Impact of periodontitis on the Oral Health Impact Profile: A systematic review and meta-analysis

Background. Periodontitis, being a chronic and multifactorial disease, affects oral health, and consequently, the patient’s quality of life (QoL). The assessment of the oral health-related quality of life (OHRQoL) is possible with the Oral Health Impact Profile-14 (OHIP-14) questionnaire comprising 7 subdomains: functional limitation, physical pain, psychological discomfort, physical disability, psychological disability, social disability, and handicap.


Introduction
Periodontitis is a chronic, multifactorial inflammatory disease associated with the dysbiosis of the bacterial biofilm in periodontal pockets, which leads to damage to the attachment apparatus through an inappropriate, usually excessive, host immune inflammatory response. 1 Risk factors affecting the initiation and progression of periodontitis can be divided into 2 categories: non-modifiable (age, gender, race, and genotype) and modifiable (poor oral hygiene, presence of periopathogens in the biocenosis of the oral cavity, nicotine dependence syndrome, selected general diseases -uncontrolled diabetes, obesity or osteoporosis -low socioeconomic status, poor dietary quality, and stress). 2 The clinical consequences of this progressive damage to periodontal tissues have an impact on the health of the oral cavity in the physiological, psychological and social aspects, which affects the patient's quality of life (QoL). 3 The current definition of oral health stresses the interaction between 3 basic elements: the impact of the extent and severity of the disease on the patient's health; the physiological functions of speaking, smiling, chewing, and swallowing; and the social functions enabling unhindered social coexistence. 4 It also draws attention to the need to determine the impact of oral health on QoL (oral health-related quality of life -OHRQoL). 4,5 The ability to measure OHRQoL allows oral health to be linked to both the subjective assessment of the patient's wellbeing and potential limitations in social life. The assessment of OHRQoL enables patient-centered care and helps to identify the needs for health promotion and prevention programs. 6 It is important, therefore, to demonstrate whether and to what extent periodontitis and its severity affects the QoL of the population exposed to it, since it is second only to caries in terms of prevalence.
A psychometric test of QoL involves the patient filling out a questionnaire, which has been analyzed and validated beforehand. In modern dental epidemiology, it is possible to make the overall assessment of OHRQoL through such multiple indicators and indices as the Oral Impact on Daily Performance (OIDP), the Oral Health Index (OHX), the Geriatric Oral Health Assessment Index (GOHAI), the Liverpool Oral Rehabilitation Questionnaire (LORQ), and the Subjective Oral Health Status Indicators (SOHSI). Another group consists of indicators that estimate the impact of specific conditions affecting the sensation of oral health, such as the Xerostomia-Related Quality of Life Scale (XeQoLS), the Dentin Hypersensitivity Experience Questionnaire (DHEQ) and the Quality of Life With Implant Prostheses (QoLIP-10). In addition, there are health-related quality of life (HRQoL) indicators that assess the physical, psychological and social impact of health conditions on an individual's wellbeing, e.g., the European Quality of Life (EuroQol) instrument, the 36-Item Short-Form Survey (SF-36), the World Health Organization's Quality of Life (WHOQOL) instrument, and the Sickness Impact Profile (SIP). 7 One of the most popular questionnaires for the overall OHRQoL assessment is the Oral Health Impact Profile (OHIP-49) 8 and its shortened version -OHIP-14. 9 It is based on Locker's model, which assumes a hierarchical impact of the effects of the disease and a sequential association with 7 dimensions of QoL. This questionnaire assesses the patient's perception of the impact of oral health on the social aspect of their wellbeing. 8 In the shortened version of OHIP, 14 questions are divided into 7 subdomains: functional limitation, physical pain, psychological discomfort, physical disability, psychological disability, social disability, and handicap. 9 The answers are structured on the basis of a 5-point Likert scale, indicating how often the oral cavity problems described in the questions have occurred (from 0 -never to 4 -very often). The score is the total of the answers to all questions, ranging from 0 to 56; the higher the score, the more negative the perception of the impact of oral health on the patient's QoL. The OHIP-14 questionnaire has been validated and translated into many languages, allowing its scores to be compared between specific populations. The Oral Health Impact Profile makes it possible to relate the scores to the clinical parameters of an oral health examination or sociodemographic factors. Based on the original OHIP questionnaire, attempts have been made to create specialized questionnaires on specific conditions that affect oral health, including OHIP-Esthetics, OHIP-EDENT (for edentulous patients), OHIP-TMDs (for patients with temporomandibular disorders), OHIP-PD (for patients with the periodontal disease), 10 and OHIP-CP (for patients with chronic periodontitis). 11 The last 2 questionnaires are designed for people with periodontitis, and the questions included in them concern the most common symptoms and their impact on the patient's wellbeing.
The aim of the study was to conduct a literature review on the impact periodontitis has on the most frequently used OHRQoL assessment indicator, which is the shortened OHIP-14. The review includes the best studies that assessed the impact of periodontitis along with its severity and extent on the scores of this profile as well as the relationship between the subdomains of this indicator and the diagnosis of periodontitis. The combined effect of these studies was assessed in a meta-analysis.

Methods
The following databases were thoroughly surveyed for articles published through May 1, 2020: MEDLINE; Scopus; Google Scholar; and the Polish Medical Bibliography (PBL). Bibliographic lists were searched after the following terms had been entered: 'periodontitis' , 'periodontal disease' , 'quality of life' , 'oral-health-related quality of life' , and 'oral health impact profile' , which corresponds to the Medical Subject Headings (MeSH) key word database.
The initial identification included the titles and abstracts of articles on the impact of periodontitis on QoL. This resulted in 1,346 titles. Selection was then made by rejecting the titles of articles not related to the literature review. Subsequently, the number of titles was reduced by selecting original observational studies published in English, German, Russian, or Polish. Further selection required both reviewers (BPC and TK) to read the full text of a given article and take into account the following criteria: -the use of the shortened OHIP-14 version in a national language; -the application of the definition of periodontitis based on the measurements of pocket depth (PD) and clinical attachment loss (CAL) or bone loss (BL) in radiographs; -a control group, defined as individuals with a clinically healthy periodontium or with gingivitis, not with periodontitis; -a full periodontal examination of all teeth with at least 4 measurement points; -observations among adults aged 18-70 years; -a cross-sectional or case-control study including confounding variables; -the estimation of the strength of the relationship between periodontitis and OHIP-14 with the mean difference (MD), relative risk (RR) or odds ratio (OR); alternatively, available data to calculate one of these measures; and -for studies from the same country, the later one was taken into account, with a preference for data from a national survey.
The following information was gathered about the studies eligible for the final analysis: the authors; the country and year of publication; the number and age of the participants; the definition of periodontitis; the estimation of the link along with the confounding variables taken into consideration; a significant demonstration of the relationship between the OHIP-14 subdomains and periodontitis; and the authors' final conclusions.
To determine the combined effect of studies reporting on the impact of periodontitis on the assessed Quality of Life Index (QLI), a model with a variable effect was chosen (the DerSimonian and Lard method). For the combined studies, the total RR or raw mean difference (RMD) was determined along with the corresponding confidence interval (CI). A significance level of p < 0.05 was adopted. The heterogeneity of the selected studies was assessed using the Q test and I 2 statistics, assuming a significance level of p < 0.05. The cumulative impact of the severity of periodontitis (mild, moderate or severe) on the OHIP-14 values was analyzed as well. The publication bias was assessed with Egger's test (with a significance level of p < 0.05) and by creating a funnel plot. All analyses were carried out using the analysis kit of Statistica, v. 13.1 (StatSoft, Kraków, Poland). Russia (n = 1; Drachev et al. 64 ). Forty-three studies did not meet the inclusion criteria and were not part of the review and meta-analysis. The reasons for eliminating these studies were as follows: the use of the Community Periodontal Index of Treatment Needs (CPITN) or the Community Periodontal Index (CPI) to assess the clinical condition of periodontitis (n = 15) [13][14][15]18,20,24,25,28,31,33,44,47,48,55,56 ; no adopted definition of periodontitis (n = 7) 21,26,34,45,57,62,63 ; the use of an outdated definition of periodontitis from 1999 or before (n = 7) 16,17,23,38,40,51,52 ; the assessment of individuals who were too old or too young (n = 5) 12,30,43,49,64 ; no control group (n = 5) 19,27,58,60,61 ; a lack of complete OHIP data (n = 3) 36,46,59 ; and an inadequate methodology of the OHRQoL indicator. 50 Ten studies (8 cross-sectional and 2 case-control; 8 local and 2 national) were eligible for the review and metaanalysis (Table 1). 22,29,32,35,37,39,41,42,53,54 All of them showed a significant impact of the diagnosis of periodontitis (defined by the measurements of PD and CAL) on the deterioration of OHIP-14. In addition, the studies indicated that the association between the prevalence of periodontitis and OHRQoL was directly modified by the stage and extent of periodontal tissue damage 32,37,54 relative to the number of preserved teeth. 42 The quality of the selected articles is evidenced by the number of variables taken into account, related to both periodontitis and OHRQoL, which distorted these observations. The most frequently assessed confounding factors were age (n = 8), gender (n = 8), the markers of the socioeconomic status (n = 7), nicotinism (n = 6), and the number of teeth (n = 6). Eight observations recorded the effect of periodontitis on the OHIP-14 subdomains. 22,32,35,37,39,42,53,54 Seven of them showed a significant impact on psychological and physical disability, 6 of them showed an impact on psychological discomfort, social disability and handicap, and 5 on functional limitation and physical pain.
In 5 of the studies eligible for the meta-analysis, in which the relationship between periodontitis and OHIP-14 was expressed with the adjusted odds ratio (aOR), a statistically significant relationship was shown, although not a very strong one. 22,29,32,37,39 The strongest correlation was in the Chinese report -aOR for severe periodontitis was 1.63 with 95% CI ranging from 1.41 to 1.98. 37 The relatively weakest correlation, although also statistically significant, was found in the New Zealand study, in which OR, adjusted by only 3 confounding variables, was 1.49 with a fairly wide 95% CI: 1.01-2.19. 39 In our own metaanalysis of these 5 studies, involving a total of 1,869 individuals with periodontitis and 2,805 individuals in the control group from 4 countries (New Zealand, the UK, Brazil, and China), the relationship between the prevalence of periodontitis and an increase in the OHIP-14 index had a cumulative OR of 1.33 (95% CI: 1.25-1.43); p < 0.001 (Fig. 2). The analysis of heterogeneity of these 5 studies did not confirm this relationship (Q = 3.66; p = 0.45; I 2 = 0%; 95% CI: 0-78.6).
In the case of 7 studies eligible for the meta-analysis in which the difference between the periodontitis group and the control group was assessed through a difference in the mean values of the OHIP-14 index, in each study, the mean OHIP-14 value was significantly higher with exposure to periodontitis. 22,32,35,41,42,53,54 These differences were quite diverse, ranging from 9.7 points for severe periodontitis in a recent English study 32 to only 2.8 points in an Australian study. 41 The meta-analysis of these 7 studies, including a total of 1,981 individuals with periodontitis and 3,472 individuals in the control group from 7 countries (Australia, Jordan, Sweden, Brazil, Israel, Turkey, and the UK) showed a moderately significant deterioration of the OHIP-14 index by 4.2 points (95% CI: 3.10-5.31); p < 0.0001 (Fig. 3). The results were heterogeneous (Q = 36.95; p < 0.001; I 2 = 83.76%; 95% CI: 68.1-91.7).   PG -periodontitis group; C -control; P -periodontitis; G -gingivitis; lP -localized periodontitis; gP -generalized periodontitis; NA -not available; PD -pocket depth; CAL -clinical attachment loss; BL -bone loss; CDC/AAP -the Centers for Disease Control/American Academy of Periodontology; BMI -body mass index; CI -confidence interval; RR -relative risk; sP -severe periodontitis; mrP -moderate periodontitis; mP -mild periodontitis; NS -nonsignificant; OR -odds ratio; NR -not reported.
The meta-analysis of 3 studies in which the impact of the severity of periodontitis on the probability of OHIP-14 deterioration was assessed was also carried out (Fig. 4). 32,37,54 Mild periodontitis did not change the likelihood of OHRQoL deterioration, while moderate periodontitis significantly increased this probability by 64% (95% CI: 1.12-2.40); p = 0.012, and severe periodontitis increased the probability by approx. 3.5 times (95% CI: 1.32-9.73); p = 0.012. The higher OR values in this meta-analysis in relation to the one shown in Fig. 2 result from combining the unadjusted and converted measures of MD for 2 observations. 32,54 The visual analysis of the funnel plot (Fig. 5), the noticeable symmetry of the distribution of the 7 points repre senting the results of the meta-analysis shown in the forest plot in Fig. 3 and the insignificance of Egger's test (b0: −1.6; 95% CI: −5.28-2.08; p = 0.31) indicate an insignificant publication bias. Out of the 7 studies analyzed with the funnel plot, 3 of them are outside 95% CI, which indicates a large variety of these results.

Discussion
For the review and meta-analysis of studies on the link between periodontitis and QoL assessed with the most commonly used OHIP-14 index, articles of the best possible methodological quality were selected. The quality assessment of the 2 included case-control studies 32,53 according to the Newcastle-Ottawa Scale (NOS) 65 indicated an average quality (7 stars). An important inclusion criterion was the definition of periodontitis using the measurements of PD and CAL, which is consistent with the current views on the assessment of the periodontal status for epidemiological and clinical purposes. Studies that considered the impact of confounding variables by applying multiple or logistic regression were also included. The introduction of a range limit for age in which periodontitis is one of the most common oral diseases, while minimizing the impact of tooth loss, was of great importance as well.
In all of these studies, a significant deterioration of OHRQoL under the influence of periodontitis was found, and this impact increased with the severity and extent of periodontopathy. This was reflected in the meta-analysis -periodontitis increased the probability of OHRQoL deterioration by 33%, and a significant impact of moderate and severe periodontitis was 64% and 358%, respectively. No reference was found in the available literature for our own meta-analysis. The unambiguity of these findings is distorted only by the heterogeneity of studies in which the correlation measure was expressed with MD (Fig. 3). This was probably due to the type of research conducted as well as to subjective differences in emotional patterns for experiencing discomfort, disability and pain among populations from different continents (East and West Asia, Oceania, Europe, and South America). Three systematic literature reviews (SLRs) on the impact of periodontitis on OHRQoL are not as unambiguous as our review. [66][67][68] In the first one, Al-Harthi et al. indicated that 3 out of 4 studies showed a significant relationship between periodontitis and OHIP-14. 66 This was not confirmed in the Australian study by Marinõ et al. 40 In another review, Buset et al. selected 15 studies on this relationship and its significance was confirmed in 12 of them. 67 No such relationship was found by Marinõ et al., 40 Montero-Martín et al. 47 or Bandéca et al. 15 In the latest available SLR on the impact of periodontitis on OHIP-14, a significant relationship was found in 18 out of 22 of the selected observational studies. 68 This was not confirmed in the study by Bandéca et al., 15 another Brazilian study by Batista et al., 18 an Indian study by Sanadhya et al.,25 or the study in our own review, regarding the observations among homeless people in Hong Kong (the study was excluded on the 3 rd stage of selection). 69 All of those observations, in which this significant relationship was not found, combined 2 methodological characteristics -the assessment of the condition of the  periodontium and a definition of the periodontal disease based on CPI as well as a partial protocol for oral health testing. It seems that an incorrect determination of the rate of prevalence for periodontitis and a lack of the assessment of the periodontal status for all teeth may interfere with observations regarding the impact of periodontitis on OHRQoL.
The impact of periodontitis on the deterioration of OHRQoL is quite clearly explained by the influence of clinical symptoms such as gingival redness and swelling, bleeding while brushing the teeth, gingival recessions, often associated with the excessive sensitivity of the exposed dentin, tooth mobility, pathological tooth migration (PTM), or recurrent halitosis on the dysfunctions of the stomatognathic system, red complex esthetics and self-assessment. These symptoms should be associated only with the periodontal disease; thus, in observational studies, they should be controlled for by taking into account many local factors that may distort them, e.g., caries, mouth and facial pain, an increased tooth sensitivity, or the loss of tooth functionality (less than 10 interdental contact points) and severe tooth loss (less than 10 teeth).
In the last 3 studies included in the final part of our lite rature review, 32,35,37 a significant relationship was confirmed between all subdomains of the OHIP-14 index and periodontitis. The consideration of earlier studies indicated psychological limitation (embarrassment, difficulty in being relaxed) and physical limitation (difficulty in toothbrushing and eating) as subdomains most frequently associated with periodontitis. In turn, pain, discomfort and functional limitation were indicated relatively rarely. The predominance of the impact of periodontitis on the psychological and social subdomains of OHRQoL was also noted by Buset et al. in their literature review. 67 This indicates the predominance of problems experienced by periodontal patients which very often go unnoticed during treatment, which is usually provided only to inhibit functional limitation, eliminate pain and improve an esthetic effect. However, in the patients' perception, the most important problems that periodontitis may cause are psychological discomfort, embarrassment, stress, problems in interpersonal relations, and even difficulties in everyday activities. This shows the need for even greater holistic planning as regards periodontal therapy goals. Lawrence et al. found that the severity of periodontitis led to an increase in the values of only some subdomains, not to an increase in all subdomains that showed incorrect values. 39 This may indicate that the progression of periodontitis affects the quality of particular OHRQoL subdomains and does not affect them all. He et al., on the other hand, showed that with an increase in the severity of periodontitis, there is a significant increase in the highest-scoring answers to the questions in all subdomains. 37 These differences may be caused by national, racial or socioeconomic discrepancies.
The literature review and the resulting meta-analysis have a number of limitations. Firstly, there is a risk of selection bias. Only in 5 of the selected studies was randomized sampling performed. 29,37,41,42,54 A lack of the random selection of individuals for an epidemiological inquiry runs the risk of it being unrepresentative of the general or local population. This particularly concerns surveying individuals in academic centers. Secondly, not all confounding variables can be controlled for. In 4 studies, only 1-3 confounding variables were taken into account. 35,39,41,53 Many of these studies show that general diseases associated with periodontitis and OHRQoL, such as diabetes, metabolic syndrome, osteoporosis, and local clinical conditions, e.g., caries, odontogenic pain and discomfort associated with prosthetic restorations, are not controlled for. Thirdly, the nature of an observational study in the form of a cross-sectional or case-control study does not unequivocally establish that exposure to periodontitis results in a reduced sense of QoL, since OHRQoL before periodontopathy is unknown. In addition, cross-sectional studies are highly exposed to the risk of selection bias. Fourthly, there is a possibility of observer bias. This is often the case when there is no process of calibrating examiners, or assessing inter-or intra-examiner reproducibility, especially in the clinical assessment of CAL. In the studies included in our review, such a procedure was performed only in 5 of them. 22,29,35,37,54 The possibility of making a diagnostic error can come from adopting different definitions of periodontitis based on PD and CAL measurements. Currently, the definition of periodontitis by Eke et al., 70 subsequently adopted by the Centers for Disease Control/American Academy of Periodontology (CDC/AAP), is considered to be the gold standard for epidemiological periodontal research. This literal definition was used only in the study by He et al. 37 Finally, there are limitations due to the use of the OHIP-14 index and varying methodologies for a subjective psychometric test. The questions contained in this questionnaire do not relate directly to the condition of the periodontium, which may lead to incorrect answers, especially when the questionnaire is self-administered. To reduce memory bias, one should limit the timeframe of the events being asked about. The best solution is to appoint a second examiner for a psychometric test who is blinded to the clinical condition of the periodontium, or to conduct such tests before the clinical examination. In the vast majority of studies eligible for this review, the psychometric test process was not described in sufficient detail.
This literature review and innovative meta-analysis provides further evidence on the relationship between periodontitis and OHRQoL, with a strong influence of the severity and extent of periodontopathy of a dose-effect type. The impact of periodontitis on the psychological and social aspects of the patient's perception of the disease has been shown, which should be taken into account for treatment purposes. Recommendations for further studies on the relationship between periodontitis and OHRQoL indicators include conducting more high-quality European (including Polish) studies in this area, using QoL indicators profiled for periodontitis, using a comprehensive protocol for oral health testing as well as defining the prevalence rate, severity and extent of periodontitis according to CDC/AAP, controlling for the widest possible spectrum of local and general confounding factors, and conducting randomized, controlled studies on the impact of non-surgical and/or surgical treatment on OHRQoL indicators.