Periodontitis, assessed using periodontal treatment as a surrogate marker, has no association with a first myocardial infarction in a Swedish population

1 INTRODUCTION

Although cardiovascular disease (CVD) is among the leading causes of death, incidence has been decreasing in the Western world while accelerating in middle- to low-income countries.1 In 2015, an estimated 423 million people were suffering from and 18 million deaths were attributed to a CVD.2 In the last decade, inflammation has emerged as a central mechanism in the development and acceleration of atherosclerosis.3, 4

Periodontitis is a bacterially induced, chronic inflammatory disease, affecting both the connective tissue and bone supporting the teeth.5 In 2010, the prevalence of severe periodontitis was approximately 11% in the USA6 and 9% in Western Europe.7

Epidemiological studies have found periodontitis to be positively associated with CVD and independently associated with myocardial infarction (MI).8-14 Although no causal relationship has been established, chronic inflammation caused by periodontitis has been suggested to accelerate the progress of coronary atherosclerosis and raise the risk of plaque rupture, leading to acute coronary syndromes.8, 15

In some developed countries, including Sweden, periodontal and cardiovascular health have improved, 16 which may have weakened the association between these conditions. To further evaluate the association between periodontal and cardiovascular health in a large population with a high standard of dental care and CVD prevention, we used periodontal treatment in national registry records as a surrogate marker for periodontitis over a 3-year period to investigate whether a first myocardial infarction and periodontitis were associated.

2 MATERIALS AND METHODS 2.1 Study design and setting

The present case-control study was based on nationwide data retrieved from registries maintained by the Swedish National Board of Health and Welfare (NBHW) and Statistics Sweden. The study population has been described previously (Nordendahl et al. 2018).17 In brief, cases comprised 51,880 individuals with a first MI event (fatal or non-fatal, International Classification of Diseases [ICD]-10 I.21 codes) recorded in the Swedish National In-Patient Register (IPR) between January 1, 2011 and December 31, 2013. Controls comprised 246,978 individuals (five controls per case) randomly selected from the national population registry and matched for age, gender, and geographic area at the time of the MI diagnosis among the cases (index date). Exclusion criteria for both groups was a prior MI. The present study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines18 for reporting observational studies.

2.2 Measurement of exposure and outcome

The exposure was periodontitis, assessed as periodontal treatment, derived from the Swedish National Dental Health Registry (DHR). As the DHR records no information on periodontal disease severity, periodontal treatment (type and frequency) was used as a surrogate marker to reflect severity. We identified the types of periodontal treatment according to the treatment codes in the DHR, which contains information on all dental treatments for adults above age 19 years in Sweden since 2008. Thus, we could extract information on all dental treatments in the 3 years before the index date (the date of the first MI event) for each case and the five matched controls. Diagnosis codes accompanied all treatment codes in the DHR.

2.3 Type of periodontal treatment

The case-definition of periodontal treatment was classified in two groups: (i) supragingival curettage or (ii) scaling and root planing (SRP) and/or periodontal surgery (PS). Dental treatment was grouped into four categories based on the treatment and diagnosis codes in the DHR (see Supplement 1):

No record of dental treatment

No record of periodontal treatment

One or more records of supragingival curettage

One or more records of SRP and/or PS

The therapy that was most advanced determined the classification of an individual; for example, if the patient received both supragingival curettage and SRP and/or PS, treatment was classified as the most advanced.

2.4 Number of periodontal treatments

Cases and controls were grouped into one of the following categories, based on the annual frequency of SRP and/or PS during the 3 years preceding the index date.

One treatment with SRP without PS

Two treatments with SRP without PS

Three or more treatments with SRP and/or one or more PS

The outcome was the first MI (fatal or non-fatal) recorded in the IPR between January 1, 2011 and December 31, 2013.

2.5 Registry sources

Dental treatment data for the present study were retrieved from the DHR. Public and private dental caregivers are required to submit information on all dental treatment of adults above age 19 to the DHR; the DHR is also connected to the administrative system which the Swedish social insurance system [Försäkringskassan] uses to determine reimbursements to dental caregivers for treatment. About 99% of all Swedish dental caregivers submit information on dental treatments to the DHR database.19 The Dental and Pharmaceutical Benefits Agency in Sweden maintains the national code system for defining treatments in the DHR.

Diabetes was determined based on an IPR diagnosis (ICD-10 codes: E.10-E.14), or the administration of glucose lowering therapy recorded in the Swedish Prescribed Drug Register20 (SPDR; Anatomical Therapeutic Chemical [ATC] codes: A10A or A10B). Type 1 diabetes was defined as ICD code E10 in combination with ATC code A10AB for short-acting insulin 6 months before the index date. If the criteria for type 1 diabetes were not fulfilled, type 2 diabetes was assumed. Information from ATC codes in the SPDR has been validated.20

Statistics Sweden supplied residence information. Participants were categorized in one of three regions: (i) North (Värmland, Dalarna, Gävleborg, Västernorrland, Jämtland, Västerbotten, and Norrbotten counties), (ii) Middle (Stockholm, Uppsala, Södermanland, Östergötland, Örebro and Västmanland counties), or (iii) South (Jönköping, Kronoberg, Kalmar, Gotland, Blekinge, Skåne, Halland, and Västra Götaland counties).

The Longitudinal Integration Database for Health Insurance and Labour Market Studies supplied information on annual income and education. Four educational classifications were defined: (1) ≤ 9 years of education, (2) 10 to 12 years of education, (3) college/university or PhD degree, and (4) missing data. Annual income was classified into three groups: (1) ≤ SEK 99,000 (USD 11,682); (2) SEK 100 to 299,000 (USD 11,800 to 35,282), and (3) ≥ SEK 300,000 (USD 35,400).

The Cause of Death Registry supplied date of death.

2.6 Statistical analysis

Statistical comparisons included the t-test for continuous variables and chi-squared tests for categorical variables. Odds ratios (ORs) were estimated using conditional logistic regression modeling with 95% confidence intervals (CI).

The association between periodontitis, using periodontal treatment as a surrogate marker, and MI are presented in two different models: (1) crude and (2) adjusted—adjusted for the matched variables plus income, education, and diabetes.

SAS* system statistical software was used for all analyses. A two-sided P-value of < 5% indicated statistical significance.

2.7 Ethics approval

The Regional Ethics Committee in Stockholm approved the study (Daybook no. [Dnr]: 2015/279-31/1). The study was conducted according to the principles outlined in the Helsinki Declaration.

3 RESULTS

Tables 1 and 2 present clinical characteristics for cases and controls. Cases were 72.6 ± 13 years old and controls, 72.3 ± 13 years old; the proportion of males:females was (62/38) in the case group and (61/39) in the control group. Cases more often had diabetes, fewer teeth, lower education, and lower income compared to controls (Table 1).

TABLE 1. Characteristics of cases with a first myocardial infarction (n = 51,880) and their controls (n = 246,978) Variables Cases Controls P-value Gender Women 19,773 (38.0) 95,342 (39.0) 0.037 Men 32,107 (62.0) 151,636 (61.0) Mean age (y, mean [± SD]) 73 (±13) 72 (±13) < 0.001 Geographic areaa North 11,834 (22.8) 56,150 (22.7) Middle 16,652 (32.1) 79,538 (32.2) 0.874 South 23,394 (45.1) 111,290 (45.1) Education Primary school 22,130 (42.7) 95,432 (38.6) High school 20,377 (39.3) 94,333 (38.2) < 0.001 College/University/PhD 8542 (16.4) 53,725 (21.8) Missing data 831 (1.6) 3488 (1.4) Income (USD) ≤ 11,682 7567 (14.8) 33,328 (13.6) 11,800–35,282 38,232 (74.5) 178,173 (72.9) < 0.001 ≥ 35,400 5508 (10.7) 32,992 (13.5) No. of teeth (mean [± SD]) 20 (±9) 21 (±8) < 0.001 Comorbidities Diabetesb 9931 (19.1) 26,707 (10.8) < 0.001 Type 1c 1181 (2.3) 1924 (0.8) Type 2d 8753 (16.9) 24,899 (10.1) Diabetes treatmente Insulin (A10A) 5033 (9.7) 10,590 (4.3) Oral (A10B) 6110 (11.8) 18,821 (7.6) Previous CVD 25,374 (48.9) 57,100 (23.1) <0.001 Heart failure 8265 (15.9) 12,716 (5.2) <0.001 Atrial fibrillation 6609 (12.7) 23,465 (9.5) <0.001 Angina pectoris 14,102 (27.2) 15,272 (6.2) <0.001 Stroke 6286 (12.1) 21,754 (8.8) <0.001 CVD drug treatment 35,355 (68.2) 139,054 (56.3) <0.001 Antihypertensive 27,395 (52.8) 104,063 (42.1) <0.001 Statins 14,156 (27.3) 51,871 (21.0) <0.001 Low-dose aspirin 17,866 (34.4) 57,143 (23.1) <0.001 Beta blockers 19,552 (37.7) 65,785 (26.6) <0.001 Data are presented as numbers (%) unless otherwise noted; P-values compare cases and controls. TABLE 2. Total and annual numbers of periodontal treatments in the 3 years preceding the index event, a first myocardial infarction (MI); P-values compare cases (n = 51,880) with controls (n = 246,978) Periodontal treatment in the 3 years preceding a first MI Cases n (%) Controls n (%) P-value Total numbers No record of dental treatment 10,076 (19.4) 40,637 (16.5) No record of periodontal treatment 24,174 (46.6) 121,467 (49.2) < 0.001 One or more supra-gingival curettagea 7686 (14.8) 35,948 (14.6) One or more SRP and/or one or more PSb 9944 (19.2) 48,926 (19.8) Annual numbers One SRP 7294 (14.1) 36,833 (14.9) Two SRP 1656 (3.2) 7561 (3.0) < 0.001 Three or more SRP and/or one or more PSb 773 (1.5) 3469 (1.4)

Among those with no record of dental treatment, cases had a lower education (primary school; 56% versus 59%; P < 0.001) and a higher frequency of diabetes (24% vs 14%; P < 0.001) compared to controls.

SRP and/or PS was more common among controls than cases (19.8% vs 19.2%; Table 2). Supragingival curettage, however, was more common among cases than controls (14.8% versus 14.6%), as was an annual frequency of three or more SRP and/or one or more PS (1.5% vs 1.4%; Table 2).

A conditional logistic regression analysis found no association between SRP and/or PS and MI (OR = 1.02; 95% CI: 1.00–1.05; Table 3). The association between supragingival curettage and a first MI remained relatively unchanged after additional adjustments (OR = 1.06; 95% CI 1.03–1.09; Table 3).

TABLE 3. The association between periodontal treatment and a first myocardial infarction (MI), Odds Ratios (ORs), and 95% confidence intervals (CIs) Periodontal treatment in the 3 years preceding a first MI n Crude OR (95%CI) Adjusteda OR (95% CI) No record of periodontal treatment 24,174 Ref Ref No record of dental treatment 10,076 1.25 (1.21–1.28) 1.15 (1.12–1.18) One or more supra-gingival curettageb 7686 1.07 (1.04–1.10) 1.06 (1.03–1.09) One or more SRP and/or one or more PSc 9944 1.02 (0.99–1.05) 1.02 (1.00–1.05)

The crude association between a high yearly frequency of SRP (≥ 3 visits) and/or ≥ 1 PS per year and a first MI was higher (OR = 1.16; 95% CI 1.02–1.30) but somewhat attenuated after additional adjustments (OR = 1.14; 95% CI 1.00–1.29; Table 4).

TABLE 4. The association between number of visits for periodontal treatment and a first myocardial infarction (MI), Odds Ratios (ORs), and 95% confidence intervals (CIs) Periodontal treatment in the 3 years preceding a first MI n Crude OR (95% CI) Adjusteda OR (95% CI) One treatment with SRP 7294 Ref Ref Two treatments with SRP 1656 1.11 (1.02–1.22) 1.10 (1.00–1.20) Three treatments with SRP and/or one or more PSb 773 1.16 (1.02–1.30) 1.14 (1.00–1.29)

There was an increased risk for a first MI among individuals with no record of dental treatment compared with individuals who had received periodontal treatment; after additional adjustments, the risk decreased somewhat (crude OR = 1.25; 95% CI 1.21–1.28; adjusted OR = 1.15; 95% CI 1.12–1.18; Table 3).

4 DISCUSSION

The present study found no increased risk of a first-time MI among individuals with periodontitis, assessed as SRP and/or PS during the 3-year period prior to the MI. However, among patients who underwent three or more treatments of SRP and PS, prevalence of a first-time MI was slightly increased, although not significant. This finding indicates that an association between severe periodontitis, assessed as three or more SRP and/or PS, and increased incidence of MI in a contemporary Swedish population may exist, but at a marginal level.

The findings of the present study are in contrast to a large number of observational studies, conducted in Sweden and elsewhere, that have found an association between periodontitis and CVD.11-14, 21-24 Methodological factors, diverse study populations, and varying definitions of periodontitis might be one explanation. The present study, however, puts forth the suggestion that a well-established and accessible healthcare system, including dental care, may reduce the association.

The most widely accepted mechanism of an association between periodontitis and CVD is a leakage of proinflammatory mediators from local inflammation in the periodontium into the systemic circulation, which would cause low-grade systemic inflammation and, thus, have a pro-atherosclerotic effect.15 Several studies have reported higher systemic inflammation in patients with periodontitis (see Lockhart et al. 2012 for review).8 Thus, it is possible that an association between periodontitis and CVD is less pronounced in a population where a majority have access to regular dental care, including dental hygiene and periodontal treatment, and who thus have less periodontal inflammation.

In the literature, periodontitis has been defined in several ways; for example, as tooth loss, through radiographic bone loss, and in clinical assessments.25-29 The present study used information and frequency of periodontal treatment submitted to national registries as a surrogate marker for periodontitis. SRP and PS are performed in order to reduce the inflammation of deepened pockets. SRP is recommended in individuals with pathological pocket depths of ≥ 5 mm and PS, when pocket depths of ≥ 6 mm persist.30

Periodontal treatment, which decreases oral inflammation and systemic inflammation, improves endothelial function and could lower the risk of MI.31 Therefore, the use of periodontal treatment procedures as a surrogate marker to reflect the severity of periodontitis might explain the lack of association in our study. Previous reports have suggested that preventive treatment for periodontal disease, such as SRP, can reduce the development of CVD.10, 31-33 Holmlund et al. (2017) demonstrated that individuals who did not respond to periodontal treatment had a higher risk of developing CVD than when such treatment was effective, indicating that successful periodontal treatment may affect the progression of subclinical CVD.10 Peng et al. (2017) suggested that advanced periodontal therapy might lower the rate of CVD, especially of MI and heart failure; however, this study was confined to diabetic patients.33 Thus, whether preventive measures to reduce periodontal disease will decrease the development or recurrence of CVD is still unclear.34

In the present study, cases with a high annual frequency of SRP in combination with PS presented a slightly increased risk of MI (OR 1.14). Even though this result was non-significant, it indicates that an association could be present in individuals with a severe level of periodontitis. One might speculate that these individuals do not respond to periodontal treatment and therefore still have a high degree of periodontal inflammation.

The present nationwide study was conducted in Sweden, which has a well-established social healthcare system and a high standard of dental care. The dental status of Swedish patients who visit a dentist regularly is good: edentulousness is low,16 and the incidence of periodontitis has decreased in recent decades.16 Almost 99% of all dental treatments performed in Sweden are registered in the DHR, which provides comprehensive, nationwide dental treatment information.19 Information on all hospital discharges in Sweden are registered in the IPR.35 These nationwide databases allow investigations of the entire population regarding dental care and CVD.

In medical research, observational studies are important as experimental studies cannot always be performed. Because periodontitis develops over several decades, a longitudinal cohort study was judged to be less suitable since the only available data on dental health was during the years 2010-2013. On a national basis regarding information on the level of dental health, registry data comprise the best available information for reflecting periodontal health in Sweden. It must be stated that, at this time, although it is only a surrogate marker for periodontitis, periodontal treatment is the best existing indicator of periodontal health on a national level.

In the present study, cases that had received no dental treatments during the years 2010-2013 had the highest risk of experiencing a first MI. The reason for this is unknown; however, in this group, well-known risks factors for periodontitis and MI, such as lower levels of education, lower family finances, and higher prevalence of diabetes, were present.36, 37

The results of the present study suggest that preventive dental care and healthcare that is accessible and of a high standard may weaken the association between periodontitis and MI.

In summary, the present study indicates that the association between periodontitis, assessed using periodontal treatment as a surrogate marker, and the incidence of a first MI seems to be less strong in Sweden than previously reported.

4.1 Strengths and limitations

The strength of the present study is that it was based on data from Swedish national healthcare registries which contain all patients diagnosed with a first MI event (fatal or nonfatal) during 2011 to 2013. The study also includes approximately 99% of all dental treatments performed in Sweden during the 3 years preceding the index date. This means that the study population is quite large, and since it includes all patients who had a first MI during the index period, the risk of selection bias is non-existent.

One obvious and important limitation of the study is whether the exposure factor, periodontal treatment, assessed by coded information in the registry correlates with presence of the actual periodontal disease; unfortunately, this cannot be determined. However, as prospective studies on the relation between MI and periodontitis have ethical consideration, are the utilized treatment codes for periodontal treatment, on a national basis, the best available registry information to reflect periodontal health in Sweden. Further, the dental health register is utilized by all active dental personnel throughout Sweden. Thus, the unique possibility of being able to describe the national situation has the attendant disadvantages of limited detailed information and use of surrogate markers. At this time, no systematic validation has been done to determine how well the reported treatment codes correspond to actual treatment. To manage these limitations, strict criteria for periodontal treatment were used as a surrogate marker of periodontitis. There is always a risk of introducing differential or non-differential misclassification bias. It could be claimed that using periodontal treatment as a proxy for the exposure periodontitis is less robust. However, it is unlikely that an individual receiving three or more advanced periodontal treatments (SRP and PS) annually does not have periodontitis. In this setting, if it occurred, it would appear as a non-differential misclassification bias, a bias toward null. In other words, misclassification would be equal in both groups, leading to a faded result.

MI is an acute manifestation of atherosclerotic vascular disease, which affects the whole body. Controls without MI may have had cerebrovascular accidents, stable coronary artery disease, or peripheral artery disease, among others and the same applies to the cases, which must be considered limitations of the study.

Another limitation is that the available registries don't include information regarding smoking and other MI risk factors such as hypertension and dyslipidemia. However, considering the fact that smoking is an important risk factor for both periodontitis and coronary heart disease, it is unlikely that the lack of information about this confounder could lead to an attenuation of the association.

Other limitations in non-randomized studies such as the present study are that relevant unmeasured characteristics sometimes differ between cases and controls; however, the matching process we used limited this by controlling for age, sex, and geographic area of residence, and by adjusting for education and income as a proxy for socioeconomic status. The matching was done by birth year—not by months—which explain the slight difference in mean age (72.6 versus 72.3 years) that was statistically significant, but likely not clinically relevant. Furthermore, matching was done before excluding previous MI, with a slight remaining difference in age and gender when only including first MI cases.

5 CONCLUSION

The present nationwide case-control study using information extracted from Swedish national registries found no association between periodontitis, assessed with periodontal treatment as the surrogate marker, and a first MI.

ACKNOWLEDGMENTS

We thank Tobias Svensson and Associate Professor Fredrik Granath at the Clinical Epidemiology Unit, Karolinska Institutet, Solna, Sweden, for valuable help with the statistical analyses.

CONFLICTS OF INTEREST

The authors report no conflicts of interest.

AUTHOR CONTRIBUTIONS

E. Nordendahl, contributed to conception, design, data acquisition, and interpretation; performed all statistical analyses; and drafted and critically revised the manuscript. M. Fored, B. Kjellström, A. Norhammar, and A. Gustafsson contributed to conception, design, data acquisition and interpretation; and drafted and critically revised the manuscript. A. Ekbom contributed to conception and design; and critically revised the manuscript. All authors gave their final approval and agree to be accountable for all aspects of the work.

留言 (0)

沒有登入
gif