Consequences of early extraction of compromised first permanent molar: a systematic review

We registered our research topic at the Center for Reviews and Dissemination (registration number CRD42015020275). We developed a research protocol that included a search strategy, inclusion/exclusion criteria, and a data extraction plan. The search started in May 2015 and ended in 2017. The search strategy comprised key words used separately or in various combinations (“first permanent molar tooth”, “early extraction”, “early loss”, “results” and “effects”), and was performed in three search engines (PubMed, Scopus, and Science Direct) from 1960 to 2017. No language restriction was imposed.

The following inclusion criteria were applied: a clinical trial, case-control, cross-sectional or cohort study design; patient age 5 years (time when FPM eruption starts) to 15 years (time of complete eruption of second permanent molar [SPM]) [6] at the time of extraction; and extraction of FPM due to caries or hypomineralization. The exclusion criteria were: other study design, such as a case report; patient aged older than 15 years or younger than 5 years [6] at the time of extraction; extraction of a tooth other than the FPM; and extraction of FPM as an orthodontic treatment. One study in the Turkish language [7] was translated to English by a native speaker of the Turkish language.

Data were extracted from eligible papers by the three investigators working independently and included the following: site and country, study design, setting, duration, sample size, age of the patient population at the time of extraction, duration of follow-up, and methods used to evaluate the consequences of early extraction of the FPM. Only information specifically related to our research was extracted from the eligible articles.

A meta-analysis was carried out to compare the frequency of spontaneous space closure following FPM extraction in the maxilla with that of the mandible using Review Manager Software (Rev Man 5.1, Cochrane Collaboration). The Mantel-Haenszel method was used to combine the studies for calculation of summary odds ratios and 95% confidence intervals [8]. To decide whether the results of the separate studies could be combined meaningfully, a statistical test of homogeneity was carried out. Based on the chi-square test, an inconsistency coefficient (I² statistic) was computed where a value of more than 50% indicated moderate heterogeneity and a value of more than 75% indicated high heterogeneity [9]. The odds ratios were pooled with a random effects model for heterogeneous studies. Odds ratios with their 95% confidence limits for the individual studies and a summary estimate of effect were graphically displayed in a forest plot.

The three investigators evaluated the strengths and limitations of the considered articles separately and then discussed them together. Any disagreement was resolved by consensus. The strengths and limitations of the studies were evaluated according to the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) checklist. This checklist consists of 22 items that should be reported in the title, abstract, introduction, methods, results and discussion sections of published studies, and used to assess the main observational studies, i.e., cohort, case-control and cross-sectional studies [10]. All reported items, especially in the methods and results sections of the included articles, were considered a strength of the study, while all missing items were considered a weak point. The strength of each article was scored according to the STROBE checklist: 1–7 (poor strength), 8–15 (moderate strength), and 16–22 (high strength). We also used the system devised by Shekelle et al. to grade the evidence and classify the strength of the recommendations in the included articles [11] (Additional file 1).

Our search strategy yielded 1602 hits, comprising 148 in PubMed, 194 in Scopus, and 1260 in Science Direct. Duplicates were removed, leaving 1554 titles. Of these titles, 68 studies were approved for evaluation; their abstracts were screened and those that were beyond the scope of our research were excluded, leaving 29 studies. References of full-text articles were checked for related articles. Five articles were added from the references. The full texts of these studies were acquired and screened according to their inclusion and exclusion criteria. The remaining articles were excluded due to either orthodontic treatment or older age at the time of extraction. No additional information was obtained by contacting authors for their unpublished data. The final result was eleven articles (Fig. 1).

Eight of the included studies were cross-sectional [12‐19], two were case-control [7, 20], and one was prospective and observational [21]. One study [12] was excluded because the data for malocclusion as an outcome of FPM extraction were invalid; in addition, the criteria used to classify normal and abnormal malocclusion were not uniform or acceptable. Table 1 shows the characteristics and outcomes of the included studies.

The consequences of early FPM extraction reported in the selected papers were as follows: effects on post extraction space; effects on development and eruption of the SPM and third molar; caries and/or fillings in adjacent teeth; effects on incisors; and effects on skeletal development.

Six articles reported changes in the post extraction space. One was a case-control, split-mouth study [7] that analyzed linear and angular changes in the SPM and second premolar adjacent to the extraction site in cephalometric radiographs. The FPM was extracted at a mean patient age of 10.5 years. The investigators measured the angles between the long access of the SPM or second premolar and Frankfort horizontal plane in the maxilla and the occlusal plane in the mandible. They also measured the distances between the distal surface of the upper SPM or second premolar and the pterygomaxillary fissure in the maxilla and the distance between the lower SPM or second premolar and the ramus of the mandible. They reported that the FPM extraction space was closed mostly by the SPM rather than by the second premolar. SPM angles and distances increased by 15.65° and 0.18 mm, respectively, in the maxilla and by 4.62° and 3.83 mm in the mandible, while second premolar angles and distances decreased by 5.68° and 2.58 mm in the maxilla and 4.63° and 1.82 mm in the mandible when compared with the angles and distances before extraction. The results also showed a significant change in the means of all angles and distances on the extraction side after one year compared with readings at the time of extraction (P < 0.05). This comparison was performed by superimposition of lateral cephalometric radiographs. However, no statistically significant difference in any of the readings was found on the non-extraction side (P > 0.05) [7].

Jälevik and Möller [14] investigated the effect of extraction of hypomineralized FPMs in a group of children aged 5.6–12.7 years. They found that the overall rate of space closure was 85.2%. This higher rate was more frequent in the maxilla (81.6%) than in the mandible (75%), but the difference was not statistically significant (P = 0.518).

Rãducanu et al. [16] reported that post-extraction migration occurred in the following ways: over-eruption of opposing teeth, horizontal migration of neighboring teeth, space reduced by tipping, dual drift (horizontal and vertical), or complete space closure. The age at which FPM extraction was performed had a significant influence on the post-extraction space in both the maxillary (P = 0.02) and mandibular (P < 0.001) arches. The results of this research confirmed that spontaneous space closure does not occur in children who undergo FPM extraction after the age of 11 years. The authors reported that the rate of space closure in the mandible (50%) was significantly greater than that in the maxilla (33.3%; P = 0.024).

Teo et al. [18] evaluated spontaneous space closure after FPM extraction in children aged 7–13 years. They classified space closure into five categories. The first category was complete space closure between the contact points of the SPM and the second premolar. However, the other four categories had remaining space ranging from 1 mm to 5 mm or/and angulation or rotation of the SPM or second premolar. Complete space closure in the upper arch was significantly (P < 0.001) more likely to be achieved than in the lower arch (in 94% versus 66% of cases, respectively).

Rahhal [12, 21] studied the timing of extraction of severely decayed upper FPMs. In his sample of children (mean age 10.5 years) who underwent extraction of an upper FPM, Rahhal found that 84.6% of upper SPMs had complete space closure without any orthodontic intervention and only 15.4% of upper SPMs erupted 1 mm distal to the second premolar.

In a more recent study, Teo et al. [17] reported on space closure in a group of children of mean age 9.2 years at the time of extraction. They studied three confounding radiographic factors related to early lower FPM extraction and its relationship to space closure (Table 1). The three factors were: “1, the second premolar is engaged in the bifurcation of the second primary molar; 2, the SPM is mesially angulated in relation to the FPM; and 3, presence of the third molar”. The presence of all three factors was associated with significantly better space closure (P < 0.001), and a combination of the second and third factors resulted in the most favorable outcomes (P < 0.001).

A meta-analysis was performed for the three studies [14, 16, 18] that reported space closure in both arches separately to compare the frequency of spontaneous space closure of the SPM in the maxilla with that in the mandible. Space closure was more likely to occur in the maxilla than in the mandible, but the difference was not statistically significant (odds ratio 2.06, 95% confidence interval 0.46–9.28; P = 0.35). The heterogeneity between the three studies was moderate (I² = 72%) and statistically significant (P = 0.03). One study found that space closure was significantly greater in the maxilla than in the mandible (94% and 66%, respectively; P = 0.0001) [18]. The second study also found that space closure was greater in the maxilla than in the mandible (81.6% and 75%, respectively) but the difference was not statistically significant (P = 0.518) [14]. In contrast, the third study found that space closure was significantly greater in the mandible than in the maxilla (50% and 33.3%, respectively; P = 0.024) [16]. All these results were illustrated in a forest plot (Fig. 2).

A paper by Telli and Aytan [7] discussed the effect of FPM extraction on the development and eruption of the SPM. They observed that root closure and eruption of the SPM were accelerated on the extraction side when compared with the non-extraction side according to panoramic X-ray (Table 1).

Ay et al. [13] studied the changes in angle and position of the mandibular third molar after unilateral mandibular FPM extraction in a group of children aged younger than 16 years. They used the Pell and Gregory classification for the third molar in relation to the mandibular ramus and its depth of impaction as measured on panoramic X-ray [22]. This classification was described as follows: “Class I, the crown is near the anterior border of the mandibular ramus; Class II, the crown is one-half covered by the ramus; Class III, the crown is completely within the mandibular ramus” (Additional file 2) and as “Class A, the occlusal surface of the impacted tooth is level or nearly level with the SPM; Class B, the occlusal surface is between the occlusal plane and the cervical line of the SPM; Class C, the occlusal surface is below the cervical line of the second molar” (Additional file 3). Angulation of the third molar was measured between the occlusal plane of the first and second premolars and the occlusal surface of the third molar. The angles were classified as follows: vertical, < 11°; mesioangular or distoangular, 11°–70°; and horizontal, > 70°. Ay et al. found that 72% of third molars on the extraction side had their crowns near the anterior border of the mandibular ramus versus 18.7% on the non-extraction side. In addition, 76.6% of third molars on the extraction side had their occlusal surfaces at or near the level of the SPM compared with 46.7% on the non-extraction side. They also found that 81.3% of third molars were in a vertical position on the extraction side versus 34.6% on the non-extraction side. These results indicate that third molars on the extraction side underwent accelerated eruption when compared with the non-extraction side. The difference between the extraction and non-extraction sides was consistently statistically significant (P < 0.001) [13].

Yavuz et al. [19] studied the effects of early FPM extraction (before the age of 12 years) on development and eruption of the third molar in adolescents with mean age of 15.35 years. They reported that 17% of third molars had erupted on the extraction side compared with only 6.6% of third molars on the contralateral side; the difference was statistically significant (P < 0.05; Table 1).

Oliver et al. [15] studied the relationship between FPM extraction and the prevalence of caries and/or filling of adjacent teeth. The extractions occurred at the age of 11–12 years. The authors found significantly fewer decayed and/or filled proximal surfaces in arches with a missing FPM (7.9%) than in arches where the FPM was present (15.3%; P < 0.001). However, there was a significantly greater percentage of decayed and/or filled occlusal surfaces in the adjacent SPM and premolars in the arches with an extracted FPM (34.1%) when compared with arches in which the FPM was present (23.5%; P < 0.001; Table 1).

Normando and Cavacami [20] discussed the effects of FPM extraction on the incisors in subjects aged approximately 11 years. Bilateral loss of the lower FPM resulted in pronounced lingual tipping (P = 0.004) and retrusion (P = 0.03) of the mandibular incisors. However, there was no pronounced effect on inclination of the maxillary incisors or their anteroposterior position (Table 1).

Normando and Cavacami [20] also discussed the effects of FPM extraction on the skeleton. A decrease in the gnathion to sella-nasion angle, (i.e., the point indicating anteroinferior growth of the mandible) (P = 0.05), counter clockwise rotation of the occlusal plane (P = 0.0003), and a mild decrease in lower anterior face height (P = 0.048) occurred as a result of bilateral loss of the lower FPM. There was no significant effect on the maxillomandibular relationship in the anteroposterior direction (Table 1).

Two of the included articles [17, 18] suggested that the ideal time for FPM extraction is when development is at Demerjian stage E [23]. The stages of SPM development according to Demirjian are: “Stage D crown developed, Stage E early bifurcation, Stage F late bifurcation, and Stage G root development almost complete” [18]. In addition, one of the articles examined spontaneous space closure for SPMs that were developed to stage E or F. The spontaneous space closure rate was found to be 58% when the SPM was at stage E and 56.5% when it was at stage F. However, the difference was not statistically significant (P = 1.00) [17].

None of the included studies reported sample size or power calculations. Only one study clearly described the study design [16]. The setting in which the study was carried out was not reported in two of the articles [7, 12, 20]. Six of the studies clearly mentioned patient age at the time of extraction and their age when the consequences of extraction were evaluated. However, two articles did not mention patient age at the time of extraction [12, 16, 20]. The sequelae were closely related to age at the time of extraction. Two articles [16, 21] did not mention age at the time of evaluation of the consequences of extraction. In addition, three of the studies did not clearly state their duration of follow-up [19‐21]. Two studies did not undertake case ascertainment in their study participants [16, 20]; however, the other nine studies addressed case ascertainment by intraexaminer or interexaminer reproducibility or by using precise methods for assessment of the outcomes. Seven studies were retrospective in design and extracted information from patient records, which is considered a limitation [13, 16‐18]. Interexaminer or intraexaminer reliability tests were performed in all except three articles [7, 15, 21]. Confounding factors such as age, gender, and geographic location were discussed by only one study [20] that matched patients for age, while four studies [14, 16‐18] reported the age of their patients but did not match them in this regard. On the other hand, two studies [7, 15, 21] included age in their inclusion criteria. Finally, only one study reported and analyzed radiographic factors, which comprised: the second premolar being engaged in the bifurcation of the second primary molar; the SPM being mesially angulated in relation to the FPM; and the third molar being present [17] (Table 2).

Scoring of the included articles was done according to the STROBE checklist. All of the articles had moderate strength except for two [12, 21] that were found to have weak strength. All the included studies were Category III, i.e., yielded evidence obtained from nonexperimental descriptive studies, such as comparative studies, correlation studies, cohort studies, and case-control studies, according to the system devised by Shekelle et al. to grade evidence [11]. In addition, all articles were scored as Class C, based on category III evidence according to the method used by Shekelle et al. to classify the strength of the recommendations (Table 3).