|Year : 2022 | Volume
| Issue : 3 | Page : 275-282
Development of descriptive mathematical models for different domains of gingival recession using multiple linear regression analysis
Vishakha Grover, Shifali Mahajan, Gurjeet Kaur, Ashish Jain
Department of Periodontics, Dr. Harvansh Singh Judge Institute of Dental Sciences and Hospital, Panjab University, Chandigarh, India
|Date of Submission||26-Apr-2021|
|Date of Decision||29-Oct-2021|
|Date of Acceptance||12-Dec-2021|
|Date of Web Publication||02-May-2022|
Department of Periodontics, Dr. Harvansh Singh Judge Institute of Dental Sciences and Hospital, Panjab University, Chandigarh
Source of Support: None, Conflict of Interest: None
| Abstract|| |
Context: Gingival recession (GR) is a common finding seen in the periodontics clinic. It has a significant functional and esthetic impact on the patient's dentition and quality of life. Aims: The current study aimed to develop the descriptive mathematical models for different domains of GR based on the data obtained from the North Indian population. Settings and Design: Cross-sectional observational study. Materials and Methods: Consecutive 130 participants were enrolled between June and August 2019. Complete case history and thorough oral examination were carried out including assessment of periodontal variables, for example, pocket depth (PD), gingival marginal level, clinical attachment level (CAL), simplified oral hygiene index, and gingival index. Prediction models for different domains of GR, namely Miller's class, severity, extent, and distribution of recession were made, and further, the best-fitted model on the basis of “coefficient of determination (R2)” was analyzed. Statistical Analysis Used: Multiple linear regression. Results: Nine factors, i.e., mean CAL, mean PD, tooth mobility, abrasion, width of attached gingiva, number of teeth present, age, type of brush, and socioeconomic status showed a significant association with different domains of the GR. In addition, a high degree of overlap was observed among factors associated with different domains of the GR. Conclusion: Diverse clinical (mean CAL, mean PD, tooth mobility, and abrasion), biological (width of attached gingiva, number of teeth present, and age), and environmental factors (type of brush and socioeconomic status) were found to have a significant association with the occurrence of GR in the North Indian population. Owing to the multifactorial etiology of GR, the identification of susceptible patients based on the presence of risk factors is an essential step in developing action plans for the prevention of the disease.
Keywords: Descriptive models, diagnosis, gingival recession, multiple linear regression, risk factor, treatment planning
|How to cite this article:|
Grover V, Mahajan S, Kaur G, Jain A. Development of descriptive mathematical models for different domains of gingival recession using multiple linear regression analysis. J Indian Soc Periodontol 2022;26:275-82
|How to cite this URL:|
Grover V, Mahajan S, Kaur G, Jain A. Development of descriptive mathematical models for different domains of gingival recession using multiple linear regression analysis. J Indian Soc Periodontol [serial online] 2022 [cited 2022 May 21];26:275-82. Available from: https://www.jisponline.com/text.asp?2022/26/3/275/344499
| Introduction|| |
Due to the increasing demand of patients for an esthetically pleasing smile, gingival recession (GR) has come to the forefront in clinical periodontal practice. It is a common clinical condition in which the gingival margin is located apical to the cementoenamel junction with concomitant exposure of the root surface. GR may lead to sensitivity and root caries. However, even without the disease and discomfort, it is a very significant esthetic concern for many patients.
GR is a clinical situation that is highly prevalent in adult populations., Richmond et al. showed that GR was recognized as an important factor in oral health perception. The etiology of GR is multifactorial including various anatomic, physiological, pathological, and traumatic factors that enhance the likelihood of the appearance of GR.,, In addition, a few deleterious oral habits such as smoking and lip piercing also been implicated in the initiation of GR., Gingival biotype, oral hygiene habits, and periodontal disease have been most widely known, studied, and documented as significant factors that cause GR. However, each individual may have a different set of combinations of etiological factors. A careful look at different cross-sectional and longitudinal studies revealed that while a specific factor may be related to the prevalence of GR, whereas others could be more close to the prognosis of existing GR or the prevention of GR in predisposing sites.
Further, frequent observations to identify the role of various risk factors revealed considerable differences exist between study populations.,, For example, the study done in the Yemeni adult population showed a high prevalence of GR (60.5%) and identified sex, age, poor oral hygiene, destructive periodontitis, malposition, aberrant frenum, and khat chewing as significant. In the French population, age, gender, plaque index, and tobacco consumption were found to be independent risk factors for the extent and severity of GR with the number of missing teeth.
Risk assessment models or algorithms take the form of a series of patient-specific data entries representing the constellation of accepted risk factors for periodontal disease, which are then subjected to some form of data analysis. Based on such models, a more sophisticated assessment of the clinical condition including quantification of disease severity commonly associated with a specific diagnosis, a general prognosis, and treatment interventions typically associated with a periodontal condition as modified by risk have been reported. The information collected by such tools would probably help in evolving long-term strategies to prevent the occurrence of GR by identifying the individualized set of risk factors for GR and shall enable us to predict the rate of success of the therapeutic measures. However, not much work has been carried out in the context of developing mathematical models for GR to understand the susceptibility profile of the Indian population. GR is a complex entity that can be studied systematically under diverse domains such as class, severity, extent, and distribution, for more objective, better assessment and clearer understanding. Hence, the objective of this study is to develop descriptive mathematical models for different domains of GR as seen in the North Indian population.
| Materials and Methods|| |
A cross sectional observational study has been conducted at Dr. HSJ Institute of dental sciences, Panjab University, Chandigarh from June to August 2019. One hundred and thirty consecutive patients reporting to the outpatient department of the dental institute for dental checkups and treatment were enrolled in the study. The study has been approved by the Institutional Ethical Committee (PUIEC/2018/105/A/09/01) and has been conducted in accordance with the Helsinki Declaration. The written informed consent was obtained from each subject. A complete history was taken and clinical examination was done by a single trained and calibrated dentist with the help of a sterilized mouth mirror, periodontal probe, and disposable gloves under artificial light for each participant. Code sheet for recording both dependent and independent factors has been provided as [Supplementary Table 1]. GR has been studied under four different domains in the present study defined as follows:
- Class of recession (MR): Miller's Classification of Recession
- Severity of recession (SR): Recorded with the help of UNC-15 probe in mm unit
- Extent of recession (ER)
- ER-Qualitative: Location of recession (None, localized, and generalized)
- ER-Quantitative: Number of teeth involved taken into consideration.
- Distribution of recession
- Jaw involved in recession (JR): Maxilla and Mandible
- Surface involvement of recession (SuR): Labial, lingual, and proximal surfaces
- Position of recession (PR): Anterior and posterior teeth
- Site of recession (SiR): Left and right side of teeth.
For questionnaire data, information was obtained by structured questionnaire with interview method that included the recording of data pertaining to age, sex, education, medical and dental history, history of dental trauma, history of smoking and tobacco chewing, type of brush, brushing method, frequency and duration, proximal cleaning, previous history of dental and orthodontic therapy, dental and gingival trauma and medical history. The socioeconomic status of participants was determined by using Modified Kuppuswamy's Scale.
For clinical examination, a width of the attached gingiva was examined by stretching the lip and recorded as adequate or inadequate width of the attached gingiva based on visual observation. To determine the gingival biotype, a periodontal probe was placed in the gingival sulcus, and the presence of transparency was observed. The presence of trauma from occlusion (TFO) and tooth mobility was recorded dichotomously with the help of the Fremitus test and standard method of tooth mobility measurement, respectively. To classify the level of frenum attachment, a tension test was used. The number of teeth present in the oral cavity was noted by the examiner. The presence or absence of malocclusion and abrasion were recorded dichotomously in the study. Simplified Oral Hygiene (OHI-S) Index and Gingival Index were recorded as a standard method., Pocket depth (PD), gingival marginal level, and clinical attachment level (CAL) were recorded for all the teeth present in the oral cavity at six sites per tooth with the help of a UNC-15 probe.
Descriptive data were reported as frequencies, percentages, and ranges. A multiple linear regression method was used to identify the relationship between different variables affecting GR. Based on the “coefficient of determination (R2)” value, the most closely fit model was chosen for further analysis. Analysis of variance was utilized to identify the variance in the association of dependent variables with independent variables. Best models (B) of different domains of GR based on “R2” were computed for Class of recession (B-MR), B-SR, (B-ER-Qualitative and B-ER-Quantitative), and distribution of recession (B-JR, B-SuR, B-PR, and B-SiR). All statistical analyses were carried out in the IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY: IBM Corp.
| Results|| |
Demographic data regarding 130 study participants are shown in [Table 1] with the coding assigned in the study. Based on “R2”, only the best-fitted models are described in [Table 2],[Table 3],[Table 4],[Table 5]. In the tables, a positive coefficient B value indicated that as the value of the independent variable increase, the mean of the dependent variables also tends to increase and vice-versa.
|Table 2: Model of miller's class of recession (dependent variable) - (miller's class of recession)|
Click here to view
|Table 3: Model of severity of recession (dependent variable) - (severity of recession)|
Click here to view
|Table 4: Model of location of recession (dependent variable) - qualitatively (extent of recession - qualitatively) and – Quantitatively (extent of recession - quantitatively)|
Click here to view
|Table 5: Model of Distribution of Recession (dependent variable) - jaw involved in recession, surface of tooth involved in recession, position of recession, and site of recession|
Click here to view
| Discussion|| |
The contemporary dental practice is consistently adopting evidence-based clinical practices., Risk assessment, prevention, diagnosis, and management of GR is a significant and worthy of critical need of evidence-based protocols, especially, looking at the prevalence of the clinical condition. The focus of the current investigation was to identify specific risk factors and develop predictive models for GR, based on the composite effects of multiple risk factors for the Indian population. Multifactorial models have been reported to be of more utility as compared to individual factor-based models. Many statistical methods, for example, logistic regression, linear discriminate analysis, classification and regression tree analysis, Poisson's regression analysis have been commonly utilized, but we have used multiple linear regression methods in our study. The regression methods allow researchers to find factors that contribute to the overall status of response variables (GR, in this case) and also permit to estimate the magnitude of the effect of various factors.
The study findings identified five factors, namely mean CAL, abrasion, tooth mobility, a width of attached gingiva, mean PD, which have a statistically significant relationship with B-MR. CAL is the hallmark of periodontal disease. The positive association between the CAL and GR suggests a higher prevalence of recession in patients suffering from periodontal disease and these findings are in accordance with the previous literature. Abrasion also showed a significant positive association with recession. Forced tooth brushing and level of filament stiffness of brush may put the gingival tissues at risk from mechanical trauma. Beck and Koch reported toothbrush abrasion as an important reason for attachment loss evident as GR. Tooth mobility is an indicator of the functional status of the periodontium and increased mobility has been well known as a factor influencing the severity and rate of progression of periodontal disease. Bernimoulin and Curilovié found no correlation between GR and tooth mobility, which is contrary to our findings. Further, they have reported an influence of TFO on GR and increased mobility. Conflicting relationships of GR and the width of attached gingiva have been documented in past studies conducted in different populations. In our study, a width of attached gingiva revealed a significant positive association with GR, suggesting a higher prevalence of GR in the patients having inadequate width of attached gingiva. These findings are congruent with multiple studies.,, However, two different studies emphasized the significance of maintenance of good oral hygiene in areas of lack of an adequate zone of attached gingiva to prevent GR.,
In the study, the mean PD displayed a significant negative association with GR which in the author's opinion is an oversimplified relationship as both pockets and GR are site-specific entities, but the current analysis of the factors is based on combined mean parameters as observed in the whole oral cavity. Clinical attachment loss around the teeth may clinically appear as deep pocketing or as GR or concurrent presence of the two on different teeth within the same dentition. The relationship between PD, GR, and CAL is complex owing to multiple shared etiologic and biologic factors with diverse effects impacting these three parameters. In most instances of severe GR, minimal PD is observed as the etiology of the recession may completely be rooted in mechanical trauma associated with aggressive or faulty tooth brushing and patients with such habits and history tend to maintain meticulous oral hygiene with minimal dental plaque accumulation or inflammation of the gingival tissues. Thus, periodontal pocket formation is generally not seen in such patients. In other case scenarios where thin gingival biotype is present, there occurs a greater prevalence of GR rather than increased PD. The apparent disparity with current understanding and other findings of the study shall be interpreted in light of the fact that there appears an intricate interplay of a variety of factors associated with tooth brushing (force, frequency, pressure, duration, technique, etc.) along with biologic topographical factors such as gingival biotype, inadequate width of attached gingiva, etc., which bears a significant impact to decipher an individual overall case-based and site-based relationship between mean PD and GR.
There was observed an overlap of factors significantly related to the B-SR with factors impacting class of recession, namely mean CAL, abrasion, and tooth mobility. In addition, the number of teeth present in the oral cavity showed a positive relationship with the severity of GR. The findings are contrary to an investigation carried out by Sarfati et al. who reported a positive association with several missing teeth and emphasized upon more periodontitis, more attachment loss, and consequently more GR. However, this may be true for clinical situations where exclusive periodontitis associated GR is being analyzed, whereas it may be a variable association in case of other etiologic factors such as mechanical trauma or faulty tooth brushing are playing a predominant role. Many studies have indicated a two-way association between GR severity and periodontitis severity, rather than a causal relationship. The gingival tissue inflammation in response to the plaque microbiota is considered as the principal biologic feature shared by GR and periodontitis.
For B-ER-Quant., five factors, namely mean CAL, abrasion, a number of teeth present, mean PD, and socioeconomic status were found to be statistically significant. Mean CAL, abrasion, and mean PD were significant for both previous domains. Socioeconomic status appeared as a significant factor and was negatively associated with GR. This finding was concordant with previous literature suggesting a higher prevalence of GR in lower socioeconomic status patients, irrespective of age, and this association may be related to the lack of awareness and illiteracy among lower socioeconomic status population given by Susin et al. For B-ER-Qual., six factors, namely mean CAL, age, abrasion, number of teeth present, and type of brush were found to be statistically significant. In 2011, Amran and Ataa also observed the prevalence of GR increases with age probably due to the longer period of exposure to the agents that caused GR to be associated with local and systemic alterations for the individual. Multiple studies ascertained the same findings pertaining to age and GR.,,,, In 2012, Minaya-Sánchez et al. reported that the number of sites (or the risk to develop sites) with GR increased with increasing age. GR was positively associated with the type of brush and reflected more prevalence in the patient who used manual toothbrushes as compared to powered toothbrushes. However, a recent systematic review did not support the association between tooth brushing and noninflammatory GR due to insufficient data in this context. Tooth brushing factors such as duration and frequency of brushing, technique, brushing force, frequency of replacement of brushes, and bristle hardness may be associated with rates or extent of gingival trauma and thus GR has been implied by many earlier investigations. The rest associated factors with this domain were shared with other domains of the GR in this study viz. MR, SR, and ER.
For DR domain, five factors, namely mean CAL, age, number of teeth present, type of brush, and mean PD were found to be statistically significant in the context of the B-JR, whereas for B-SuR, only two factors were found to be statistically significant, viz. mean CAL and age. Four factors were found to be statistically significant for association with the B-PR namely mean CAL, age, abrasion, and type of brush. Further, only three factors, namely mean CAL, age, and a number of teeth present were found to be an insignificant positive association for B-SiR. Based on this data, there was seen a high degree of overlap among factors associated with different domains of the GR.
The present study has several strengths. Beyond traditional local factors, our analysis included diverse biologic and environmental factors and explored their role as risk factors to cause GR. To our knowledge, these variables generally have not been included in multivariate models dealing with recession defects. Further, exploring the eight different domains of GR makes this study quite distinguishable and different from other similar studies. The inherent limitations of the investigation are the cross-sectional design and the exploratory multivariate approach where the results of analysis depend on the only variables that are included in the model. Further, as the study sample was a small convenience sample obtained from a dental institute, the findings have limited external validity and should be ascertained in the large population-based study for better insights. This model has been conceived as a dynamic tool that is capable of progressively incorporating additional and reliable data to overcome its current limitations. Longitudinal prospective investigations are needed to ascertain and validate the currently developed mathematical models on an epidemiologic basis to pave the path for predictive modeling for GR in the Indian population.
| Conclusion|| |
Overall, the present models of included variables identified nine factors, namely mean CAL, mean PD, a width of attached gingiva, tooth mobility, abrasion, type of brush, number of teeth present, socioeconomic status, and age showed a significant association with different domains of the GR, whereas sex, previous dental and orthodontic therapy, previous history of gingival trauma, medical history, smoking, and tobacco chewing habits, the texture of bristles, type of brushing method, frequency and duration of brushing, proximal cleaning, malocclusion, gingival biotype, TFO, frenum attachment level, CI-S, OHI-S, GI did not show an association with the occurrence of GR. Further, there is essentially a need for validation studies of the proposed models in the Indian population with GR to ascertain the working of the models in different domains. However, the present study provides comprehensive yet compact descriptive mathematical models regarding the association of GR based on the composite collection of periodontal, systemic, and environmental variables and can serve as a foundation for the development of predictive tools for different domains of GR in the Indian population.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Guimaraes GM, Aguiar EG. Prevalence and type of gingival recession in adults in the city of Divinopolis, MG, Brazil. Braz J Oral Sci 2012;11:357-61.
Amran AG, Ataa MA. Statistical analysis of the prevalence, severity and some possible etiologic factors of gingival recessions among the adult population of Thamar city, Yemen. RSBO 2011;8:305-13.
Smith RG. Gingival recession. Reappraisal of an enigmatic condition and a new index for monitoring. J Clin Periodontol 1997;24:201-5.
Richmond S, Chestnutt I, Shennan J, Brown R. The relationship of medical and dental factors to perceived general and dental health. Community Dent Oral Epidemiol 2007;35:89-97.
Alldritt WA. Abnormal gingival form. Proc R Soc Med 1968;61:137-42.
Wennström JL, Lindhe J, Sinclair F, Thilander B. Some periodontal tissue reactions to orthodontic tooth movement in monkeys. J Clin Periodontol 1987;14:121-9.
Sarfati A, Bourgeois D, Katsahian S, Mora F, Bouchard P. Risk assessment for buccal gingival recession defects in an adult population. J Periodontol 2010;81:1419-25.
Kapferer I, Benesch T, Gregoric N, Ulm C, Hienz SA. Lip piercing: Prevalence of associated gingival recession and contributing factors. A cross-sectional study. J Periodontal Res 2007;42:177-83.
Nwhator OS, Ogunbodede E, Adedigba M, Sagay D. Prognostic indicators of gingival recession in Nigeria: Preliminary findings. TAF Prev Med Bull 2010;9:187-94.
Susin C, Haas AN, Oppermann RV, Haugejorden O, Albandar JM. Gingival recession: Epidemiology and risk indicators in a representative urban Brazilian population. J Periodontol 2004;75:1377-86.
Marini MG, Greghi SL, Passanezi E, Sant'ana AC. Gingival recession: Prevalence, extension and severity in adults. J Appl Oral Sci 2004;12:250-5.
Kye W, Davidson R, Martin J, Engebretson S. Current status of periodontal risk assessment. J Evid Based Dent Pract 2012;12:2-11.
Jain S, Kaur H, Aggarwal R. Classification systems of gingival recession: An update. Indian J Dent Sci 2017;9:52-9. [Full text]
Kumar BP, Dudala SR, Rao AR. Kuppuswamy's socio-economic status scale – A revision of economic parameter for 2012. Int J Res Dev Health 2013;1:2-4.
Reddy RT, Vandana KV, Prakash S. Gingival biotype – A review. Indian J Dent Adv 2017;9:86-91.
Singh DK, Jalaluddin M, Rajeev R. Trauma from occlusion: The overstrain of the supporting structures of the teeth. Indian J Dent Sci 2017;9:126-32. [Full text]
Gaberthüel TW, Mörmann W. The angiographic tension test in mucogingival surgery. J Periodontol 1978;49:395-9.
Greene JC, Vermillion JR. The simplified oral hygiene index. J Am Dent Assoc 1964;68:7-13.
Löe H. The gingival index, the plaque index and the retention index systems. J Periodontol 1967;38:l610-6.
Egger M, Davey-Smith G, Altman D. Systematic Reviews in Health Care: Meta-Analysis in Context. 2nd
ed. BMJ Publishing Group: John Wiley & Sons; 2008.
Ismail AI, Bader JD. Evidence-based dentistry in clinical practice. JDR Clin Trans Res 2004;135:78-83.
Powell LV. Caries prediction: A review of the literature. Community Dent Oral Epidemiol 1998;26:361-71.
Beck JD, Koch GG. Characteristics of older adults experiencing periodontal attachment loss as gingival recession or probing depth. J Periodontal Res 1994;29:290-8.
Bernimoulin J, Curilovié Z. Gingival recession and tooth mobility. J Clin Periodontol 1977;4:107-14.
Stoner JE, Mazdyasna S. Gingival recession in the lower incisor region of 15-year-old subjects. J Periodontol 1980;51:74-6.
Dodwad V. Aetiology and severity of gingival recession among young individuals in Belgaum district in India. Ann Dent Univ Malaya 2001;8:1-6.
Nguyen-Hieu T, Ha Thi BD, Do Thu H, Tran Giao H. Gingival recession associated with predisposing factors in young Vietnamese: A pilot study. Oral Health Dent Manag 2012;11:134-44.
Wennström JL. Lack of association between width of attached gingiva and development of soft tissue recession. A 5-year longitudinal study. J Clin Periodontol 1987;14:181-4.
Kennedy JE, Bird WC, Palcanis KG, Dorfman HS. A longitudinal evaluation of varying widths of attached gingiva. J Clin Periodontol 1985;12:667-75.
Van der Velden U, Abbas F, Armand S, Loos BG, Timmerman MF, Van der Weijden GA, et al.
Java project on periodontal diseases. The natural development of periodontitis: Risk factors, risk predictors and risk determinants. J Clin Periodontol 2006;33:540-8.
Albandar JM, Kingman A. Gingival recession, gingival bleeding, and dental calculus in adults 30 years of age and older in the United States, 1988-1994. J Periodontol 1999;70:30-43.
Holtfreter B, Schwahn C, Biffar R, Kocher T. Epidemiology of periodontal diseases in the Study of Health in Pomerania. J Clin Periodontol 2009;36:114-23.
Rios FS, Costa RS, Moura MS, Jardim JJ, Maltz M, Haas AN. Estimates and multivariable risk assessment of gingival recession in the population of adults from Porto Alegre, Brazil. J Clin Periodontol 2014;41:1098-107.
Minaya-Sánchez M, Medina-Solís CE, Vallejos-Sánchez AA, Marquez-Corona MD, Pontigo-Loyola AP, Islas-Granillo H, et al.
Gingival recession and associated factors in a homogeneous Mexican adult male population: A cross-sectional study. Med Oral Patol Oral Cir Bucal 2012;17:e807-13.
Rajapakse PS, McCracken GI, Gwynnett E, Steen ND, Guentsch A, Heasman PA. Does tooth brushing influence the development and progression of non-inflammatory gingival recession? A systematic review. J Clin Periodontol 2007;34:1046-61.
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]