Link to other data assessments
Prepared by: Mark O'Doherty, UKCRC Centre of Excellence for Public Health, Queens University Belfast, United Kingdom
Contents:
1. Introduction
Family history of premature CHD, stroke and diabetes may be a powerful predictor of incident CVD and diabetes events, including premature onset and the family history data collected in CHANCES will assist in refining the magnitude of this increased risk.
The cohorts were asked to report on their collection of data on family history of CHD, stroke and diabetes so that a comparison could be made of the types and quality of the procedures used and measurements employed by each cohort. An overview of the responses to the questionnaire, and also information we have collected from published manuscripts, personal communication and cohort websites can be found in attached Excel file .
Detailed information on family history is available from the majority of cohorts except the EPIC Elderly cohorts and RES (wiki cohort descriptions indicate that family history may be available for CHD).
Family history data definitions vary considerably between the cohorts. However, this report focuses on three specific items:
- Age: i.e. what is the age limit that defines a relative having CHD, stroke, diabetes. The definition of "premature" cardiovascular disease varies, typically premature CHD and stroke can be defined as having a cardiovascular event aged 65 or less for women and 55 or less for men, but this would not necessary exclude other cohorts that lack this information.
- Affected relative: did the baseline questionnaire ask for information on parents, siblings, offspring or just 1st degree relatives?
- Definition: what disease definition was used in the questionnaire; e.g. myocardial infarction, coronary heart disease, angina pectoris, thrombosis of heart?
2. Coronary heart disease (CHD)
Most cohorts have reported in their responses to the questionnaire that information on family history of CHD was collected at baseline (ESTHER, HAPIEE, MORGAM, NHS, Tromsø, and Zutphen). Exceptions to this would be NIH_AARP, SENECA, and SHARE, which do not have family history data at baseline. EPIC Elderly Denmark has indicated on their wiki cohort description that family history is not available; EPIC Elderly Bilthoven and RES have indicated on their wiki cohort descriptions that family history is available. Conversely, from its wiki cohort description, EPIC Elderly Greece has indicated that information on family history is available, but it is not ready to be used and it is therefore not inserted into their database.
Information on self-reported family history of CHD was generally recorded on the baseline questionnaires. Some cohorts have allowed for responses for each particular family member, e.g. mother, father, sibling (ESTHER, NHS, and Zutphen); whereas others did not allow for such distinctions, e.g. mother and father grouped together or all first degree relatives (HAPIEE, MORGAM (some individual cohorts may have more specific information on family history that is not harmonized to MORGAM), and Tromsø). Nevertheless, many of the cohorts did specify an age limit for premature onset (e.g. before the age of 60: HAPIEE and Zutphen); or have asked for the age at which the family member had a CHD event (ESTHER and NHS). MORGAM cohorts have used different age limits for premature CHD and it is possible to select out the groups of cohorts that comply with a given family history definition. Tromsø was the only cohort at baseline which did not set an age limit and did not request an age-of-onset within their baseline questionnaire (baseline: T2).
Recontact information on family history of CHD was reported by some of the cohorts, except for ESTHER, HAPIEE, MORGAM (some only) and Zutphen. NHS repeated similar questions to those used at baseline, for their recontacts in 1984, 1996, and 2008 (age brackets used at each recontact: <50, 50-59, 60-69, 70+) and collected family history information for brothers and sisters in 1996, and for brothers, sisters and offspring in 2008. Tromsø repeated a similar question to baseline at T3. T4 and T5 changed slightly with the inclusion of a question on family history of CHD before the age of 60.
3. Stroke
Five cohorts have reported that information on family history of stroke was collected at baseline (ESTHER, HAPIEE, MORGAM, Tromsø, and Zutphen). Exceptions to this include NIH_AARP, NHS, SENECA, and SHARE, which do not have family history data at baseline. RES has not provided a questionnaire response yet, nor have they indicated on their wiki cohort description whether family history is available. EPIC Elderly Bilthoven and EPIC Elderly Denmark have indicated on their wiki cohort description that family history is not available. Conversely, from its wiki cohort description, EPIC Elderly Greece has indicated that information on family history is available, but it is not ready to be used and has therefore not been included into their database.
As with CHD, information on family history of stroke was recorded on the baseline questionnaires, but none of the cohorts with family history have validated these events. Some cohorts have allowed for responses for each particular family member, e.g. mother, father, sibling (ESTHER, Tromsø, and Zutphen), whereas others did not allow for such distinctions, e.g. mother and father grouped together or all first degree relatives only (e.g. HAPIEE and MORGAM (some individual cohorts may have more specific information on family history that is not harmonized to MORGAM)). Many of these cohorts did specify an age limit for premature onset (before the age of 60: HAPIEE and Zutphen), or asked for the age at which the family member had had a stroke (ESTHER). MORGAM accepted that different individual cohorts had used different age limits for premature stroke in their own individual questionnaires, and this has been reflected in their own harmonized variables where the limit of 65 years for both men and women was used to define a premature event. Tromsø was the only cohort at baseline which did not set an age limit, and did not request an age of onset within their baseline questionnaire (baseline: T2).
Recontact information on family history of stroke was reported by some of the cohorts, except for ESTHER, HAPIEE, MORGAM (where individual cohorts may have follow-up for family history recorded), and Zutphen. Tromsø repeated a similar question to baseline at T3, T4 and T5. This changed slightly from T4 onwards to include mother, father, brothers, sisters and own children separately. NHS did not record information on family history of stoke at baseline (1976), but collected information from their recontacts in 1996 and 2008. Age brackets were used at each recontact (<50, 50-59, 60-69, 70+) and data were itemized separately for brothers and sisters in 1996, and for mothers, fathers and siblings in 2008.
4. Diabetes
Most cohorts have reported that information on family history of diabetes was collected at baseline (ESTHER, HAPIEE, Tromsø, and Zutphen). Exceptions to this would be MORGAM (only some individual cohorts may have information on family history of diabetes), NIH_AARP, NHS, SENECA, and SHARE, which do not have family history data at baseline. RES has not provided a questionnaire response yet, nor have they indicated on their wiki cohort description whether family history is available. EPIC Elderly Bilthoven has indicated on their wiki cohort description that information on family history is available, whilst EPIC Elderly Denmark has indicated on their wiki cohort description that family history is not available. Conversely, EPIC Elderly Greece has indicated that information on family history is available from their wiki cohort description, but is not ready to be used, and is therefore not included in their database.
As with CHD and stroke, information on family history of diabetes was asked in the baseline questionnaires, but none of the cohorts with family history data have validated these events. Some cohorts have allowed for responses for each particular family member, e.g. mother, father, sibling (ESTHER, Tromsø, and Zutphen), whereas others did not allow for such distinctions, e.g. parents and siblings grouped together (HAPIEE). HAPIEE and Zutphen specified an upper age limit for a diagnosis of "premature" diabetes, of 60 years old, whereas ESTHER asked for the age at which the family member was diagnosed. Tromsø was the only cohort at baseline which did not set an age limit, and did not request an age of diagnosis within their baseline questionnaire (T2).
Only the Tromsø cohort recorded recontact information on family history of diabetes, repeating a similar question to that used at baseline at T3, T4 and T5. This changed slightly from T4 onwards to include mother, father, brother, sister, and own children.
As with family history of stroke, NHS began collecting information on family history of diabetes in their recontacts in 1982, 1988, and 1992. At these recontacts, data was collected on whether a participant’s mother, father, brother, or sister had ever had diabetes, but no age limits or age at diagnosis was recorded. The NIH_AARP cohort also asked a question on family history of diabetes at their first recontact (1996-1997), and as with NHS, data was collected on whether a participant’s mother, father, brother, or sister had ever had diabetes, but no age limits or age at diagnosis was recorded.
5. Proposed harmonized variables
Based on a review of this information, a set of proposed harmonized variables have been suggested, and these have been defined in page Family history variables.
We have proposed variables that combine first degree relatives (parents, siblings, or offspring) together for CHD, stroke and diabetes. A combined variable for each disease was proposed because the family history for each affected relative is not available from all cohorts. However, a variable for each particular family member (mother, father, brother, sister, son, daughter) can be created similar to the combined variable if required in future analyses. We have also proposed a variable which can act as a filter variable to limit analysis to only those first degree relatives with premature onset of CHD or stroke. Different cohorts have used different age limits for premature CHD and stroke in their questionnaires, but 60 years old was selected as the cut-point for premature onset because those cohorts with age limits used 60 years, but others with the age of onset recorded separately or with age brackets used may be able to code this variable also.
Family history CHD-Stroke-Diab.xls (application/vnd.ms-excel)
Family history CHD-Stroke-Diab.xls (application/vnd.ms-excel)
Family history CHD-Stroke-Diab.xls (application/vnd.ms-excel)
