Four-Rooted Maxillary First Molars: A Systematic Review and Meta-Analysis

Objectives The majority of human maxillary first molars is usually described as having three roots, but different morphologies were documented in several studies and case reports. One very rare and less investigated anatomical anomaly is the occurrence of four radicular structures in the upper first molars. This communication aimed to define the prevalence of four-rooted maxillary first molars on a meta-analytical basis. The external and internal morphology of these teeth was described by the collection of published case reports. Materials and Methods Six electronic databases were accessed to collect case reports dealing with four-rooted maxillary first molars, as well as population-based cone-beam computed tomography (CBCT) studies. Afterward, the publications were selected according to predefined inclusion/exclusion criteria and evaluated using the Joanna Briggs Institute Critical Appraisal tool. The teeth of the chosen case reports were then independently analyzed by two dental professionals according to different dental classifications. Furthermore, the population studies were meta-analyzed to calculate the global and regional prevalence of four-rooted maxillary molars. Results Included were forty-nine population-based CBCT studies containing 26663 maxillary first molars. Upon these data, the global incidence of four-rooted maxillary molars was meta-analytically determined as 0.047% (95%-CI:0.011–0.103%). In combination with the case reports, it was pointed out that this anomaly is distributed worldwide. Furthermore, forty-eight case reports were included containing fifty-three maxillary molars with four roots. The analyzed teeth exhibited Versiani´s pulpal chamber floor Types A and B. The majority of four-rooted maxillary first molars were classified as Type I regarding Christie's configuration. But, also 7.54% of the altered teeth could not be described by this classification. 62.34% exhibited four root canals, but also variations with five, six, or seven canals were identified. Furthermore, a significant difference was found in the occurrence rate between male and female patients. Conclusion Due to the worldwide occurrence, dental professionals should be aware of this rare anomaly to avoid treatment errors, especially during endodontic or surgical therapies.


Introduction
For centuries, anatomists have investigated dental anatomy to describe each human tooth by determining its coronal and radicular structures. Human maxillary first molars are generally characterized to possess four or five cusps (the additional mesiolingual cusp of Carabelli) and three roots [1,2]. ese roots were designated due to their anatomical position as mesiobuccal, distobuccal, and palatal root, usually exhibiting four root canals (mesiobuccal root with a second canal, whereas the other roots commonly contain only one canal) [3,4]. However, numerous studies and case reports also presented maxillary first molars with various anatomical alterations resulting in an enormous diversity in the number of roots and canals. For example, rare cases of single, two-, or even five-rooted maxillary first molars were described [3,[5][6][7] as well as three-rooted versions with multiple canals [8][9][10].
Another rare aberration was described by ews et al. 1979 [11], who identified radiographically four separate roots during endodontic treatment. is unlikely morphological anomaly was classified upon the radicular shape and the degree of root separation by Christie et al. in 1991 [12].
eir characterization was based mainly on studying endodontic treated or extracted maxillary molars and identified three different radicular configurations. Type I maxillary molars have widely divergent, long, and tortuous palatal roots with "cow-horn" shaped buccal roots. Type II maxillary molars have four, shorter, parallel running roots with blunt apices. By definition, "a type III maxillary molar is constricted in root morphology with the mesiobuccal, mesiopalatal, and distopalatal canal encaged in a web of root dentin" [12]. Furthermore, Baratto-Filho et al. in 2002 demonstrated an endodontic case with fused mesiobuccal and mesiopalatal roots and suggested an additional class IV [13] (Figure 1).
However, Versiani et al. in 2012 indicated by studying four-rooted maxillary second molars with micro-CT the Christie's configuration as not feasible, because fusions might occur on a different root level [14]. erefore, they defined a classification introducing a new type III with less divergent and short palatal roots along with widely divergent buccal roots. In addition, Christie's Types II and III were combined [12,14]. However, another classification for this very rare anatomical anomaly was designed by naming the additional palatal root as radix mesio-or distolingualis based on its direct affinity to the pronounced part of the crown [15]. Moreover, Carlsen and Alexandersen described maxillary molars exhibiting three buccal roots and characterized the additional radicular structure as radix paramolaris [16] ( Figure 1).
Interestingly, all mentioned classifications depended on the investigation of the second or even third maxillary molars and were commonly used for the description of maxillary first molars without any scientific proof [12,14,15]. However, due to the rarity of four-rooted maxillary first molars, no anatomical study could be found in dental literature [3]. erefore, this study systematically collected case reports on maxillary first molars with four roots to study their specific anatomy. In addition, population-based cone-beam computed tomographic (CBCT) studies were analyzed to assess the unknown worldwide prevalence and distribution of these teeth.

Literature Search Strategy.
is systematic review, case report collection, and meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [17]. A literature search was conducted between August 2019 and November 2019 by both authors. Six electronic databases were searched (Google Scholar, PubMed, BASE (Bielefeld Academic Search Engine), SciELO (Scientific Electronic Library Online), AJOL (African Journals OnLine), and DHBD (Circumpolar Health Bibliographic Database)) for population-based studies using CBCT imaging for the investigation of maxillary first molars root anatomy according to specific terms (MeSH terms: "maxillary first molar" "cone-beam computed tomography"). Publications in Chinese, English, German, Italian, Japanese, Portuguese, Russian, Spanish, and Turkish were identified. Afterward, the references for these studies were hand-searched. e selection procedure of the records followed a three-step evaluation. At first, titles and abstracts were accessed and characterized as relevant or irrelevant in agreement with predefined inclusion and exclusion criteria (Suppl . Table 1). Afterward, the full texts (if available) of the selected articles were evaluated according to the mentioned criteria. e resulting articles were critically assessed and evaluated independently by both authors based on the JBI questionnaire [18]. Furthermore, few authors were contacted via e-mail to receive missing information on their studies according to the inclusion criteria and the JBI questions.
Additionally, the mentioned databases were accessed to identify case reports concerning maxillary molars with four roots (MeSH terms:"maxillary first molar" "four roots," "maxillary first molar" "two palatal roots," "maxillary first molar" "Radix mesiolingualis," "maxillary first molar" "Radix distolingualis"). Publications in Chinese, English, Farsi, Portuguese, Spanish, and Turkish were identified. Afterward, the reference lists of the identified case reports were hand-searched. As mentioned above, at first, the titles and abstracts were evaluated according to predefined inclusion and exclusion criteria (Suppl. Table 1). Second, the available full texts were analyzed upon these criteria. Finally, both authors defined the case report teeth upon Christie's radicular configuration regarding the published radiographs [12]. Afterward, when available, the intraoperative photographs of the teeth were classified according to Versiani's pulpal chamber floor type [14] (Figure 2). Disagreements were resolved through consensus.

Statistical Analysis.
e summary of the selected (population-based CBCT) studies and the calculation of the pooled proportion of teeth with 4 roots were carried out by a random-effects model with inverse-variance weights. Since some CBCT studies showed no record of upper first molars with 4 roots, the proportions of individual studies were first transformed for the calculation of the pooled proportion (Freeman-Tukey Double Arcsine Transformation). e graphical representation of the proportions of the individual studies and the fraction pooled share with a 95% confidence interval was carried out with a forest plot. In addition, the dispersion of the studies was illustrated in a funnel plot in which the number of teeth of each study was plotted against the proportion of teeth with 4 root canals. Furthermore, Cohen's kappa was calculated to quantify the degree of compliance of two assessors in answering the JBI questions. All statistical tests were carried out on two sides at the significance level 0.05. Stata/IC 16.1 for Unix (StataCorp 4905 Lakeway Drive, College Station, TX 77845, USA) was used for data preparation and analysis.

Included Studies.
e electronic and manual search identified 117 relevant studies for the population-based CBCTstudies. Sixty-eight studies were excluded by accessing the title, abstract, and/or full text ( Figure 3). e JBI questions [18] Q5 and Q9 were considered not applicable, and Q6 was eliminated due to the predefined CBCT technique like previous dental studies described [19]. e results of the Cohen kappa interrater reliability for the publications investigated by the JBI questionnaire were 0.883 (Q1), 0.795 (Q2), 0.979 (Q4), 0.645 (Q7), and 0.764 (Q8). e resulting 54 studies had an average score of 77.2%. Five studies were excluded due to their low score (≤50%), whereas 38 papers had a high (≥75%) and 11 a moderate (≥50%) score. e finally selected 49 studies included 26663 investigated maxillary first molars (Table 1).
From 140 selected case reports concerning four-rooted maxillary first molars, 93 publications were excluded based on the predefined inclusion and exclusion criteria. One population-based CBCT study also documented a micro-CT of a four-rooted molar and was included ( Figure 3). e finally selected 48 studies included 53 four-rooted maxillary molars ( Table 2). In combination with additional Type A Type B Type C Type D Figure 2: Schematic drawing of Versiani's configuration of canal orifices in four-rooted maxillary second molars [14]. e orifices were classified in relation to the pulpal chamber floor as Type A (irregular quadrilateral-shaped), Type B (trapezoid-shaped), Type C (lozengeshaped), and Type D (kite-shaped).
Christie [12,13] Versiani [14] Carlsen and Alexandersen [15,16] Type I   Type I  Type II  Type III   Type II  Type III  Type IV   BV  PV  BV  PV [12][13][14][15][16]. Christie classified fourrooted maxillary molars in Type I to Type III based on the radicular shape and the degree of root separation [12]. Type IV was added by Baratto-Filho [13] with a fused mesiobuccal and mesiopalatal root. Versiani's modified radicular classification [14]. Carlsen and Alexandersen defined and named additional radicular structures upon their buccal or palatal location and according to their affinity to the dental crown. BV:buccal view, PV:palatal view, MV:mesial view, RPM:radix paramolaris, RML:radix mesiolingualis, RDL:radix distolingualis, RML/DL:radix mesiolingualis/distolingualis. information in the CBCT studies, the sample size of the analyzed teeth could be increased. So, 75 maxillary first molars with four roots could be investigated regarding their amount of root canals and 67 teeth to study the gender of the patients. On the other hand, the radicular subtype (46 teeth), Versiani's pulpal chamber floor type (33 teeth), and the distinction of the right or left molar (49 teeth) had a less amount of studied subjects due to missing data in the case reports.

Anatomical Description of Four-Rooted Maxillary
Molars. e analysis of the case reports (Table 2), according to Christie's radicular structure, demonstrated that 37.7% of the authors used this classification. By using Christie's accurate description for the radiographic identification, we characterized 52.83% as Type I, 18, 87% as Type II, 18, 78% as Type III, 1, and 88% as Type IV, and 7.54% could not be described by this classification (Table 3). Versiani's introduced pulpal chamber floor type was only used by Magnucki et al. in 2018 [61]. Our analysis classified the geometrical location of the root canal orifices in 45.45% as Type A (irregular quadrilateral), 51.52% as Type B (trapezoid), and 3.03% as Type D (kite-shaped). Type C could not be found in

Discussion
e knowledge of oral anatomy, its anomalies, and their frequencies is fundamental for successful dental therapies. erefore, this systematic review assessed the prevalence of maxillary first molars with four roots in humans and documented a very low global occurrence rate of 0.047%. Nevertheless, these morphologically altered teeth were found worldwide in documented case reports and population studies, except for Eastern Europe, Australia, and mostly Africa. Probably, a higher study activity in the field of dental anatomy could fill these areas, as it is mainly the lack  of publication (e.g., 250 investigated teeth in Australia), which led to these unaddressed geographic regions. Population-based CBCT studies with high amounts of sample size can identify even very rare anatomical anomalies and are therefore a suitable scientific tool [4,19,57]. Besides an appropriate number of studied teeth, a reproducible methodology and detailed description of demographic factors are recommended as a guideline for these studies [4,19]. e present study linked the occurrence of fourrooted maxillary first molars with regionally subgrouped populations based on the demonstrated association of rare morphological variations and ethnicity in dental literature [19,112]. e highest occurrence rate was found in the Greek-Turkish population with a significantly higher prevalence than in all other populations. e lowest rates were identified in Eastern Europe and on the American  Figure 4). However, an anthropological conclusion reconstructing the human prehistoric colonization upon the teeth size as previously shown for C-shaped mandibular second molars [19] could not be demonstrated for four-rooted maxillary first molars.
However, the rarity of four-rooted maxillary first molars underlined by this review (in 26663 investigated maxillary first molars, only 47 exhibited four roots) caused the complete absence of studies concerning this topic [3]. erefore, a systematic collection of case reports regarding . e two gray curves mark the 95% confidence interval, assuming a true percentage of 0.047. e left gray line marks the lower limit of the 95% CI and runs almost parallel to the y-axis; i.e., the lower boundary is 0. Overall, the percentage of teeth with 4 roots is small (0.047%); therefore, a large part of the studies shared 0%. treatment protocols of maxillary molars with four roots was considered to be an appropriate study design. With this technique, the number of 53 teeth could be identified. is amount can be compared to the most extensive published studies on four-rooted maxillary second molars with either 22 [12] or 25 investigated teeth [14]. One of the further strengths of the present review is the combination of CBCT studies and case reports, which results in, e.g., 77 investigated teeth concerning the number of root canals. us, upon this data, an anatomical description of four-rooted maxillary molars should be possible. e radicular structure of four-rooted maxillary first molars is commonly described with Christie's classification (Figure 1), although it was designed mainly upon maxillary second molars [12]. In the present review, 37.7% of the case report`s authors used Christie's characterization. But, even after the application of Christie's defined radiographic regulation, 7.54% of the teeth could not be classified upon the different types. ese unclassified teeth had either fusion between roots [94] or three buccal (two mesiobuccal and one distobuccal) Radixes [88,91]. erefore, Christie's classification cannot be transferred from maxillary second molars with four roots to four-rooted maxillary first molars without neglecting some teeth.
Regarding fusions, this review agreed with Versiani et al. in 2012, who pointed out that Christie's configuration is not feasible because fusions might occur in different levels of all roots (Figure 1) [14]. On the other hand, Versiani's suggested classification also ignored the description of three buccal and one palatal root. But, Versiani investigated fourrooted maxillary second molars, and these teeth exhibit probably no third buccal radicular structure. e main question is whether a second mesiobuccal root could be classified as a mesiopalatal, even if both structures can be found in five-rooted maxillary teeth [113].
However, the definition of Carlsen and Alexandersen, which is rarely used in scientific publications concerning maxillary molars with four roots, included the possibility of three buccal roots but was designed in vitro and on untreated teeth [15,16]. But, their anatomical characterization described the radicular structures based on coronal anomalies [15], which might be decayed under clinical circumstances. In summary, all standard classifications were not entirely feasible or established for the description of four-rooted maxillary first molars. To cover all variants of four-or even five-rooted maxillary molars [113] or O-shaped teeth [114], new classifications have to be designed, which should also support the clinical and radiographic diagnostic.
In relation to Versiani's pulpal chamber floor classification, which defined the localization of the root canal orifices geometrically, the analyzed case reports documented the main allocation on Type A (irregular quadrilateral) and Type B (trapezoid) (Figure 2). With respect to the sample size of 33 teeth, this data corresponds to findings in fourrooted maxillary second molars where Types A and B were exhibited mainly [14]. However, the majority of maxillary first molars with four roots showed one root canal per root, which also correlated with findings in four-rooted maxillary second molars [14]. Five root canals were identified in 27.27% of the cases, but also six or seven canals were found. It must be taken into account that the present review depends mainly on clinical case reports and that in vivo CBCT or in vitro µCT studies would have higher reliability.
Interestingly, the present review identified a significant difference in the occurrence of four roots in maxillary first molars between males and females. Comparable data for four-rooted maxillary second molars have not been published due to the rarity of these teeth. erefore, and under consideration of the sample size of 68 analyzed teeth, the presented data could indicate a morphological discrepancy between the genders, which should be further investigated by worldwide CBCT population studies as mentioned above.
ese future anatomical studies should also focus on coronal anomalies (like a pronounced cusp of Carabelli) [15] as well as on the occurrence of enamel pearls in four-rooted maxillary molars. Coronal factors were often mentioned to support the diagnostic of radicular alterations but have to be scientifically proven. In the present review, only 15.1% of the case reports [11,69,82,87,96,100,102,108] documented further anatomical variations combined with the exhibition of four roots (Table 2). However, dental professionals should be aware of this rare anomaly to avoid treatment errors that result oftentimes in endodontic retreatments [93,95,98,101,115].

Conclusion
e occurrence rate for 4-rooted maxillary first molars in humans is 0.047%. By collecting population-based CBCT studies and case reports, this quadrangular anomaly was described with mainly four root canals. e root canal orifices on the pulpal chamber floor are located in trapezoid or irregular quadrilateral shape.

Data Availability
All underlying data can be found in the manuscript.

Supplementary Materials
Supplementary Table 1:the predefined inclusion and exclusion criteria for the literature selection of populationbased CBCT studies and case reports concerning fourrooted maxillary first molars. Supplementary Table 2: the investigated CBCT studies in regional subgroups to evaluate the worldwide occurrence rate of maxillary first molars with four roots. e final results were illustrated in Figure 4. e analyzed regional subgroups are in alphabetical order Australia (Suppl.