Relationship Between Herpesviruses and Periodontal Disease Progression

Abstract

AIM:

To investigate the role of Epstein-Barr virus (EBV), cytomegalovirus (CMV) and anaerobic bacteria in the progression of periodontitis.

METHODS:

Eighty-one adults with generalized moderate to severe periodontitis were randomly assigned to: oral hygiene or scaling and root planing +/− placebo or polyunsaturated fatty acids fish oil. Subgingival plaque samples collected from three healthy and three disease sites at weeks 0, 16 and 28 and from sites demonstrating disease progression were analyzed for EBV, CMV, P. gingivalis (Pg), T. forsythia (Tf), and T. denticola (Td) DNA using quantitative polymerase chain reaction.

RESULTS:

CMV was detected in 0.3% (4/1454) sites. EBV was present in 11.5% of healthy sites (89/728) and 27.6% disease sites (201/726; p < 0.0001), but was in low copy number. Disease progression occurred in 28.4% of participants (23/81) and developed predominantly at sites identified as diseased (75/78; 96.2%). CMV and EBV were not associated with disease progression (p = 0.13) regardless of treatment. In contrast, disease sites were associated with higher levels of Pg, Td, Tf and total bacteria, and sites that exhibited disease progression were associated with an abundance of Td and Tf (p < 0.04)

CONCLUSION:

Disease progression was associated with Gram-negative anaerobic bacteria; not EBV or CMV.

INTRODUCTION

Periodontal disease is a chronic inflammatory infection that affects the supporting tissues of the teeth. This highly prevalent infection is caused primarily by increases in Gram-negative bacteria that elicit dysregulated host defense responses, dysbiosis and tissue destruction (Eke et al., 2015, Sanz et al., 2015, Eke et al., 2016). Although a chronic bacterial infection, periodontitis paradoxically displays cyclical periods of exacerbation and remission suggesting that other microorganisms may be involved (Goodson et al., 1982, Lindhe et al., 1983, Teles et al., 2008, Hajishengallis et al., 2017, Gonzalez et al., 2018).

Several studies in recent years point to a putative role of select human herpesviruses (HHVs) in the etiology of periodontitis (Saygun et al., 2004a, Slots et al., 2006, Slots, 2010, Slots, 2015, Zhu et al., 2015, Hajishengallis et al., 2017). These viruses are reported to interact with periodontal pathogens (Imai et al., 2012), infect inflammatory defense cells, and incite cytokine release (Sabeti et al., 2012), which are purported to accelerate periodontal disease progression (Slots, 2007, Slots, 2010). Consistent with this theory, cytomegalovirus (CMV) and Epstein-Barr Virus (EBV) have been reported to be more prevalent in deep pockets (Kato et al., 2013) and diseased sites associated with aggressive forms of periodontal disease (Kubar et al., 2004, Kubar et al., 2005, Saygun et al., 2008, Sharma et al., 2012). However, many of these virological findings are based on qualitative analyses, pooled samples and nested polymerase chain reactions (PCR), which could allow varied interpretations of the results. Contrasting evidence from other labs using more quantitative analyses indicate that EBV and CMV prevalence and viral loads are not correlated with probing depth or attachment loss (Rotola et al., 2008,Sunde et al., 2008, Dawson et al., 2009, Nibali et al., 2009, Stein et al., 2013).

Thus a critical question remains, “Are HHVs significant contributors to periodontal disease or its progression?” A longitudinal study that examines viral presence and viral loads at sites that demonstrate disease progression (DP) would help address this question. Accordingly, our aim was to test the hypothesis that presence and quantity of EBV and/or CMV are associated with the progression of periodontitis in an effect of treatment study design.

MATERIAL AND METHODS

Participants.

This study represents a secondary analysis of a randomized, placebo-controlled, double-blind, 6-month longitudinal intervention study where participants received oral hygiene instructions (OHI) or scaling and root planing (SRP), combined with either dietary supplementation of fish oil capsules (n-3 polyunsaturated fatty acids, PUFA) or placebo (corn-soybean oil). Participants were recruited from persons receiving care at the University of Kentucky College of Dentistry as well as the surrounding counties. Inclusion criteria included ≥ 18 years of age who were in good general health, (excluding the case definition) and had ≥ 20 erupted teeth. Participants had to have the diagnosis of chronic moderate to severe periodontitis (Armitage, 2004, Armitage, 1999) with inclusion criteria of five qualifying sites in two quadrants with a minimum of two affected teeth in each quadrant and each qualifying site having probing depth (PD) ≥ 5 mm, clinical attachment loss (AL) of ≥ 3 mm, and bleeding upon probing (BOP) score of ≥ 2 (0 = one, 1 = pinpoint, 2 = interdental bleeding, 3 = spontaneous/heavy bleeding). Exclusion criteria were periodontal therapy in the past two years, a history of alcoholism; salivary gland, kidney or liver dysfunction; granulomatous diseases; inflammatory bowel disease; diabetes or were undergoing or had undergone organ transplant or cancer therapy. Pregnancy or lactation, use of antibiotics within the past three months or immunosuppressant medication within the last 6 months, need for antibiotics for infective endocarditis prophylaxis during dental procedures, symptoms of acute illness (i.e., fever, sore throat, body aches, and diarrhea), orthodontic appliances or presence of an oral mucosal inflammatory condition (e.g., aphthous, lichen planus, leukoplakia, and oral cancer) also were exclusion criteria. The study was performed at the University of Kentucky between August 2005 and August 2009 and was approved by the University Institutional Review Board. All participants provided written informed consent and received incentives (i.e., monetary compensation and a clinical examination) as part of the study protocol.

Clinical Examination and Plaque Collection.

Periodontal examinations included a plaque index (PI) and measures of PD, BOP, and clinical AL that were performed at weeks 0, 16 and 28 by one investigator (DRD) as previously described (Dawson et al., 2009). Subgingival dental plaque was obtained from three diseased sites (≥ 5 mm PD) and three healthy sites (≤ 3 mm PD) from each participant at each visit. Sites demonstrating BOP and ≥ 2 mm AL compared to prior study visits, were defined as DP and sampled as described in Dawson et al. 2009 (Dawson et al., 2009). All plaque specimens were stored at −80°C and assayed within six months of collection.

Therapy.

Participants were randomly assigned to receive either OHI or SRP, with either dietary PUFA supplementation or placebo capsules using a randomization schedule constructed before study initiation. OHI and SRP were performed by a calibrated dental hygienist during the first 30 days following the baseline visit. Participants returned at week 8 (from baseline) for dental plaque sampling, and had a clinical examination, sample collections, and repeat OHI or SRP by the treatment hygienist at week 16. The final study visit (week 28) consisted of clinical periodontal examination and sample collection. A calibrated periodontist (DRD), who was blinded to the participant group assignments, performed the clinical evaluations. SRP was performed as we have described previously (Preshaw et al., 2008).

Study Design:

The study was a randomized clinical trial based on a 2×2 factorial design (OHI versus SRP and PUFA vs placebo) with an estimated sample size of 25 patients per treatment combination and primary endpoint DP. This estimated sample size had 80% power to detect at the 0.05 level of significance a main effect provided the percentage of patients with a DP in the control arm (placebo or PHI) was at least 40% and the corresponding treatment arm (PUFA or SRP) reduced DP to less than 15%. The current study is adjuvant to this trial focusing on the presence of herpesviruses in those patients who completed the study protocol.

Real-time PCR.

DNA was extracted and real-time PCR performed for CMV, EBV, P. gingivalis (Pg), T. forsythia (Tf), T. denticola (Td) and Fusobacterium nucleatum (Fn) DNA as described previously (Dawson et al., 2009). Fn DNA, which served as a control in the analyses, was readily detected in all PCR amplified plaque samples, albeit four samples (three disease and one healthy site) from one patient were unavailable for analysis.

Statistical Analysis.

Demographic variables were compared between the OHI and SRP groups using a Chi-square test or Fisher’s exact test for categorical responses and a two-sample t-test for interval level responses. Changes from baseline in the periodontal indices were compared between the treatment groups using analysis of covariance with adjustments for age and race. A generalized linear mixed model (GLMM) with a logit link was used to correlate the presence of EBV in a plaque site taking into account the status of the site (diseased versus healthy), the treatment, time of measurement (weeks 0, 16 and 28) and the nesting of sites within health status. Mean total bacteria (recorded on a log scale) and the percentage of the total due to Td, Pg, or Tf were compared using a corresponding linear mixed model (LMM). Odds ratios (and 95% confidence intervals) to measure the association between viral DNA in plaque were based on these GLMM models. To determine the relationship between the prevalence of EBV DNA and increasing BOP, AL, or PD a Cochran-Armitage test for linear trend and a receiver operator characteristic curve (ROC curve for nonlinear trend) were constructed. Viral and bacterial relationships at diseased sites between DP and non-DP groups during study timeline were analyzed using logistic regression for number of sites positive for EBV and using a zero inflated Poisson regression for EBV amount because of the excess of zero counts. Total bacteria, Pg, Td and Tf were analyzed using a two way analysis of variance (ANOVA) with log transformation for normality. All models included the covariates of week, DP, non-DP patient groups and their interaction. If the interaction term was significant, post hoc comparisons were made using a Tukey adjustment. All analyses were performed using the PC SAS 9.1 (SAS Institute Inc., Cary, NC, USA) with statistical significance determined at the 0.05 level.

RESULTS

Demographics.

Eighty-one of 101 enrolled participants completed the six-month longitudinal study and provided samples for analysis. Table 1 shows the characteristics of the participants by treatment group. Participants were predominantly Caucasian, 63.8% male, with a minority identified as smokers (15/81, 18.5%). Clinical measures showed the presence of localized severe periodontitis with generalized moderate periodontitis in 90% of participants, with 10% having generalized severe periodontitis. The groups were similar demographically and clinically at baseline, except for a lower PI in the SRP PUFA group. Table 2 compares the demographic and clinical characteristics of the DP and non-DP groups.

Table 1.

Characteristics of the study population by treatment group at baseline.

	N	OHI PL	OHI PUFA	SRP PL	SRP PUFA	p-value
Age, years mean (SD)	22–69	46.48 (8.95)	44.71 (12.23)	41.07 (12.37)	41.70 (8.03)	0.30a
Gender						0.54b
Female	30	11 (44.00%)	9 (42.86%)	5 (33.33%)	5 (25.00%)
Male	51	14 (56.00%)	12 (57.14%)	10 (66.67%)	15 (75.00%)
Ethnicity						0.87c
Caucasian		9 (36.00%)	10 (47.62%)	5 (33.33%)	8 (40.00%)
African-American		7 (28.00%)	4 (19.05%)	2 (13.33%)	3 (15.00%)
Asian		1 (4.00%)	2 (9.52%)	1 (6.67%)	3 (15.00%)
Other		8 (32.00%)	5 (23.81%)	7 (46.67%)	5 (25.00%)
DP						0.59b
Non DP		18 (72.00%)	17 (80.95%)	9 (60.00%)	14 (70.00%)
DP		7 (28.00%)	4 (19.05%)	6 (40.00%)	6 (30.00%)
Baseline AL mean (SD)		2.09 (1.40)	1.59 (1.04)	2.11 (1.43)	1.59 (0.83)	0.15a
Baseline PD mean (SD)		3.56 (0.63)	3.27 (0.61)	3.67 (0.80)	3.42 (0.59)	0.29a
Baseline BOP mean (SD)		1.35 (0.75)	1.13 (0.47)	1.65 (0.73)	1.28 (0.63)	0.33a
Baseline PI mean (SD)		0.73 (0.34)*	0.71 (0.27)	0.72 (0.33)	0.65 (0.30)	0.036a

^aANOVA (Log transformed outcome),

^bChi-square,

^cFisher’s exact.

AL: attachment loss; BOP: bleeding on probing; DP: disease progression; OHI: oral hygiene instructions only; PD: pocket depth; PI: plaque index; PL: placebo; PUFA: n-3 polyunsaturated fatty acids fish oil capsules; SRP: scaling and root planing; SD: standard deviation

Table 2.

Demographics by disease progression (DP).

	Non-DP n=58	DP n=23	p-value
Mean Age, years (SD)	42.31 (10.70)	47.70 (8.75)	0.023c
Gender n (%)			0.44a
Female	23 (39.66%)	7 (30.43%)
Male	35 (60.34%)	16 (69.57%)
Ethnicity n (%)			0.92b
Caucasian	22 (37.93%)	10 (43.84%)
African-American	12 (20.69%)	5 (21.74%)
Asian	6 (10.34%)	1 (4.35%)
Other	18 (31.03%)	7 (30.43%)
Treatment Type			0.59a
OHI-PL	18 (31.03%)	7 (30.43%)
OHI-PUFA	17 (29.31%)	4 (17.39%)
SRP-PL	9 (15.51%)	6 (26.08%)
SRP-PUFA	14 (24.13%)	6 (26.08%)
Baseline AL mean (SD)	1.52 (1.04)	2.65 (1.23)	<0.0001d
Baseline PD mean (SD)	3.34 (0.53)	3.78 (0.81)	0.0204b
Baseline BOP mean (SD)	1.30 (0.69)	1.41 (0.61)	0.40d
Baseline PI mean (SD)	0.69 (0.31)	0.71 (0.30)	0.98d

^aChi-square test,

^bFisher’s exact test,

^cT-test,

^dWilcoxon rank sum test

AL: attachment loss; BOP: bleeding on probing; DP: disease progression; OHI: oral hygiene instructions only; PD: pocket depth; PL: placebo; PUFA: n-3 polyunsaturated fatty acids fish oil capsules; SRP: scaling and root planing; SD: standard deviation

Detection of CMV, EBV and bacterial species.

CMV was detected by quantitative PCR in 0.3% of subgingival samples (4/1454); three were from disease sites (mean viral load 870.2 copies/site) and one from a healthy site (186 copies/site). EBV was detected in 290 sites throughout the study duration; significantly more often in disease sites (27.6%, 201/726 sites) than healthy sites (12.2%, 89/728 sites) and at each time point, respectively (30.4% vs. 11.9% at baseline; 27.9% vs. 14.0% at week 16; 24.7% vs. 9.9% at week 28; p < 0.0001). When present, the median load of EBV detected at disease sites (9.9 copies per site, range 1 to 1,157,961 copies/site) did not differ from the median load detected at healthy sites (11.3 copies per site, range 1 to 142,248 copies/site). Only 2.2% of samples from healthy sites and only 3.5% from disease sites contained more than 100 copies of EBV DNA. There was no relationship with smoking and the presence of EBV or CMV (data not shown). Clinical parameters (AL, BOP, PD) at diseased sites were similar whether EBV was present or not (Table 3). Also, EBV was not associated with deeper probing depths across groups. Linear mixed model analyses showed that the mean total bacteria count (on log scale) was significantly higher in disease versus healthy sites (p < 0.0001). Bacterial (Pg, Td, or Tf) percentages also were higher in disease versus healthy sites (p < 0.0001 at all three time points; Table 4).

Table 3.

Relationship of EBV with clinical parameters at diseased, healthy and disease progression sites.

		Diseased Sites n=726			Healthy Sites n=728			DP Sites n=78
		EBV+	EBV−	p-value a	EBV+	EBV−	p-value a	EBV+	EBV−	p-valuea
PD (mm)	Mean	5.36	5.47	0.74	2.54*	2.32	0.004	5.06	5.91*	0.01
	SD	1.24	4.62		0.72	0.70		1.42	1.54
AL (mm)	Mean	2.55	3.02	0.18	0.70	0.68	0.88	5.09	5.26	0.67
	SD	1.63	4.90		1.39	1.26		1.38	1.82
BOP (0–3)	Mean	1.82	2.12	0.38	0.34	0.24	0.25	1.46	1.59	0.56
	SD	1.16	4.79		0.87	0.68		0.76	1.10

^ap-values refer to within groups comparisons as determined by T-test

AL: attachment loss; BOP: bleeding on probing; DP: disease progression; EBV: Epstein-Barr virus; mm: millimeters; PD: pocket depth; SD: standard deviation

Table 4.

Relationship of EBV with total and specific bacteria at diseased, healthy and disease progression sites

		Diseased Sites n=726			Healthy Sites n=728			DP Sites n=78
		EBV +	EBV −	p-valuea	EBV +	EBV −	p-valuea	EBV +	EBV −	p-valuea
Total bacteria	Mean	3.73 X 107	2.59 X 107	0.12	3.18 X 106	2.23 X 106	0.29	1.51 X 107	1.22 X 107	0.47
	SD	7.68 X 107	3.47 X 107		8.36 X 106	5.48 X 106		1.42 X 107	1.93 X 107
Pg %	Mean	6.37%	6.69%	0.76	3.43%	2.74%	0.45	8.89%	9.92%	0.67
	SD	7.17%	8.40%		8.86%	5.04%		9.80%	10.81%
Td %	Mean	2.56%*	1.71%	0.02	0.61%	0.71%	0.73	2.92%	3.18%	0.77
	SD	3.41%	2.27%		2.05%	2.29%		4.19%	3.45%
Tf %	Mean	3.47%	3.46%	0.97	1.66%	1.73%	0.90	6.92%	6.38%	0.71
	SD	3.28%	4.04%		3.08%	4.41%		6.02%	6.28%

^ap-values refer to within groups comparisons as determined by T-test

Pg: P. gingivalis; Tf: T. forsythia; Td: T. denticola

Disease Sites Characteristics.

Bacterial and viral changes at disease sites were analyzed between DP and non-DP patient groups at each time point, week 0 (baseline), 16 and 28 (Table 5). Presence of EBV positive at disease sites were similar between groups throughout the study timeline. However, EBV amount was higher in non-DP patient group at baseline compared to DP patient group (p <0.0001). EBV amount decreased at disease sites from baseline to 28 weeks in both groups (p <0.0001). Total bacterial amounts at disease sites were similar between DP and non-DP patient groups at weeks 0, 16 and 28. Both groups presented significant reduction in total bacterial amounts at disease sites from baseline to week 28 (p=0.0097). No differences were detected for specific periodontal bacteria, Pg, Td, Tf at disease sites between groups at different time points (Table 5).

Table 5.

Bacterial and viral changes at disease sites between Non-DP and DP patient groups during study timeline.

	Disease Sites						p-value
	Non-DP patients			DP Patients
	0	16	28	0	16	28
EBV + n (%)	25 (43.10%)	24 (41.83%)	23 (41.07%)	11 (47.83%)	9 (39.13%)	8 (42.11%)	> 0.05 a
EBV + (amount)	6709.95 (50707.10)	222.32 (1569.89)	19.69 (121.72)	1274.64 (6098.00)	41.46 (155.14)	357.50 (1494.29)	<0.0001b
Total Bacteria	3.59 X 107 (4.07 X 107)	1.91 X 107 (2.73 X 107	2.71 X 107 (4.73 X 107)	6.70 X 107 (1.39 X 108)	2.79 X 107 (2.77 X 107)	2.01 X 107 (2.30 X 107)	0.0097 c
Pg (%)	8.66 (9.19)	4.70 (6.24)	5.26 (7.63)	8.54 (8.03)	5.36 (5.70)	8.60 (9.28)	>0.05 c
Td (%)	2.78 (3.26)	1.99 (2.69)	1.58 (2.51)	2.68 (3.60)	1.57 (2.49)	1.51 (1.57)	>0.05 c
Tf (%)	4.20 (3.96)	2.41 (2.31)	2.58 (3.44)	4.30 (3.36)	4.37 (5.19)	4.94 (4.55)	>0.05 c

Pg: P. gingivalis; Tf: T. forsythia; Td: T. denticola

^aLogistic regression

^bZero-inflated Poisson regression: Week, groups and week*groups comparisons, p <0.0001, post-hoc comparisons were all significant p <0.0001

^c2-way ANOVA model (outcome log transformed): Week variance was significant, (p-value=0.0016). Week 0 was significantly different from Week 28 overall (p-value=0.0097). DP and week* patients groups were not significant.

All descriptives are presented as mean (SD) except for EBV+ amount

Characteristics of Disease Progression (DP).

DP developed in 23 of 81 (28.4%) participants (mean 47.70 years, range 37–69 years; Table 2), Three of the 23 had generalized moderate periodontitis at baseline, the remainder had localized severe periodontitis with generalized moderate periodontitis. DP showed no preference by ethnicity or smoking status (nine were smokers), and occurred regardless of treatment group at 78 sites of 29 teeth; a rate of <0.9% of the monitored sites during the six month study. Patients with DP sites were older and had greater PD and clinical AL at baseline compared to those who did not demonstrate DP (p < 0.03). DP developed predominately from disease sites, i.e., 75 of 78 (96.2%). Nineteen DP sites occurred in mandibular teeth and 18 in maxillary teeth. The majority of DP sites (55/78, 70.5%) occurred in eight patients who showed continuing AL and BOP in 20 teeth.

EBV was detected in the plaque of 13 of the 23 (56.5%) subjects who demonstrated DP, and in 29 of 78 sites (37.2%). Nineteen of the 29 EBV positive sites were found in eight participants showing AL on several teeth. Overall, patients in the DP group demonstrated EBV in a similar proportion of DP sites (29/78, 37.2%) as non-DP sites (65/206, 31.4%; p = 0.42), both of which occurred significantly more often than in healthy sites 14.07% (29/206; p = 0.004 and 0.002, respectively). The amount of EBV at DP sites (median 1 copy) was not significantly different from that detected at non-DP sites (median 9 copies; p > 0.05). In contrast, the amount of total bacteria was significantly higher in the diseased sites (2.59 X 10⁷ ± 3.47 X 10⁷) and DP sites (1.22 X 10⁷ ± 1.93 X 10⁷) than the healthy sites (2.33 X 10⁶ ± 5.48 X 10⁶; p < 0.0001 in each comparison).

In terms of bacterial species, Td and Tf were found in significantly higher percentages in both diseased and DP sites than healthy sites (p < 0.0001). Also, Td was significantly more abundant in DP sites than healthy sites (p = 0.028), and Pg was significantly more abundant at diseased sites than healthy (p < 0.0029). Table 4 shows that the bacteria and virus interactions were similar within disease, DP and healthy sites, except that EBV-positive disease sites presented higher Td % compared to EBV-negative sites (p=0.02). Similar amounts of Pg, Td or Tf were detected in disease and DP sites (p > 0.05), and CMV was not detected in any DP site.

Treatment effect.

Treatment group and the presence of EBV were not associated with the development of DP sites over the course of the study (p > 0.05). However, a higher prevalence of EBV was detected in sites after SRP treatment (data not shown), whereas SRP reduced the mean total bacteria (log₁₀ = 0.59) compared with OHI (p = 0.045) based on the 16 and 28 weeks follow-up data. Notably, there was no treatment group effect of SRP versus OHI with or without placebo/PUFA on the percentage of total bacteria involving Td, Tf, or Pg. Also, there was no treatment effect on viral loads.

DISCUSSION

This observational study aimed to investigate the role of specific HHVs in the progression of periodontal disease. The prevalence and viral loads of EBV and CMV were quantitatively analyzed from diseased, healthy and DP sites from a cohort of patients who had chronic moderate to severe periodontitis over a six month period following OHI or SRP treatment, with or without the use of n-3 PUFA. Our findings demonstrate that 1) CMV was seldom detected in healthy, disease or DP sites, 2) EBV was more prevalent at disease and DP sites than healthy sites, albeit EBV prevalence was similar at disease sites that progressed as those that did not progress, and viral loads were low and not significantly different at disease, DP or healthy sites, and 3) specific Gram-negative anaerobic bacteria were significantly higher in diseased sites and DP sites than in healthy sites. Thus, these findings support rejecting the hypothesis that EBV and/or CMV are associated with the progression of periodontal disease, and lend support to the association of Gram-negative anaerobic bacteria in this process.

In this study, CMV was seldom detected in subgingival plaque regardless of the health of the periodontal site. Only four of 1454 samples sites (0.3%) were CMV positive, i.e., three from disease sites and one from a healthy site. When CMV was present, the copy numbers were low. The low prevalence and low copy number of CMV at periodontal sites in our patient population is consistent with the very low prevalence of CMV in chronic periodontitis (0–0.3%) reported by others (Dawson et al., 2009, Rotola et al., 2008, Nibali et al., 2009) and questions the role of CMV in periodontal disease, despite the findings from several cross-sectional studies indicating otherwise (Contreras & Slots, 2001, Slots et al. 2003, Yapar et al., 2003, Saygun et al., 2004b, Saygun et al., 2005, Imbronito et al., 2008, Grenier et al., 2009, Chalabi et al., 2010). It should be noted that a reason for the inconsistent prevalence of HHVs in periodontal sites described in the literature likely relates to the variability in the methodologies employed. Studies with higher CMV prevalence are associated with nested PCR and pooled samples. This method is known to encounter spurious amplification (Don et al., 1991) and can overestimate the prevalence at individual sites (Botero et al., 2008). For these reasons, we obtained site-specific samples and used quantitative PCR to focus on whether the prevalence and/or amount of virus increased at progressing sites, which if observed would suggest replication of the virus.

Similar to other studies we found that EBV was more prevalent at disease sites than healthy sites at all-time points (Contreras & Slots, 2001, Yapar et al., 2003, Saygun et al., 2004b, Saygun et al., 2005, Imbronito et al., 2008, Dawson et al., 2009, Grenier et al., 2009, Nibali et al., 2009, Chalabi et al., 2010, Stein et al., 2013, Kato et al., 2013). However, an increased prevalence of EBV at disease sites does establish an etiological role for EBV in periodontal disease. If EBV is etiological, then one would expect the prevalence and copy number of EBV to be increased at DP sites. However, the prevalence of EBV at disease and DP sites, and the viral loads at disease, DP and healthy sites were not significantly different in this study population. When detected, EBV DNA copy numbers were low (i.e., median <10 copy/site). Low copy numbers of EBV in periodontal tissues, saliva and plaque of periodontitis patients is a consistent finding in the literature (Saygun et al., 2004b, Kubar et al., 2005, Miller et al., 2005, Rotola et al., 2008, Dettori, 2011). Together the findings question the importance of EBV in periodontitis and its progression. Also, our findings that there were increased amount of total bacteria and increased percentage of red complex bacteria (Pg, Td, or Tf) in disease and DP sites unrelated to EBV contrast with the suggestion that EBV is a cofactor in periodontal disease (Contreras et al., 1999a, Contreras et al., 1999b, Contreras & Slots, 2001, Slots, 2007, Kato et al., 2015). We did not detect EBV co-existence with red complex bacteria (Pg, or Tf), although EBV was in a higher percentage at sites containing Td. Thus, our findings do not appear to support an important role of EBV in periodontal disease progression with respect to viral load and prevalence, although other mechanisms potentially involving EBV-related immune alterations specific to DP and Td were not explored.

Although two HHVs did not demonstrate a role in progression of periodontitis in this study, several important findings were observed. First, 28% of patients demonstrated at least one DP site within six months of follow-up, and the rate of progression was ~1% of sites. Progression was associated with older age, more severe disease at baseline, and clustered around select patients. Interestingly, there was no arch predilection, nor an association with the type of treatment received. Thus, these findings suggest that practitioners can potentially expect that at least 25% of their patients over the age of 40 years who have periodontitis to experience disease progression despite the delivery of the standard of care. This provides a perspective compared to the natural history of periodontal disease when treatment is not provided (Kocher et al., 2000, Nomura et al., 2017, Ramseier et al., 2017). Further, our findings suggest that the standard of care is effective for most sites (i.e., 99% of sites do not progress in six months), yet practitioners should be vigilant for sites that are likely to progress. Further, more needs to be done to identify the risk factors involved in disease progression, and how best to prevent disease progression. As practitioners target more precise therapies, it is important to realize that Pg, Td, and Tf are strongly associated with disease progression as documented by our findings and many others (Haffajee et al., 2008, Teles et al., 2008, Farias et al., 2012, Nomura et al., 2017); thus these periodontal pathogens should be targeted for elimination, if progression of periodontitis is to be halted.

We acknowledge our study findings may present limitations in interpretation. First, there was a 19.8% drop out rate. Although the dropout rate in our study population is at the threshold (20%) for acceptable dropout rates for clinical trials, we understand it could influence the validity of our results considering patients who dropout might have been EBV positive (Miller et al., 2005, Dettori, 2011). Second, there were more male than female participants, so our findings may not be representative of the population as a whole. Third, treatment was provided, i.e., OHI vs. SRP with and without PUFA to all study participants, which may have influenced disease progression and virus prevalence; however, the clinical characteristics and EBV distribution of each treatment group were similar and we did not detect a treatment effect for SRP vs. OHI or with the use or non-use of PUFA. Lastly, it has been reported in the literature that socioeconomic status, geographic and ethnic differences can affect EBV and CMV prevalence (Hanookai et al., 2000, Botero et al., 2008, Rotola et al., 2008, Dawson et al., 2009, Nibali et al., 2009, Stein et al., 2013). Availability of this information may have affected the interpretation of the data.

In conclusion, this is the first study to quantify CMV, EBV and anaerobic periodontal pathogens at sites with progressing periodontal disease. Our results are consistent with Pg, Td and Tf having more important roles in disease progression than EBV and CMV. Thus, treatment efforts should continue to focus on eliminating the bacterial pathogens in disease susceptible individuals.

CLINICAL RELEVANCE:

Scientific Rational for Study:

We studied the presence and abundance of EBV, CMV and anaerobic bacteria in plaque from healthy, diseased and periodontal sites that exhibited disease progression (i.e., advancing periodontal pockets, clinical attachment levels and bleeding on probing) over 28 weeks.

Principal Findings:

Although EBV was detected more often in sites with deep periodontal pockets, EBV and CMV levels were not associated with sites of disease progression. In contrast, there was an association between Gram-negative anaerobic bacteria and disease progression.

Practical Implications:

Periodontal therapy should continue to target Gram-negative anaerobic bacteria to minimize disease progression.