**3. Results**

A total of 57 and 41 patients were eligible to participate in the study in the test and control groups respectively (Figure 2), which were subsequently reduced to 44 and 41 patients due to intra-operative exclusion criteria (Table 1).

**Figure 2.** Flowchart for test and control groups based on STROBE recommendations.

#### *3.1. Short-Term Evolution of Pain*

This analysis could be performed for 44 and 40 patients in the test and control groups respectively. There was no significant difference (*p* > 0.05) between both groups at each time point (T0, 24 h, 7 d) (Figure 3). However, a significant reduction in pain was observed between each time point (*p* < 0.0001).

#### *3.2. Treatment Success*

Follow-up for global success evaluation was possible for 79.5% and 75% of the patients, with a median follow-up of 24 and 20.5 months, and a mean follow-up of 25.9 and 25.6 months, respectively, in test and control groups. The characteristics of the population in both groups are presented in Table 2. The logistic regression (univariate) regarding patient distribution between both groups based on these variables revealed the following significant differences: pre-operative pain to percussion (*p* = 0.049), patient age (*p* = 0.0004) and presence of a carious cavity (*p* < 0.0001).

**Figure 3.** Boxplots of short-term pain evolution for test and control groups; the dots correspond to outliers.



Five patients experienced *clinical* failure <12 months, four in test group (managed by root canal treatment) and one in control group, which led to tooth extraction due to vertical root fracture.

For patients with follow-up ≥12 months, the evaluation of the radiographs by two independent observers resulted in a consensus PAI scoring, characterized by a Cohen's kappa coefficient of 0.72 between the two evaluation sessions. The PAI evolution for each patient according to their follow-up duration is available in the Supplementary Material (Tables S1 and S2). Representative examples of the radiographic evolution can be observed in Figure 4a–h for test group and Figure 4i–p for control group.

**Figure 4.** (**<sup>a</sup>**–**h**) test group—4 clinical cases of pulpotomy. (**<sup>a</sup>**,**b**) #37 Healthy status quo, follow-up at 46 months; (**<sup>c</sup>**,**d**) #36 peri-apical healing, follow-up at 20 months; (**<sup>e</sup>**,**f**) #36 peri-apical healing, follow-up at 16 months; (**g**–**h**) #47 peri-apical aggravation, follow-up at 20 months. (**i**–**p**) control group—4 clinical cases of root canal treatment. (**i**,**j**) #26 Healthy status quo, follow-up at 47 months; (**k**,**l**) #47 peri-apical status quo, follow-up at 50 months; (**<sup>m</sup>**,**<sup>n</sup>**) #36 peri-apical aggravation, follow-up at 24 months; (**<sup>o</sup>**,**p**) #37 peri-apical healing, follow-up at 12 months.

No significant difference was observed between test and control groups regarding *clinical* (*p* = 0.32) or *global* success (*p* = 0.347). However, a significant difference in *radiographic* success was observed (*p* = 0.014) (Table 3).

**Table 3.** Characteristics of cases included in treatment success evaluation (*N* = 65).


 excluding clinical failures <12 months. \*\* Significant difference (*p <* 0.017).

\*

None of the variables listed in Table 2 were found to have a significant effect on either *global*, *radiographic* or *clinical* success (univariate analysis; *p* > 0.017).

Among the *clinical* failures in the test group (<12 months, *n* = 4; ≥12 months, *n* = 3), six cases presented evidence of pulp vitality in all canals. A pronounced inflammation was observed, indicated by intense bleeding. One case was re-treated in another practice.

#### *3.3. Histological Analysis*

One case classified as global success in the test group at 45 months had to be extracted for restorative reasons, i.e., non-restorable fracture (Figure 5a–c). The tooth was processed for histological evaluation (Figure 5d–g), which revealed a vital pulp tissue present in the root canals without any sign of inflammation and with an odontoblastic layer lining the canal walls. It must be noted that difficulties in sectioning were encountered. While demineralized, the collagen remained hard in some areas, preventing serial sectioning throughout the whole sample.

**Figure 5.** Histological analysis of a case considered as global success, extracted for restorative reason, performed by Domenico Ricucci. (**a**) #27 pre-operative situation, PAI score:1; (**b**) #27 healthy status quo, follow-up at 45 months; (**c**) radiographic and clinical images of extracted tooth; (**d**–**g**) histological sections (H&E staining) at various magnifications, revealing the presence of a vital pulp tissue in the root canals without any sign of inflammation or tissue disruption. An odontoblastic layer lining the canal walls can be observed.

## **4. Discussion**

The first major finding was the efficiency of the pulpotomy procedure as permanent treatment in terms of short-term pain relief. A comparable pain reduction at 24 h and 7 d between both treatment procedures was indeed observed. This is in accordance with a randomized clinical study comparing full pulpotomy, selective pulpectomy (restricted to one canal) and full pulpectomy [25]. Another randomized trial reported an even higher and significant pain reduction in the pulpotomy group as compared to the root-canal treatment [26]. The evolution of pain at 24 h and 7 d follows a previously described trend following root-canal treatment, with a moderate drop within one day, and a substantial reduction to minimal levels in 7 days [20]. Hence, pulpotomy as permanent treatment of irreversible pulpitis using tricalcium-silicate cements is at least as efficient as the gold standard procedure in terms of short-term pain relief. Pain relief is an essential component of an endodontic procedure, with pathologies that can lead to very high levels of pain [1,17–19], in the same range as those observed in other painful diseases such as renal colic [27,28]. Such levels of dental pain can severely affect patient quality of life during the acute phase, and were, also, shown to be a major cause of acute medical admission following unintentional paracetamol overdose [29,30].

The second major finding was the equivalent efficiency of the pulpotomy procedure as permanent treatment regarding *global* success, as compared to the gold standard procedure. A higher trend was even observed in the test group (77%) but not significantly (*p* > 0.017) compared to the control group (67%). This is in line with a similar observation made in a randomized clinical trial [12]. The latter is currently the most robust study available on the topic, with confirmed irreversible pulp status, >12 months mean follow-up and including a control group.

The incidence (11.4%) of *clinical* failure observed within the first 12 months following the test procedure is in line with the data available in the literature [4–7]. The overall *clinical* success rate was 80% in our work, all *clinical* failures were related to pain and not to other clinical signs, and lead to re-intervention via standard root-canal treatment. It is important to mention that the location of root canals required the use of magnification, which was possible in all cases. Moreover, the evidence of pulp vitality associated with pronounced inflammation observed in most failed cases connects to the possible interest of quantifying biomarkers such as MMP-9, TNFα or IL-8 as predictors of the procedure prognosis [4,31–33].

The significant difference in *radiographic* success observed in favor of the test group (94% vs. 69%; *p* < 0.017) could be considered unexpected. It cannot be excluded that the higher percentage pre-operative PAI ≥ 3 in the control group affects such finding, but this variable was not identified as statistically significant (*p* > 0.05). Nevertheless, the trend is consistent with the one reported in the literature at two years [13], even considering cases with pre-existing periapical involvement. This trend, illustrated in Figure 4c,d may indicate the positive effect of preserving a living, immunocompetent pulp tissue within the root canal on periapical healing. It is worth noticing that while the presence of a periapical radiolucency (PAI > 3) is not expected in case of an inflamed pulp, some cases were classified as PAI = 4 in the present work and in another [6]. This is due to the fact that the PAI score was not considered in the inclusion/exclusion criterion, which are mainly clinical. Since the radiographic interpretation is performed by independent observers, some cases with PAI > 3 can therefore be reported.

The survival of healthy radicular pulp tissue in successful cases following a pulpotomy as permanent treatment remains unknown. Histology is indeed required to determine the exact status of the pulp tissue, and only few short-term studies provided such information for pulpotomies as permanent treatment in case of irreversible pulpitis. A two-month report on 12 molars reported the presence of healthy radicular pulp devoid of inflammatory signs [34] and a 10-month case report made similar observations in a premolar [35]. In the present work, the histology performed on the successful case provided similar evidence but at 45 months follow-up. In the absence of histology, the observation of a hard tissue barrier underneath the capping material has been considered by some authors as indicator of pulp vitality, as mentioned in a recent review [36]. However, the latter, also, underlines the low reliability of such observation. It is indeed not trivial to determine accurately on a radiograph if an actual barrier has formed, since there is no standardized definition nor are the X-ray images taken in a reproducible manner. Despite the level of evidence regarding the ability of materials to induce mineral bridge formation, it was reported that either pure calcium hydroxide powder or tricalcium-silicate cements are likely the most appropriate materials to cover the pulp [37,38], which was performed here (Biodentine, Septodont, Saint-Maur-des-Fossés, France). The remaining pulp tissue must be free of inflammation, which can be assessed clinically by obtaining a hemostasis. This was a prerequisite in the present work, since it was demonstrated in the context of a direct pulp capping following carious exposure that the degree of pulp bleeding was associated with treatment success [39]. Furthermore, compared to carious exposures, the success rate of pulp capping in traumatic exposures, associated with little or no pulp inflammation, is known to be higher [40]. Nevertheless, it must be noted again that clinical criteria are unfortunately associated with little level of evidence to correlate to histological pulp status [41]. More recently, other intra-operative parameters were subject to discussion, such as the time required to achieve hemostasis [42] or the type and concentration of irrigants [43], without identifying clear trends with regards to treatment outcome. In addition, it was recently shown that

in presence of low levels of pulp inflammation (evaluated by MMP-9 quantification), the irrigant (saline vs. 2.5% NaOCl) had little impact on the direct pulp capping outcome [44].

One of the limitations of the study is the modest sample size in each group, but it is relatively in the same range as most available studies on the topic of vital pulp therapy of irreversible pulpitis [4–6], except one works with a much larger sample size [12]. In terms of follow-up, apart from the latter providing results at five years, the present study provides the longest duration with an average of 25 months. The recall rate is reasonable given the difficulties associated with this specific topic, although it should ideally be higher given the limited sample size. Emergency patients are indeed quite difficult to follow over time, as they often consult irregularly and/or tend to return to their regular dentist once the emergency is over, both in test and control groups. The number of patients was balanced between test and control groups. Both groups were also shown to be equivalent in terms of initial pain intensity (*p* > 0.05), but not for three other variables. Among these, none had a significant impact on treatment success (*p* > 0.017). However, based on the differences of distribution between both groups, we cannot exclude the existence of a selection bias. The latter, combined with the limited sample size, could account for the lower success rate in the control group. It can also explain the lower *global* success rate in that group compared to those reported in the literature [45,46]. However, these reported success rates shall be compared with those of irreversible pulpitis with caution, as they correspond to cases with "vital pulp", without further distinction regarding pulpal diagnosis. To our knowledge, the endodontic treatment success rates available regarding irreversible pulpitis are 65.8% at five years [12], which is lower than the rates usually reported for vital pulp cases as a whole.

Another limitation of the work is the restriction to molar cases only. This was determined based on two main criteria: The ease to identify the anatomical transition between coronal and radicular pulp, and the level of treatment complexity. The first criteria was in fact a methodological decision in order to make the procedure more easily reproducible. The second criteria was related to clinical relevance, since both European and American endodontic associations have identified molar teeth as a criteria for treatment complexity [47,48]. Hence, it was the point of the present work to consider the more conservative pulpotomy strategy (test group) as one potentially applied by a larger population of dentists, whereas the gold standard root canal treatment may be restricted to specialized endodontists. It was interesting to assess whether a strict protocol for performing the vital pulp therapy procedure (facilitated by use of magnification) could compensate for the lesser experience of non-specialist practitioners.

Regarding study design, it has been described in the past that while randomized clinical trials are ideal for evaluating the effect of drugs, it is not easily the case for surgical interventions such as endodontic procedures [49]. A double-blind system or the use of a placebo is generally not possible. Moreover, the preference of using "best available therapy" control group rather than a placebo has been described [50] with reference to the World Medical Association Declaration of Helsinki. In the context of irreversible pulpitis, the "best available therapy" would be defined as root canal treatment, but no study to date has considered practitioner experience as an important part of the definition. Both AAE and ESE clearly identify "best available therapy" as one performed by specialized endodontists [47,48]. Hence, it was the purpose of this study to take this aspect into account in the group design, with specialists performing the procedures in the control group, and non-specialists in the test group. In order to be as relevant as possible, specialists working in private practice were selected for control cases, which therefore resulted in a non-randomized study design. While the latter certainly introduces biases on the one hand, it potentially increases the usefulness of the study on the other hand, which is to be taken into account in a study quality aspect [51].
