Family history (FHx) of premature coronary artery disease (CAD) is a risk factor for development of incident cardiovascular disease. However the association between FHx and outcomes in patients with established CAD is unclear. We followed 84,373 patients with angiographic CAD enrolled in the inclusive Alberta Provincial Project for Outcomes Assessment in Coronary Heart Disease registry between April 2002 and March 2013. Overall, 25,566 (30%) self-reported an FHx of CAD, defined as a first-degree relative with premature CAD (men, age <55 years; women, age <65 years). We tested the association between FHx and all-cause mortality using multivariable Cox proportional hazards regression. After adjusting for baseline differences in clinical characteristics, indication, and extent of CAD, FHx was associated with reduced all-cause mortality over a median 5.6 years in follow-up (hazard ratio [HR] 0.77 [95% CI 0.73 to 0.80]). The magnitude of this protective association was weaker in those with versus without a previous myocardial infarction (HR 0.87 [95% CI 0.81 to 0.93] versus 0.72 [0.69 to 0.76], interaction p <0.0001) and slightly stronger in those presenting with versus without an acute coronary syndrome (HR 0.74 [0.70 to 0.79] versus 0.80 [0.75 to 0.85], interaction p = 0.08). There was attenuation of association with increasing age, but FHx remained protective even in those aged older than 80 years (HR 0.86 [0.77 to 0.95]). In conclusion, in patients with angiographic CAD, self-reported FHx of premature CAD is associated with improved long-term survival rate, independent of clinical characteristics, mode of presentation, and extent of disease. Further investigation of potential patient- and system-level mediators of this seemingly paradoxical relation is required.
Family history (FHx) of premature coronary artery disease (CAD) is a traditional risk factor for developing incident CAD, as well as a predictor of adverse cardiovascular events in those without established cardiovascular disease. The relative risk for CAD in first-degree relatives of affected persons ranges from 2 to 12 times that of the general population. Risk increases with the number of primary relatives affected and at younger ages of onset in the probands. Current cardiovascular risk prediction models consider FHx to be an important risk modifier: the Reynolds Risk Score reclassifies 20% to 30% of patients from intermediate to high risk because of the presence of FHx of CAD, and the Framingham Offspring Study showed that FHx doubles the risk of developing major adverse cardiovascular outcomes. However, the prognostic importance of positive FHx once CAD is established remains unclear. Some studies show that it is associated with adverse outcomes, whereas some studies show it is protective. We sought to study the influence of FHx on all-cause mortality in patients with established CAD in a prospective registry with long-term follow-up.
Methods
The Alberta Provincial Project for Outcomes Assessment in Coronary Heart Disease (APPROACH) Program is an inclusive prospective registry of patients who underwent coronary angiography in Alberta, Canada. Database and data collection methods have previously been described in detail. In brief, data collected include demographic characteristics, cardiac risk factors, major clinical co-morbidities, previous cardiac events, indication for angiography, ejection fraction, coronary anatomy, and cardiac medications at the time of angiography.
For this analysis, we included patients enrolled after a first coronary angiogram, regardless of indication, between April 2002 and March 2013. An FHx of premature CAD was identified by patient self-report at the time of enrollment and was defined as 1 or more first-degree relatives with known premature CAD (men, age <55 years; women, age <65 years). In the Newcastle Family History Study, self-report of an FHx of premature CAD in a first-degree relative has been found to be reasonably accurate with sensitivity more than 80% and specificity of about 90%. Because we were interested in the association between FHx and outcomes in patients with established CAD, we included only patients with ≥1 stenosis of ≥50% at angiography.
The primary outcome was all-cause mortality, ascertained by quarterly linkage with Alberta Vital Statistics, up to March 31, 2014. The Conjoint Health Region Ethics Board at the University of Calgary approves APPROACH data collection and linkage, as well as the waiver of individual consent required to provide an inclusive prospective registry.
Data are reported as mean ± SD or as total numbers and percentage. We compared baseline clinical, demographic, and angiographic characteristics of those with versus without FHx using the Fisher’s exact test for categorical variables and the unpaired t test for continuous variables. We used multivariable Cox proportional hazard regression models to quantify the association between FHx and mortality, after verifying the proportional hazards assumption with log–log plots. Models were adjusted for all clinical, demographic, and angiographic variables listed in Tables 1 and 2 . In addition, we assessed for effect measure modification by stratifying adjusted analyses based on age at presentation (in 10-year bands), gender, indication for angiography (acute coronary syndrome [ACS] vs other), and history of a myocardial infarction (MI) or revascularization procedure (percutaneous coronary intervention [PCI] or coronary artery bypass grafting [CABG]). Finally, we also performed a propensity-matched analysis as a sensitivity test. We calculated the propensity to FHx of premature CAD using logistic regression, including all baseline variables in Tables 1 and 2 , and estimating the association between FHx and mortality in a 1:1 cohort matched on propensity score. Analyses were performed with SAS 9.2 (SAS, Cary, NC) and were 2-tailed with a significance level of 0.05.
| Variable | Total cohort (N=84,373) | Family History of Premature Coronary Artery Disease | P | |
|---|---|---|---|---|
| Yes (N=25,566) | No (N=58,807) | |||
| Mean age, year (SD) | 64.0±11.8 | 60.9±11.1 | 65.3±11.8 | <0.0001 |
| Men | 61,696 (73%) | 18,609 (73%) | 43,087 (73%) | 0.1480 |
| Age >75 years | 16,789 (20%) | 3,067 (12%) | 13,722 (23%) | <0.0001 |
| Cerebrovascular disease | 5,330 (6%) | 1358 (5%) | 3972 (7%) | <0.0001 |
| Diabetes mellitus | 21,629 (26%) | 6,124 (24%) | 15,505 (26%) | <0.0001 |
| Dialysis | 1,197 (1%) | 193 (1%) | 1,004 (2%) | <0.0001 |
| Hypertension | 58,083(69%) | 18,095(71%) | 39,988 (68%) | <0.0001 |
| Hyperlipidemia | 61,820 (73%) | 19,977 (78%) | 41,843 (71%) | <0.0001 |
| Malignancy | 3,333 (4%) | 864 (3%) | 2,469 (4%) | <0.0001 |
| Prior myocardial infarction | 18,725 (22%) | 4,318 (17%) | 14,407 (25%) | <0.0001 |
| Prior percutaneous coronary intervention | 4,824 (6%) | 1,233 (5%) | 3,591 (6%) | <0.0001 |
| Prior coronary bypass | 2,852 (3%) | 726 (3%) | 2,126 (4%) | <0.0001 |
| Peripheral vascular disease | 6,821 (8%) | 1,875 (7%) | 4,946 (8%) | <0.0001 |
| Gastrointestinal disease | 5,698 (7%) | 1,753 (7%) | 3,945 (7%) | 0.4299 |
| Liver disease | 849 (1%) | 183 (1%) | 666 (1%) | <0.0001 |
| Congestive Heart failure | 11,124 (13%) | 2,333 (9%) | 8,791 (15%) | <0.0001 |
| Renal disease (Creatinine>200 mmol/L) | 3,796 (5%) | 641 (3%) | 3,155 (5%) | <0.0001 |
| Pulmonary disease | 13,349 (16%) | 3,484 (14%) | 9,865 (17%) | <0.0001 |
| Current smoker | 24,082 (29%) | 8,525 (33%) | 15,557 (27%) | <0.0001 |
| Previous smoker | 28,972 (34%) | 9,049 (35%) | 19,923 (34%) | <0.0001 |
| Left Ventricular Ejection Fraction | <0.0001 | |||
| >50% | 43,709 (52%) | 15,421 (60%) | 28,288 (48%) | |
| 35-50% | 13,726 (16%) | 4,169 (16%) | 9,557 (16%) | |
| <35% | 10,520 (13%) | 1,392 (6%) | 9,128 (16%) | |
| Not done / missing | 16,418 (20%) | 4,584 (18%) | 11,834 (20%) | |
| Medications | ||||
| Aspirin | 69,813 (83%) | 22,246 (87%) | 47,567 (80%) | <0.0001 |
| Warfarin | 3,392 (4%) | 834 (3%) | 2,558 (4%) | <0.0001 |
| Ticlopidine or Clopidogrel | 37,292 (44%) | 12,898 (51%) | 24,394 (42%) | <0.0001 |
| Beta-blocker | 55,103 (65%) | 17,813 (70%) | 37,290 (63%) | <0.0001 |
| Angiotensin converting enzyme inhibitor | 41,198 (49%) | 13,166 (52%) | 28,032 (48%) | <0.0001 |
| Calcium Channel Blocker | 10,924 (13%) | 3,211 (13%) | 7,713 (13%) | 0.0270 |
| Angiotensin II receptor blocker | 7,761 (9%) | 2,624 (10%) | 5,137 (9%) | <0.0001 |
| Long acting Nitrates | 13,579 (16%) | 4,005 (16%) | 9,574 (16%) | 0.0255 |
| Statins | 49,048 (58%) | 16,454 (64%) | 32,594 (55%) | <0.0001 |
| Total cohort (N=84,373) | Family History of Cardiovascular Disease | P | ||
|---|---|---|---|---|
| Yes (N=25,566) | No (N=58,807) | |||
| Indication | ||||
| Acute coronary syndrome | 50,319 (59.6%) | 15,025 (58.8%) | 35,294 (60.0%) | 0.0007 |
| Stable angina pectoris | 22,293 (26.4%) | 7,940 (31.1%) | 14,353 (24.4%) | <0.0001 |
| Heart failure | 2,556 (3.0%) | 605 (2.4%) | 1,951 (3.3%) | <0.0001 |
| Other | 9,124 (10.8%) | 1,963 (7.7%) | 7,161 (12.2%) | <0.0001 |
| Coronary characteristics | <0.0001 | |||
| Borderline | 13,726 (16.3%) | 4,099 (16.0%) | 9,627 (16.4%) | |
| 1 non proximal left anterior descending | 18,957 (22.5%) | 6,032 (23.6%) | 12,925 (22.0%) | |
| 1 proximal left anterior descending | 3,724 (4.4%) | 1,197 (4.7%) | 2,527 (4.3%) | |
| 2 non proximal left anterior descending | 15,894 (18.8%) | 4,935 (19.3%) | 10,959 (18.6%) | |
| 2 proximal left anterior descending | 3,605 (4.3%) | 1,101 (4.3%) | 2,504 (4.3%) | |
| 3 vessel disease | 21,919 (26.0%) | 6,415 (25.1%) | 15,504 (26.4%) | |
| Left main | 6,548 (7.8%) | 1,787 (7.0%) | 4,761 (8.1%) | |
| Outcomes | ||||
| Percutaneous coronary intervention within 1 year | 35,590 (42.2%) | 11,420 (44.7%) | 27,170 (41.1%) | <0.0001 |
| Coronary bypass within 1 year | 13,906 (16.5%) | 4,147 (16.2%) | 9,759(16.6%) | 0.1782 |
| Median follow-up, years Median [IQR] Mean(std) | 5.6 (5.7%) 5.8±3.3 | 5.4 (4.9%) 5.5±2.8 | 5.7 (6.2%) 6.0±3.5 | <0.0001 |
| Death (crude mortality rate, %/year) | 3.0 | 1.9 | 3.4 | <0.0001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
