MORGAM Manual

MORGAM derived variables Form: 31

© National Institute for Health and Welfare and the MORGAM Project investigators Last updated: 22 March 2016 For more information, please contact Kari Kuulasmaa (firstname.lastname@thl.fi)

The purpose of this form is to provide the definitions and format for the derived variables which are available at the MORGAM Date Centre (MDC) for MORGAM Manuscript Groups. Data in the format specified here can be provided for every member of each MORGAM cohort. The form will be appended (without changing the version number) whenever new derived variables become available.

This format should not be used for transferring the cohort data from the MORGAM Participating Centres to the MORGAM Data Centre, because all these variables can be derived from the data provided on the MORGAM Data Transfer Formats.

Contents:

• Format specification
  • Form identification, key and check items
  • Age
  • Blood pressure
  • Blood lipids
  • Body composition
  • Smoking
  • Schooling years
  • Risk scores
  • Disease history
  • CHD1: First CHD Event Type 1 during follow-up
  • CHD3: First CHD Event Type 3 during follow-up
  • CHD4: First CHD Event Type 4 during follow-up
  • CHD5: First CHD Event Type 5 during follow-up
  • CHD6: First CHD Event Type 6 during follow-up
  • STROKE1: First Stroke Event Type 1 during follow-up
  • STROKE3: First fatal or non-fatal likely cerebral infarction during follow-up
  • STROKE4: First fatal or non-fatal possible cerebral infarction during follow-up
  • STROKE5: First fatal or non-fatal likely haemorrhagic stroke during follow-up
  • STROKE6: First Stroke Event Type 6 during follow-up
  • CVD1: First fatal or non-fatal CHD Event Type 1 or Stroke Event Type 1 during follow-up
  • CVD3: First fatal or non-fatal CHD Event Type 1 or likely cerebral infarction during follow-up
  • CVD4: First fatal or non-fatal CHD Event Type 3 or likely cerebral infarction during follow-up
  • CVD5: First CHD Event Type 1 or Stroke Event Type 1 or heart failure during follow-up
  • CVD6: First CHD Event Type 1 or Stroke Event Type 1 or heart failure or peripheral vascular disease during follow-up
  • CVD7: First fatal or non-fatal CHD Event Type 6 or Stroke Event Type 1 during follow-up
  • CVD8: First fatal or non-fatal CHD Event Type 4 or likely cerebral infarction during follow-up
  • CVD9: First fatal or non-fatal CHD Event Type 3 or Stroke Event Type 1 during follow-up
  • Death during follow-up
  • Diabetes during follow-up
  • HF1: Heart failure during follow-up
  • AF1: Atrial fibrillation during follow-up
  • PVD1: Peripheral vascular disease during follow-up
  • CANCER1: Cancer during follow-up
  • VTE1: Venous thromboembolism during follow-up
• Columns of the format specification
• Definitions of the variables

Columns of the format specification

ITEM NAME

name used for the item in the MDC.

SPECIFICATION AND CODES

specification and values of the variable. More details can be found in the section "Definitions of the variables" below, or by following the hyperlink in column CHARACTERS.

CHARACTERS

indicates the data format in which the variable is available for the Manuscript Groups. In the cases where the value of the variable is fixed, the value is also given in this column.

COMMENTS

date when the variable was added or the latest revision.

Format specification

Definitions of the variables

The definitions of each derived variable, with hyperlinks to relevant data items of the Data Transfer Formats, are given below.

Form Identification, key and check items:

Items FORM and VERSN

These items identify this data format.

Items CENTRE...SERIAL

These are key items used for merging the different records of the same individual. They should be the same for all records of the same individual.

See MORGAM Participating Centres and cohorts for details.

Item RUA: Reporting Unit Aggregate

See MORGAM Reporting Unit Aggregates for details.

Item ROUNDS: Repeat measurements

Note that item ROUNDS is a cohort and not individual level item, and the same code is used for all members of the cohort in the certain contact round, even if some of them did not respond or were not even contacted. ROUNDS is a key item, together with items CENTRE...SERIAL.

Item EVENT is used in some of the Data Transfer Formats to identify different end-point events on the same individual, and therefore it also is a key item. However, it is irrelevant for this data format.

Items MBIRTH and SEX

These are the same as items MBIRTH and SEX of the "Data transfer format: MONICA survey data", and they should be used for double checking the key items. If you combine data for this individual from from various sources, please check the consistency of items MBIRTH and SEX.

Age

These age group variables were derived from item EAGE of the "Data transfer format: MONICA survey data"

AGE1 is calculated as a difference between the date of birth (MBIRTH of the "Data transfer format: MONICA survey data") and the date of examination (DEXAM of the "Data transfer format: MONICA survey data"). The missing day of birth in MBIRTH was taken as 15. In cases where the month of the birth was missing, it was taken as 06.

Blood pressure

SYSTC1 is the first systolic blood pressure value, derived from items SYST1 and RZ1 of the "Data transfer format: MONICA survey data" as:

SYSTC2 is the second systolic blood pressure value, defined similarly as SYSTC1, but using items SYST2 and RZ2 of the "Data transfer format: MONICA survey data":

SYSTM is systolic blood pressure, derived from items SYSTC1 and SYSTC2 as:

Note that when only one measurement value is available, that is used for SYSTM.

DIASTC1 is the first diastolic blood pressure value, defined similarly as SYSTC1, but using items DIAST1 and RZ1 of the "Data transfer format: MONICA survey data" as:

DIASTC2 is the second diastolic blood pressure value, defined similarly as SYSTC1, but using items DIAST2 and RZ2 of the "Data transfer format: MONICA survey data" as:

DIASTM is diastolic blood pressure, defined similarly as SYSTM, but using items DIASTC1 and DIASTC2:

BP1 is derived from SYSTC1 and DIASTC1 as:

BP1 is derived from SYSTC2 and DIASTC2 as:

BPM is derived from SYSTM and DIASTM as:

HIGHBP1 is derived using self reported blood pressure drug use (DRUGS of form 20) and SYSTM and DIASTM as:

if (DRUGS = 1 or 3) or (140.0 < SYSTM < 999.9) or (90 < DIASTM < 999.9) then HIGHBP1 = 1, if (DRUGS = 2, 8 or 9) and (0 < SYSTM <= 140) and (0 < DIASTM <= 90) then HIGHBP1 = 2. In all the other cases, HIGHBP1 = 9.

DRUG_HYPERT is derived from items HIBP and DRUGS of the "Data transfer format: MONICA survey data" as:

Laboratory results

Total serum cholesterol to two decimal places in mmol/l, derived from items CHOL and CHOLDL of the "Data transfer format: MONICA survey data"as:

*The result is rounded to the nearest two decimals.

The MPCs provided the cholesterol data either in units mmol/l or mg/dl, depending on which units were used in the laboratory. Item CHOLA converts all measurements to units mmol/l.

Total serum cholesterol to one decimal place in mg/dl, derived from items CHOLDL and CHOL of the "Data transfer format: MONICA survey data"as:

*The result is rounded to the nearest one decimal.

The MPCs provided the cholesterol data either in units mmol/l or mg/dl, depending on which units were used in the laboratory. Item CHOLB converts all measurements to units mg/dl.

HDL serum cholesterol to three decimal places in mmol/l, derived from items HDL and HDLDL of the "Data transfer format: MONICA survey data" as:

*The result is rounded to the nearest three decimals.

HDL serum cholesterol to one decimal place in mg/dl, derived from items HDLDL and HDL of the "Data transfer format: MONICA survey data" as:

*The result is rounded to the nearest one decimal.

RCHOL to two decimal places is derived from items CHOLA and HDLA as:

*The result is rounded to the nearest two decimals.

NONHDL to two decimal places is derived from items CHOLA and HDLA as:

*The result is rounded to the nearest two decimals.

NONHDLB to one decimal place is derived from items CHOLB and HDLB as:

*The result is rounded to the nearest one decimal.

LDL to two decimal places is derived from items TRIGLA, CHOLA and HDLA as:

*The result is rounded to the nearest two decimals.

Note: LDL is calculated from the Friedewald formula which is considered to be invalid if

1. TRIGLA > 4.52 mmol/l or
2. there was less than 12h fasting prior to blood sampling,

and inaccurate if TRIGLA is between 2.5 and 4.52 mmol/l (see for example Sniderman et al. 2003). Due to the inaccuracy of the formula, negative values of LDL are possible; these have been left in place and should be handled before data analysis.

LDL to one decimal place in mg/dl is derived from items TRIGLB, CHOLB and HDLB as:

*The result is rounded to the nearest one decimal.

Note: LDL is calculated from the Friedewald formula which is considered to be invalid if

1. TRIGLB > 400 mg/dl or
2. there was less than 12h fasting prior to blood sampling,

and inaccurate if TRIGLB is between 220 and 400 mg/dl (see for example Sniderman et al. 2003). Due to the inaccuracy of the formula, negative values of LDLB are possible; these have been left in place and should be handled before data analysis.

Serum triglycerides to two decimal places in mmol/l, derived from items TRIGL and TRIGLDL of the "Data transfer format: Baseline data addition 2"as:

*The result is rounded to the nearest two decimals.

Serum triglycerides in mg/dl, derived from items TRIGLDL and TRIGL of the "Data transfer format: Baseline data addition 2"as:

Plasma fibrinogen to two decimal places in g/l, derived from items FIBR and FIBRDL of the "Data transfer format: Baseline data addition 2"as:

DRUG_HYPERL is derived from items HICH and CHRX of the "Data transfer format: MONICA survey data" as:

Body composition

BMI to two decimal places in kg/m², derived from items HEIGHT and WEIGHT of the "Data transfer format: MONICA survey data"as:

*The result is rounded to the nearest two decimals.

WHR to two decimal places, derived from items WAIST and HIP of the "Data transfer format: MONICA survey data" as:

*The result is rounded to the nearest two decimals.

Smoking

Daily cigarette smoking status of the person at examination, derived from item CIGS of the "Data transfer format: MONICA survey data" as:

This item is relevant for ex-daily cigarette-smokers and provides an estimate of the age when they stopped smoking daily. It is derived from items MBIRTH and STOP of the "Data transfer format: MONICA survey data" as follows:

YBIRTH is defined as the last four characters of MBIRTH, i.e. the calendar year of birth. STOPAGE is then defined as:

CIGYRS estimates the number of years of daily cigarette smoking prior to the examination. It does not take into account possible non-smoking periods between smoking periods or smoking after the examination. CIGYRS is derived from items CIGS, EVERCIG, EAGE and CIGAGE of the "Data transfer format: MONICA survey data" and the derived variable STOPAGE as follows:

Note: The data for CIGYRS is misleading for the cohorts where the data for items CIGAGE or STOPAGE are missing commonly. For such cohorts, CIGYRS has value "0" for the never-daily-smokers, but it is "99" for all current and/or past daily smokers (see distributions of items CIGAGE and STOPAGE).

Current smoking status of the person at examination, derived from items CIGS, CIGAR, PIPE, CIGARSM and PIPESM of the "Data transfer format: MONICA survey data"as:

Schooling years

SCHOOL3 is derived from SCHOOL taking into account the birth year of the subject (extracted from MBIRTH), SEX and CENTRE.

SCHOOL3 is coded 9 if schooling years are missing (SCHOOL=99).

The procedure for determining SCHOOL3 can be presented as follows:

1. The calculations are done separately for each CENTRE and SEX.
2. The subjects are sorted by the year of birth. Years with only few subjects are combined.
3. Schooling years are processed in moving windows of seven years. The moving window for year of birth BYEAR contains subjects who were born in year BYEAR-3, BYEAR-2, BYEAR-1, BYEAR, BYEAR+1, BYEAR+2 or BYEAR+3. Tertiles are calculated in each window.
4. Because the tertiles for certain year are integers, there four possible schemes to define the cut points: the lower tertile may belong to either to the lowest category or  to the middle category and the upper tertile may belong to either to the middle category or  to the highest category. For each scheme the percentages of subjects in each category are calculated and the schemes are scored according to these percentages.  The score is the sum of the following penalty terms:
   • Absolute differences between the percentages and the desired percentage 33.33 %.
   • Threshold penalty if the extreme categories have less than 15 % of subjects.
   • Linear penalty if the extreme categories have less than 25 % of subjects.
   • Absolute differences between the percentages and the percentages of the previous window.
   • Small threshold penalty if the resulting cut points are smaller the cut points of the previous window. (We know that the schooling years should an increasing trend)
5. The scheme with lowest score is chosen and the cut points are defined.
6. The cut points are used to determine the value of SCHOOL3 for each subject.

Risk scores

The Framingham 2008 risk score FRAMINGHAM08 is derived using variables AGE1, SEX, CHOLB, HDLB, DRUG_HYPERT, SYSTM, DSMOKER and BASEDIAB1 according to Appendix of article: D'Agostino RB Sr., Vasan RS, Pencina MJ, Wolf PA, Cobain M, Massaro JM, Kannel WB. General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation. 2008;117:743–753. [PubMed]

FRAMINGHAM08 is coded 9.999 if any of the above mentioned variables are missing.

Note: The Framingham 2008 risk score is defined for people aged 30-74 years. FRAMINGHAM08 is determined for all subjects of the cohorts despite of age or disease history at baseline. This should be taken into account in data analysis.

The Framingham 1998 risk score FRAMINGHAM98 is derived using variables AGE1, SEX, CHOLB, HDLB, SYSTM, DIASTM, DSMOKER and BASEDIAB1 according to Appendix of article: Wilson PW, D'Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circulation. 1998;97:1837-1847.  [PubMed]

Variables CHOLB and HDLB were categorized for the score calculation according to the limits in Table 6 in the article:

Hypertension categories were created using both SYSTM and DIASTM:

If SYSTM=999.9 or DIASTM=999.9 then the other one is used to  define the hypertension category except for the first category where both variables are needed. If SYSTM and DIASTM give different categories, the higher category is selected.

FRAMINGHAM98 is coded 9.999 if any of the variables CHOLB, HDLB, DSMOKER and BASEDIAB1 or hypertension category are missing.

Note: The Framingham 1998 risk score is originally defined for people aged 30-74 years. FRAMINGHAM98 is determined for all subjects of the cohorts despite of age or disease history at baseline. This should be taken into account in data analysis.

The risk score formulated by the SCORE project is derived using variables AGE1, SEX, CHOLA, SYSTM and DSMOKER according to Appendix A of article: Conroy RM, Pyorala K, Fitzgerald AP, Sans S, Menotti A, De Backer G, De Bacquer D, Ducimetiere P, Jousilahti P, Keil U, Njolstad I, Oganov RG, Thomsen T, Tunstall-Pedoe H, Tverdal A, Wedel H, Whincup P, Wilhelmsen L, Graham IM. Estimation of ten-year risk of fatal cardiovascular disease in Europe: the SCORE project. Eur Heart J. 2003;24:987-1003. [PubMed]

The coefficients for high risk areas were used for MPCs (item CENTRE) 35 (Krakow), 36 (Warsaw), 45 (Kaunas), 47 (Novosibirsk), 81 (Estonia), 83 (Novosibirsk-HAPIEE) and 84 (Czech Republic). For all other RUAs the low risk coefficients were used.

SCORE is coded 9.999 if any of the variables CHOLA, SYSTM and DSMOKER are missing. SCORE is also coded 9.999 if AGE1<20 because SCORE formula is unidentifiable for ages <20 years.

Note: Because of the inaccuracy of the SCORE formula, values higher than 1 are possible; these have been left in place and should be handled before data analysis. The risk score by SCORE project was originally defined for people aged 45-64 years. SCORE is determined to all subjects of the cohorts despite of age (>20 years) or disease history at baseline. This should be taken into account in data analysis.

The ACC/AHA risk score ACC_AHA is derived using variables AGE1, SEX, CHOLB, HDLB, DRUG_HYPERT, SYSTM, DSMOKER and BASEDIAB1 according to Appendix 7 of article: Goff DC Jr, Lloyd-Jones DM, Bennett G, Coady S, D'Agostino RB, Gibbons R, Greenland P, Lackland DT, Levy D, O'Donnell CJ, Robinson JG, Schwartz JS, Shero ST, Smith SC Jr, Sorlie P, Stone NJ, Wilson PW. 2013 ACC/AHA guideline on the assessment of cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation. 2014;129(25 Suppl 2):S49–73. [PubMed]

ACC_AHA is coded 9.999 if any of the above mentioned variables are missing.

Note: The ACC/AHA risk score is defined for people aged 40-79 years. ACC_AHA is determined for all subjects of the cohorts despite of age or disease history at baseline. This should be taken into account in data analysis.

Disease history

BASEMI1 is derived from items HISMI1 and HISMI2 of the "Data transfer format: additional baseline data" (Form 21) as:

BASEMI1 =9 for everybody in the Cohorts where specific data for HISMI1 and HISMI2 are not available (e.g. Cohort 3 of MPC 57 and all cohorts of MPC 36).

BASEMI2 is derived from items HISMI1, HISMI2 and HISUC of the "Data transfer format: additional baseline data" (Form 21) as:

When specific data for HISMI1 and HISMI2 are not available, BASEMI2 uses data on HISUC, which indicates Coronary Heart Disease but the data are unspecific to separate between angina pectoris, MI and possibly cardiac revascularization (e.g. Cohort 3 of MPC 57 and all cohorts of MPC 36).

BASESTR1 is derived from items HISSTR1 and HISSTR2 of the "Data transfer format: additional baseline data" (Form 21) as:

BASECVD1 is derived from items BASEMI1 and BASESTR1 as:

BASECVD1 = 9 for everybody who did not have a stroke in the Cohorts where specific data on documented or self-reported history of MI are not available (e.g. Cohort 3 of MPC 57 and all cohorts of MPC 36).

BASECVD2 is derived from items BASEMI2 and BASESTR1 as:

When specific data on documented or self-reported MI are not available, BASECVD2 uses data on self-reported data on MI, angina pectoris or possibly cardiac revascularization, where these three conditions cannot be separated from each other (e.g. Cohort 3 of MPC 57 and all cohorts of MPC 36).

This history item is defined for excluding prevalent cases for incident diabetes.

BASEDIAB1 includes both type 1 and type 2 diabetes and it is derived from items HISDIAB, HIST1DM1, HIST1DM2, HIST2DM1, HIST2DM2, DIAB1_STATUS and DIAB1_EXIT as:

• BASEDIAB1 = 1, if
  • HISDIAB, HIST1DM1, HIST1DM2, HIST2DM1, or HIST2DM2 = 1, or
  • DIAB1_STATUS = 1 when DIAB1_EXIT is before DEXAM or DIAB1_EXIT=99999999.
• BASEDIAB1 = 2, if
  • HISDIAB, HIST1DM1, HIST1DM2, HIST2DM1, and HIST2DM2 ≠ 1, and
  • some of the following = 2:
    • HISDIAB, HIST1DM1, HIST1DM2, HIST2DM1, or HIST2DM2;
    • DIAB1_STATUS when DIAB1_EXIT is before DEXAM.
• BASEDIAB1 = 9 otherwise.

This history item is defined for excluding prevalent cases for incident heart failure.

BASEHF1 is derived from items HISHF1, HISHF2, HF_STATUS and HF_EXIT as:

• BASEHF1 = 1, if
  • HISHF1 or HISHF2 = 1, or
  • HF_STATUS = 1 when HF_EXIT is before DEXAM or HF_EXIT=99999999.
• BASEHF1 = 2, if
  • HISHF1 and HISHF2 ≠ 1, and
  • some of the following = 2:
    • HISHF1, HISHF2, or
    • HF_STATUS when HF_EXIT is before DEXAM.
• BASEHF1 = 9 otherwise.

This history item is defined for excluding prevalent cases for incident atrial fibrillation.

BASEAF1 is derived from items HISAF1, HISAF2, AF_STATUS and AF_EXIT as:

• BASEAF1 = 1, if
  • HISAF1 or HISAF2 = 1, or
  • AF_STATUS = 1 when AF_EXIT is before DEXAM or AF_EXIT=99999999.
• BASEAF1 = 2, if
  • HISAF1 and HISAF2 ≠ 1, and
  • some of the following = 2:
    • HISAF1, HISAF2, or
    • AF_STATUS when AF_EXIT is before DEXAM.
• BASEAF1 = 9 otherwise.

This history item is defined for excluding prevalent cases for incident peripheral vascular disease.

BASEPVD1 is derived from items HISPVD1, HISPVD2, PVD_STATUS and PVD_EXIT as:

• BASEPVD1 = 1, if
  • HISPVD1 or HISPVD2 = 1, or
  • PVD_STATUS = 1 when PVD_EXIT is before DEXAM or PVD_EXIT=99999999.
• BASEPVD1 = 2, if
  • HISPVD1 and HISPVD2 ≠ 1, and
  • some of the following = 2:
    • HISPVD1, HISPVD2, or
    • PVD_STATUS when PVD_EXIT is before DEXAM.
• BASEPVD1 = 9 otherwise.

This history item is defined for excluding prevalent cases for incident venous thromboembolism.

BASEVTE1 is derived from items HISVTE1, HISVTE2, VTE_STATUS and VTE_EXIT as:

• BASEVTE1 = 1, if
  • HISVTE1 or HISVTE2 = 1, or
  • VTE_STATUS = 1 when VTE_EXIT is before DEXAM or VTE_EXIT=99999999.
• BASEVTE1 = 2, if
  • HISVTE1 and HISVTE2 ≠ 1, and
  • some of the following = 2:
    • HISVTE1, HISVTE2, or
    • VTE_STATUS when VTE_EXIT is before DEXAM.
• BASEVTE1 = 9 otherwise.

This history item is defined for excluding prevalent cases for incident cancer.

BASECAN1 is derived from items HISCAN1, HISCAN2, CANCER_STATUS and CANCER_EXIT as:

• BASECAN1 = 1, if
  • HISCAN1 or HISCAN2 = 1, or
  • CANCER_STATUS = 1 when CANCER_EXIT is before DEXAM or CANCER_EXIT=99999999.
• BASECAN1 = 2, if
  • HISCAN1 and HISCAN2 ≠ 1, and
  • some of the following = 2:
    • HISCAN1, HISCAN2, or
    • CANCER_STATUS when CANCER_EXIT is before DEXAM.
• BASECAN1 = 9 otherwise.

CHD 1: First fatal or non-fatal CHD Event Type 1 during follow-up

CHD Event Type 1 is a fatal or non-fatal coronary event (Form 22) with any of the following diagnostic categories or event type:

• EVTYPE = 3 and 
  • DGNCAT = 1 (definite AMI)
  • DGNCAT = 2 (possible AMI or coronary death)
  • DGNCAT = 3 (definite or possible, not specifiable)
  • DGNCAT = 4 (unstable angina pectoris)
  • DGNCAT = 5 (unclassifiable) and SURVIV = 2 only
• EVTYPE = 2 or 4 (cardiac revascularization)

For a given measurement round (ROUNDS), CHD1 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round).  The follow-up starts at DEXAM at that round  and ends at EXDATEC of Form 27 or EXDATE of Form 25, whichever is earlier. The definition ignores possible events that took place before DEXAM. Therefore, persons with history of acute coronary events should be excluded from the the analysis, using relevant disease history variables.

CHD1 is coded 1 if the person had a CHD Event Type 1 during the follow-up and the first such event was fatal (SURVIV = 2).

CHD1 is coded 2 if the person had a CHD Event Type 1 during the follow-up and the first such event was non-fatal (SURVIV = 1).

CHD1 is coded 3 if the person did not have a CHD Event Type 1 during the follow-up.

CHD1 is coded 8 (irrelevant) if there was no follow-up for non-fatal events for the person (EXDATEC = 77777777 or 88888888).

CHD1 is coded 9 when there are no follow-up data available.

CHDDATE1 is the date of exit from the follow-up for a CHD1 event for a given measurement round (ROUNDS). The follow-up starts at DEXAM in that round and CHDDATE1 is derived from EVDATE of Form 22, EXDATEC of Form 27 and EXDATE of Form 25.

If CHD1=1 or 2, then CHDDATE1 = EVDATE of the coronary event.

If CHD1= 3, then CHDDATE1 is the earlier one of EXDATEC and EXDATE.

Note: In case if the follow-up for non-fatal events was ended prior to the follow-up for fatal events (either due to upper age-limit or due to coverage of calendar period by the event register which was used for the follow-up), the follow-up of fatal CHD1 events  is also considered only up to that time.  An upper age limit for the follow-up of nonfatal events for LTU-KAUa, POL-WARa and RUS-NOVa was 65 years. For POL-WARa the follow-up of fatal events ended on 31.12.1994.

If CHD1 is 8, CHDDATE1 is coded as 88888888.

CHDDATE1 is coded as 99999999 if CHD1 is 9 or EVDATE  is not a valid date. 

CHDTIME1 gives the length of the follow-up for a CHD1 event in days for a given measurement round (ROUNDS). It is the difference between the CHDDATE1 and DEXAM (Form 20).

If CHD1 is 8, CHDTIME1 is coded as 88888.

CHDTIME1 is coded as 99999  if CHD1 is 9 or CHDDATE1 = 99999999.

CHDAGE1 gives the age at the time of exit from the follow-up of a CHD1 event for a given measurement round (ROUNDS). It is the sum of the age at the examination (AGE1) and CHDTIME1/365.25.

If CHD1 is 8, CHDAGE1 is coded as 888.88.

CHDAGE1 is coded as 999.99   if CHD1 is 9 or CHDTIME1 = 99999.

CHD 3: First fatal or non-fatal CHD Event Type 3 during follow-up

CHD Event Type 3 is a fatal or non-fatal coronary event (Form 22) with any of the following diagnostic categories or event type:

• EVTYPE = 3 or 4  (acute coronary event) and
  • DGNCAT = 1 (definite AMI)
  • DGNCAT = 2 (possible AMI or coronary death)
  • DGNCAT = 3 (definite or possible, not specifiable)
  • DGNCAT = 4 (unstable angina pectoris)
  • DGNCAT = 5 (unclassifiable) and SURVIV = 2 only

The difference between CHD3 and CHD1 is that CHD3 does not include  revascularization (EVTYPE =  2).

For a given measurement round (ROUNDS), CHD3 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round).  The follow-up starts at DEXAM at that round  and ends at EXDATEC of Form 27 or EXDATE of Form 25, whichever is earlier. The definition ignores possible events that took place before DEXAM. Therefore, persons with history of acute coronary events should be excluded from the the analysis, using relevant disease history variables.

CHD3 is coded 1 if the person had a CHD Event Type 3 during the follow-up and the first such event was fatal (SURVIV = 2).

CHD3 is coded 2 if the person had a CHD Event Type 3 during the follow-up and the first such event was non-fatal (SURVIV = 1).

CHD3 is coded 3 if the person did not have a CHD Event Type 3 during the follow-up.

CHD3 is coded 8 (irrelevant) if there was no follow-up for non-fatal events for the person (EXDATEC = 77777777 or 88888888).

CHD3 is coded 9 when there are no follow-up data available.

CHDDATE3 is the date of exit from the follow-up for a CHD3 event. It  is derived from EVDATE of Form 22, EXDATEC of Form 27 and EXDATE of Form 25.

If CHD3=1 or 2, then CHDDATE1 = EVDATE of the coronary event.

If CHD3= 3, then CHDDATE3 is the earlier one of EXDATEC and EXDATE.

Note: In case if the follow-up for non-fatal events was ended prior to the follow-up for fatal events (either due to upper age-limit or due to coverage of calendar period by the event register which was used for the follow-up), the follow-up of fatal CHD3 events  is also considered only up to that time.  An upper age limit for the follow-up of nonfatal events for LTU-KAUa, POL-WARa and RUS-NOVa was 65 years. For POL-WARa the follow-up of fatal events ended on 1994-12-31.

If CHD3 is 8, CHDDATE3 is coded as 88888888.

CHDDATE3 is coded as 99999999 if CHD3 is 9 or EVDATE  is not a valid date. 

CHDTIME3 gives the length of the follow-up for a CHD3 event in days. It is the difference between the CHDDATE3 and DEXAM (Form 20).

If CHD3 is 8, CHDTIME3 is coded as 88888.

CHDTIME3 is coded as 99999  if CHD3 is 9 or CHDDATE3 = 99999999.

CHDAGE3 gives the age at the time of exit from the follow-up of a CHD3 event. It is the sum of the age at the examination (AGE1) and CHDTIME3/365.25.

If CHD3 is 8, CHDAGE3 is coded as 888.88.

CHDAGE3 is coded as 999.99   if CHD3 is 9 or CHDTIME3 = 99999.

CHD 4: First fatal or non-fatal CHD Event Type 4 during follow-up

CHD Event Type 4 is a fatal or non-fatal coronary event (Form 22) with any of the following diagnostic categories or event type:

• EVTYPE = 3 or 4 (acute coronary event) and  
  • DGNCAT = 1 (definite AMI)
  • DGNCAT = 2 (possible AMI or coronary death)
  • DGNCAT = 3 (definite or possible, not specifiable)
  • DGNCAT = 4 (unstable angina pectoris)

The difference between CHD4 and CHD1 is that CHD4 does not include revascularization (EVTYPE =  2)  or unclassifiable death (EVTYPE = 3 and DGNCAT = 5) .

For a given measurement round (ROUNDS), CHD4 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round).  The follow-up starts at DEXAM at that round  and ends at EXDATEC of Form 27 or EXDATE of Form 25, whichever is earlier. The definition ignores possible events that took place before DEXAM. Therefore, persons with history of acute coronary events should be excluded from the the analysis, using relevant disease history variables.

CHD4 is coded 1 if the person had a CHD Event Type 4 during the follow-up and the first such event was fatal (SURVIV = 2).

CHD4 is coded 2 if the person had a CHD Event Type 4 during the follow-up and the first such event was non-fatal (SURVIV = 1).

CHD4 is coded 3 if the person did not have a CHD Event Type 4 during the follow-up.

CHD4 is coded 8 (irrelevant) if there was no follow-up for non-fatal events for the person (EXDATEC = 77777777 or 88888888).

CHD4 is coded 9 when there are no follow-up data available.

CHDDATE4 is the date of exit from the follow-up for a CHD4 event. It  is derived from EVDATE of Form 22, EXDATEC of Form 27 and EXDATE of Form 25.

If CHD4=1 or 2, then CHDDATE4 = EVDATE of the coronary event.

If CHD4= 3, then CHDDATE4 is the earlier one of EXDATEC and EXDATE.

Note: In case if the follow-up for non-fatal events was ended prior to the follow-up for fatal events (either due to upper age-limit or due to coverage of calendar period by the event register which was used for the follow-up), the follow-up of fatal CHD4 events  is also considered only up to that time.  An upper age limit for the follow-up of nonfatal events for LTU-KAUa, POL-WARa and RUS-NOVa was 65 years. For POL-WARa the follow-up of fatal events is ended on 31.12.1994.

If CHD4 is 8, CHDDATE4 is coded as 88888888.

CHDDATE4 is coded as 99999999 if CHD4 is 9 or EVDATE  is not a valid date. 

CHDTIME4 gives the length of the follow-up for a CHD4 event in days. It is the difference between the CHDDATE4 and DEXAM (Form 20).

If CHD4 is 8, CHDTIME4 is coded as 88888.

CHDTIME4 is coded as 99999  if CHD4 is 9 or CHDDATE4 = 99999999.

CHDAGE4 gives the age at the time of exit from the follow-up of a CHD4 event. It is the sum of the age at the examination (AGE1) and CHDTIME4/365.25.

If CHD4 is 8, CHDAGE4 is coded as 888.88.

CHDAGE4 is coded as 999.99   if CHD4 is 9 or CHDTIME4 = 99999.

CHD 5: First fatal or non-fatal CHD Event Type 5 during follow-up

CHD Event Type 5 is a fatal or non-fatal coronary event (Form 22) with  the following diagnostic categories or event type:

• EVTYPE = 3 or 4 (acute coronary event) and  
  • DGNCAT = 1 (definite AMI)

For a given measurement round (ROUNDS), CHD5 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round).  The follow-up starts at DEXAM at that round  and ends at EXDATEC of Form 27 or EXDATE of Form 25, whichever is earlier. The definition ignores possible events that took place before DEXAM. Therefore, persons with history of acute coronary events should be excluded from the the analysis, using relevant disease history variables.

CHD5 is coded 1 if the person had a CHD Event Type 5 during the follow-up and the first such event was fatal (SURVIV = 2).

CHD5 is coded 2 if the person had a CHD Event Type 5 during the follow-up and the first such event was non-fatal (SURVIV = 1).

CHD5 is coded 3 if the person did not have a CHD Event Type 5 during the follow-up.

CHD5 is coded 8 (irrelevant) if there was no follow-up for non-fatal events for the person (EXDATEC = 77777777 or 88888888).

CHD5 is coded 9 when there are no follow-up data available.

CHDDATE5 is the date of exit from the follow-up for a CHD5 event. It  is derived from EVDATE of Form 22, EXDATEC of Form 27 and EXDATE of Form 25.

If CHD5=1 or 2, then CHDDATE5 = EVDATE of the coronary event.

If CHD5= 3, then CHDDATE5 is the earlier one of EXDATEC and EXDATE.

Note: In case if the follow-up for non-fatal events was ended prior to the follow-up for fatal events (either due to upper age-limit or due to coverage of calendar period by the event register which was used for the follow-up), the follow-up of fatal CHD5 events  is also considered only up to that time.  An upper age limit for the follow-up of nonfatal events for LTU-KAUa, POL-WARa and RUS-NOVa was 65 years. For POL-WARa the follow-up of fatal events ended on 1994-12-31.

If CHD5 is 8, CHDDATE5 is coded as 88888888.

CHDDATE5 is coded as 99999999 if CHD5 is 9 or EVDATE  is not a valid date. 

CHDTIME5 gives the length of the follow-up for a CHD5 event in days. It is the difference between the CHDDATE5 and DEXAM (Form 20).

If CHD5 is 8, CHDTIME5 is coded as 88888.

CHDTIME5 is coded as 99999  if CHD5 is 9 or CHDDATE5 = 99999999.

CHDAGE5 gives the age at the time of exit from the follow-up of a CHD5 event. It is the sum of the age at the examination (AGE1) and CHDTIME5/365.25.

If CHD5 is 8, CHDAGE5 is coded as 888.88.

CHDAGE5 is coded as 999.99   if CHD5 is 9 or CHDTIME5 = 99999.

CHD 6: First fatal or non-fatal CHD Event Type 6 during follow-up

CHD Event Type 6 is a fatal or non-fatal coronary event (Form 22) with any of the following diagnostic categories or event type:

• EVTYPE = 3 or 4  (acute coronary event) and
  • DGNCAT = 1 (definite AMI)
  • DGNCAT = 2 (possible AMI or coronary death)
  • DGNCAT = 3 (definite or possible, not specifiable)
  • DGNCAT = 5 (unclassifiable) and SURVIV = 2 only

The difference between CHD6 and CHD3 is that CHD6 does not include unstable angina pectoris (DGNCAT = 4). CHD6 is relevant for cohorts in which unstable angina pectoris can be separated reasonably well from possible MI. Typically, this concerns cohorts of the troponin era where the diagnosis of non-fatal myocardial infarctions follows the Universal Definition (see e.g. Thygesen K1, Alpert JS, White HD; Joint ESC/ACCF/AHA/WHF Task Force for the Redefinition of Myocardial Infarction. Universal definition of myocardial infarction. Eur Heart J. 2007;28:2525-38).

For a given measurement round (ROUNDS), CHD6 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round).  The follow-up starts at DEXAM at that round  and ends at EXDATEC of Form 27 or EXDATE of Form 25, whichever is earlier. The definition ignores possible events that took place before DEXAM. Therefore, persons with history of acute coronary events should be excluded from the the analysis, using relevant disease history variables.

CHD6 is coded 1 if the person had a CHD Event Type 6 during the follow-up and the first such event was fatal (SURVIV = 2).

CHD6 is coded 2 if the person had a CHD Event Type 6 during the follow-up and the first such event was non-fatal (SURVIV = 1).

CHD6 is coded 3 if the person did not have a CHD Event Type 6 during the follow-up.

CHD6 is coded 8 (irrelevant) if there was no follow-up for non-fatal events for the person (EXDATEC = 77777777 or 88888888).

CHD6 is coded 9 when there are no follow-up data available.

CHDDATE6 is the date of exit from the follow-up for a CHD6 event. It  is derived from EVDATE of Form 22, EXDATEC of Form 27 and EXDATE of Form 25.

If CHD6=1 or 2, then CHDDATE6 = EVDATE of the coronary event.

If CHD6= 3, then CHDDATE6 is the earlier one of EXDATEC and EXDATE.

Note: In case if the follow-up for non-fatal events was ended prior to the follow-up for fatal events (either due to upper age-limit or due to coverage of calendar period by the event register which was used for the follow-up), the follow-up of fatal CHD6 events  is also considered only up to that time.  An upper age limit for the follow-up of nonfatal events for LTU-KAUa, POL-WARa and RUS-NOVa was 65 years. For POL-WARa the follow-up of fatal events ended on 1994-12-31.

If CHD6 is 8, CHDDATE6 is coded as 88888888.

CHDDATE6 is coded as 99999999 if CHD6 is 9 or EVDATE  is not a valid date. 

CHDTIME6 gives the length of the follow-up for a CHD6 event in days. It is the difference between the CHDDATE6 and DEXAM (Form 20).

If CHD6 is 8, CHDTIME6 is coded as 88888.

CHDTIME6 is coded as 99999  if CHD6 is 9 or CHDDATE6 = 99999999.

CHDAGE6 gives the age at the time of exit from the follow-up of a CHD6 event. It is the sum of the age at the examination (AGE1) and CHDTIME6/365.25.

If CHD6 is 8, CHDAGE6 is coded as 888.88.

CHDAGE6 is coded as 999.99   if CHD6 is 9 or CHDTIME6 = 99999.

STROKE 1: First fatal or non-fatal Stroke Event Type 1 during follow-up

Stroke Event Type 1 is a fatal or non-fatal stroke event (Form 23) with any of the following diagnostic categories:

• DGNCAT = 1 (definite)
• DGNCAT = 9 (unclassifiable)

For a given measurement round (ROUNDS), STROKE1 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round).  The follow-up starts at DEXAM at that round  and ends at EXDATES of Form 28 or EXDATE of Form 25, whichever is earlier. The definition ignores possible events that took place before DEXAM. Therefore, persons with history of stroke should be excluded from the the analysis, using relevant disease history variables.

STROKE1 is coded 1 if the person had a Stroke Event Type 1 during the follow-up and the first such event was fatal (SURV28 = 2).

STROKE1 is coded 2 if the person had a Stroke Event Type 1 during the follow-up and the first such event was non-fatal (SURV28 = 1).

STROKE1 is coded 3 if the person did not have a Stroke Event Type 1 during the follow-up.

STROKE1 is coded 8 (irrelevant) if there was no follow-up for non-fatal events for the person (EXDATES = 77777777 or 88888888).

STROKE1 is coded 9 when there are no follow-up data available.

STRDATE1 is the date of exit from the follow-up for a STROKE1 event. It is derived from EVDATE of Form 23, EXDATES of Form 28 and EXDATE of Form 25.

If STROKE1=1 or 2, then STRDATE1 = EVDATE of the stroke event.

If STROKE1= 3, then STRDATE1 is the earlier one of EXDATES and EXDATE.

Note: In case if the follow-up for nonfatal events was ended prior to the follow-up for fatal events (either due to upper age-limit or due to coverage of calender period by the event register which was used for the follow-up), fatal follow-up is also considered only up to that time.  The upper age limit for the follow-up of nonfatal events for LTU-KAUa, POL-WARa was 65 years. For  RUS-NOVa it was 75 years. For POL-WARa, follow-up of fatal events ended on 31.12.1994.

If STROKE1 is 8, STRDATE1 is coded as 88888888.

STRDATE1 is coded as 99999999 if STROKE1 is 9 or EVDATE is not a valid date .

STRTIME1 gives the length of the follow-up for a STROKE1 event in days. It is the difference between the STRDATE1 and DEXAM (Form 20).

If STROKE1 is 8, STRTIME1 is coded as 88888.

STRTIME1 is coded as 99999 if STROKE1 is 9 or STRDATE1=99999999.

STRAGE1 gives the age at the time of exit from the follow-up of a STROKE1 event. It is the sum of age at the examination (AGE1) and STRTIME1/365.25.

If STROKE1 is 8, STRAGE1 is coded as 888.88.

STRAGE1 is coded as 999.99 if  STROKE1 is 9 or STRTIME1 is 99999.

STROKE3: First fatal or non-fatal likely cerebral infarction during follow-up

Several stroke end-points are needed because the type of stroke event is sometimes unspecified and different analyses may require different ways to handle these stroke events. The definitions of STROKE3 and STROKE4 differ in strokes whose type is not specified (code 436 in ICD-8 and ICD-9 and code I64 in ICD-10): these strokes are excluded  in STROKE3 and included in STROKE4. Note STROKE4 allows also analysis where stroke events whose type is not specified by ICD codes (STROKE4 = 5 or STROKE4 = 6) are not considered as events of interest but as censorings.

Likely cerebral infarction is a fatal or non-fatal event based on the specific MORGAM diagnosis of cerebral infarction (item CI)  when that is available, and on the clinical or death certificate diagnosis when CI is undetermined. The definition is less reliable than an event definition based on validated diagnosis, but unlike the validated diagnosis, it is available for all cohorts.

More specifically, the diagnosis is based on the data of Data Transfer Format: Stroke Events (Form 23) and Data Transfer Format: Follow-up Data ( Form 25). Potential events are those whose diagnostic category is "definite stroke" or "unclassifiable" (i.e. DGNCAT of Form 23 = 1 or 9):

• If the diagnosis was cerebral infarction (i.e. CI = 1), then the event is a likely cerebral infarction. 
• If the diagnosis was not cerebral infarction (i.e. CI = 2), then the event is not a likely cerebral infarction. 
• If the data was insufficient for the diagnosis (i.e. CI = 9) but the clinical or death diagnoses are available, then the event is a likely cerebral infarction if:
  • any of the clinical diagnoses available indicates cerebral infarction (i.e. CLIND1 or CLIND2 or CLIND3 has value:
    • 432, 433 or 434 in ICD-8;
    • 433 or 434 in ICD-9; or
    • I63 in ICD-10; or
  • the event is fatal and the underlying, direct or antecedent cause of death indicates cerebral infarction (i.e. SURV28 = 2 and DEATHDU, DEATHDA, DEATHDB or DEATHDC has one of the above ICD-values).

For a given measurement round (ROUNDS), STROKE3 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round).  The follow-up starts at DEXAM at that round  and ends at EXDATES of Form 28 or EXDATE of Form 25, whichever is earlier. The definition ignores possible events that took place before DEXAM. Therefore, persons with history of stroke should be excluded from the the analysis, using relevant disease history variables.

STROKE3 is coded 1 if the person had a likely cerebral infarction during the follow-up and the first such event was fatal (SURV288 = 2).

STROKE3 is coded 2 if the person had a likely cerebral infarction during the follow-up and the first such event was non-fatal (SURV28 = 1).

STROKE3 is coded 3 if the person did not have a likely cerebral infarction during the follow-up.

STROKE3 is coded 8 (irrelevant) if there was no follow-up for non-fatal events for the person (EXDATES = 77777777 or 88888888).

STROKE3 is coded 9 when there are no follow-up data available.

STRDATE3 is the date of exit from the follow-up for a STROKE3 event. It is derived from EVDATE of Form 23, EXDATES of Form 28 and EXDATE of Form 25.

If STROKE3 = 1 or 2, then STRDATE3 = EVDATE of the stroke event..

If STROKE3 = 3, then STRDATE3 is the earlier one of EXDATES and EXDATE.

Note: In case if the follow-up for nonfatal events was ended prior to the follow-up for fatal events (either due to upper age-limit or due to coverage of calender period by the event register which was used for the follow-up), fatal follow-up is also considered only up to that time.  The upper age limit for the follow-up of nonfatal events for LTU-KAUa, POL-WARa was 65 years. For  RUS-NOVa it was 75 years. For POL-WARa, follow-up of fatal events ended on 31.12.1994.

If STROKE3 is 8, STRDATE3 is coded as 88888888.

STRDATE3 is coded as 99999999 if STROKE3 is 9 or EVDATE is not a valid date.

STRTIME3 gives the length of the follow-up for a STROKE3 event in days. It is the difference between the STRDATE3 and DEXAM (Form 20).

If STROKE3 is 8, STRTIME3 is coded as 88888.

STRTIME3 is coded as 99999 if STROKE3 is 9 or STRDATE3 = 99999999.

STRAGE3 gives the age at the time of exit from the follow-up of a STROKE3 event. It is the sum of age at the examination (AGE1) and STRTIME3/365.25.

If STROKE3 is 8, STRAGE3 is coded as 888.88.

STRAGE3 is coded as 999.99 if  STROKE3 is 9 or STRTIME3 is 99999.

STROKE4: First fatal or non-fatal possible cerebral infarction during follow-up

Several stroke end-points are needed because the type of stroke event is sometimes unspecified and different analyses may require different ways to handle these stroke events. The definitions of STROKE3 and STROKE4 differ in strokes whose type is not specified (code 436 in ICD-8 and ICD-9 and code I64 in ICD-10): these strokes are excluded  in STROKE3 and included in STROKE4. Note STROKE4 allows also analysis where stroke events whose type is not specified by ICD codes (STROKE4 = 5 or STROKE4 = 6) are not considered as events of interest but as censorings.

Possible cerebral infarction is a fatal or non-fatal event based on the specific MORGAM diagnosis of cerebral infarction (item CI)  when that is available, and on the clinical or death certificate diagnosis when CI is undetermined. The definition makes a difference between validated diagnosis, diagnosis that is not validated but is based on ICD codes and diagnosis that is validated and is not specified by ICD codes.

More specifically, the diagnosis is based on the data of Data Transfer Format: Stroke Events (Form 23) and Data Transfer Format: Follow-up Data ( Form 25). Potential events are those whose diagnostic category is "definite stroke" or "unclassifiable" (i.e. DGNCAT of Form 23 = 1 or 9):

1. If the diagnosis was cerebral infarction (i.e. CI = 1), then the event is a possible cerebral infarction . 
2. If the diagnosis was not cerebral infarction (i.e. CI = 2), then the event is not a likely cerebral infarction. 
3. If the data was insufficient for the diagnosis (i.e. CI = 9) but the clinical or death diagnoses are available, then the event is a possible cerebral infarction based on ICD codes if:
   • any of the clinical diagnoses available indicates cerebral infarction (i.e. CLIND1 or CLIND2 or CLIND3 has value:
     • 432, 433 or 434 in ICD-8;
     • 433 or 434 in ICD-9; or
     • I63 in ICD-10; or
   • the event is fatal and the underlying, direct or antecedent cause of death indicates cerebral infarction (i.e. SURV28 = 2 and DEATHDU, DEATHDA, DEATHDB or DEATHDC has one of the above ICD-values).
4. If the data was insufficient for the diagnosis (i.e. CI = 9) but the clinical or death diagnoses are available, then the event is a possible cerebral infarction not specified by ICD codes if:
   • any of the clinical diagnoses available indicates cerebral infarction (i.e. CLIND1 or CLIND2 or CLIND3 has value:
     • 436 in ICD-8;
     • 436 in ICD-9; or
     • I64 in ICD-10; or
   • the event is fatal and the underlying, direct or antecedent cause of death indicates cerebral infarction (i.e. SURV28 = 2 and DEATHDU, DEATHDA, DEATHDB or DEATHDC has one of the above ICD-values).

For a given measurement round (ROUNDS), STROKE4 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round).  The follow-up starts at DEXAM at that round  and ends at EXDATES of Form 28 or EXDATE of Form 25, whichever is earlier. The definition ignores possible events that took place before DEXAM. Therefore, persons with history of stroke should be excluded from the the analysis, using relevant disease history variables.

STROKE4 is coded 1 if the person had a possible cerebral infarction during the follow-up based on validated diagnosis and the first such event was fatal (CI = 1 and SURV288 = 2).

STROKE4 is coded 2 if the person had a possible cerebral infarction during the follow-up based on validated diagnosis and the first such event was non-fatal (CI = 1 and SURV28 = 1).

STROKE4 is coded 3 if the person had a possible cerebral infarction during the follow-up based on ICD codes and the first such event was fatal (CI = 9 and SURV288 = 2 and ICD-codes are as listed in the item 3. in the list above).

STROKE4 is coded 4 if the person had a possible cerebral infarction during the follow-up based on ICD codes and the first such event was non-fatal (CI = 9 and SURV28 = 1 and ICD-codes are as listed in the item 3. in the list above).

STROKE4 is coded 5 if the person had a possible cerebral infarction during the follow-up which was not specified by ICD codes and the first such event was fatal (CI = 9 and SURV288 = 2 and ICD-codes are as listed in the item 4. in the list above).

STROKE4 is coded 6 if the person had a possible cerebral infarction during the follow-up which was not specified by  ICD codes and the first such event was non-fatal (CI = 9 and SURV28 = 1 and ICD-codes are as listed in the item 4. in the list above).

STROKE4 is coded 7 if the person did not have a possible cerebral infarction during the follow-up.

STROKE4 is coded 8 (irrelevant) if there was no follow-up for non-fatal events for the person (EXDATES = 77777777 or 88888888).

STROKE4 is coded 9 when there are no follow-up data available.

STRDATE4 is the date of exit from the follow-up for a STROKE4 event. It is derived from EVDATE of Form 23, EXDATES of Form 28 and EXDATE of Form 25.

If STROKE4 = 1, 2, 3, 4, 5, or 6 then STRDATE4 = EVDATE of the stroke event..

If STROKE4 = 7, then STRDATE4 is the earlier one of EXDATES and EXDATE.

Note: In case if the follow-up for nonfatal events was ended prior to the follow-up for fatal events (either due to upper age-limit or due to coverage of calender period by the event register which was used for the follow-up), fatal follow-up is also considered only up to that time.  The upper age limit for the follow-up of nonfatal events for LTU-KAUa, POL-WARa was 65 years. For  RUS-NOVa it was 75 years. For POL-WARa, follow-up of fatal events ended on 31.12.1994.

If STROKE4 is 8, STRDATE4 is coded as 88888888.

STRDATE4 is coded as 99999999 if STROKE4 is 9 or EVDATE is not a valid date.

STRTIME4 gives the length of the follow-up for a STROKE4 event in days. It is the difference between the STRDATE4 and DEXAM (Form 20).

If STROKE4 is 8, STRTIME4 is coded as 88888.

STRTIME4 is coded as 99999 if STROKE4 is 9 or STRDATE4 = 99999999.

STRAGE4 gives the age at the time of exit from the follow-up of a STROKE4 event. It is the sum of age at the examination (AGE1) and STRTIME4/365.25.

If STROKE4 is 8, STRAGE4 is coded as 888.88.

STRAGE4 is coded as 999.99 if  STROKE4 is 9 or STRTIME3 is 99999.

STROKE5: First fatal or non-fatal likely haemorrhagic stroke during follow-up

The definitions of STROKE3 and STROKE4 cover cerebral infarctions, but another endpoint, STROKE5, is needed for haemorrhagic strokes. Strokes whose type is not specified (code 436 in ICD-8 and ICD-9 and code I64 in ICD-10) are not included in STROKE5.

More specifically, the diagnosis is based on the data of Data Transfer Format: Stroke Events (Form 23) and Data Transfer Format: Follow-up Data ( Form 25). Potential events are those whose diagnostic category is "definite stroke" or "unclassifiable" (i.e. DGNCAT of Form 23 = 1 or 9):

•  If the diagnosis was subarachnoid haemorrhage (i.e. SAH = 1) or intracerebral haemorrhage (ICH = 1), then the event is a likely haemorrhagic stroke.
•  If the diagnosis was not subarachnoid haemorrhage (i.e. SAH = 2) and not intracerebral haemorrhage (ICH = 2), then the event is not a likely haemorrhagic stroke.
•  Otherwise, if the data was insufficient for the diagnosis (i.e. SAH = 9 or ICH = 9 or both) but the clinical or death diagnoses are available, then the event is a likely haemorrhagic stroke if:
  • any of the clinical diagnoses available indicates haemorrhagic stroke (i.e. CLIND1 or CLIND2 or CLIND3 has value:
    • 430 or 431 in ICD-8;
    • 430 or 431 in ICD-9; or
    • I60 or I61 in ICD-10; or
  • the event is fatal and the underlying, direct or antecedent cause of death indicates haemorrhagic stroke (i.e. SURV28 = 2 and DEATHDU, DEATHDA, DEATHDB or DEATHDC has one of the above ICD-values).

STROKE5 is coded 1 if the person had a likely haemorrhagic stroke during the follow-up and the first such event was fatal (SURV28 = 2).

STROKE5 is coded 2 if the person had a likely haemorrhagic stroke during the follow-up and the first such event was non-fatal (SURV28 = 1).

STROKE5 is coded 3 if the person did not have a likely haemorrhagic stroke during the follow-up.

STROKE5 is coded 8 (irrelevant) if there was no follow-up for non-fatal events for the person (EXDATES = 77777777 or 88888888).

STROKE5 is coded 9 when there are no follow-up data available..

STRDATE5 is the date of exit from the follow-up for a STROKE5 event. It is derived from EVDATE of Form 23, EXDATES of Form 28 and EXDATE of Form 25.

If STROKE5 = 1 or 2, then STRDATE5 = EVDATE of the stroke event..

If STROKE5 = 3, then STRDATE5 is the earlier one of EXDATES and EXDATE.

Note: In case if the follow-up for nonfatal events was ended prior to the follow-up for fatal events (either due to upper age-limit or due to coverage of calender period by the event register which was used for the follow-up), fatal follow-up is also considered only up to that time.  The upper age limit for the follow-up of nonfatal events for LTU-KAUa, POL-WARa was 65 years. For  RUS-NOVa it was 75 years. For POL-WARa, follow-up of fatal events ended on 31.12.1994.

If STROKE5 is 8, STRDATE5 is coded as 88888888.

STRDATE5 is coded as 99999999 if STROKE5 is 9 or EVDATE is not a valid date.

STRTIME5 gives the length of the follow-up for a STROKE5 event in days. It is the difference between the STRDATE5 and DEXAM (Form 20).

If STROKE5 is 8, STRTIME5 is coded as 88888.

STRTIME5 is coded as 99999 if STROKE5 is 9 or STRDATE5 = 99999999.

STRAGE5 gives the age at the time of exit from the follow-up of a STROKE5 event. It is the sum of age at the examination (AGE1) and STRTIME3/365.25.

If STROKE5 is 8, STRAGE5 is coded as 888.88.

STRAGE5 is coded as 999.99 if  STROKE5 is 9 or STRTIME5 is 99999.

STROKE6: First fatal or non-fatal Stroke Event Type 6 during follow-up

Stroke Event Type 6 is a fatal or non-fatal stroke event (Form 23) with any of the following diagnostic categories:

• DGNCAT = 1 (definite)
• DGNCAT = 3 (self-reported)
• DGNCAT = 9 (unclassifiable)

The definitions of STROKE1 and STROKE6 differ in strokes which are self-reported (DGNCAT = 3): these strokes are excluded from STROKE1 and included in STROKE6.

For a given measurement round (ROUNDS), STROKE6 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round).  The follow-up starts at DEXAM at that round  and ends at EXDATES of Form 28 or EXDATE of Form 25, whichever is earlier. The definition ignores possible events that took place before DEXAM. Therefore, persons with history of stroke should be excluded from the the analysis, using relevant disease history variables.

STROKE6 is coded 1 if the person had a Stroke Event Type 6 during the follow-up and the first such event was fatal (SURV28 = 2).

STROKE6 is coded 2 if the person had a Stroke Event Type 6 during the follow-up and the first such event was non-fatal (SURV28 = 1).

STROKE6 is coded 3 if the person did not have a Stroke Event Type 6 during the follow-up.

STROKE6 is coded 8 (irrelevant) if there was no follow-up for non-fatal events for the person (EXDATES = 77777777 or 88888888).

STROKE6 is coded 9 when there are no follow-up data available.

STRDATE6 is the date of exit from the follow-up for a STROKE6 event. It is derived from EVDATE of Form 23, EXDATES of Form 28 and EXDATE of Form 25.

If STROKE6 = 1 or 2, then STRDATE6 = EVDATE of the stroke event.

If STROKE6 = 3, then STRDATE6 is the earlier one of EXDATES and EXDATE.

Note: In case if the follow-up for nonfatal events was ended prior to the follow-up for fatal events (either due to upper age-limit or due to coverage of calender period by the event register which was used for the follow-up), fatal follow-up is also considered only up to that time.  The upper age limit for the follow-up of nonfatal events for LTU-KAUa, POL-WARa was 65 years. For  RUS-NOVa it was 75 years. For POL-WARa, follow-up of fatal events ended on 31.12.1994.

If STROKE6 is 8, STRDATE6 is coded as 88888888.

STRDATE6 is coded as 99999999 if STROKE6 is 9 or EVDATE is not a valid date.

STRTIME6 gives the length of the follow-up for a STROKE6 event in days. It is the difference between the STRDATE6 and DEXAM (Form 20).

If STROKE6 is 8, STRTIME6 is coded as 88888.

STRTIME6 is coded as 99999 if STROKE6 is 9 or STRDATE6=99999999.

STRAGE6 gives the age at the time of exit from the follow-up of a STROKE6 event. It is the sum of age at the examination (AGE1) and STRTIME6/365.25.

If STROKE6 is 8, STRAGE6 is coded as 888.88.

STRAGE6 is coded as 999.99 if STROKE6 is 9 or STRTIME6 is 99999.

CVD1: First fatal or non-fatal CHD Event Type 1 or Stroke Event Type 1 during follow-up

CVD1 is derived from CHD1, STROKE1, CHDAGE1 and STRAGE1 as follows:
 

CVDDATE1 is the date of exit from the follow-up for a CVD1 event. It is derived from CHD1, STROKE1, CHDDATE1 and STRDATE1 as follows:

CVDTIME1 gives the length of the follow-up for a CVD1 event in days. It is derived from CHD1, STROKE1, CHDTIME1 and STRTIME1 as follows:

CVDAGE1 gives the age at the time of exit from the follow-up of a CVD1 event. It is derived from CHD1, STROKE1, CHDAGE1 and STRAGE1 as follows:

CVD3: First fatal or non-fatal CHD Event Type 1 or likely cerebral infarction during follow-up

CVD3 is derived from CHD1, STROKE3, CHDAGE1 and STRAGE3 as follows:
 

CVDDATE3 is the date of exit from the follow-up for a CVD3 event. It is derived from CHD1, STROKE3, CHDDATE1 and STRDATE3 as follows:

CVDTIME3 gives the length of the follow-up for a CVD3 event in days. It is derived from CHD1, STROKE3, CHDTIME1 and STRTIME3 as follows:

CVDAGE3 gives the age at the time of exit from the follow-up of a CVD3 event. It is derived from CHD1, STROKE3, CHDAGE1 and STRAGE3 as follows:

CVD4: First fatal or non-fatal CHD Event Type 3 or likely cerebral infarction during follow-up

CVD4 is derived from CHD3, STROKE3, CHDAGE3 and STRAGE3 as follows:
 

CVDDATE4 is the date of exit from the follow-up for a CVD4 event. It is derived from CHD3, STROKE3, CHDDATE3 and STRDATE3 as follows:

CVDTIME4 gives the length of the follow-up for a CVD4 event in days. It is derived from CHD3, STROKE3, CHDTIME3 and STRTIME3 as follows:

CVDAGE4 gives the age at the time of exit from the follow-up of a CVD4 event. It is derived from CHD3, STROKE3, CHDAGE3 and STRAGE3 as follows:

CVD5: First CHD Event Type 1 or Stroke Event Type 1 or heart failure during follow-up

CVD5 is derived from CVD1, HF1, CVDAGE1 and HFAGE1 as follows:
 

CVDDATE5 is the date of exit from the follow-up for a CVD5 event. It is derived from CVD1, HF1, CVDDATE1 and HFDATE1 as follows:

CVDTIME5 gives the length of the follow-up for a CVD5 event in days. It is derived from CVD1, HF1, CVDTIME1 and HFTIME1 as follows:

CVDAGE5 gives the age at the time of exit from the follow-up of a CVD5 event. It is derived from CVD1, HF1, CVDAGE1 and HFAGE1 as follows:

CVD6: First CHD Event Type 1 or Stroke Event Type 1 or heart failure or peripheral vascular disease during follow-up

CVD6 is derived from CVD5, PVD1, CVDAGE5 and PVDAGE1 as follows:
 

CVDDATE6 is the date of exit from the follow-up for a CVD6 event. It is derived from CVD5, PVD1, CVDDATE5 and PVDDATE1 as follows:

CVDTIME6 gives the length of the follow-up for a CVD6 event in days. It is derived from CVD5, PVD1, CVDTIME5 and PVDTIME1 as follows:

CVDAGE6 gives the age at the time of exit from the follow-up of a CVD6 event. It is derived from CVD5, PVD1, CVDAGE5 and PVDAGE1 as follows:

CVD7: First fatal or non-fatal CHD Event Type 6 or Stroke Event Type 1 during follow-up

CVD7 is derived from CHD6, STROKE1, CHDAGE6 and STRAGE1 as follows:
 

This cardiovascular end-point covers definite or possible myocardial infarction, stroke and unclassifiable death and is relevant for cohorts in which unstable angina pectoris can be separated reasonably well from possible MI. Typically, this concerns cohorts of the troponin era where the diagnosis of non-fatal myocardial infarctions follows the Universal Definition (see e.g. Thygesen K1, Alpert JS, White HD; Joint ESC/ACCF/AHA/WHF Task Force for the Redefinition of Myocardial Infarction. Universal definition of myocardial infarction. Eur Heart J. 2007;28:2525-38). For such cohorts this end-point is close to the one used for the 2013 ACC/AHA Guideline on the Assessment of Cardiovascular Risk (Goff Jr DC, Lloyd-Jones DM, Bennett G, O’Donnell CJ, Coady S, Robinson J, D’Agostino RB, Schwartz JS, Gibbons R, Shero ST, Greenland P, Smith SC, Lackland DT, Sorlie P, Levy D, Stone NJ, Wilson PWF, 2013 ACC/AHA Guideline on the Assessment of Cardiovascular Risk, Journal of the American College of Cardiology (2014), doi: 10.1016/j.jacc.2013.11.005.)

CVDDATE7 is the date of exit from the follow-up for a CVD7 event. It is derived from CHD6, STROKE1, CHDDATE6 and STRDATE1 as follows:

CVDTIME7 gives the length of the follow-up for a CVD7 event in days. It is derived from CHD6, STROKE1, CHDTIME6 and STRTIME1 as follows:

CVDAGE7 gives the age at the time of exit from the follow-up of a CVD7 event. It is derived from CHD6, STROKE1, CHDAGE6 and STRAGE1 as follows:

CVD8: First fatal or non-fatal CHD Event Type 4 or likely cerebral infarction during follow-up

CVD8 is derived from CHD4, STROKE3, CHDAGE4 and STRAGE3 as follows:
 

CVDDATE8 is the date of exit from the follow-up for a CVD8 event. It is derived from CHD4, STROKE3, CHDDATE4, STRDATE3 as follows:

CVDTIME8 gives the length of the follow-up for a CVD8 event in days. It is derived from CHD4, STROKE3, CHDTIME4 and STRTIME3 as follows:

CVDAGE8 gives the age at the time of exit from the follow-up of a CVD8 event. It is derived from CHD4, STROKE3, CHDAGE4 and STRAGE3 as follows:

CVD9: First fatal or non-fatal CHD Event Type 3 or Stroke Event Type 1 during follow-up

CVD9 is derived from CHD3, STROKE1, CHDAGE3 and STRAGE1 as follows:
 

CVDDATE9 is the date of exit from the follow-up for a CVD9 event. It is derived from CHD3, STROKE1, CHDDATE3, STRDATE1 as follows:

CVDTIME9 gives the length of the follow-up for a CVD9 event in days. It is derived from CHD3, STROKE1, CHDTIME3 and STRTIME1 as follows:

CVDAGE9 gives the age at the time of exit from the follow-up of a CVD9 event. It is derived from CHD3, STROKE1, CHDAGE3 and STRAGE1 as follows:

Death: Death during follow-up

For a given measurement round (ROUNDS), DEATH is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round). The follow-up starts at DEXAM at that round.

DEATH is coded 1 if the person died during the follow-up (EXREAS of Form 25 = 1).

DEATH is coded 2 if the person did not die during the follow-up (EXREAS of Form 25 = 2, 3, 4, 5).

DEATH is coded 9 when there are no follow-up data available due to any reason.

For the analysis of DEATH as an end-point, the date, follow-up time and age at the time of exit from the follow-up of DEATH are the same as MORTDATE, MORTTIME and MORTAGE.

For a given measurement round (ROUNDS), CHD2 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round). The follow-up starts at DEXAM at that round.

CHD2 is coded 2 if the person did not die due to coronary heart disease during the follow-up.

CHD2 is coded 9 when there are no follow-up data available due to any reason.

Note that CHD2 = 2 includes those who survived during the follow-up or who died due to cause other than coronary heart disease. For the analysis of CHD2 as an end-point, the date, follow-up time and age at the time of exit from the follow-up of CHD2 are the same as MORTDATE, MORTTIME and MORTAGE.

For a given measurement round (ROUNDS), STROKE2 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round). The follow-up starts at DEXAM at that round.

STROKE2 is coded 1 if the person died due to stroke during the follow-up (SURV28 of Form 23 = 2 and DGNCAT of Form 23 = 1 or 9).

STROKE2 is coded 2 if the person did not die due to stroke during the follow-up.

STROKE2 is coded 9 when there are no follow-up data available due to any reason.

Note that STROKE2 = 2 includes those who survived during the follow-up or who died due to cause other than stroke. For the analysis of STROKE2 as an end-point, the date, follow-up time and age at the time of exit from the follow-up of STROKE2 are the same as MORTDATE, MORTTIME and MORTAGE.

For a given measurement round (ROUNDS), CVD_SCORE is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round). The follow-up starts at DEXAM at that round.

CVD_SCORE is coded 1 if the person died due to cardiovascular causes using end-point definition of SCORE during the follow-up. The following conditions and ICD-codes as underlying cause of death (DEATHDU) are considered as cardiovascular death:

• SURVIV of Form 22 = 2 and DGNCAT of Form 22 = 1, 2, 3 or 5
• SURV28 of Form 23 = 2 and DGNCAT of Form 23 = 1 or 9
• 400-404, 427-429 or 440-443 in ICD-8,
• 401-405, 426-429 or 440-443 in ICD-9 and
• I10-I15, I44-I51, I70-I73 in ICD-10.

CVD_SCORE is coded 2 if the person did not die due to cardiovascular causes during the follow-up.

CVD_SCORE is coded 9 when there are no follow-up data available due to any reason.

Note that CVD_SCORE = 2 includes those who survived during the follow-up or who died due to cause other than cardiovascular. For the analysis of CVD_SCORE as an end-point, the date, follow-up time and age at the time of exit from the follow-up of CVD_SCORE are the same as MORTDATE, MORTTIME and MORTAGE.

Note: CVD_SCORE was motivated by the risk score of SCORE project endpoint.

For a given measurement round (ROUNDS), CANCER is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round). The follow-up starts at DEXAM at that round.

• 140-208 in ICD-8,
• 140-209 in ICD-9 and
• C00-C97 in ICD-10.

CANCER is coded 2 if the person did not die due to cancer during the follow-up.

CANCER is coded 9 when there are no follow-up data available due to any reason or person has died but DEATHDU = NNN.

Note that CANCER = 2 includes those who survived during the follow-up or who died due to cause other than cancer. For the analysis of CANCER as an end-point, the date, follow-up time and age at the time of exit from the follow-up of CANCER are the same as MORTDATE, MORTTIME and MORTAGE.

MORTDATE is the date of exit from the mortality follow-up and is equal to EXDATE of Form 25.

If DEATH=9, then MORTDATE = 99999999.

MORTTIME gives the length of the mortality follow-up in days. It is the difference between the MORTDATE and DEXAM (Form 20). When DEATH is 9, MORTTIME is coded as 99999.

MORTAGE gives the age at the time of exit from the mortality follow-up. It is the sum of the age at the examination (AGE1) and (MORTTIME/365.25). When DEATH is 9, MORTAGE is coded as 999.99.

DIAB1: First documented clinical diagnosis of type 2 diabetes during follow-up

Documented clinical diagnosis of type 2 diabetes is specified by item DIAB1_STATUS of Form 84.

For a given measurement round (ROUNDS), DIAB1 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round). The follow-up starts at DEXAM at that round and the definition ignores possible events that took place before DEXAM. Therefore, persons with history of diabetes should be excluded from the the analysis using relevant disease history variables.

DIAB1 is coded 1 if the person had a documented clinical diagnosis of diabetes during the follow-up.

DIAB1 is coded 2  if the person did not have a documented clinical diagnosis of diabetes during the follow-up.

DIAB1 is coded 8 (irrelevant) if  DIAB1_EXIT is before DEXAM or DIAB1_FU = 2.

DIAB1 is coded 9 when there are no follow-up data available.

DIABDATE1 is the date of exit from the follow-up for a DIAB1 for a given measurement round (ROUNDS). The follow-up starts at DEXAM  at that round and DIABDATE1 is derived from DIAB1_EXIT of Form 84.

If DIAB1=1 or 2, then DIABDATE1 = DIAB1_EXIT.

If DIAB1=8, DIABDATE1 is coded as 88888888.

DIABDATE1 is coded as 99999999 if DIAB1 is 9 (or DIAB1_EXIT is not a valid date).

DIABTIME1 gives the length of the follow-up for a DIAB1 in days for a given measurement round (ROUNDS). It is the difference between the DIABDATE1 and DEXAM (Form 20).

If DIAB1 is 8, DIABTIME1 is coded as 88888.

DIABTIME1 is coded as 99999 if DIABDATE1=99999999.

DIABAGE1 gives the age at the time of exit from the follow-up of a DIAB1 for a given measurement round (ROUNDS). It is the sum of age at the examination (AGE1) and DIABTIME1/365.25.

If DIAB1 is 8, DIABAGE1 is coded as 888.88.

DIABAGE1 is coded as 999.99 if DIABTIME1 is 99999.

DIAB2: First glucose or HbA1c measurement based diagnosis of type 2 diabetes during follow-up

Glucose or HbA1c measurement based diagnosis of type 2 diabetes is specified by item DIAB2_STATUS of Form 84.

For a given measurement round (ROUNDS), DIAB2 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round). The follow-up starts at DEXAM at that round and the definition ignores possible events that took place before DEXAM. Therefore, persons with history of diabetes should be excluded from the the analysis using relevant disease history variables.

DIAB2 is coded 1 if the person had a documented clinical diagnosis of diabetes during the follow-up.

DIAB2 is coded 2  if the person did not have a documented clinical diagnosis of diabetes during the follow-up.

DIAB2 is coded 8 (irrelevant) if  DIAB2_EXIT is before DEXAM or DIAB2_FU = 2.

DIAB2 is coded 9 when there are no follow-up data available.

DIABDATE2 is the date of exit from the follow-up for a DIAB2 for a given measurement round (ROUNDS). The follow-up starts at DEXAM  at that round and DIABDATE2 is derived from DIAB2_EXIT of Form 84.

If DIAB2=1 or 2, then DIABDATE2 = DIAB2_EXIT.

If DIAB2=8, DIABDATE2 is coded as 88888888.

DIABDATE2 is coded as 99999999 if DIAB2 is 9 (or DIAB2_EXIT is not a valid date).

DIABTIME2 gives the length of the follow-up for a DIAB2 in days for a given measurement round (ROUNDS). It is the difference between the DIABDATE2 and DEXAM (Form 20).

If DIAB2 is 8, DIABTIME2 is coded as 88888.

DIABTIME2 is coded as 99999 if DIABDATE2=99999999.

DIABAGE2 gives the age at the time of exit from the follow-up of a DIAB2 for a given measurement round (ROUNDS). It is the sum of age at the examination (AGE1) and DIABTIME2/365.25.

If DIAB2 is 8, DIABAGE2 is coded as 888.88.

DIABAGE2 is coded as 999.99 if DIABTIME2 is 99999.

DIAB3: First self-reported type 2 diabetes during follow-up

Self-reported type 2 diabetes is specified by item DIAB3_STATUS of Form 84.

For a given measurement round (ROUNDS), DIAB3 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round). The follow-up starts at DEXAM at that round and the definition ignores possible events that took place before DEXAM. Therefore, persons with history of diabetes should be excluded from the the analysis using relevant disease history variables.

DIAB3 is coded 1 if the person had a documented clinical diagnosis of diabetes during the follow-up.

DIAB3 is coded 2  if the person did not have a documented clinical diagnosis of diabetes during the follow-up.

DIAB3 is coded 8 (irrelevant) if  DIAB3_EXIT is before DEXAM or DIAB3_FU = 2.

DIAB3 is coded 9 when there are no follow-up data available.

DIABDATE3 is the date of exit from the follow-up for a DIAB3 for a given measurement round (ROUNDS). The follow-up starts at DEXAM  at that round and DIABDATE3 is derived from DIAB3_EXIT of Form 84.

If DIAB3=1 or 2, then DIABDATE3 = DIAB3_EXIT.

If DIAB3=8, DIABDATE3 is coded as 88888888.

DIABDATE3 is coded as 99999999 if DIAB3 is 9 (or DIAB3_EXIT is not a valid date).

DIABTIME3 gives the length of the follow-up for a DIAB3 in days for a given measurement round (ROUNDS). It is the difference between the DIABDATE3 and DEXAM (Form 20).

If DIAB3 is 8, DIABTIME3 is coded as 88888.

DIABTIME3 is coded as 99999 if DIABDATE3=99999999.

DIABAGE3 gives the age at the time of exit from the follow-up of a DIAB3 for a given measurement round (ROUNDS). It is the sum of age at the examination (AGE1) and DIABTIME3/365.25.

If DIAB3 is 8, DIABAGE3 is coded as 888.88.

DIABAGE3 is coded as 999.99 if DIABTIME3 is 99999.

DIAB4: First documented clinical diagnosis or self-reported type 2 diabetes during follow-up

Documented clinical diagnosis or self-reported type 2 diabetes is derived from items DIAB1 and DIAB3.

For a given measurement round (ROUNDS), DIAB4 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round). The follow-up starts at DEXAM at that round and the definition ignores possible events that took place before DEXAM. Therefore, persons with history of diabetes should be excluded from the the analysis using relevant disease history variables.

DIAB4 is derived from items DIAB1 and DIAB3 as:

DIABDATE4 is the date of exit from the follow-up for a DIAB4 for a given measurement round (ROUNDS). The follow-up starts at DEXAM  at that round and DIABDATE4 is derived from DIABDATE1 and DIABDATE3.

If DIAB4=1, then DIABDATE4 = DIABDATE1.

If DIAB4=2, then DIABDATE4 = DIABDATE3.

If DIAB4=3, then DIABDATE4 is the later of DIABDATE1 and DIABDATE3.

If DIAB4=8, then DIABDATE4 is coded as 88888888.

If DIAB4=9, then DIABDATE4 is coded as 99999999.

DIABTIME4 gives the length of the follow-up for a DIAB4 in days for a given measurement round (ROUNDS). It is the difference between the DIABDATE4 and DEXAM (Form 20).

If DIAB4 is 8, DIABTIME4 is coded as 88888.

DIABTIME4 is coded as 99999 if DIABDATE4=99999999.

DIABAGE4 gives the age at the time of exit from the follow-up of a DIAB4 for a given measurement round (ROUNDS). It is the sum of age at the examination (AGE1) and DIABTIME4/365.25.

If DIAB4 is 8, DIABAGE4 is coded as 888.88.

DIABAGE4 is coded as 999.99 if DIABTIME4=99999.

HF1: First heart failure during follow-up

Heart failure is specified by item HF_STATUS of Form 81.

For a given measurement round (ROUNDS), HF1 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round). The follow-up starts at DEXAM at that round and the definition ignores possible events that took place before DEXAM. Therefore, persons with history of heart failure should be excluded from the the analysis using relevant disease history variables.

HF1 is coded 1 if the person had a heart failure during the follow-up.

HF1 is coded 2  if the person did not have a  heart failure during the follow-up.

HF1 is coded 8 (irrelevant) if  HF_EXIT is before DEXAM.

HF1 is coded 9 when there are no follow-up data available.

HFDATE1 is the date of exit from the follow-up for a HF1 for a given measurement round (ROUNDS). The follow-up starts at DEXAM  at that round and HFDATE1 is derived from HF_EXIT of Form 81.

If HF1=1 or 2, then HFDATE1 = HF_EXIT.

If HF1=8, HFDATE1 is coded as 88888888.

HFDATE1 is coded as 99999999 if HF1 is 9 (or HF_EXIT is not a valid date).

HFTIME1 gives the length of the follow-up for a HF1 in days for a given measurement round (ROUNDS). It is the difference between the HFDATE1 and DEXAM (Form 20).

If HF1 is 8, HFTIME1 is coded as 88888.

HFTIME1 is coded as 99999 if HFDATE1=99999999.

HFAGE1 gives the age at the time of exit from the follow-up of a HF1 for a given measurement round (ROUNDS). It is the sum of age at the examination (AGE1) and HFTIME1/365.25.

If HF1 is 8, HFAGE1 is coded as 888.88.

HFAGE1 is coded as 999.99 if HFTIME1 is 99999.

AF1: First atrial fibrillation during follow-up

Atrial fibrillation is specified by item AF_STATUS of Form 82.

For a given measurement round (ROUNDS), AF1 is defined only for those who were examined in that round (i.e. DEXAM of Form 20 is available for the measurement round). The follow-up starts at DEXAM at that round and the definition ignores possible events that took place before DEXAM. Therefore, persons with history of atrial fibrillation should be excluded from the the analysis using relevant disease history variables.

AF1 is coded 1 if the person had a atrial fibrillation during the follow-up.

AF1 is coded 2  if the person did not have a atrial fibrillation during the follow-up.

AF1 is coded 8 (irrelevant) if  AF_EXIT is before DEXAM.

AF1 is coded 9 when there are no follow-up data available.

AFDATE1 is the date of exit from the follow-up for a AF1 for a given measurement round (ROUNDS). The follow-up starts at DEXAM  at that round and AFDATE1 is derived from AF_EXIT of Form 82.

If AF1=1, then AFDATE1 = AF_EXIT.

If AF1=8, AFDATE1 is coded as 88888888.

AFDATE1 is coded as 99999999 if AF1 is 9 (or AF_EXIT is not a valid date).