NHLBI Family Heart Study Subsamples

Last Updated August 2, 1999

Because coronary heart disease (CHD) is a multifactorial entity, the FHS Phase II data (consisting of random families, high risk families, and African-American families) have been further divided into post-hoc subsamples along the lines of the constituent domains and risk factors of CHD, for the purposes of more detailed analyses, additional biochemical assays and more extensive genotyping of the frozen, stored samples collected in Phase II.

Post-Hoc Scientific Subsamples of FHS:

1. RSAMP (Random Proband Subsample N=200 unrelated subjects)

2. ERSAMP (Extended Random Proband Sample N=454 unrelated subjects)

3. BLKSAMP (African American Sample N=263 subjs in F=102 families)

4. CHDSAMP (CHD Cases + Sib Controls N=290 sibs in S=121 sibships)

5. CHDPSAMP (CHD Sib-Pair Sibships N=341 sibs in S=82 sibships)

6. IMTSAMP (Initma Medial Thickness Sibships N=246 sibs in S=82 sibships)

7. IRSSAMP (Indiv Risk Score CHD Sibships N=1,930 sibs in S=551 sibships)

8. FCHLSAMP (Famil-Comb Hyperlipid Sibpairs N=170 sibs in S=71 sibships)

9. HYPSAMP (Hypertensive Sibpairs N=697 sibs in S=278 sibships)

[Note: It is impossible to accurately draw all of the complex intersections of these samples in two dimensions, so the above diagram is meant to be for illustrative purposes only, and is NOT drawn to scale.]

Each of these samples is described in more detail, below. Note that these 9 samples are defined independently of one another, and so they can/may overlap (in some cases, considerably). For the purposes of budgeting and management of resources, these various subsamples of FHS have historically been grouped together into various "studies" as a way of prioritizing which tests, assays, and genotyping should be done on which subsets of subjects.

Focused Study (N=681 subjects)

The Focused Study consists of three of the above overlapping subsamples:

1. RSAMP (Random Proband Subsample N=200 unrelated subjects)

4. CHDSAMP (CHD Cases + Sib Controls N=290 sibs in S=121 sibships)

6. IMTSAMP (Initma Medial Thickness Sibships N=246 sibs in S=82 sibships)

This subset of subjects was chosen early in the early formation of the FHS, while Phase II data was still being collected. It was intended to provide a source of association analyses of novel risk factor assays that were too expensive to do on the entire Phase II cohort. Because the early budget was small, the Focused Study was modest in size and could not really accommodate the N needed for definitive linkage analyses (which tend to be of lower power than association analyses). Thus, the Focused Study was designed with association as the primarily method of analysis with linkage as secondary method. The Focused Study contains a random subsample of probands, so that one can estimate the population level effects. This random sample also provides a source of controls to use in case-control analyses. The Focused Study also contains two high risk samples, for use in case-control analyses. The first contains early, validated CHD siblings and an unaffected control sib. The second, contains sibpairs who both show increased atherosclerosis as measured by intima medial carotid artery thickening, along with a control sib. See the sections below, for a more complete explanation of each subsample.

[NOTE: The Focused Study should not be considered a scientific subsample perse, but merely a construct for prioritizing assay work. Thus we recommend that you DO NOT refer to these subgroups collectively as the "Focused Study" in any publication or scientific presentation. Rather, refer to each of the three subsamples comprising the Focused Study, and give their suggested description, below.]

Extended Focused Study (N=834 subjects)

The Extended Focused Study consists of subjects in the following 3 of the above subsamples, who are NOT already part of the Focused Study.

2. ERSAMP (Extended Random Proband Sample N=454 unrelated subjects)

5. CHDPSAMP (CHD Sib-Pair Sibships N=341 sibs in S=82 sibships)

7. IRSSAMP (Indiv Risk Score CHD Sibships

FINE MAP HALF N=973 sibs in 276 sibships)

The initial RSAMP (N=200) proband subsample was expanded to include ALL Phase II random family probands, since for some analyses, N=200 did not provide enough of a basis from which to extract an adequate control sample. Also, as focus has shifted to linkage hypotheses, larger sets of sibpairs and sibships are required. Thus, two (overlapping) samples of sibships are included here. The first are sibships those that include at least one validated CHD sibpair (not necessarily early). The second is sibships with at least one sib having high individual risk for CHD based upon risk factor profile. The total IRSSAMP is 1,930 individuals was deemed too large to be able to do all assays within the budgetary constraints. Thus, it was decided to do these assays on the fine map half, since once all data have been collected and all assays run, this sample should contain those genes which are showing the greatest signals in these data.

[NOTE: As for the Focused Study, the Extended Focused Study should not be considered a scientific subsample perse, but merely a construct for prioritizing further assay work. Thus we recommend that you DO NOT refer to these subgroups collectively as the "Extended Focused Study" in any publication or scientific presentation. Rather, refer to each of the three subsamples comprising it, and give their suggested description, below.]

Many assays should now be complete or nearly complete on the combined Focused/Extended Focused Study samples.

Candidate Region Sample (N=1,726 subjects)

Candidate genes explored by the FHS have to date been done on the same sets of subjects* defined by the following 6 of the above component samples:

2. ERSAMP (Extended Random Proband Sample N=454 unrelated subjects)

3. BLKSAMP (African American Sample N=263 subjs in F=102 families)

5. CHDPSAMP (CHD Sib-Pair Sibships N=341 sibs in S=82 sibships)

6. IMTSAMP (Initma Medial Thickness Sibships N=246 sibs in S=82 sibships)

8. FCHLSAMP (Famil-Comb Hyperlipid Sibpairs N=170 sibs in S=71 sibships)

9. HYPSAMP (Hypertensive Sibpairs N=697 sibs in S=278 sibships)

Suggested description in publications:

The candidate gene region was evaluated in a sample consisting of sibships drawn from FHS families having 1) at least one CHD sibpair; 2) at least one pair of sibs both in the upper 80th percentile of the intima medial carotid wall thickness distribution; 3) Familial Combined Hyperlipidemia sibpairs; 4) at least one hypertensive sibpairs. Gene frequencies were estimated from the sample of 454 Phase II random probands. The region was also evaluated in all 102 Phase II African American families.

So far, these candidate genes have been mostly polymorphic linkage markers in a candidate region, NOT the purported "functional" mutation. Thus, the primary analyses envisioned for the candidate region is linkage (although some association analyses can also be conducted secondarily assuming some linkage disequilibrium with the marker). The sample includes the ERSAMP extended random sample (from which population gene frequencies can be estimated), as well as the entire African-American cohort, since less is known about genetic risk factors in this group, and it is not large enough to include as a separate group in the genome scan. Four samples of sibships are also included here: CHD sibpairs, IMT sibpairs, FCHL sibpairs and hypertensive sibpairs. Although it is possible that future candidate region work will be performed on a different subset of subjects, the intent is to try to do all candidate regions on the same samples, which allows for eventual multivariate analyses to include the effects of multiple small effect loci, with possible interactions.

[*Note: A few of the early candidate regions were done only on RSAMP, not ERSAMP, and did NOT include the BLKSAMP.

Some single mutation association markers of candidate genes, like ACE and AGT, were done early on in substudies drawn specifically tailored to those genes.]

Coarse Map Sample (N=1,330 subjects)

The coarse map sample is the set of subjects on whom the Utah 240 anonymous marker genome scan is being conducted. It consists of 3 of the above component subsamples:

1. RSAMP (Random Proband Subsample N=200 unrelated subjects)

5. CHDPSAMP (CHD Sib-Pair Sibships N=341 sibs in S=82 sibships)

7. IRSSAMP (Indiv Risk Score CHD Sibships

COARSE MAP HALF N=957 sibs in 275 sibships)

Suggested description in publications:

When describing the genome wide scan analysis using the Coarse Map Sample:

A genome-wide scan was performed using 240 anonymous polymorphic markers approximately equally spaced throughout the genome. This coarse map scan was done on two overlapping sets of sibships: 1) 341 sibslings from 82 sibships containing at least one early, validated CHD sibpair, and 2) a random half (coarse map sample; N=957 subjects in 275 sibships) of the sample of all sibships which have at least one member at increased risk to CHD (individual risk score sibship sample). This individual risk score sibship sample consists of 1.930 subjects from 551 sibships in which at least one sibling was above the age-sex specific 80th percentile of the Individual Risk Score for CHD (a score derived for each individual using sex-specific proportional hazards models to predict the age-of-onset of CHD, based upon carotid artery intima media thickness, lipids, body mass index, blood pressure, hypertension, diabetes, and CHD family history score). The intention is to later confirm and more finely map the most promising findings in the remaining independent half of the sample (fine map sample; N=973 subjects in 276 sibships). Gene frequencies were estimated from an independent sample of 200 random probands.

Fine Map Sample (N=1,332 subjects)

The fine map sample is the set of subjects on whom the most promising regions identified in the Utah 240 anonymous marker genome scan will be validated, and on which finer mapping will take place. It consists of 3 of the above component subsamples:

1. RSAMP (Random Proband Subsample N=200 unrelated subjects)

5. CHDPSAMP (CHD Sib-Pair Sibships N=341 sibs in S=82 sibships)

7. IRSSAMP (Indiv Risk Score CHD Sibships

FINE MAP HALF N=973 sibs in 276 sibships)

Suggested description in publications:

When describing the replication and finer mapping using the Fine Map Sample:

The fine map sample is a replication sample consisting of two overlapping sets of sibships: 1) 341 siblings from 82 sibships containing at least one early, validated CHD sibpair, and 2) 973 siblings from 276 sibships which contain at least one sibling above the age-sex specific 80th percentile of the Individual Risk Score for CHD (a score derived for each individual using sex-specific proportional hazards models to predict the age-of-onset of CHD, based upon carotid artery intima media thickness, lipids, body mass index, blood pressure, hypertension, diabetes, and CHD family history score). The entire set of such risk score sibships (1,930 subjects from 551 sibships) was divided in half, randomly, and the first half (coarse map sample) was used to conduct a coarse map genome-wide scan using 240 anonymous polymorphic markers approximately equally spaced throughout the genome. The more promising regions identified in the coarse map scan were then retested and more finely mapped in the independent fine map sample. Gene frequencies were estimated from an independent sample of 200 random probands.

The intent is to genotype markers on either side of the "hit" regions to more finely delineate the signals. Since new markers will be genotyped in addition to the original 240, the RSAMP subjects will again be needed for estimation of gene frequencies. The fine map half of the IRSSAMP will serve as an independent replication sample of the hit markers, as well as for finer mapping of the regions.

 

1. RSAMP (Random Proband Subsample N = 200 unrelated subjects)

2. ERSAMP (Extended (complete) Random Proband Sample N = 454 unrelated subjects)

Random Probands

 

Suggested description in publications:

You may call ERSAMP the Random Sample of Probands and RSAMP the Random Sub-sample of probands in publications, making ERSAMP the main "unqualified" one and RSAMP the qualified term i.e.

For ERSAMP:

The random sample of probands consists of all 456 probands from random sample Phase II families.

For RSAMP:

The random sub-sample of probands consists of 200 randomly selected probands (50 from each Field Center) from Phase II Random families.

NOTES:

Uses:

 

3. BLKSAMP (African American Sample N=263 subjects in F=102 families)

 

Suggested description in publications:

The African-American sample consists of all 263 subjects from 102 African-American families recruited from the Forsyth County, North Carolina FHS field center.

NOTES:

 

4. CHDSAMP (CHD Cases + Sib Controls ; N=290 sibs in S=121 sibships)

5. CHDPSAMP (CHD Sib-Pair Sibships; N=341 sibs in S= 82 sibships)

CHD Sibs and Sibling Controls

 

Suggested description in publications:

For CHDSAMP:

The CHD case and sib control sample consists of 121 sibships which contain early, validated, familial CHD cases (N=139); all of their late or unvalidated CHD siblings (N=35); plus the most extremely discordant unaffected sibling (N=116). The later sibling is defined as having no reported or validated CHD, and whenever possible being of the same sex as the CHD sib, at an age older than that defined for early CHD (45 years for males, 55 for females), with no reported angina or other heart disease.

For CHDPSAMP:

The CHD sibpair sample consists of 82 sibships (341 siblings) with 2 or more CHD siblings, defined as having validated CHD or ECG-MI. These sibships contain 175 CHD and 166 non-CHD siblings.

NOTES:

Uses:

 

6. IMTSAMP (Initma Medial Thickness Sibships: N=246 sibs in S=82 sibships)

IMT Sibpairs+Sib Controls

 

Suggested description in publications:

The IMT sibpair sample consists of 82 sibships (246 subjects) in which at least two siblings had carotid artery intima media thickening (above the 90th percentile of the age-specific distribution from the study distribution; N=197) along with their most extremely discordant sibling (defined as being in the lower 75% age-specific percentile, and having the smallest thickening; N=49).

NOTES:

Uses:

 

7. IRSSAMP (Indiv Risk Score CHD Sibships; N=1,930 sibs in S=551 sibships)

[Note this sample is divided randomly in half (by sibship) into two parts:

Coarse Map Sample N=957 sibs in S=275 sibships

Fine Map Sample N=957 sibs in S=276 sibships

Suggested description in publications:

For the ENTIRE IRSSAMP (both coarse and fine map samples):

The individual risk score sibship sample consists of 1.930 subjects from 551 sibships in which at least one sibling was above the age-sex specific 80th percentile of the Individual Risk Score for CHD (a score derived for each indiviudal using sex-specific proportional hazards models to predict the age-of-onset of CHD, based upon carotid artery intima media thickness, lipds, body mass index, blood pressure, hypertension, diabetes, and CHD family history score).

See the sections on Coarse Map Sample and the Fine Map Sample for descriptions of those "halves" of the IRSSAMP only.

NOTES:

Uses:

 

8. FCHLSAMP (Familial-Combined Hyperlipidemia Sibpairs N=170 sibs in 71 sibships)

 

Suggested description in publications:

The familial combined hyperlipidemia (FCHL) sibpair sample consists of FCHL sibpairs from FCHL pedigrees. FCHL affected subjects are defined as being above the age/sex specific 90th percentile in either LDL or triglyceride and are subclassified as being either type 2a (> 90th LDL but < 90th triglyceride), type 2b (> 90th in both LDL and triglycerides) or type 4 (> 90th triglycerides but < 90th LDL). FCHL pedigrees contain at least two first degree relatives both being type 2b FCHL, or having two first degree relatives having two different FCHL phenotypes (i.e. the pedigree must exhibit elevations in both lipids, even if some of its members show abnormalities in only one). There were 170 FCHL subjects from 71 FCHL sibships.

NOTES:

Uses:

 

9. HYPSAMP (Hypertensive Sibpairs N=697 sibs in 278 sibships)

 

Suggested description in publications:

The hypertensive sibship sample consists of all hypertensive siblings with at least a sibpair (defined as BP > 140/90 or on anti-hypertensive medication). There were 697 hypertensive siblings in 278 sibships.

NOTES:

Uses: