Nurse practitioners caring for patients with cardiovascular (CV) disease (CVD) commonly perform assessments of long-term CVD risk. Such risk assessments are a cornerstone of preventive care because many risk factors respond to interventions and lower the patients' long-term risk.1,2 Nurse practitioners are well positioned to have a major impact on patients' CVD risk through accurate, comprehensive assessments and corresponding interventions.
Although all care providers are taught to gather family history information, there is no systematic approach to applying such information to the patient's risk stratification.3 Possible underestimation of CVD risk misguides patients and providers, leaving them unaware of potentially harmful health risks, many of which may be modifiable.
Cardiovascular disease risk assessment warrants a high degree of accuracy because current practice guidelines assign more aggressive goals for risk factor modification to patients identified as high risk compared with patients with low or moderate risk.4 Although numerous CVD risk scoring systems have been developed and are widely used, many scores do not identify family history of premature CVD risk as a high-risk factor.5
Family history is underused in risk assessment of CVD because of 2 main factors: (1) many traditional risk scoring systems do not incorporate family history (the National Cholesterol Education Program-Adult Treatment Panel III,4 the Framingham Risk Score [FRS],5 the European Societies of Cardiology,6 and the Systematic Coronary Risk Evaluation7) and (2) clinical guidelines are vague and do not provide adequate guidance for the clinician to stratify risk by family history and thereby recommend appropriately aggressive interventions for the patient.3,8 The purpose of this integrative literature review was to examine the evidence for inclusion of family history in CVD risk assessment to improve patient risk estimation.
Literature Search Methods
A comprehensive literature search was executed in the PubMed, Cumulative Index to Nursing and Allied Health, and EMBASE databases using the Medical Subject Headings terms of family history of CVD, family history of premature CVD, risk assessment, and risk estimation. The literature search was limited to studies reported during the time frame of years 1992 through 2012. The articles selected were peer reviewed, were published in English, and included adult participants. Manual searches were also performed by reviewing the reference lists of all selected articles for additional pertinent references.
The initial search using the terms family history of CVD or family history of premature CVD displayed 416 articles. When the terms risk assessment and risk estimation were applied, the number was reduced to 85. Limiting the search to adults further reduced the number of articles to 65. All citations were reviewed independently by 2 investigators to determine whether the references were appropriate for inclusion in the review of evidence. Any disagreements were resolved by discussion with a third investigator, and differences were resolved by consensus. Of the 65 publications, 37 were retrieved in full, having eliminated all others because these were not pertinent to the topic being investigated. Thirteen articles were opinion articles or editorials, and 24 articles were identified as appropriate and contributory for inclusion in the review.
Literature Search Results
The identified articles that were appropriate to include in the evidence review are listed in the Table. There were 9 observational studies, 10 cross-sectional studies, 4 case-control cross-sectional studies, and 1 randomized controlled trial (RCT).
Literature Review Findings
The research articles in the Table are listed in chronological order of publication and categorized by study design. The populations studied, the sample sizes, and the limitations of the studies are noted.
Three of the earliest studies used observational designs that followed large patient populations for up to 26 years.9-11 These studies found that a positive family history of CVD is a powerful predictor of incident CVD, independent of other risk factors. In the Swedish Twin Registry, death from CVD in 1 twin before the age of 55 years conferred dramatic increases in risk for death from CVD in the surviving twin (8-fold and 15-fold increased risk for men and women, respectively).9 In the other 2 studies, the degree of risk associated with CVD in a first-degree relative ranged from 41% increased10 to as high as 250% increased.11 Furthermore, maternal history of CVD was demonstrated to be as important as paternal history.11 These studies did manifest limitations in their design. Specifically, the Swedish Twin Registry was a retrospective observational study relying on death certificates with diagnoses coded by the International Classification of Diseases. The Atherosclerosis Risk in Communities study10 demonstrated selection bias because the nonresponder portion of the population had lower educational achievement and tended to have less healthy lifestyles. This selection bias might overestimate or underestimate the impact of family history, in that people with higher cardiac risk estimates may not have been included in the study. The Physician Health Study and the Women's Health Study were limited, in that family history information was collected on parental CVD but not on sibling history.11
An early cross-sectional study, the Stockholm Heart Epidemiology Program, reported the synergistic effects of family history as a predictor of CVD when used in women in combination with current smoking history and a low-density lipoprotein/high-density lipoprotein quotient12 of higher than 4.0. In men, family history showed predictive synergy with diabetes mellitus.
A more recent analysis from a large-scale case-control study, the INTERHEART study,13 established parental family history of myocardial infarction (MI) as a robust predictor that was significantly associated with risk for MI, independent of 9 major CVD risk factors (abnormal lipids, smoking, high blood pressure, diabetes, abdominal obesity, psychosocial factors, physical activity, fruit and vegetable consumption, and alcohol consumption). This relationship was consistent across all world regions, income, age groups, and gender analyzed in the study, including all FRS risk groups.13
Five cross-sectional studies using coronary artery calcium (CAC) scores measured by electron beam computed tomography (EBCT) were published between 2004 and 2007 by 3 different investigator groups in 5 study populations.14-18 Each of these studies demonstrated a strong association of high CAC scores with family history. A positive sibling history was found to be more strongly associated with a high CAC score than parental history alone, and family history in both sibling and parent yielded the strongest association of all.14 Family history in a first-degree relative was more strongly associated with CAC than family history in a second-degree relative, but a positive family history in both first- and second-degree relatives had the strongest degree of association.15 In a population referred by a physician for EBCT, a positive family history of coronary disease acted synergistically with other risk factors to magnify the association with a high percentile CAC score.16 This finding was demonstrated in a population of asymptomatic women who were generally at low risk for CVD. In asymptomatic women without diabetes, the performance of the Framingham risk estimation (FRE) was evaluated by CAC score. The majority of (84%) women with significant CAC (>75th percentile) were classified as low risk, and approximately half of the low-risk women with a family history of premature coronary disease had significant CAC.17
The final cross-sectional study using CAC as its main measure is the Multi-Ethnic Study of Atherosclerosis (MESA) trial.18 This study sought to examine whether asymptomatic individuals with positive family history of premature CVD have an increased atherosclerotic burden in the low- and intermediate-risk categories based on the traditional FRS. Family history was significantly associated with coronary artery calcification in all ethnic groups. The prevalence of coronary artery calcification was higher in those with family history compared with those with no family history in the low-risk (35% vs 23%, P < 0.0001) and in the intermediate-risk (70% vs 60%, P = 0.01) categories.18 The authors state that family history in middle-aged adults should be considered for quantification of risk and incorporated with existing risk prediction algorithms for aggressive primary prevention such as intensified weight loss and pharmacotherapy (lipid-lowering, aspirin, and antihypertensive therapy).18 All 5 of the cross-sectional coronary artery calcification studies had the limitation of determining family history by self-report rather than from the objective examination of medical records.
The Johns Hopkins Sibling Study had a unique study design, in that it enrolled 102 asymptomatic women who were sisters of patients hospitalized for documented premature coronary disease.19 The 102 enrolled participants underwent assessment using the FRE. Low-risk FRE was assigned to 98 (98%) of the 102 participants, and moderate risk was assigned to the remaining 2 participants (2%). However, 40% of these participants had detectable coronary artery calcification. Significant subclinical atherosclerosis was found in 32% of the participants, and 17% had CAC scores that ranked higher than the 90th percentile.19
The Framingham Offspring Study has developed a cohort of participants whose data have been analyzed using several different study designs. Three publications have correlated newly gathered data from these participants, who are offspring of the original Framingham participants. In a cross-sectional study, carotid intima media thickness (CIMT) of 1662 offspring participants was analyzed by ultrasound. Participants who had 1 parent with premature coronary heart disease (CHD) were found to have greater CIMT values compared with participants with no parental history.20 In another look at the offspring, this time an observational design following 2302 participants for 8 years, participants who had 1 parent with premature CHD were 1.7 (women) to 2.6 (men) times more likely to experience a CVD event.21 A third offspring study examined a cohort of participants who were selected on the basis of the fact that at least 1 sibling was enrolled and followed in the Framingham Heart Study.22 Baseline risk factors and incident events for 8 years were much higher in the sibling with CVD group than in the sibling without CVD group (odds ratio, 1.55). The study also demonstrated that sibling CVD conferred increased risk for future CVD events above and beyond the established risk factors and parental CVD.
A cross-sectional case-control design was used in a multicenter Italian study that enrolled 2,016 patients with 11,696 relatives and 1,757 controls with 8,897 relatives.23 The analysis showed that, among the patients surviving their first MI before the age of 46 years, the risk for early-onset MI was higher among siblings (hazard ratio [HR], 1.7) than among parents (HR, 0.9).
In a population from Western Washington State, 107 women aged 18 to 44 years with their first acute MI were compared with 526 women similar in age who served as control participants.24 A detailed questionnaire elicited history of MI in first-degree relatives. The rate of MI was twice as high in first-degree relatives of the patients compared with the controls.
The Health Styles annual mail survey reports that family history may be especially useful in younger populations whose traditional risk factors seem benign.25 There were 4035 respondents, 60% women, with a mean age of 48.8 years. The authors report a 2- to 5-fold increased risk for early-onset CVD based on degree of severity in family history. The authors propose that, to address the barriers in existing guidelines, national guidelines should view family history as a tool that will (1) aid in the identification of people with significantly increased disease risk, attributable, in part, to genetic factors; (2) improve early detection and prevention efforts for people with increased familial risk; and (3) facilitate treatment algorithms, with messages tailored to the level of familial risk.
The inadequate performance of the FRS was further demonstrated by the Better Adherence to Therapeutic Lifestyle Change Efforts trial.26 In this analysis of 93 asymptomatic patients at a military hospital in Washington, DC (59% women; mean age, 54 years), all patients were identified as high risk for heart disease by CIMT measurement, a surrogate measure of atherosclerosis. Of these high-risk patients, 75% were categorized as low risk by the FRS. Of note, 60% of the 93 patients had a positive family history of CVD.
Four observational studies are noteworthy for their inclusion of enormous study populations. These studies range in size from nearly 20,000 randomly selected Icelanders followed for up to 19 years,27 to nearly 23,000 UK citizens in the European Prospective Investigation of Cancer study,28 to nearly 50,000 participants in the Cooper Center longitudinal study,29 to almost 4 million residents of Denmark followed for 89 million person-years.30 These studies all demonstrated similar findings of a substantial risk for the development of CHD27,28,30 (HR, 1.6-1.7) or CHD mortality29 (HR, 1.4-1.6), conferred by a positive family history of CVD.
A novel measure was used in a study from the Mayo Clinic to assess the predictive value of family history in CVD.31 In a large cohort of patients without obstructive coronary disease by coronary angiography, coronary vasoreactivity was evaluated with intracoronary bolus injections of adenosine followed by infusions of acetylcholine, with consequent measures of coronary diameters and coronary blood flow. The concept of the study relied on the notion that there can be functional abnormalities at the level of coronary microcirculation at early stages of coronary atherosclerosis in the absence of obstructive disease. The outcome measure served as a proxy for CVD events or CV death. The study reported that, in a multivariable analysis, positive family history of coronary disease was a significant independent predictor (P = 0.04) of reduced coronary flow reserve, a precursor to the development of atherosclerosis.
The feasibility of systematically collecting family history information for CVD risk was demonstrated in the primary care setting.32 In a matched-pair cluster RCT, control participants had the usual FRS-based CVD risk assessment with family history, as usually recorded in their medical records. Intervention participants had the usual CV risk assessment but also completed a family history questionnaire. There was a near universal completion rate (98%), and the number of participants found to be at high risk for CVD increased by 41% by using the family history questionnaire. Incorporating systematically collected family history information was demonstrated to be feasible and cost-effective.
Discussion
This review of the evidence clearly reveals that family history is an important independent predictor of CVD, as demonstrated in 19 separate studies.1-13,20,22,24,27-29 Of course, it is possible that this impression results from publication bias because positive associations tend to make it to publication, whereas negative findings are often left unpublished. However, a number of observations argue against publication bias and validate the important role of family history in risk estimation.
First, every study that addressed the utility of family history in CVD risk assessment uniformly agrees with the central finding that family history adds to commonly used risk prediction models. Frequently used descriptors from these studies are that family history is an independent predictor and that family history is synergistic with other risk factors in improving accuracy of risk assessment.
Second, a variety of different outcome measurements and different study designs supported the utility of family history. The outcome measures and study designs include self-reported events by mail survey,25 anatomical assessments with CAC scores14-18 and CIMT20,26 in cross-sectional surveys, case-control cross-sectional studies in patients with events,12,13,23,24 cross-sectional evaluations of siblings of patients with events,19 prospective observational studies of incident events,10,11,21,27,29 assessment of vasoreactivity during angiography,31 and an RCT.32 Despite the variety of outcome measures and study designs, every study validated the utility of family history as an independently predictive risk factor.
An area of concern with published articles on the effect of family history on CVD risk is the use of heterogeneous definitions of what constitutes a positive family history. The spectrum of definitions ranges from any CV event in a related person including second- and third-degree relatives at any age to specific cardiac events in first-degree relatives younger than 50 years. Because of the variability in definitions, practitioners are unlikely to understand the predictive value and relevance of family history findings. The lack of a precise definition is a global problem in the assessment of family history.
Naturally, the degree of predictive power wielded by family history varied greatly depending on the characteristics of the population included in the study. The INTERHEART study showed predictive value of family history, independent of other risk factors consistent across world regions, income, age, and gender.13 However, other studies emphasize differences in the contribution of family history by gender and race.10,11 Specific studies suggest that family history is of particular value in the assessment of CVD risk for both women17,19,24 and the young.15,21,24,25 These 2 groups have traditionally been scored as low risk for CVD by scoring systems using conventional risk factors. Identifying these groups as targets for early intervention is critical because the process of CVD develops for decades and intervention is most effective when applied as early as possible in the cascade of events. Identifying risk for developing CVD in these subgroups of the population could confer great benefit to the preventive efforts.
Limitations
This literature review was limited to publications in English and may have therefore missed substantial work in other languages.
There are a number of limitations in the studies reviewed and discussed. Many of the studies are subject to referral bias; that is, participants included for analysis in the studies are not the same as participants who are not included. Typically, refusal to participate in research is behavior associated with persons whose educational level is low, whose general health behavior is unfavorable, or who are too ill to participate. This implies that participants who might benefit the most from efforts at risk assessment and risk reduction are not being systematically included in studies performed to date.
Most of the publications included in this review of evidence used family history according to participant recall. No study used formal medical record reviews of family history except for sibling and offspring studies in which participants were identified by sibling or parents with a CVD event.19,22 Errors in recall are understandable and expected but do limit the reproducibility of results. Nevertheless, participant recall most closely mimics the method used to obtain family history in clinical settings and may therefore be the most appropriate method to use for clinically useful studies.
Because a broad variety of outcome measures and study designs were used to assess CV risk, such as events, coronary artery calcification, CIMT, and vasoreactivity, it is difficult to compare findings between studies. The heterogeneity of measures and study designs may be partly responsible for the wide range of statistical estimates of the impact of family history on an individual's risk; that is, the impact size of a positive family history varies enormously between reports. The heterogeneity also makes it impossible to merge data from multiple studies in a formal meta-analysis to determine the global quantitative impact of a positive family history on the degree of increased risk. However, the heterogeneity of study designs may also be a potential strength of this review, in that multiple different study designs all seem to point to the same conclusion that family history is an invaluable, potent, and independent predictor of CVD risk. Irrespective of the road taken, the journey ends with the same affirming conclusion, strengthening the validity of the findings.
Another limitation in the studies reviewed is that most of the studies were not undertaken with CVD risk and family history as the purpose of the study a priori. Most of the studies reported on data collected under the auspices of another protocol for another purpose. Some of the studies were population subsets of an original protocol. The implication is that these data may produce interesting and valuable findings, but more definitive studies should be pursued with CVD risk and family history as the prime objective to collect a truly trustworthy data set to answer the question.
Implications for Practice
Despite consistent evidence supporting family history as a known risk factor of CVD, few standards address how this information should be collected, interpreted, and applied to clinical practice. Family history is not commonly used to counsel patients about risks or preventive behaviors largely because of ambiguity in the guidelines.3,33 These findings underscore the importance for the use of family history as an accessible and inexpensive standardized tool for risk-based interventions. For example, although the third Adult Treatment Panel III of the National Cholesterol Education Program guidelines4 (2004) recommends the assessment of family history, there are currently no guidelines for clinical decision making if the patient has positive family history in the absence of other CVD risk factors.1 Because of this confusion, a great deal of variability exists in clinical practice. When CVD risk is underestimated, the clinical impact is significant because miscategorization of risk misleads providers to set less aggressive treatment goals, which, in turn, causes patients to follow recommendations that may be suboptimal for their true risk level.
A cross-sectional analysis reveals that clinicians need direction in using family history to improve health by risk assessment algorithms and guidelines.34 This recent study of Kaiser Permanente clinicians in Oregon reported that family history information would be more useful if it were integrated into algorithms and tools already used in clinical decision making, such as the FRS. In addition, the authors recommend that family history information be used to identify persons who are at an increased risk for CVD and who may be receptive to and benefit more from CVD preventive interventions than patients with no family history. The presence of family history may motivate patients who do not have CVD but are at high risk for CVD to engage in healthy, potentially protective behaviors.32
In a discussion of strategies for CVD prevention, cost-effectiveness of family history data collection as accessible and costing between $1 and $3 through Internet collection is compared with the estimated $200 for laboratory data used to measure other risk factors.35 The article recommends combining family history information with clinical assessment for more intensive intervention. Screening for family history also allows patients with borderline risk factors to be aware of their CVD risk status. Even high-risk patients should be reassured that specific lifestyle changes and preventive therapies help to reduce risk.
Risk assessment in primary care is critical because it determines and guides the intensity of therapeutic intervention. First, causal risk factors should be explored, similar to those collected for the FRS.27 Second, predisposing risk factors such as family history of CVD should be assessed. In addition, advance practice nurses who receive education in behavioral change strategies, lifestyle interventions, and management of risk factors can implement the aggressive therapy necessary for these patients such as adherence to cholesterol treatment guidelines.36 Prevention programs promoting family history evaluation have been shown to be more successful than those that do not. Furthermore, advance practice nurses can address erroneous beliefs related to family history, such as fatalism; motivate when family history is present; and adjust the severity for treatment recommendations.32
Summary and Recommendations for Future Research
Despite the strong evidence, gaps in knowledge remain. Without large RCTs that longitudinally measure hard outcomes, such as CVD events, it will be difficult to evaluate true risk prediction. Such studies are necessary to distinguish between the diagnostic or predictive value of family history and CVD. Evaluation of family history in a large sample size over time can also provide knowledge regarding gender differences and specific cut points when the risk is increased, such as with postmenopausal women.
Additional studies are also necessary to test the utility of family history when combined with existing, well-validated risk scores such as the FRS. There are numerous family history risk prediction models that have been proposed; however, the studies lack power, making it difficult to identify applicability to the population at large. In addition, many scores have added other emerging risk factors to these predictive equations, making it difficult to assess family history separately.
There are also gaps in knowledge regarding the barriers to using family history in CVD risk assessment. For example, is the root of the problem the perception of family history, knowledge of how to incorporate the information, or inadequate guidelines to follow for preventive therapy? It is unclear why there is a lag between guideline development and adoption of guidelines into clinical practice.
Review of literature for strength and quality reveals that family history is a missing component of CVD risk appraisal, and its inclusion could improve accuracy in global CVD risk estimation. Recommendations for a process change are made on the basis of the evidence that family history data are accessible, inexpensive, and personalized. The research evidence strongly supports using a clinical algorithm that recognizes family history of premature CVD as an independent high-risk factor, in addition to the FRS or incorporated within the FRS. Translation of this research into practice would help providers such as nurse practitioners to assess family history of CVD with a systematic process and to recommend appropriate aggressive preventive care for patients in this high-risk category.
The opportunity to capture patients at risk for CVD, who are otherwise overlooked by traditional methods, and to provide them the awareness to protect themselves before overt disease presents is supported by robust evidence. This evidence must not be overlooked. Furthermore, the cost of preventive care for patients detected earlier in the progression of disease would be far less than that after an acute CVD event has taken place.1
Ultimately, the implementation of a clinical algorithm for family history of premature CVD would translate the evidence to standardize preventive health practices, improve the identification of high-risk patients, and provide this susceptible population with the needed primary preventive care to combat the number 1 killer of both men and women in the United States.
Clinical Pearls
* Literature review unequivocally supports the inclusion of family history as a strong and independent risk factor of cardiovascular risk estimation.
* Although clinical guidelines recognize family history as an indicator of heightened risk, these guidelines do not provide nurse practitioners with clear direction for clinical decision making.
* A clinical algorithm for inclusion of family history would standardize preventive health practices, improve the identification of high-risk patients, and provide preventive interventions that address patients' accurate risk.
REFERENCES