External apical root resorption in African American orthodontic patients

Abstract

Objective

External apical root resorption (EARR) is a side effect of orthodontic treatment that results in root shortening. However, this condition has yet to be evaluated in African Americans. The aim of this study was to determine the EARR prevalence within this ethnicity and investigate how patient and treatment-related factors contribute to root resorption.

Methods

The records of 336 African Americans treated at the University of Alabama at Birmingham School of Dentistry Department of Orthodontics were retrospectively analyzed with Dolphin Imaging software. Pre-treatment and post-treatment panoramic radiographs were used to measure EARR. Resorption was recorded when final roots were at least 2 mm shorter after orthodontic treatment. Additionally, moderate and severe EARR was reported when 20% and 50% or more of the root structure was lost for any of the four maxillary incisors, respectively. The Pearson chi-square test was used to evaluate the associations of individual patient and treatment-related factors with EARR.

Results

The prevalence of root resorption with 2 mm or greater of root structure loss was 51.8%. The prevalence of ≥ 20% EARR was 29.8%. Only one patient displayed severe resorption (0.3%). The associations between the patient-specific and treatment-specific variables and EARR were not statistically significant (P > .05).

Conclusions

More than half of the African American patients exhibit at least 2 mm of root resorption with orthodontic treatment. However, in this ethnicity, patient-related factors such as age, gender, dental malocclusion, and skeletal classifications, as well as treatment-related factors do not indicate a significant correlation with the risk of developing EARR.

Introduction

External apical root resorption (EARR) is an important sequela of orthodontic treatment that is generally diagnosed radiographically [1]. This multifactorial condition can cause severe root shortening and threaten the longevity of the dentition if it goes unchecked. Resorption may occur on any root in the mouth but is most often seen on maxillary incisors [2].

Various factors may contribute to EARR. There are treatment specific factors that include duration of treatment, magnitude of force, amount of apical root displacement, intermittent versus continuous force application, direction of tooth movement, and type of appliance [3–7]. While previous studies have not reached a consensus regarding which treatment-related factors definitively increase the risk of root resorption, treatment duration [3, 4, 6, 8–10], extraction treatment [11, 12], and higher magnitude of forces [3, 4, 8, 13] are generally accepted to be associated with EARR. High forces often cause hyalinization, which results in the recruitment of inflammatory cytokines and cells that are capable of producing root resorption. Additionally, longer treatment duration may be explained by increased active treatment and more extensive root movement for incisors. Certain movements are considered to increase the risk or root resorption such as intrusive forces and root torqueing. In general, the more the root is moved; the more likely root resorption will occur [3, 4].

There is also variability in the individual’s susceptibility to EARR. While some patients experience severe root resorption, others undergoing similar treatment with similar malocclusions experience very little. These patient-related factors include ethnicity, genetic predisposition, systemic factors, tooth-root morphology, previous dental trauma, type of malocclusion, age, and sex [4, 6, 14, 15]. Currently, there are limited data in the literature about which races and/or ethnicities are most susceptible to develop EARR. A study by Sameshima and Sinclair found that Asian orthodontic patients exhibit a significant reduction in the overall severity of EARR compared to Hispanics and Caucasians [16]. However, the prevalence of EARR was not reported in these populations. Furthermore, there are no studies to date that have investigated incidence of this disorder in African Americans. Identifying the prevalence of EARR in different ethnicities will establish which patient groups are most susceptible and allow for more specialized care.

The goal of this study is to determine the prevalence of EARR in African Americans and to evaluate the effects of patient and treatment-related factors to the risk of developing this disorder. A better understanding of the etiologies of EARR will aid in the prevention of this idiopathic condition.

Materials and methods

This retrospective study was approved by the University of Alabama at Birmingham Institutional Review Board (IRB: # 160428005). Orthodontic records of 336 self-reported African Americans patients treated at the University of Alabama at Birmingham School of Dentistry Department of Orthodontics were included in this study. Patients were selected based on the completion of orthodontic treatment and the availability of initial and final records. To comply with the ALARA radiation principle, all radiographs included in this study were previously taken as part of the orthodontic treatment. We excluded patients with histories of craniofacial syndromes, dental trauma, or prior orthodontic treatment. Additionally, patients with evidence of open root apices in the incisors and unclear reference points on radiographs were not included in the study. We also excluded teeth with fixed prosthesis, buildup, or temporization that involved the incisal edge of the crown if it was done between initial and final radiographs because the length of the crown could have potentially been altered.

Dolphin Imaging software (Chatsworth, CA) was used to digitize pre- and post-treatment radiographs for each patient. The millimetric ruler on the cephalometric radiograph was used to calibrate the measurements of the crown width of the mandibular molars, which was then utilized to calibrate the digitized panoramic radiographs using the annotation tools on Dolphin Imaging. On the calibrated radiographs, the root and crown lengths of the maxillary incisors were measured as previously described [17, 18]. Briefly, three parallel lines were drawn for each of the four maxillary incisors (Fig. 1). The top and bottom lines go through the tooth apex and incisal edge, respectively, and the middle one bisects the lines connecting the mesial and distal cementoenamel junctions (CEJ) [17–20]. The root length and crown heights were measured on the perpendicular to these parallel lines as described in Fig. 1. While previously this method was used to determine the root to crown ratios of permanent teeth in non-orthodontic patients [17], by measuring the pre- and post-treatment radiographs, the current study evaluates the root resorption of orthodontic patients. The amount of root resorption was calculated using the following formula: ((R_i × C_f)/C_i) − R_f, where R_i and C_f represent the initial root length and crown height, respectively and; R_f and C_i the final root and crown measurements, respectively. The percentage of EARR was determined by dividing the amount of root resorption with the initial root length.

Figure 1.

The method used to measure the root and crown lengths. Ai and Af, initial and final root apexes; Mi and Mf, initial and final midpoints of the line bisecting the mesial and distal cemento-enamel junctions; Ii and IF, initial and final incisal edges; Ri and Rf, initial and final root lengths; Ci and Cf, initial and final crown heights.

Patients were identified as having EARR when 2 mm or more of root length was lost for any of the four maxillary incisors. We evaluated root resorption for each tooth type, central incisors (U1), and lateral incisors (U2). The overall prevalence of EARR was determine as the percentage of patients having at least one maxillary incisor with root resorption. EARR was also evaluated in terms of mild, moderate, and severe root resorption. Mild EARR was defined as presenting with < 20% resorption, moderate EARR was equal or greater than 20% and as severe if 50% or more resorption was observed.

Patient age, gender, treatment time, dental classification, and treatment protocol (extraction vs non-extraction) was recorded from the patient records. Skeletal classification and the relative position of maxilla to mandible (ANB), as well as upper and lower incisor proclinations (U1- SN and IMPA, respectively) and overjet (OJ) was recorded from pre-treatment cephalometric radiographs.

Statistical analysis

The sample size needed for the study was calculated for the prevalence of EARR assuming a two-sided 95% binomial confidence interval for one proportion, and an allowable 5% margin of error (i.e. confidence limits that do not exceed the prevalence by more than 5% in either direction). For our final sample size of 336 African American patients, and our observed prevalence of ≥ 20% EARR of 29.8%, the margin of error is 4.9%, and the 95% binomial confidence interval is 29.8% (24.9%, 34.7%). Observed prevalence of ≥ 20% EARR of 10%, 20%, 30%, 40%, and 50% for our final sample size yields margins of error of 3.2%, 4.3%, 4.9%, 5.2%, and 5.3%, respectively. Descriptive statistics were obtained for all patient and treatment-related factors. Logistic regression analysis and two-group t-test (continuous data) or Pearson chi-square test (nominal data) was used to evaluate the associations of individual patient and treatment-related factors with EARR. Inter-examiner reliability (for the two examiners) was assessed using the κ statistic along with its corresponding 95% confidence interval. All statistical tests were two-sided and performed using a significance level of 5%. Statistical analysis was conducted using SAS (version 9.4; SAS Institute, Cary, NC). The lone patient with severe EARR was grouped with the patients with moderate EARR for purposes of statistical analysis.

Results

The orthodontic records of 336 African American patients (199 females and 137 males) were analyzed. The patients varied in age from 9 to 60 years with the median of 14 years. The mean age for our study sample was 17.6 (±9.7) years, with females and males 19.1 (±11) and 15.5 (±6.9) years, respectively.

While 62% of the patients were classified as dental Class I, the majority (71%) displayed a Class II skeletal relationship (Table 1). The cephalometric analysis of the sample included an average ANB of 4.7°, U1-SN of 111.5°, IMPA of 98.1°, and an OJ of 3.9 mm (Table 1).

Table 1.

Characteristics of the pre-treatment and treatment-related variables of the African American patients.

Classification	Class I	Class II	Class III
Dental	209 (62.2%)	101 (30.1%)	26 (7.7%)
Skeletal	76 (22.6%)	239 (71.1%)	21 (6.3%)
	Mean	SD	Range
ANB (°)	4.7	2.8	–3.3–13
U1-SN (°)	111.5	8.0	84.4–136.4
IMPA (°)	98.1	7.7	74–118.2
Overjet (mm)	3.9	2.3	–3.6–14.3
Treatment Time (months)	28.6	8.9	8–61
Treatment Protocol
Extraction	75 (22%)
Non-extraction	261 (78%)

SD: standard deviation.

For each patient, we also examined the treatment protocol used and the duration of the orthodontic treatment. Our sample included 75 cases treated with extractions due to crowding and/or protrusion and 261 patients treated non-extraction. The treatment times averaged 28.6 months with a minimum of 8 months and a maximum of 61 months (Table 1).

To determine the prevalence of EARR, we analyzed the initial and final root lengths and crown heights of 1206 maxillary incisor teeth (635 central incisors and 570 lateral incisors) from 336 patients. The inter-examiner reliability was found to have a κ value of 0.44, which indicates moderate agreement between examiners and was statistically significant (P < .001).

We found that 146 (23.0%) of the central incisors (U1) and 132 (23.2%) of the lateral incisors (U2) displayed at least 2 mm root structure loss (Table 2). Furthermore, 9.9% of U1s and 10.5% of U2s had ≥ 20% resorption. These data show no significant differences in the prevalence of root resorption detected in each incisor type.

Table 2.

EARR prevalence.

	U1 (n = 635 central incisors)	U2 (n = 570 lateral incisors)	U1 or U2 (n = 336 patients)
EARR Overall (≥2 mm)	120 (36.4%)	111 (35.1%)	174 (51.8%)
Moderate/Severe (≥20%)a	59 (17.9%)	51 (16.1%)	99 (29.8%)
No EARR	210 (63.6%)	205 (64.9%)	162 (48.2%)

^aThe lone patient with severe EARR was grouped with the patients with moderate EARR for purposes of statistical analysis.

Of the 336 patients included in the study, more than half of them (174) were diagnosed with EARR at the end of orthodontic treatment (Table 2). This diagnosis was based on the amount of the root resorbed (at least 2 mm) from any of the four incisors measured for each patient. Additionally, a designation of moderate or severe resorption was recorded when at least 20% or 50%, respectively, of the root loss was observed following orthodontic treatment. While only one patient (0.3%) was diagnosed with severe EARR of lateral incisors (7.5 mm that resulted in 52% of root loss), we found that almost a third of them (29.5%) exhibited moderate resorption after orthodontic treatment. In our statistical analysis moderate and severe EARR were grouped together for a prevalence of 29.8% (Table 2).

Next, we looked at the demographic characteristics of our patient sample in relation to the risk of developing moderate or severe EARR with orthodontic treatment. Although, root resorption was more prevalent in males (32.8%) compared to females (27.8%), the results were not found to be statistically significant (P = .31) (Table 3). Patient’s age was also not identified as a risk factor for EARR (P = .93) (Table 3). Furthermore, patient-related factors such as dental and skeletal classifications, as well as cephalometric measurements like ANB, IMPA and overjet, were not statistically significantly associated with moderate or severe EARR in our patient group (P > .05) (Table 3). However, the link between root resorption and increased incisor proclinations (U1-SN) was bordering significance (P = .08) (Table 3).

Table 3.

Association of the patient and treatment-related factors to moderate/severe EARR.

	EARR	No EARR	P-value	Odds ratio (95% CI)
Nominal Factors	N (%)	N (%)
Gender			.31
Female	55 (27.6%)	144 (72.4%)		RG
Male	45 (32.8%)	92 (67.2%)		0.781 (0.487–1.253)
Dental Classification			.78
Class I	63 (30.1%)	146 (69.9%)		RG
Class II	28 (27.7%)	73 (72.3%)		0.899 (0.525–1.505)
Class III	9 (34.6%)	17 (65.4%)		1.227 (0.519–2.900)
Skeletal Classification			.84
Class I	24 (31.6%)	52 (68.4%)		RG
Class II	69 (28.9%)	170 (71.1%)		0.879 (0.503–1.538)
Class III	7 (33.3%)	14 (66.7%)		1.083 (0.387–3.029)
Treatment Type			.18
Extraction	27 (36.0%)	48 (64%)		1.449 (0.841–2.495)
Non-extraction	73 (28.0%)	188 (72%)		RG
Continuous Factors	Mean (SD)	Mean (SD)
Age (years)	17.7 (9.4)	17.6 (9.8)	.93	1.001 (0.977–1.025)
ANB (°)	4.5 (3.0)	4.7 (2.8)	.45	0.968 (0.892–1.052)
Overjet (°)	3.9 (2.3)	4.0 (2.4)	.67	0.978 (0.885–1.081)
U1-SN (°)	112.6 (8.1)	111.0 (8.0)	.0.8	1.027 (0.997–1.059)
IMPA (°)	98.2 (7.5)	98.0 (7.8)	.80	1.004 (0.974–1.035)
Treatment Time (months)	29.0 (8.8)	28.4 (9.0)	.53	1.008 (0.982–1.035)

CI, confidence interval; RG, reference group; SD, standard deviation.

Finally, we examined the association between certain treatment specific factors and EARR. Even though, more resorption was noted in patients who had extractions as part of their orthodontic treatment (36%) compared to the ones that did not include any extraction (28%), the relationship between treatment protocol and EARR was not statistically significant (P = .18) (Table 3). Similarly, in this patient sample, the risk of developing moderate or severe EARR was not statistically significantly associated with treatment time (P = .53) (Table 3).

Discussion

EARR is a multifactorial disorder and a common side effect of orthodontic treatment. Its prevalence and severity vary with race and/or ethnicity [16]. However, these differences are not well documented in the literature. In particular, EARR diagnosis has not been investigated in African Americans, the second largest minority population in the USA. Thus, establishing the EARR prevalence and risk factors in this population is significant, especially for Alabama, since African Americans represent the largest minority group in the state (26.8%) [21].

This retrospective study used panoramic radiographs to measure root resorption of the maxillary incisors. Calibration of the pre- and post-treatment radiographs, using an internal millimetric ruler, decreased magnification errors often associated with panoramic radiographs. However, the image quality may be inferior to other imaging techniques. Periapical radiographs and CBCT show more detailed radiographic images but are not routinely taken for every patient. Due to the difficulty in locating the cementoenamel junction, panoramic radiographs have been described to overreport EARR when compared to periapical radiographs [22]. In order to offset this claim, we excluded teeth that did not have a visible cementoenamel junction in initial and final radiographs. In a study comparing panoramic radiographs and CBCT, panoramic radiographs displayed less EARR than CBCT [23]. Higher resolution images such as CBCT would be ideal for prevalence studies on EARR but are not practical since they are not routinely taken before and after orthodontic treatment.

In this study, the inter-examiner reliability had a κ value of 0.44. According to Rosner, a κ between 0.40 and 0.75 denotes good reproducibility. In addition, our value of κ is highly statistically significant (P < .001), indicating that our inter-examiner reliability is almost certainly not due to chance [24].

Furthermore, based on our methodology, the external validity of this study is high. Our study sample was a fair representation of the African American orthodontic patients in Alabama and was free of any selection biases. Also, since this was a retrospective study, there was no Hawthorne effect that would compromise the results. Moreover, our inclusion and exclusion criteria also limited treatment and outcome biases, thus increasing the generalizability value of the results.

In order to determine EARR prevalence in our African American patients, we used the reported frequency of root resorptions in orthodontic patients to finalized our sample size [8, 25]. Since upper incisors are the most susceptible to root resorption, panoramic radiograph measurements for our study were taken from each of the four maxillary incisors [5, 22, 26]. Once prevalence was established, we looked at different patient and treatment specific factors for associations with EARR.

There is no true consensus in the literature on the proper method to characterize root resorption as either a millimetric amount or as a percentage. Levander and Malmgren considered minor root resorption as being less than or equal to either 2 mm or one-third of the initial root length [27]. They went on to characterize severe root resorption as greater than 2 mm of resorption but less than one-third of the initial root length, and any resorption exceeding one-third of the root was defined as extreme [27]. While other researchers have characterized severe EARR as exceeding 5 mm of root loss [8]. However, crown to root ratios have been shown to differ with ethnicities [17, 28, 29]. A 2 mm loss on a maxillary central incisor of a Hispanic or Asian patient with average root to crown ratios of 1.76 and 1.49, respectively, may present more clinical significance than the same amount seen on an African American or Caucasian patient with average root to crown ratios of 1.83 and 1.86, respectively [17, 28]. Therefore, we decided that percent of root loss would be a preferred metric over the millimetric amount of root loss and more clinically relevant when quantifying the severity of EARR. In our study we recorded the EARR diagnosis when at least 2 mm of root length was lost after orthodontic treatment. However, we characterized this condition as moderate when 20% or greater of root structure loss was observed. We also recorded patients with severe EARR that had lost 50% or more of their incisor roots.

While ≥ 20% resorption was seen in 29.8% of our patients, the overall prevalence of EARR in our study based on 2 mm of root structure loss was 51.8%. The sharp difference in the reported prevalence of EARR may be attributed to the fact that 2 mm is less than 20% of the average root lengths in our patient population; therefore, one would expect to see increased prevalence when the 2 mm mark is considered. Our prevalence for ≥ 2 mm of resorption is similar to the 53.68% published by Neves et al., and slightly lower than the 58% reported by Killiany, which was based on the study by Taithongchai et al [8, 25, 30].

The amount of severe root resorption was seen in only 1 out of 336 patients (0.3%) we examined. This low prevalence may be explained by our quantification of severe EARR as ≥ 50% resorption. One study cited that root resorption of greater than 5 mm occurred in 5% of the population, while resorption in excess of 6 mm was seen in 1.5% of their patients [8]. Another study reported that severe EARR defined as displaying at least 25% resorption of all four maxillary incisors was diagnosed less frequently in African Americans compared to Caucasians and Hispanics [31]. Among the 25 patients they diagnosed with severe resorption, only one was African American. Thus, ethnic differences may also account for differences seen in our study compared to the literature. It is feasible that African Americans are less likely to develop severe EARR due to decreased susceptibility.

Few studies have examined root resorption in the different ethnicities and determine its association with EARR. A study by Sameshima and Sinclair compared Asian, Caucasian, and Hispanic populations. In their study, Asian patients (N = 198) had significantly less EARR than Caucasian (N = 516) and Hispanic patients (N = 137) [16]. Out of these populations, Hispanics were found to have the highest millimetric amount of EARR [16]. Two other studies looked at severe EARR in Brazilian populations and found contrasting data on its prevalence [11, 12]. They defined severe EARR as greater than one-third of root structure loss and while the first one reported it to be 2.9% (N = 129), the second showed a much higher prevalence (14.5%, N = 1049). When using the metric of one-third of root loss (the metric used by both Brazilian groups), we found prevalence to be 3.87% for our African American sample (N = 336). A recent study from Turkey reported severe resorption for the entire dentition as 14.8%; however, this included all patients with greater than 2 mm resorption [32]. It is important to use the same metric for diagnosing EARR when comparing prevalence studies.

The literature is also inconsistent in reporting which teeth exhibit more root resorption. Nevertheless, more evidence supports that lateral incisors are at greater risk to develop EARR [5, 26, 32]. In our patient group, severe resorption was identified in lateral incisors of only one patient. However, there were no significant differences in the prevalence of EARR detected in each incisor type.

As mentioned earlier, EARR is a multifactorial disorder and various patient and treatment-related factors may contribute to its pathogenesis [33]. In this study, we recorded several pre-treatment factors including dental and skeletal classifications, and cephalometric measurements such as ANB, Overjet, U1-SN, and IMPA. Similar to the literature, Class I was the most prevalent malocclusion we observed (62% compared to 56% in Johe et al.) [34]. Furthermore, the initial cephalometric measurements of our patients (ANB: 4.7°, U1-SN: 111.5°, IMPA: 98.1° and OJ:3.9 mm) were within the norms established by Alexander and Hitchcock in 1978 and supported by Bailey and Taylor in 1998 [35, 36].

Related to the risks of root resorption, we showed that the associations between these specific patient-related factors and moderate and severe EARR were not statistically significant (P > .05). Patient demographics such as gender and age were also not associated with an increased risk of developing EARR in our study. Since the amount of apical root displacement has been shown to be a risk factor for EARR [6], we expected that patient characteristics such as increased overjet and flared upper incisors (high U1-SN) might be risk factors for EARR. However, this was not the case in our African American patients (P = .67 and .08, respectively). Other studies by Sameshima et al. and Maues et al. have shown that overjet was statistically significantly associated with EARR [12, 16]. Nevertheless, our results do not support this association when considering that increased overjet did not result in increased EARR for African Americans. This may be explained by the fact that overjet reduction, a goal of orthodontic treatment, may be achieved by various means that may not result in true apical displacement. However, this factor also relates to the orthodontic treatment protocol and a true evaluation of apical displacement would require examination of the post-treatment cephalograms, which were not included in this research project and represents a limitation of our study.

Along with patient-related variables, there are a number of treatment specific factors in the literature that have been investigated in relation to the risk of developing EARR. These include duration of treatment, magnitude of force, amount of apical root displacement, intermittent versus continuous force application, direction of tooth movement, extractions, and type of appliance [3–13]. However, in our study, we limited our investigation to duration of treatment and treatment protocol (extraction vs. non-extraction). Treatment time is generally accepted in the literature to have an association with EARR [3, 4, 6, 8, 9, 12]. This may be attributed to increased active therapy and complexity of tooth movements often resulting in longer treatment times. However, our data showed no statistically significant association between the length of orthodontic treatment and EARR (P > .05). On the other hand, we identified a stronger correlation of EARR with the extraction treatment protocol. Extractions resulted in a 14% increase in the root resorption compared to the non-extraction treatment. Nevertheless, this was not statistically significant (P = .18), which contradicts the studies by Marques et al. and Maues et al. that found extraction treatment to have statistically significant association with EARR [11, 12].

The lack of statistically significant association with the variables in this study stresses the importance that great variability exist in the susceptibility to this multifactorial condition. According to Artun et al., variance of EARR is attributed primarily to an individual’s predisposition and susceptibility, and as little as 20% of the variance can be accounted for with tooth shape and form [5]. However, in their systematic review, Weltman et al. determined that tooth morphology was not related to an increase risk of root resorption [5]. On the other hand, Neves et al. reported recently that rhomboid roots may result in decreased susceptibility to EARR compared to dilacerated, triangular or pipette-shaped ones [30]. Nevertheless, they conclude that this is not a determinant factor of EARR severity. Root morphology, along with a number of other factors we did not evaluated, may constitute confounding variables not included in our analysis, thus representing a limitation of our study.

As discussed previously, all teeth undergoing active orthodontic tooth movement experience root resorption; however, some individuals experience clinically significant root resorption. Harris et al. first established a genetic component with EARR [14]. Hartsfield et al. found that 50–66% of the variation in EARR can be attributed to heritability [37]. Various studies have identified a number of single nucleotide polymorphisms associated with an increased risk of EARR [38–41]. Since the expression of a variant may be influenced by ethnicity, its phenotype or disease risk may also vary by ethnicity [42–44]. Furthermore, inheritance for susceptibility to EARR may not only be polygenic in nature; it can also be influenced by epigenetic interactions including orthodontic force [1].

Further research is needed to identify an individual’s predisposition for root resorption and deliver more specialized care for our patients. Since African Americans are underrepresented in medical and dental research, it is important that future studies make provisions for increasing their enrollment. For clinicians, this will translate into delivering a personalized treatment protocol with fewer side effects.

Conclusions

1. More than half of African American patients exhibit at least 2 mm of root resorption with orthodontic treatment.

2. Moderate EARR, consisting of 20% or more reduction in root length after orthodontic treatment, occurs in one third of African Americans.

3. Severe EARR is very rare in African American orthodontic patients.

4. The associations between investigated treatment and patient variables in African Americans are not statistically significant with moderate and severe EARR.

Funding

This study was supported in part by grants from the American Association of Orthodontists Foundation [2016 Postdoctoral Fellowship Award]; and the National Institutes of Health [UL1TR003096].

Conflict of interest

The authors have no conflicts of interest to declare.

Data availability

The data underlying this article cannot be shared publicly due to ethical concerns (for the privacy of individuals that participated in the study). The data will be shared on reasonable request to the corresponding author.