*3.4. Meta-Analysis*

Pooled POP data (mean ± standard deviation [SD]) experienced by patients 24 h and 48 h after RCF with BCS or RBS are presented in Figures 2 and 3, respectively. Six studies did not report the mean POP ± SD 48 h after RCF and these were not included in the respective forest plot [28,30,33–36]. Since in the study by Graunaite et al. 2018 [31] asymptomatic patients were treated, as opposed to all the other studies in which patients were symptomatic, the sensitivity analysis was performed in all analyses where that study was considered, to see if the results changed with the exclusion of asymptomatic subjects. Pooled data analyses indicate that POP was significantly lower in patients who underwent RCF with BCS compared with RBS at 24 h (SMD = −0.20; *P* = 0.04) and 48 h (SMD = −0.26; *P* = 0.0005) after treatment. After the sensitivity analysis, by excluding Graunaite et al. 2018, the results did not change significantly.


**Figure 3.** Forest plot of POP level 48 h after RCF with BCS vs. RBS.

The analgesic intake did not significantly differ between the BCS and RBS groups 24 h after RCF (RR = 0.46; *P* = 0.14; Figure 4). Five studies were not included in this forest plot due to a lack of available data [11–13,33,36]. The incidence of flare-up was also not significantly different between the BCS and RBS groups (OR = 0.32; *P* = 0.24; Supplementary Materials Figure S1). After the sensitivity analysis, by excluding Graunaite et al. 2018, the results did not change significantly.

**Figure 4.** Forest plot of analgesic intake 24 h after RCF with BCS vs. RBS.

The next two diagrams underline the pain prevalence and severity of the BCS group over the RBS group in 24 h (Supplementary Materials Figure S2) and 48 h (Supplementary Materials Figure S3) after RCF.

The probability of "No pain" 24 h after treatment was 1.12× higher in the BCS group vs. the RBS group (OR = 1.12; 95% CI, 0.77–1.64; *P* = 0.86), while the same was observed for "Moderate pain" probability (OR = 1.21; 95% CI, 0.61–2.38; *P* = 0.59). There was no heterogeneity in the study effect for the BCS and RBS groups (*I*2 = 0%; *P* = 0.86, and *I*2 = 0%; *P* = 0.89, respectively), indicating perfect consistency in the results. The probability of "Mild pain" and "Severe pain" was 1.2× and 1.7× higher in the RBS group vs. the BCS group, respectively (OR = 0.83; 95% CI, 0.54–1.25; *P* = 0.37, and OR = 0.59; 95% CI, 0.08–4.58; *P* = 0.62). There was also no heterogeneity in the study effect in either group (*I*2 = 0%; *P* = 0.97, and *I*2 = 0%; *P* = 0.61, respectively; Supplementary Materials Figure S2).

A diagram of pain characteristics in the BCS and RBS groups 48 h after treatment is presented in Supplementary Materials Figure S3. The studies that did not report pain characteristics 48 h after treatment were excluded. The probability of "No pain" 48 h after treatment was 1.21× higher in the BCS group vs. the RBS group (OR = 1.21; 95% CI, 0.60–2.42; *P* = 0.60; heterogeneity: *I* 2 = 18%; *P* = 0.30). "Mild pain" was more commonly reported in the RBS group vs. the BCS group 48 h after treatment (OR = 0.82; 95% CI, 0.40–1.68; *P* = 0.59; heterogeneity: *I* 2 = 1%; *P* = 0.40), while "Moderate pain" was equally likely in both groups 48 h after treatment (OR = 1.00; 95% CI, 0.14–7.27; *P* = 1.00; heterogeneity: *I* 2 = 0%; *P* = 0.33). "Severe pain" was reported only once in both groups (Supplementary Materials Figure S3). After the sensitivity analysis, by excluding Graunaite et al. 2018, the results did not change significantly neither for data after 24 h nor for 48 h after RCT.

The effect of the obturation technique (WVT vs. SCT) on POP based on data 24 h and 48 h after RCF is presented in Supplementary Materials Figures S4 and S5. The studies that did not report 48-h data were excluded. There was a numerical difference in POP in favor of BCS for both WVT and SCT subgroups 24 h after RCF (OR = 0.85; *P* = 0.41). The probability of POP was numerically higher in the RBS subgroup that underwent the WVT technique (*P* = 0.49) in 1.2× and lower in the BCS subgroup that underwent the SCT technique (*P* = 0.64; Supplementary Materials Figure S4). There was evidence of lower POP in both WVT and SCT subgroups of the BCS group 48 h after RCF ( *P* = 0.33). The probability of POP did not differ according to the obturation technique in the BCS group, but was 1.3× higher in the RBS group 48 h after RCF (OR = 0.78; *P* = 0.33; Supplementary Materials Figure S5).

The probability of POP by the pulp status (vital [V] and non-vital [NV] pulp) 24 h and 48 h after RCF is presented in Supplementary Materials Figures S6 and S7, respectively. The studies that did not report 48-h data were excluded. There was evidence of a nonsignificant difference in POP in favor of RBS in the V subgroup ( *P* = 0.50) and a trend for lower POP in NV teeth within the BCS group 24 h after RCF ( *P* = 0.16). However, there was no statistically significant overall effect of V vs. NV pulp on POP (OR = 0.84; *P* = 0.45; Supplementary Materials Figure S6). The probability of POP was 2× higher in the V pulp of the BCS group, and there was evidence of lower POP in the V subgroup of the RBS group 48 h after RCF (OR = 2.01; *P* = 0.18). There was also a non-significant trend for lower POP in NV teeth within the BCS group (OR = 0.92; *P* = 0.84; Supplementary Materials Figure S7). After the sensitivity analysis, by excluding Graunaite et al. 2018, the results did not change significantly.

Pooled data from the three studies that reported the retreatment show a trend for a difference in POP in favor of RBS (OR = 1.20; *P* = 0.65) 24 h after RCF (Supplementary Materials Figure S8). The meta-analysis determined that there was a 48% level of heterogeneity within the included nine studies. Only one study had analyzed a consistent number of cases [31]. Supplementary Materials Figure S9 presents POP in retreatment groups 48 h after RCF. Unfortunately, the limited numbers of studies and retreated cases are insufficient to determine any effect of BCS vs. RBS on POP (OR = 1.34).

There was no significant difference in POP between the BCS and RBS groups when the treatment was carried out over single (OR = 0.77; *P* = 0.28) or multiple visits 24 h after RCF (OR = 0.81; *P* = 0.50; Supplementary Materials Figure S10). A lack of data precluded the equivalent analysis of POP 48 h after RCF.

## **4. Discussion**

This meta-analysis of nine pooled RCTs indicates that POP was significantly lower after RCF with BCS compared with RBS. However, none of the RCTs individually reported any significant effect of BSC vs. RBS on POP [28–36].

In this analysis, we found that BCS was non-significantly correlated with reduced analgesic intake vs. RBS, an observation that was also reported by one of the included studies [35] using the warm-obturator filling technique. However, two of the other included RCTs [32,33] demonstrated comparable analgesic intake in the control and experimental groups after RCF using SCT.

This systematic review found a non-significant trend of reduced flare-up in the BCS group vs. the RBS group. This is supported by one of the included RCTs that reported a significant reduction of flare-up following RCF with BCS vs. RBS [35]. However, another reported an equal occurrence between groups [31].

Regarding the warm and cold filling techniques, SCT with BCS has previously been associated with higher POP, while WVT with RBS has been associated with the lowest POP scores [28]. However, our pooled analysis suggests that there is a non-significant trend in favor of BCS.

Our results indicate that POP was lower in the V pulp when filled with RBS and in the NV pulp when filled with BCS. However, we found no additional background literature to place this in context.

Our results also evidence a non-significant difference in POP in favor of RBS at retreatment. Moreover, the only RCT [31] included to report this parameter indicates no difference between filling techniques in POP, following the retreatment procedures.

According to our results, the trend for lower POP following RCF with BCS vs. RBS filling technique was observed across single and multiple visit treatments. However, there are no studies in the literature to provide additional context.

The main limitations of this review are inter-study variability and inconsistency, as well as a lack of clinically relevant outcomes. Furthermore, as mentioned, different scales for pain measurement were used in different studies. Though the authors made efforts to resize all the scales to a 1–10 scale, it is difficult to understand if this had a relevance in the results. Of course, for future studies, it is recommended to use only scales for which there is an overall consensus. Therefore, the findings presented here need to be confirmed by further well-designed studies and should be interpreted with caution.
