4.2.1 The assessment of periodontal condition
Trained and calibrated dentists, with the assistance of a dental nurse, conducted almost similar clinical oral examinations during the Health 2000 survey, the Follow-up Study on Finnish Adults’ Oral Health, and the Health 2011 Survey. The clinical examination was performed using a fibre-optic light (Novar), a fibre-optic headlamp (Tekmala), a mouth mirror, and a WHO periodontal probe (Plandent, nro. 19577) with a probing force of 20 g (calibrated using a letter scale) in a portable dental treatment unit (Dentronic Mini-Dent®, Planmeca) with a portable chair. The total number of teeth was checked and recorded during the clinical oral health examinations. If the subjects had at least one natural tooth observed during the clinical oral examination, they were considered dentate.
The participants’ periodontal condition was assessed by measuring the periodontal pocket depth on four sites per tooth (distobuccal, midbuccal, midpalatal/lingual, mesiopalatal/lingual), except for third molars and tooth remnants. The deepest pocket depth for every tooth was recorded as follows: no pathologically deepened periodontal pocket; periodontal pocket with a depth of 4−5 mm; and periodontal pocket with a depth of ≥6 mm. In the analyses, the periodontal pocket depth of 4‒5 mm and ≥6 mm were combined in one category as periodontal pocket depth of ≥4 mm. In the Health 2000 Survey, the inter-examiner and intra-examiner reliability were assessed for pocket measurement, which showed a percentual agreement of 77% between examiners (k-value of 0.41), and repeatability showed a k-value of 0.83 (Vehkalahti, 2004).
For the cross-sectional studies (studies I−II), the number of teeth with deepened (≥4 mm) periodontal pockets (a count variable ranging from 0 to 28) was the outcome variable. For the longitudinal (incidence) studies (studies III−IV), participants were considered periodontally healthy if they had no teeth with deepened (≥4 mm) periodontal pockets (Dataset I); or alternatively, if they had ≤1 tooth with deepened (≥4 mm) periodontal pockets or two adjacent teeth with deepened (≥4 mm) periodontal pockets (Dataset II). At follow-up (2004 [study III] and 2011 [study IV]), outcome variables were defined as follows: a count variable—the number of teeth with deepened periodontal pockets at follow-up (Dataset I); and a dichotomous variable—the presence of deepened periodontal pockets in at least two non-adjacent teeth at follow-up (Dataset II).
4.2.2 Alcohol use measures
Information about participants’ alcohol use was collected through a self-report questionnaire during the Health 2000 Survey and the Health 2011 Survey. The questionnaire comprised items such as the amount of alcohol consumed, the frequency of alcohol intake, the maximum amount of alcohol consumed on one
occasion during the last 12 months, and the frequency and the amount of intake of different alcoholic beverages (for example, wine, spirits, and beer/cider/long drink).
The amount of alcohol use (g/week) in 2000 was calculated based on the participants’ responses about their last months’ average amount of the intake of different alcoholic beverages (wine, spirits, and beer/cider/long drink) per week.
These responses were pooled together after converting them to alcohol (g/week), assuming that wine contains 13%, spirits 44%, and beer 4.5% of alcohol. The alcohol use (g/week) variable was used as a continuous and a categorical variable in studies I, III and IV. In study I, the categories were as follows: 0; 0.1−23.8; 23.9−79.4; and 79.5−1303.5 g/week. Likewise, in study III, alcohol use was categorized into equal quartiles of alcohol consumed: 0; 0.1–30.8; 30.9–105.3; and ≥105.4 g/week. In study IV, alcohol use (g/week) was categorized as: 0; 0.1–18.9; 19.0–76.8; and >76.8 g/week. In addition to alcohol use (g/week) in 2000, alcohol use (g/year) in 2011 was used as a continuous variable in study IV.
Additionally, the average beverage-specific intake per week during the last month for each beverage was used as an explanatory variable in study II. The participants reported their intake, and this was converted into average g/week by the THL, according to the portion conversions used at the time of the survey. The average wine intake per week during the last month was recorded as: not at all; less than one glass (8‒12 cl); 1‒4 glasses; from 0.5‒3 bottles; 3‒5 bottles; and >5 whole bottles.
Answer options for spirit intake per week during the last month were as follows: not at all; < 1 glass (4 cl); 1‒6 glasses; half a bottle‒2 half-liter bottles; 2 half-liter bottles−4 half-liter bottles; and ≥4 half-liter bottles. The average weekly intake of beer/cider/long drinks during the last month was recorded in terms of the number of bottles: not at all; and the number of bottles. This variable was used as a continuous and a categorical variable. The average weekly intake of beer/cider/long drinks was categorized into three equal groups: not at all, ≤ 23.8 g/week; and >23.8 g/week.
Another question assessed participants’ alcohol use frequency during the last 12 months, with response options: never; seldom (ranging from once a year to a couple of times a month); and often (at least once a week) (studies I and III). In 2000 and 2011, more-specific questions about the frequency of use (alcohol in general, wine, spirits, and beer/cider/long drinks) were posed to the participants. The responses
for the beverage-specific frequency of use in 2000 and 2011 were: never; about once a month to once/twice a year; a couple of times a month to once a week; and 2–7 times a week.
The questionnaire also assessed alcohol use over the risk limit by asking the participants whether they had consumed the maximum amount of alcohol (over seven portions for men and over four portions for women) during any one drinking occasion within the last 12 months (no/yes). One portion equals one bottle (1/3 l) of beer, a glass (12 cl) of mild wine, a glass (8 cl) of strong wine, or a glass (4 cl) of spirits or other strong alcohol. Alcohol use over the risk limit variable was used as an explanatory variable in studies I and III.
Further information about alcohol use measures can be obtained from the Health 2000 (THL, Health 2000 Questionnaire 1, 2000) and the Health 2011 questionnaires (THL, Health 2011 Questionnaire 1, 2011).
Apart from self-reported alcohol use, a clinical measure of alcohol intake—in the form of serum GGT levels—served as an explanatory variable. The serum samples were collected, and the serum concentration of GGT in IU/L was analyzed using a kinetic method (IFCC/ECCLS, Konelab, Thermo Electron Oy, Finland). This measure was used in study II.
4.2.3 Covariates Socio-demographic factors
Population registers provided information about the age and gender of the participants. Age in 2000 was used as a continuous variable in studies I−IV, except in study I, where it was also used as a dichotomous variable (30−49 years and 50−65 years).
Socioeconomic position (SEP)
The level of education recorded during the home interview in the Health 2000 Survey indicated the SEP of the participants, and this was categorized into three categories (basic; intermediate; and higher). The basic level of education comprised those with no high-school education or no formal vocational qualifications. The intermediate level comprised those who had completed either high school or vocational training. The higher level of education comprised those who had a diploma or a degree from a higher institution (polytechnic, university). These three categories were used in studies I‒IV. For stratified analyses in studies I−II, the basic and intermediate education levels were combined.
Oral hygiene and oral health behaviors
The presence of dental plaque indicated the oral hygiene status of the participants during the Health 2000 Survey and the Health 2011 Survey. In 2000, plaque was assessed using a modified Silness and Löe method (Silness & Löe, 1964) from one surface of three index teeth: the buccal surface of the most posterior tooth on the right side of the maxilla, the lingual surface of the most posterior tooth on the left side of
the mandible, and the buccal surface of the lower left canine (lower right canine, if the lower left canine was absent). The presence of plaque was scored as no plaque;
visible plaque on gingival margins only; and visible plaque also elsewhere. The highest score from any index tooth described the oral hygiene status of the participant. These three categories were used in studies I‒IV. In 2011, the presence of plaque was assessed from the buccal surface of every tooth. The number of teeth with plaque indicated the oral hygiene status of the participants in 2011. The presence of plaque in 2011 was used as a continuous variable in study IV.
Information about dental attendance pattern and toothbrushing frequency was obtained during the health interview. The participants responded to a question about their dental attendance pattern in 2000 and 2011 “Do you usually go to a dentist?”
with the response options: regular check-ups; only when you have a toothache or some other trouble; and never. Response options two and three were combined, yielding two categories: regular check-ups versus irregular check-ups. The final two categories were used in studies I‒III. For study IV, the dental attendance pattern in 2000 and 2011 was combined and the final categories were: regular check-ups in 2000, regular ups in 2011, regular ups in 2000 and in 2011, and irregular check-ups in 2000 and in 2011.
The question about toothbrushing frequency in 2000 and 2011 was “How often do you usually brush your teeth?” and the response options were: more than twice a day; twice a day; once a day; less frequently than every day; and never. Response options one and two, as well as four and five, were combined, resulting in three categories: twice a day or more; once a day; and less frequently. The final three categories in the Health 2000 Survey were used in studies I‒IV. In study IV, toothbrushing frequency in 2011 was also used.
Body mass index
The body mass index (BMI) (kg/m2) of the participants was calculated based on the information about height and weight obtained primarily during the clinical health examination. If this was not available, the information from the questionnaire was used. BMI in 2000 was used as a continuous variable in studies I‒IV. The change in BMI from 2000 to 2011 was also used as a continuous variable in study IV.
Smoking
The smoking habits of the participants were asked during the home interview. In studies I−II, the smoking habit in 2000 was assessed by a question: “Do you currently smoke (cigarettes, cigars, or pipe)?” The options were daily smokers; occasional smokers; and non-smokers (those who had never smoked, those who had smoked less than 100 times in their whole life but did not smoke during the study, and those who had quit smoking at least one month ago). For studies I−II, daily smokers and occasional smokers were considered current smokers. In study III, the smoking habit in 2000 was used and the categories were: daily smokers; occasional smokers; quit 1–
12 months previously; quit over a year previously; and non-smokers. In study IV, the smoking habit in 2011 was used and the categories were: daily smokers; occasional smokers; quit 1–12 months previously; quit over a year previously; and non-smokers.
For stratified analyses, daily smokers; occasional smokers; those who quit 1–12 months previously; and quit over a year previously were combined into one category (studies III–IV).
C-reactive protein
The information about serum C-reactive protein (CRP) concentrations (mg/l) was obtained through laboratory measurements. The serum samples obtained from the survey were collected in small tubes using pipettes. The samples were first clotted by placing them on the table, then centrifuged and frozen at −20°C. The CRP levels were then quantified using an automated analyzer (Optima, Thermo Fisher Scientific Oy, Vantaa, Finland) and an ultrasensitive immunoturbidimetric test (Ultrasensitive CRP, Orion Diagnostica, Espoo, Finland). This variable was used as a continuous variable in study II.
Medications/drugs
The information about the use of lipid-lowering drugs and the use of NSAIDs was recorded during the interview. The use of lipid-lowering drugs (yes; no; or not known) in 2000 was used as a confounding variable in studies I‒III, and the use of lipid-lowering drugs (yes; no; or not known) in 2011 was used as a confounding variable in study IV. The use of NSAIDs (yes; no; or not known) in 2000 was used as a confounding variable in studies I−II.