Efficacy of electronic apex locators in comparison with intraoral radiographs in working length determination- a systematic review and meta-analysis

The eligibility criteria were designed according to PICO (Patient/Population, Interventions, Comparison, Outcomes). The patient population included permanent human teeth with closed apices undergoing root canal treatment. The intervention included EALs or Endo motor with integrated apex locator; comparator included 2-dimensional intraoral periapical radiographs and 3-dimensional imaging (CBCT). The primary outcome was postoperative pain, WL accuracy, and adequacy, and secondary outcome included the number of radiographs, time required, and associated adverse effects. The study question was: Is there any difference in postoperative pain and working length determined using apex locators compared to other imaging modalities?

Included studies involved fully developed human permanent teeth; clinical studies; studies that provide comprehensive data about the measured distance between the file tip employed for EAL measurement and the exact location of the apical constriction (AC); manual assessment of the working length using multiple frequency EALs and determination of the working length during rotary file preparation.

Studies done on teeth with open apex, primary teeth, teeth with resorption, perforated, resected teeth; endodontically treated teeth; studies with an observing of the file through the apex; case reports, reviews, and observational studies; histological evaluation of apical anatomy; studies in which the distance of the file tip used for EAL measurement to the AC is not specified or given as a range, first and second generation EAL; identification of landmarks other than the constriction/ minor foramen; foreign language articles without English translation were excluded.

A computerized literature search in seven databases was undertaken: MEDLINE PubMed, MEDLINE via OVID, LILACS, Embase, Scopus, Google Scholar, Cochrane Library from 1990 until October 2023, using words ‘‘Radiography’’, ‘‘Working length’’, and ‘‘Electronic Apex Locator’’ through PubMed to find the Medical Subject Headings terms for each word (Table 1). Additionally, a cross-reference search was done for included studies, other systematic reviews, and meta-analysis through connectedpapers.com.

Among the authors, two of them assessed the research headings and abstracts independently and in duplicate using Rayyan software [22]. For articles with full text, another two authors screened the studies individually and in duplicate to meet the inclusion and exclusion criteria. Differences in opinion were resolved through discussion with the arbiter to reach a consensus.

Data extraction form was designed with information about the study, purpose, sample, intervention, comparator, outcome measurement, results, adverse events, and author's conclusion (Table 2). Excluded studies and grounds for omission are reported in Table 3.

RoB evaluation was carried out by authors in pairs of two employing the RoB 1 tool [34]. Analysis and data entry was done using Review Manager 5.4.1 software.

Postoperative pain was expressed in the VAS scale as continuous data. If the scales used for postoperative pain were similar, results were described as mean differences (MD) with 95% confidence intervals (CI). If the scales were different, standardized mean difference (SMD) with 95% CI was used to pool data for similar outcomes from different trials. For the data that is continuous, the results were laid out as MD with 95% CI or SMD with 95% CI to incorporate data. The results for the data that is dichotomous, were laid out as risk ratios (RR) with 95% CI. Apart from reporting adverse effects, the rest of the secondary results were represented in the form of MD with 95% CI. Adverse effects were expressed qualitatively.

The meta-analysis was carried out with RevMan 5.4.1 which used SMD, random effect model, and inverse difference method. For WL determination mean difference, random effect model, statistical heterogeneity using the visual method, I² and Chi-square, and inverse difference method were used. All the results were presented for difference comparison. The I² statistic was used to evaluate the percent variation among studies due to heterogeneity.

The certainty of evidence for each outcome was assessed using the GRADE framework through GRADEpro GDT [21].

One thousand four hundred forty-three references were found in the electronic database search and 4 additional studies from a non-systematic search of Google Scholar, contacting corresponding authors to attain full texts, relevant systematic reviews, and cross-referencing of included studies. Following the removal of duplicate entries, 1042 references underwent initial screening based on their titles and abstracts, resulting in the exclusion of 1012 references. From 30 full-text articles, 10 studies were excluded (Table 3) [19, 35‐43]. One ongoing trial [44]was identified however, from 8 studies required data couldn’t be obtained and hence await classification [45‐52]. In accordance with the specified inclusion criteria, 11 studies met the requirements for inclusion in the systematic review [23‐33], with 8 of these studies being further incorporated into the meta-analysis [23, 24, 26‐29, 31‐33]. Figure 1 shows the study selection process in detail.

Among the eleven studies considered, ten studies were recorded in scholarly, peer-reviewed publications, while one among them was a dissertation [32]. Only two studies reported their funding details [28, 32]. The rest of the studies neither reported nor disclosed any funding details. All trials were parallel-group randomized controlled trials, with no cross-over trials.

In terms of patients, the least of the sample size was thirty [26], and the largest sample size was 220 (34). The maximum age of the participants taken was 75 years [32], and the lowest was 5 years [26]. All the selected studies had patients with permanent teeth except for one study, which had both primary and permanent dentition [26]. However, the results of the permanent teeth were only taken for analysis. One study [30] failed to provide any information about the age group.

Three studies [23, 27, 33] only included teeth with single roots and patent canals; however, another research [28] included both single-rooted and multirooted teeth. In addition, two research on multirooted teeth alone [24, 30] and two investigations on molars [25, 26] were also done. Root canals were considered in two studies; however, it was not stated whether the teeth were single or multirooted or whether they were mandibular or maxillary [29, 31].

Most included studies recruited healthy participants who were advised for root canal treatment. Patients with cardiac pacemakers, periapical radiolucency, curved roots, incomplete root formation, root resorption, Expectant mothers, and individuals with pre-existing systemic ailments were not considered for inclusion. One study did not mention any exclusion criteria [24].

Five studies [24‐26, 28, 32] had high susceptibility to bias and the rest of the 6 trials had unclear susceptibility to bias because they each had at least two unclear bias domains [23, 27, 29‐31, 33] (Fig. 2a and b). Random sequence generation was reported in 7 trials [24, 25, 27‐29, 32, 33] whereas, only 3 trials provided allocation concealment details [24, 28, 32]. One study provided details on the concealment of information from both participants and personnel [32] and blinding was reported in 6 studies [24, 25, 27, 29, 31, 32]. For the attrition bias, 3 trials [26, 28, 32] had drop-outs and hence had high susceptibility to bias. Among the entire set of studies, only one had a registered study plan and disclosed all the pre-planned results [29].

No other pertinent prejudices were identified in any of the 11 studies, rendering them at minimal risk of bias. All the incorporated studies had either higher risk or unclear risk of bias. Therefore, sensitivity analysis was not done as planned. Publication bias was not assessed since the meta-analysis didn’t include 10 or more studies.

The systematic review conducted by Amin et al. concluded that the accuracy of CBCT compared to EAL couldn’t be determined due to significant heterogeneity but suggested using pre-existing CBCT scans for WL determination [64]. A study concluded that the precision of EAL was comparable to the radiographic method. However, EAL and digital radiographic methods could reduce radiation dose exposure [24]. Another systematic review by Martins et al. reported inadequate scientific evidence and a considerable risk of bias. They suggested that WL determination using EAL could perform better than radiography alone, reducing patient radiation exposure. However, it was also recommended to perform at least one radiographic assessment to identify potential errors in electronic devices [56].

In this review, three out of eleven studies assessed the postoperative pain following WL determination using either EAL or radiographs [33, 62] which was the primary outcome. Kara-Tuncer & Gerek, Naeem et al., and Khan et al. 2021 concluded no significant disparity among the EAL and digital radiography groups in the postoperative pain dissipation period [23, 24, 33].

The adequacy and accuracy of WL determination were interpreted as continuous and dichotomous data [25‐29, 31‐33]. Studies done by Rathore et al. and Vanitha & Sherwood were a part of the systematic review but weren’t incorporated in the meta-analysis as the readings presented were for individual canals [25, 26]. However, both the studies were included qualitatively and the RoB assessment was done. Substantial heterogeneity was observed (Heterogeneity: Tau² = 0.06; Chi² = 2.06, df = 1 (P = 0.15); I² = 51%). For the dichotomous data, five studies evaluated the adequacy of working length between EAL and Radiographs [27‐31]. However, the study done by Saraf et al. presented data with respect to six different EALs and thus not incorporated in the meta-analysis [30].

Appropriate utilization of EAL alone could eliminate the requirement for an additional radiograph for diagnosis to determine WL. Patients who don't need to repeatedly be exposed to radiation due to mental, medical, or dental issues may benefit from this procedure. According to previous research, the use of EALs decreased the frequency of taking radiographs, reducing treatment time, effort, and radiation exposure to patients [29, 65, 66]. Furthermore, EALs have the potential to reduce the incidence of overextension of root canal procedures, which may result in postoperative pain and difficulty in maintaining the apical stop. Estimation of WL prior to radiographic verification maintains the correct working length for termination of obturation thus preventing overestimation of root canal length resulting in postoperative pain. The results of a study suggested that the apical foramen was accurately located by apex locators when used correctly and that only one preoperative radiograph was required [67]. They also help the operator in suspecting root fractures, resorptions, and perforations [58].

There is significant evidence that shows the point of termination of root canal instrumentation and obturation affects the outcome of endodontic treatment. However, there is insufficient data on the outcome of endodontic therapy in relation to EAL assessment. As a result, the radiographic WL measurement remains crucial for clinical purposes, and the Ideal application of EALs would be to reduce the number of radiographic exposures by estimating the WL accurately before taking any radiographic measurements [68]. By considering both radiographic and electronic measurements, as well as the accuracy of EALs and the morphology of the root apex, the final working length (FWL), would be determined [65].

The meta-analysis suggested that radiographs were more accurate and EALs were more adequate in determining WL. Nonetheless, deriving dependable conclusions wasn’t feasible given the extremely low certainty of evidence owing to concerns about bias, inconsistencies, and imprecision. As a result, the superiority of any intervention over another couldn’t be determined. The findings of the research should be interpreted with caution and further clinical trials are needed to confirm the results.

Since the available evidence is of very low certainty, more randomised controlled trials assessing the effect of working length on postoperative pain, working length accuracy and adequacy need to be conducted. Also, more trials on the effect of integrated apex locators on the working length accuracy and adequacy needs to be performed. Furthermore, well- executed RCTs need to be carried out on different generations of EALs and radiographic methods such as digital radiography and CBCT. Moreover, the studies included in the review did not assess the cost-effectiveness, reduction in radiation exposure, and the number of radiographs except one study [29].

Due to the unavailability of the full text, eight articles weren’t incorporated in the research synthesis. The studies were searched across libraries, Research Gate, and Google Scholar, apart from efforts made to contact authors by email. The extensiveness of the search was limited since the grey literature was not explored. None of the included studies fulfilled the secondary outcome of adverse events.

To explore the possible effect of losses to follow-up on the effect estimates for the primary outcomes, sensitivity analyses, subgroup analysis and other factors for heterogeneity was planned in the protocol. Nevertheless, only one or two studies were included under most of the comparisons and thus the sensitivity and subgroup analysis could not be conducted.

The publication bias could not be assessed since the meta-analysis didn’t include 10 or more studies. The primary outcome, postoperative pain could have several contributory reasons. However, this review does not include all the reasons since the primary studies did not report the cause of the pain. The results of unpublished data have been included in this meta-analysis, since the overall evidence is of very low quality; we assume it wouldn’t have affected the results.

Future research should focus on:

The current review emphasizes the requirement for well-executed RCTs with trial reports adhering to the guidelines from CONSORT 2010 [69] and incorporating the results generated through fundamental consortiums like Core Outcome Measures in Effectiveness Trials (COMET) [70]. This will contribute to the current body of evidence, allowing researchers to formulate more dependable findings.