, although it has a higher initial success rate. Cost‐effectiveness analysis (CEA) is a critical EE tool widely employed to identify oral health interventions that provide the greatest health benefits within a given budget (Mariño et al. 2020). Payers, including insurers and patients, are particularly concerned with the costs of initial treatments and potential re‐treatments; hence, understanding the long‐term implications of these treatment decisions is essential for optimising resource allocation and improving patient outcomes within the healthcare system (Schwendicke and Herbst 2023).

Among studies on the cost‐effectiveness of endodontic treatments, most have relied on Markov simulation models, with success rates extrapolated from the literature rather than derived from real‐world clinical outcomes (Emara et al. 2020; Schwendicke et al. ²⁰¹³; Schwendicke et al. ²⁰¹⁴). The lack of robust clinical data with mid‐ or long‐term outcomes and the heterogeneity in study designs significantly hinders health economic analyses in endodontics (Lucena et al. 2017; Duncan et al. ²⁰²⁰; El Karim et al. ²⁰²¹; Nagendrababu et al. ²⁰²¹). Thus, there is a need for clinical trials that assess and compare both clinical and economic outcomes of vital pulp therapies, including SCR and pulpotomy, over extended follow‐up periods.

This two‐armed, parallel‐group randomised clinical trial aimed to compare pulp survival in teeth with extensive caries treated with SCR or full pulpotomy, and to evaluate the comparative effectiveness and cost‐effectiveness of these two treatment approaches.

2. Materials and Methods

This randomised controlled trial was reported according to the Preferred Reporting Items for Randomised Trials in Endodontics (PRIRATE) 2020 guidelines (Nagendrababu et al. 2020), and the economic evaluation was reported in accordance with the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) 2022 guidelines (Husereau et al. 2022).

2.1. Study Design and Setting

This study is a 3‐year follow‐up of a previously piloted two‐arm, exploratory randomised controlled trial carried out at a tertiary hospital in Singapore (Chua et al. 2023). Patients with vital, mature permanent teeth with deep or extremely deep primary or secondary caries extending at least 75% into the dentine were screened for eligibility, and informed consent was obtained. Only medically healthy patients (American Society of Anesthesiologists physical status classification I or II) aged 21–65 years, with restorable teeth and mature apices and a provisional diagnosis of reversible pulpitis, were included in the study. Teeth were randomised to receive either SCR or full pulpotomy treatment, with up to 60 teeth in each group. Allocation was performed by a research assistant not involved in the treatment. Randomised patient codes were generated using a computer‐based permuted block randomisation approach and concealed in opaque envelopes, which were only opened immediately prior to treatment allocation. The sample size calculation, patient enrollment, randomisation, and treatment protocol have been previously described in detail (Chua et al. 2023). Outcomes were assessed over a 3‐year period.

Preoperative parameters collected included the patient's age, gender (male or female), tooth type (incisor, canine or molar), caries status (primary or secondary), depth of caries (deep or extremely deep), location of caries (occlusal, proximal or both), and tenderness to percussion and palpation. Deep caries was defined as caries reaching the inner quarter of dentine, with a radiographically detectable zone of hard or firm dentine between the lesion and the pulp. Extremely deep caries referred to lesions that penetrated the full thickness of dentine, where pulp exposure was considered unavoidable during operative treatment (Duncan et al. 2019). Intraoperative parameters recorded included the time to achieve haemostasis (for the pulpotomy group), type of restoration (amalgam or composite resin) and the treatment time.

All teeth were reviewed at 6 months, 1 year and 3 years post‐treatment. Data collected included subjective symptoms, pulp status based on response to pulp tests (cold test and electric pulp test), responses to percussion and palpation tests, probing depths, tooth mobility, integrity of the coronal restorations, and presence or absence of recurrent caries. Unplanned treatments prior to scheduled review appointments, and indicated treatments (e.g., RCT and restoration repair) were recorded as well. Periapical radiographs were taken before and after initial treatment as well as at each review visit using the long cone paralleling technique. The radiographs were evaluated for widened periodontal ligament space or development of periapical radiolucencies at the periradicular region of treated teeth, as well as signs of restoration defects or recurrent caries.

Tooth outcomes were categorised as either ‘survived’ or ‘RCT or extraction needed’ using baseline and follow‐up clinical and radiographic data (Chua et al. 2023). Teeth were classified as ‘RCT needed’ if they were symptomatic (e.g., experienced spontaneous pain, tenderness to percussion, swelling or sinus tract) and/or if radiographs showed a periapical lesion with a radiolucency exceeding twice the width of the periodontal ligament space. Conversely, asymptomatic teeth with a widened periodontal ligament space less than twice its normal width were classified as ‘survived.’ Teeth with catastrophic fractures that were deemed unrestorable were categorised as ‘extraction needed’. Positive response to pulp sensibility tests was considered a proxy of pulpal health for teeth treated with SCR. For teeth treated with full pulpotomy, a lack of response to sensibility tests was not considered pathological unless accompanied by symptoms or radiographic signs of disease.

3. Cost Effectiveness Analysis

A trial‐based economic evaluation was carried out. Cost‐effectiveness analysis was performed with the key outcome as the avoidance of RCT or extraction. The time horizon was 3 years. The incremental cost‐effectiveness ratio (ICER) quantifies the additional cost required to achieve an additional unit of health benefit, such as an additional year of tooth retention (Drummond et al. 2005):

$$\text{ICER}=\left({\text{Cost}}_1–{\text{Cost}}_2\right)/\left({\text{Effectiveness}}_1-{\text{Effectiveness}}_2\right)$$

In this context, Cost₁ and Effectiveness₁ represent the cost and health outcome of the intervention being evaluated (i.e., pulpotomy), while Cost₂ and Effectiveness₂ correspond to the cost and outcomes of the comparator (i.e., SCR), respectively.

Both the healthcare system and societal perspectives were adopted. Total costs, direct costs and indirect costs were tabulated and reported in 2020 Singapore dollars (SGD). The exchange rate between SGD and USD was 1 SGD = 0.75 USD as of December 31, 2020 (Oanda Currency Converter). From the healthcare system perspective, direct medical costs included unsubsidised treatment fees charged in the National University Centre for Oral Health Singapore (NUCOHS), the cost of initial treatment (SCR/pulpotomy), consultation fees during review visits, the cost of radiographs and restorations, as well as the cost of any additional procedures during the follow‐up period. From the societal perspective, additional indirect cost components were considered beyond direct medical costs. Indirect costs included productivity losses which were tabulated by multiplying the total treatment time (in hours) by the median hourly wage in each year derived from the median monthly salary in Singapore from 2020 to 2023 (Ministry of Manpower, n.d.). The total treatment time was calculated by summing the time spent at the dental clinic to complete treatment, the waiting time per visit, and the travel time per round trip to the clinic. For initial treatments, the time taken for SCR and full pulpotomy was recorded for each patient. Follow‐up visits were standardised to account for travel and waiting times, with a duration of 1 h allocated for review visits and restoration repairs, and 2 h for pulpectomies or RCTs. Transport costs were set at $3 for each round trip to the hospital. All cost parameters were converted to SGD in the base year (2020) using the Consumer Price Index. A discount rate of 3% for both outcomes and costs (Agency for Care Effectiveness ²⁰¹⁹) was applied.

3.1. Statistical Analyses

All statistical analyses were performed using R version 4.2 (R Core Team), with statistical significance set at p < 0.05. The unit of analysis was the tooth. A modified intention‐to‐treat (mITT) approach was adopted (Chua et al. 2023). All teeth were first randomised to either the SCR or pulpotomy treatment arm. Teeth that did not complete the assigned treatment at baseline (e.g., non‐vital pulp encountered in the pulpotomy group) were excluded from the analysis (Figure 1). Descriptive statistics were used to summarise patient demographics and baseline characteristics. Multiple imputation was carried out to account for missing data on outcomes and costs incurred at the third‐year follow‐up. The imputation model included baseline characteristics (demographics, clinical data, cost components at each review interval), and outcomes such as RCT or extraction avoidance, preoperative (caries depth, percussion, etc.), and intra‐operative variables (restorative material, treatment time, etc.). Thirty imputed datasets were generated to account for variability. Each dataset was analysed independently, and results were pooled using Rubin's Rules (Little and Rubin 1987) to produce final estimates with confidence intervals. Pearson's chi‐square test assessed the homogeneity between treatment groups. Kaplan–Meier survival curves and cluster‐weighted log‐rank tests were used to compare survival over time. Cox proportional hazards regression was applied to examine survival outcomes, with clustering adjusted using a robust variance estimator.

FIGURE 1.

Study flowchart. This diagram outlines the number of subjects at each stage, including the reasons for exclusion, randomisation into the two treatment groups (SCR and pulpotomy), and follow‐up data at various intervals (6 months, 1 year and 3 years).

For the cost‐effectiveness analysis, a generalised linear model with a robust sandwich estimator was applied to account for clustering due to patients contributing multiple teeth, with Gamma distributions for costs and logistic regressions for the probability of avoiding RCT or extraction. The model adjusted for gender, age group and caries depth to account for variations in biological response, healing capacity, and differences in disease severity that could influence treatment outcomes. Incremental costs to avoid an additional RCT or extraction for pulpotomy, compared to SCR, were reported. As part of a sensitivity analysis, Bayesian models were fitted using the brms package in R, which utilises Stan for Hamiltonian Monte Carlo (HMC) sampling, with weakly informative priors Normal (0, 2) for fixed effects to balance robustness and computational stability. Markov Chain Monte Carlo (MCMC) sampling ran 4 chains with 4000 iterations (1000 warmup), and convergence was assessed via trace plots, R‐hat statistics (< 1.01), and effective sample sizes (> 1000). Primary inferences were drawn from the mITT analysis, without imputation for missing data (i.e., complete case analysis). Results of the cost‐effectiveness analysis from the imputed models were reported in the Appendix S1. In addition, subgroup analyses were conducted according to lesion depth (deep and extremely deep caries) for both survival and cost‐effectiveness outcomes.

Deterministic sensitivity analyses were carried out to assess the sensitivity of the model to variations in input parameters. One way sensitivity analyses were carried out for median annual income, treatment durations, and the unit costs of treatments such as RCT, SCR and pulpotomy. Median income was varied by setting the low value at 80% and the high value at 120% of the base‐case value for each year (2020–2023). To account for uncertainty in cost parameters, unit costs for RCT, SCR and pulpotomy were varied by ±20%. For treatment duration, the time spent on RCTs was varied by ±50% (1.00 to 3.00 h) to capture procedural variability, which can be influenced by case complexity, operator skill, and anatomical factors, impacting both treatment costs and productivity loss. Similarly, the duration of review visits was adjusted by ±20% to account for differences in post‐treatment monitoring and patient‐specific follow‐up needs. In addition to variations in cost and time, the effect difference between pulpotomy and SCR in avoiding RCT or extraction was varied by ±30% (0.081–0.151) to assess its influence on cost‐effectiveness.

To characterise uncertainty around mean cost‐effectiveness estimates, Monte Carlo bootstrapping with 5000 replications was performed to generate scatter plots on the cost‐effectiveness plane. A cost‐effectiveness acceptability curve (CEAC) was then constructed to illustrate the probability that the intervention was cost‐effective at each willingness‐to‐pay (WTP) threshold.

4. Results

A total of 101 patients were recruited with informed consent and randomised into SCR and full pulpotomy groups, with 11 patients contributing more than one tooth (Figure 1). Of 120 teeth, six teeth were excluded at baseline because they did not complete the allocated treatment and one tooth was excluded as it was associated with severe periodontal disease, leaving 58 teeth in the SCR group and 55 teeth in the pulpotomy group that completed treatment. All 58 teeth in the SCR group and 55 teeth in the pulpotomy group were included in the modified intention‐to‐treat analysis. At the end of the 3‐year follow‐up period, 44/58 teeth (75.9%) in the SCR group and 37/55 teeth (67.3%) in the pulpotomy group were reviewed (Figure 1). There were no significant differences in baseline characteristics, including lesion depth distribution or restorative needs, between the two groups. However, the time taken for completion of pulpotomy (90.0 [22.5] min) was significantly longer than that for SCR (50 [33.75] min) (p < 0.001) (Table S1). The majority of lesions were classified as extremely deep, accounting for 70.7% in the SCR group and 72.7% in the pulpotomy group (Table S1).

4.1. Pulpal and Restorative Failure

Four out of 37 teeth in the full pulpotomy group (10.8%) required further intervention of RCT or extraction at the end of the follow‐up period, compared to 13/44 teeth (29.5%) in the SCR group. The details of these cases are described in Table S2. With multiple imputation, the survival rate of full pulpotomy declined from 98.1% at 1 year to 89.1% at 3 years, while SCR showed a decrease from 89.7% to 74.1% over the same period. The survival rates at the end of the follow‐up period for both groups, with and without multiple imputation of data that was missing due to loss to follow‐up, are presented in Table 1. Overall, 79.0% of teeth survived without requiring further intervention over a period of 3 years. With multiple imputation, this figure was 81.4%. A statistically significant difference was noted in the survival rates between the SCR and pulpotomy groups (p < 0.05) (Figure 2). In the multivariable‐adjusted Cox regression, full pulpotomy had a higher survival rate at 3 years [Adjusted HR 2.58 (95% confidence interval (CI): 0.80–8.30)] (Table 2). In addition to pulpal failure, both the SCR and pulpotomy groups experienced a comparable number of restorative failures across the follow‐up period (Table S3).

TABLE 1.

Survival rates with and without multiple imputation.

Characteristic	Pulpotomy (N = 55)	SCR (N = 58)	p
Without Multiple imputation
Outcome at 3 years
RCT/extraction required	4 (10.8%)	13 (29.5%)	0.039*
Surviving	33 (89.2%)	31 (70.5%)
Missing	18	14
With multiple imputation
Outcome at 3 years
RCT/extraction required	6 (10.9%)	15 (25.9%)	0.041*
Surviving	49 (89.1%)	43 (74.1%)

^* p < 0.05.

FIGURE 2.

Kaplan–Meier survival curve presenting cumulative survival probabilities of SCR and pulpotomy with and without multiple imputation. The solid lines indicate the estimated survival probabilities, while the dotted lines represent the 95% confidence intervals. The ‘+’ marks denote censored observations. The table below the graph shows the number of teeth at risk at each time point, with the cumulative number of events indicated in parentheses.

TABLE 2.

Cox regression of pulp survival comparing pulpotomy and SCR at 3 years.

Variable	Unadjusted HR (95% CI)	p	Adjusted HR (95% CI)	p
Allocated treatment group
Pulpotomy	Ref.	—	Ref.	—
Selective caries removal	2.80 (0.91–8.61)	0.098	2.58 (0.80–8.30)	0.144
Gender
Female	Ref.	—	Ref.	—
Male	0.80 (0.32–2.00)	0.641	0.69 (0.28–1.67)	0.427
Age group
≤ 40 years	Ref.	—	Ref.	—
> 40 years	0.40 (0.15–1.11)	0.097	0.43 (0.16–1.19)	0.124
Tooth type
Incisor/Canine	Ref.	—
Molar	0.62 (0.07–5.16)	0.663
Premolar	1.12 (0.14–8.99)	0.919
Caries status
Primary caries	Ref.	—
Secondary caries	0.55 (0.15–1.98)	0.375
Caries depth
Deep	Ref.	—	Ref.	—
Extremely deep	2.05 (0.57–7.34)	0.292	2.22 (0.64–7.83)	0.247
Percussion sensitivity
Not tender	Ref.	—
Tender	1.92 (0.41–9.03)	0.423
Restoration type
Amalgam	Ref.	—
Composite resin	1.04 (0.36–3.01)	0.946
Treatment time
≤ 60 min	Ref.	—
61–90 min	0.55 (0.20–1.53)	0.273
> 90 min	0.61 (0.18–2.13)	0.453

In the subgroup analysis of extremely deep lesions, full pulpotomy demonstrated statistically significantly higher survival rates compared to SCR (p < 0.05) (Figure S1). In contrast, in the subgroup of deep lesions, no statistically significant differences in survival were observed between SCR and full pulpotomy (p > 0.05) (Figure S2).

4.2. Cost Effectiveness Analysis

Full pulpotomy incurred higher costs but also had greater effectiveness in avoiding RCT and extraction compared to SCR, regardless of whether multiple imputations were carried out. In the complete case analysis, adopting a societal perspective, the incremental cost of full pulpotomy compared to SCR per additional RCT/extraction avoided was $927.61, with a mean cost difference of $158.34 (95% CI: −108.47 to 425.15). From a healthcare system perspective, the incremental cost of full pulpotomy compared to SCR per additional RCT/extraction avoided was $916.09, with a mean cost difference of $156.37 (95% CI: −88.37 to 401.11). The effect difference was 0.171 (95% CI: −0.006 to 0.347). After accounting for missing data using multiple imputations, the incremental cost of full pulpotomy per RCT/extraction avoided was $1269.44 from a societal perspective and $1361.86 from a healthcare system perspective. Results from the analysis indicate that the main driver of the ICER is the unit costs of treatment, which accounted for 78.2% of the total cost. A breakdown of cost components by percentage contribution is provided in Table S4. Similar results were observed using Bayesian models (Table S5). In extremely deep lesions, full pulpotomy dominated SCR across all models, being less costly and more effective (Table S6).

For the deterministic sensitivity analysis (Table 3) adopting the societal perspective, the ICER increased to $1325.16 per RCT or extraction avoided when the difference in the number of RCTs/extractions avoided between full pulpotomy and SCR decreased from 0.171 to 0.120. Conversely, the ICER decreased to $713.55 per RCT or extraction avoided when this difference increased from 0.171 to 0.222. For unit costs, the ICER increased to $1275.25 per RCT or extraction avoided when pulpotomy unit costs increased by 20%, and decreased to $575.09 per RCT or extraction avoided when pulpotomy unit costs decreased by 20%. In contrast, the ICER changed to $730.70 and $1129.40 per RCT or extraction avoided when the unit costs of SCR increased and decreased by 20%, respectively.

TABLE 3.

One way sensitivity analyses.

Variable	Parameter values			Incremental direct cost per RCT or extraction avoided
				Societal perspective		Healthcare system perspective
	Base	Low	High	Low	High	Low	High
Income—2020	4534.00	3627.20	5440.80	922.49	933.27	—	—
Income—2021 a	4680.00	3744.00	5616.00
Income—2022 a	5070.00	4056.00	6084.00
Income—2023 a	5197.00	4157.60	6236.40
RCT unit cost	As recorded	20% lower	20% higher	859.86	996.43	837.60	996.11
SCR unit cost	As recorded	20% lower	20% higher	1129.40	730.70	1142.40	699.70
Pulpotomy unit cost	As recorded	20% lower	20% higher	575.09	1275.25	537.18	1285.58
Time taken for RCT	2.00 h	1.60 h	2.40 h	942.82	912.48	—	—
Time spent on reviews	1.00 h	0.80 h	1.20 h	925.47	930.25	—	—
Effect diference between pulpotomy and SCR (avoiding RCT/extraction)	0.171	0.120	0.222	1325.16	713.55	1308.69	704.68

^a Income‐related values presented in this table reflect values prior to CPI adjustment.

At lower WTP thresholds, SCR shows a higher probability of being cost‐effective; however, as WTP increases, the probability of full pulpotomy becoming more cost‐effective increases (Figures 3 and 4). When the WTP to avoid RCT/extraction exceeds $930, the probability of pulpotomy being cost‐effective rises above 50% (Figure 4).

FIGURE 3.

Cost‐effectiveness plane showing scatterplot of incremental costs and incremental RCT/extractions avoided. (a) Incremental total costs. (b) Incremental direct medical costs. These cost‐effectiveness planes display 5000 bootstrapped simulations estimating the joint distribution of incremental costs and effectiveness (measured as RCT/extractions avoided). Each blue point reflects one simulation outcome, and the red dot marks the base‐case ICER. Incremental costs are calculated as: pulpotomy − SCR. The clustering of points in the northeast quadrant indicates that pulpotomy is generally more effective and more costly.

FIGURE 4.

Cost‐effectiveness acceptability curve. (a) Change in probability of being cost‐effective with higher WTP (societal perspective); (b) Change in probability of being cost‐effective with higher WTP (healthcare system perspective); The CEAC illustrates the probability that pulpotomy is more cost‐effective than SCR at varying WTP thresholds per additional RCT or extraction avoided. To generate the CEACs, 5000 bootstrap replications of the incremental cost and effectiveness estimates were simulated based on trial data. Each simulation yields one pair of incremental cost and effect values (pulpotomy − SCR). For each WTP value (i.e., how much one is willing to pay to avoid one RCT or extraction), the proportion of bootstrap samples where the incremental cost‐effectiveness ratio (ICER) falls below the WTP is calculated and plotted. At lower WTP thresholds, SCR shows a higher probability of being cost‐effective; however, as WTP increases, the probability of pulpotomy becoming more cost‐effective increases. When the WTP to avoid RCT/extraction exceeds $930 (for societal perspective) and $920 (for healthcare system perspective), the probability of pulpotomy being cost‐effective rises above 50%.

5. Discussion

In this randomised controlled trial, where more than two‐thirds of the cases in the material involved extremely deep caries, full pulpotomy demonstrated higher success rates in avoiding RCT or extraction compared to SCR at 36 months. The pulp survival rate of full pulpotomy in our study fell from 98.1% at 1 year to 89.1% at 3 years. This aligns with the radiographic success rates reported in a systematic review by Cushley et al. (2019), declining from 95.4% to 88.39% over a similar follow‐up period. In contrast, the clinical success rates in Cushley et al.'s systematic review were consistently higher, decreasing from 97.4% to 93.97%. While Cushley et al. separately defined clinical success as the absence of symptoms (e.g., spontaneous pain, tenderness to percussion, swelling or sinus tract) and radiographic success as absence of apical periodontitis or resolution of existing lesions, our study defined ‘survival’ as a combined outcome. Survival in our study required the tooth to be both asymptomatic and without a periapical lesion, based on clinical and radiographic criteria. Moreover, it is important to note that the studies included in the review primarily involved pulpotomies performed in teeth diagnosed with symptomatic irreversible pulpitis, as opposed to teeth with reversible pulpitis in our study. Challenges associated with accurate diagnosis of pulpal inflammation could limit direct comparisons of outcomes. Nonetheless, a consistent trend of declining success rates over time was observed across studies. The pulp survival rate for full pulpotomy in our study was lower than the 2–3‐year pulp survival rate of 96.9% reported in the meta‐analysis by Aguilar and Linsuwanont (2011). However, the sample sizes of the primary studies included in the systematic review were relatively small (n ≤ 21) (Aguilar and Linsuwanont 2011). Beyond pulpal failures, 5/37 (13.5%) of teeth in the pulpotomy group experienced restoration breakdown or recurrent caries. This finding underscores the critical need for patient's self‐care and plaque control, placement of coronal restorations that can be predictably maintained by patients, and regular follow‐up monitoring to address potential late failures and maximise the longevity of pulpotomy‐treated teeth. While partial pulpotomy preserves additional tooth structure and retains more coronal pulp allowing for sensibility testing during follow‐up, variation in the depth of pulp removal introduces heterogeneity that may confound the interpretation of outcomes in a trial. Full pulpotomy was therefore selected to standardise the procedure across operators, provide a definitive anatomical endpoint at the canal orifices, and increase the likelihood of removing irreversibly inflamed pulp tissue (Taha and Abdulkhader 2018). While studies suggested that both full pulpotomy and partial pulpotomy demonstrated favourable clinical and radiographic outcomes in teeth diagnosed with irreversible pulpitis, the differences were not statistically significant (Baranwal et al. 2022; Ramani et al. ²⁰²²; Jassal et al. ²⁰²³). However, these studies differ from the present trial which included teeth diagnosed with reversible pulpitis. To date, no randomised controlled trials have compared partial versus full pulpotomy in teeth with reversible pulpitis (Louzada et al. 2025). Nonetheless, a previous systematic review reported high success rates for partial pulpotomy in such cases, with a 1‐year success rate of 98% (95% CI, 96–99), comparable to that of full pulpotomy under similar diagnostic conditions (Elmsmari et al. 2019).

The reduction in the SCR pulpal survival rate from 89.7% at 1 year to 74.1% at 3 years observed in our study highlights ongoing concerns regarding its long‐term effectiveness in managing extensive caries, of which the majority were extremely deep lesions. Hashem et al. (2019) similarly observed a decline in SCR success from 84% at 1 year to 72% at 2 years in deep lesions extending at least 75% into dentine, underscoring the need for longer follow‐up. While some systematic reviews have suggested that SCR may be as effective as non‐selective caries removal and carries a lower risk of pulp exposure (Barros et al. 2020; Li et al. ²⁰¹⁸), these conclusions should be interpreted with caution. A more recent review by Schwendicke et al. (2021) identified substantial limitations in the existing evidence base, including high risk of bias, small sample sizes, and imprecise effect estimates, resulting in low certainty for most comparisons.

The risk of failure was significantly higher in the SCR group, with a hazard ratio of 2.58, indicating that teeth treated with SCR were 2.58 times more likely to require RCT or extraction than those treated with full pulpotomy. These results differ from those of a randomised controlled non‐inferiority trial by Rechithra et al. ²⁰²³ who reported comparable 1‐year success rates for selective removal to soft dentine (95.45%) and full pulpotomy (95.65%) in teeth with reversible pulpitis, suggesting both are viable options for deep lesions. However, unlike their study which included only deep caries, our trial involved a greater proportion of extremely deep lesions. The higher failure rate in the SCR group may reflect the limitations of SCR in managing extremely deep lesions, where caries penetrates the entire thickness of the dentine radiographically, a feature associated with increased risk of bacterial penetration into tertiary dentine and more extensive pulpal inflammation (Bjørndal et al. 2019; Demant et al. ²⁰²¹). Such pulpal inflammation may not be resolved through selective excavation alone. Additionally, clinically firm dentine can remain histologically infected (Ricucci et al. 2019) with residual bacteria. Despite a good marginal seal, this may contribute to delayed pulpal breakdown and treatment failure. This study's findings are consistent with findings from Taha et al. (2024), who reported 12‐month pulp survival rates of 82.5% for SCR and 98.4% for Total Caries Removal (TCR) with immediate vital pulp therapy in teeth diagnosed with reversible pulpitis. The difference in liner materials used may also have contributed to the observed outcomes. In the SCR group, Resin‐Modified Glass Ionomer Cement (RMGIC) was used as a liner, while Biodentine (Septodont) was used in the full pulpotomy group. While calcium silicate materials such as Biodentine and MTA have demonstrated superior biological and sealing properties compared to calcium hydroxide (Taha and Abdulkhader 2018; Tan et al. ²⁰²⁰), evidence remains mixed for SCR. Several clinical studies and systematic reviews have shown no significant differences in pulp vitality with different materials placed following SCR (da Rosa et al. ²⁰¹⁹; Hashem et al. ²⁰¹⁹; Patankar et al. ²⁰²⁵; Singh et al. ²⁰¹⁹). Nevertheless, while Hashem et al. (2019) found no statistically significant difference in 12‐ and 24‐month success rates between teeth with reversible pulpitis restored with Biodentine and GIC as indirect pulp capping materials, three‐dimensional radiographic evaluation revealed superior periapical healing with Biodentine.

Our study's findings contribute to the growing body of evidence informing the management of extremely deep carious lesions. In our cohort, SCR was associated with a higher need for RCT, with 29% of extremely deep cases resulting in pulpal failure requiring RCT, compared to only 5% in the full pulpotomy group (Figure S1). Additionally, the average total cost was higher for SCR ($851 vs. $760). The subgroup analysis of teeth with extremely deep caries showed that SCR is dominated by full pulpotomy for teeth with extremely deep caries as it is less effective and more costly to manage (Table S6). Given the small number of teeth with deep caries only (n < 18 in each intervention group), we were unable to perform the subgroup analysis for teeth with deep caries only. In addition, our study findings align with growing evidence that pulp exposure during caries excavation is no longer considered an unfavourable prognostic indicator for long‐term pulp vitality when properly managed with bioactive materials.

Our study found that full pulpotomy was more cost‐effective than SCR in managing caries at a WTP threshold of $928 or above to avoid one tooth extraction or RCT within a patient population with the majority of lesions classified as extremely deep. Treatment failures often lead to more invasive procedures such as RCT, extractions or prosthetic replacements. These interventions are not only more costly in terms of direct medical costs, but are also more time‐consuming, potentially having a negative impact on patients' oral health–related quality of life (QoL) (Wigsten et al. 2021). Previous economic evaluations have indicated that SCR is more cost‐effective than stepwise excavation for deep caries management, largely attributed to reduced failure rates and the efficiency of a single‐visit approach (Jardim et al. 2023; Schwendicke et al. ²⁰¹³). While we did not directly compare pulpotomy with RCT, Asgary and co‐workers conducted a multi‐center trial comparing pulpotomy and RCT in individuals aged 9–65 years, finding no significant difference in clinical or radiographic outcomes within 2 years but lower costs associated with full pulpotomy for both initial treatment and failure management (Asgary et al. 2017). Similarly, Naved et al. (2024) modelled the cost‐effectiveness of pulpotomy versus RCT for irreversible pulpitis in mature permanent teeth, highlighting pulpotomy as a more economically viable option at lower WTP thresholds. Further studies should be carried out to assess the cost‐effectiveness of pulpotomy relative to other vital pulp therapies, as well as conventional RCT to validate these findings.

This is the first study to compare SCR and full pulpotomy over a longer follow‐up period of 3 years and the first trial‐based economic evaluation comparing these two approaches. This method provided high internal validity and precise cost data through micro‐costing, accounting for labour, materials and indirect costs (Gray 2006; Hughes et al. ²⁰¹⁶). Unlike modelling approaches such as Markov models, which rely on extrapolated data and input parameters to simulate long‐term outcomes, trial‐based evaluations offer real‐world evidence accounting for patient heterogeneity. This trial‐based evaluation attempts to fill the gap in real‐world data by providing comparative evidence over 3 years, a timeframe sufficient to assess meaningful differences in cost‐effectiveness between SCR and full pulpotomy. The inclusion of a societal perspective also enables a comprehensive evaluation of both direct and indirect costs, making the findings relevant for policy and economic evaluations.

However, the study has a few limitations. The 3‐year follow‐up, while presenting informative mid‐term data, might be insufficient to capture long‐term restorative outcomes, such as restoration failures necessitating retreatments. Future cost‐effectiveness analyses using data with longer follow‐up periods and larger cohorts can be carried out to address this gap. While extrapolation using decision‐analytic modelling could offer insight into long‐term outcomes, it was not performed here due to the limited number of events and uncertainty in estimating transition probabilities beyond 3 years. Additionally, the study faced challenges with loss to follow‐up. To address missing data in costs, predictors, and outcomes, both complete case analyses and multiple imputation were conducted assuming data were missing completely at random and missing at random respectively. Multiple imputation accounts for uncertainty in imputed values by generating multiple datasets and providing more robust and accurate parameter estimates (Briggs et al. 2003; Faria et al. ²⁰¹⁴). Moreover, Bayesian models were fitted and Monte Carlo bootstrapping was carried out to address uncertainty in view of the modest sample size. In addition, the reliance on unsubsidised treatment charges as a proxy for costs, while reflective of patient expenses, may not fully capture provider‐specific costs. Variations in healthcare coverage and patient demographics also present challenges to generalising these findings to broader populations or diverse healthcare settings. Another limitation was the lack of stratification for carious lesion depth at randomisation, which might have influenced the distribution of deep and extremely deep cases between treatment arms. While there was no significant difference between groups at baseline, the overall cohort was skewed, with more than two‐thirds of enrolled teeth classified as extremely deep lesions. Subgroup analyses by lesion depth may be underpowered due to smaller sample sizes. Future studies with stratified randomisation by lesion depth can thus be carried out.

The absence of an established WTP threshold for avoiding root canal treatment or extraction in dentistry poses a key limitation in interpreting the cost‐effectiveness of full pulpotomy relative to SCR. Without a defined benchmark, it is challenging to determine whether the calculated ICERs fall within an acceptable or justifiable range. This gap limits the direct application of our findings to policy or reimbursement decisions. Nonetheless, the CEAC generated in this study offers a useful framework, illustrating the probability of pulpotomy being cost‐effective across a range of hypothetical WTP thresholds. While some studies have reported WTP estimates for avoiding tooth loss, these are context‐specific and may not reflect broader population values (Ghahramani et al. 2022; Shahkoohi et al. ²⁰²⁵). Therefore, our findings should be interpreted with caution and viewed as an initial step towards developing more comprehensive, value‐based frameworks in endodontic care. Future research can also explore the use of preference‐based measures to support clearer economic guidance for clinical and policy decision‐making. Incorporating patient‐reported outcome measures would enhance the understanding of the impact on QoL, further aligning evaluations with patient‐centred care.

6. Conclusion

In conclusion, full pulpotomy demonstrated greater effectiveness than SCR in avoiding further intervention over a 3‐year period in teeth with extensive caries, of which the majority were extremely deep lesions. Further long‐term studies and broader health economic evaluations are warranted to guide clinical decision‐making and policy development.

Author Contributions

Jill Jing Rui Chew: conceptualisation, methodology, formal analysis, investigation, writing – original draft, writing – review and editing. Sharon Hui Xuan Tan: conceptualisation, methodology, formal analysis, writing – review and editing, supervision. Sarah Kho Xian Chua: investigation, writing – review and editing. Jeen Nee Lui: writing – review and editing, supervision. Yu Fan Sim: formal analysis, writing – review and editing. Victoria Soo Hoon Yu: conceptualisation, methodology, writing – review and editing, supervision, funding acquisition.

Ethics Statement

The study protocol was approved by an institutional Research Ethics Committee (DSRB 2018/01360) and conformed to the Helsinki Declaration of 1975.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Baseline and intra‐operative characteristics of teeth that completed treatment.

Table S2: Clinical and radiographic findings of teeth requiring RCT.

Table S3: Clinical and radiographic findings of teeth requiring replacement of restoration.

Table S4:. Breakdown of cost.

Table S5: Comparison of estimates from alternative generalised linear models (GLM).

Table S6: Comparison of estimates from alternative GLMs (extremely deep caries).

Figure S1:. Kaplan–Meier survival curve presenting cumulative survival probabilities of SCR and pulpotomy in extremely deep caries.

Figure S2: Kaplan–Meier survival curve presenting cumulative survival probabilities of SCR and pulpotomy in deep caries.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author, Sharon Hui Xuan Tan. The data are not publicly available due to an ethical restriction (patient confidentiality) that was imposed by the ethics review board.