American Journal of Epidemiology Vol. 152, No. 9 : 884-888
Copyright © 2000 by The Johns Hopkins University School of Hygiene and Public Health
ORIGINAL CONTRIBUTIONS |
Assessing the Impact of Classical Risk Factors on Myocardial Infarction by Rate Advancement Periods
1 Institute of Epidemiology and Social Medicine, University of Münster, Münster, Germany.
2 Department of Epidemiology and Biostatistics, University of South Carolina, Columbia, SC.
3 Department of Epidemiology, University of Ulm, Ulm, Germany.
4 GSF-Institute of Epidemiology, Neuherberg, Germany.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The risk or rate advancement period (RAP) proposed by Brenner et al. (Epidemiology 1993;4:229–36) conveys information on the impact of a risk factor on the age dimension of chronic disease occurrence and may thus facilitate communication of epidemiologic findings. The RAP expresses how much sooner a given risk or rate of disease occurrence is reached among exposed than among unexposed individuals. The purpose of the present analysis was to derive estimates of RAPs for cardiovascular risk factors in relation to incident nonfatal and fatal myocardial infarction in middle-aged men of the Monitoring Trends and Determinants in Cardiovascular Diseases (MONICA) Augsburg cohort, Germany, between 1984 and 1995. RAPs were estimated based on Cox proportional hazards models. After multivariate adjustment, hypertension, smoking, and dyslipidemia were associated with RAPs of 8, 11, and 11 years, respectively, conditional on infarction-free survival to baseline and absence of competing risks. The RAP may be interpreted as that, on average, smokers are expected to advance their risk of myocardial infarction approximately 11 years compared with never/former smokers; for example, 50-year-old smokers are expected to carry the same risk of infarction as 61-year-old nonsmokers. The authors encourage the use and evaluation of the RAP as an effective risk communication tool in actual counseling situations. Am J Epidemiol 2000;152:884–8.
cohort studies; methods; myocardial infarction; risk
Abbreviations: MONICA, Monitoring Trends and Determinants in Cardiovascular Diseases; RAP, risk or rate advancement period.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The risk or rate advancement period (RAP) as developed by Brenner et al. (1
) has been proposed as a novel and informative measure of risk factor impact on chronic disease occurrence. Similar to the concept of years of life lost, the risk or rate advancement period estimates the impact of a risk factor on the timing of disease occurrence; that is, the risk is phrased in terms of premature disease risk or rate among exposed individuals. The purpose of this study was to estimate the impact of the classical cardiovascular risk factors hypertension, smoking, and dyslipidemia on the prematurity of occurrence of myocardial infarction. We applied the concept of risk or rate advancement periods in the Monitoring Trends and Determinants in Cardiovascular Diseases (MONICA) Augsburg cohort and critically appraised properties and usefulness as a potential risk communication tool.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Study design
The design of the study has been described in detail before (2
). In brief, the MONICA Augsburg cohort was initiated in 1984–1985 in the city of Augsburg, Germany, and two adjacent counties. Participants aged 25–64 years were interviewed with respect to medical history, alcohol intake, smoking habits, and medication use, including antihypertensive drugs. Systolic and diastolic blood pressure, body weight, and height were subsequently measured. A nonfasting, venous blood sample was drawn and analyzed with respect to serum total cholesterol and high density lipoprotein cholesterol. These analyses focus on the 1,014 men aged 45–64 years (83 eligible men had missing data on a relevant variable and were excluded) who comprise 50 percent of the entire male cohort.
Follow-up was conducted from 1984 through 1995, and cases of nonfatal and fatal myocardial infarction were ascertained via the MONICA Augsburg coronary event registry. A coronary heart disease event was considered as incident if it was the first event during follow-up in a person reporting no history of heart attack in the 1984–1985 survey. The World Health Organization MONICA diagnostic categories (derived from electrocardiogram, enzyme, symptom, and necropsy findings) included as coronary heart disease events in this cohort study are definite and possible nonfatal acute myocardial infarction and fatal coronary heart disease (combining definite and possible fatal coronary events and unclassifiable deaths). Detailed descriptions of the definitions and applications of these diagnostic categories have been published (). A total of 81 cases of incident myocardial infarction occurred among 1,014 men over the course of 11.5 years.
The classical cardiovascular risk factors considered in these analyses were defined as follows. Subjects having blood pressure values of 
160 mmHg systolic or 95 mmHg diastolic were defined as having hypertension. Those who were aware of their hypertension and were taking medication against hypertension were also considered hypertensive. The total cholesterol/high density lipoprotein cholesterol ratio was derived and categorized as 5.5 (dyslipidemia) and <5.5. Individuals who reported current regular smoking of one or more cigarettes per day were classified as smokers.
The present analyses extend the previously published report on the association of the risk factors hypertension, cigarette smoking, and dyslipidemia with incident nonfatal and fatal myocardial infarction occurrence over 8 years (2). In this report, we have shown that hypertension, cigarette smoking, and the high total cholesterol/high density lipoprotein cholesterol ratio were statistically significant, independent predictors of myocardial infarction in the male MONICA Augsburg cohort.
For comparison purposes, we used published parameter estimates to derive the RAP of smoking from other European cohort studies focusing on incident coronary heart disease that included middle-aged men and were conducted at a similar time.
Statistical methods
The derivation of the risk and rate advancement periods has been described in detail by Brenner et al. (1). Briefly, the RAP describes the advancement in time of the risk or rate of a chronic disease among subjects exposed to some risk factor in the absence of competing risks; that is, it expresses how much sooner a given risk or rate of disease occurrence is reached among exposed than among unexposed individuals.
Exposures conferring an increased risk will result in a positive value of the RAP (i.e., disease risk will be advanced to a younger age), while exposures conferring a protective effect will result in a negative value (i.e., disease risk will be postponed to an older age).
The fundamental assumption underlying the RAP concept that the disease under study exhibits a monotonic increase in disease rates (risk) with age is met by many chronic diseases including atherosclerosis. The definition of the RAP implies that it is inversely related to the age gradient of the disease rate (risk). For computation of a RAP, age needs to be coded continuously. Interactions between exposure and age can be accommodated but, for the purpose of these analyses, the situation of no interaction was used.
Estimation of rate advancement periods.
RAPs may be derived from any linear model of the general form
 |
 |
For our analyses, we chose dichotomous risk factors; however, continuous risk factors may also be used. Since our study was a cohort study with person-time data, we applied a Cox proportional hazards model.
With hypertension as an example, from a multivariate model of the simplest form
 |
 |
Estimation of confidence intervals.
Confidence intervals were calculated as described by Brenner et al (1). With the notation introduced above, large sample approximations of 95 percent confidence intervals of RAP were obtained as
 |
 |
). |
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Table 1 shows baseline characteristics of the male cohort grouped by the presence or absence of the risk factors hypertension, cigarette smoking, and dyslipidemia. A clear positive relation between hypertension and dyslipidemia and between smoking and dyslipidemia and a weak negative relation between hypertension and smoking were apparent.
|
The risk of myocardial infarction associated with hypertension, cigarette smoking, and dyslipidemia is shown in table 2 in terms of both hazard rate ratios and rate advancement periods. Results are presented by a multivariate model. Each of the three risk factors was independently associated with a two- to threefold risk of experiencing a nonfatal or fatal myocardial infarction.
|
Estimates of the RAP for hypertension, smoking, and dyslipidemia were 9.4, 11.3, and 14.2 years in models containing one risk factor at a time in addition to age (table 2). These estimates were slightly reduced but still remained substantial after simultaneous adjustment for the other risk factors.
Table 3 lists the RAP of incident coronary heart disease associated with current regular smoking calculated from published coefficients of a variety of other cohort studies among middle-aged men. Most of these studies included a larger number of participants and events, resulting in more narrow confidence intervals for the RAP estimates. All studies adjusted simultaneously for systolic blood pressure and total cholesterol and some for additional risk factors. Across these studies, the RAP estimates of incident coronary heart disease associated with smoking ranged from 6.4 years to 10.5 years.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Rate advancement periods as applied here express the risk associated with the three classical risk factors hypertension, cigarette smoking, and dyslipidemia in terms of the timing of occurrence of myocardial infarction. Literally, an adjusted RAP for hypertension of 8.2 years is an age difference. Hypertensives who have a certain risk or rate of infarction at one age would be expected to have had the same risk or rate of disease at a later age (+8.2 years) had they been normotensive, given that all other covariates are equal and in the absence of competing risks. Furthermore, our results suggest that the average time period by which the age-related increase of risk of myocardial infarction could be postponed among men with dyslipidemia if these men had normal values is more than 11 years. Men smoking regularly (one or more cigarettes per day, i.e., on average nine cigarettes per day in our population) will reach the same infarction risk level as nonsmoking men about 10.5 years earlier.
Like other measures of risk factor impact, the RAP of one risk factor of disease should be carefully controlled for other, potentially confounding risk factors to quantify the net impact of the risk factor of interest. In our example, simultaneous inclusion of all three risk factors reduced the RAP estimates to some extent, which mainly results from the positive association among risk factors (see table 1). Since the addition of covariates into a multivariate model can influence the parameter estimate of both the risk factor of interest and the age variable, both of which are used in the computation of the RAP, prediction of the direction of the resulting change in the RAP estimate is somewhat less straightforward, however, than prediction for measures based on a single parameter, such as the relative risk.
We compared the RAP estimate for cigarette smoking for the men of the MONICA Augsburg cohort with that of other prospective European studies of incident coronary heart disease among middle-aged men (5–7). With various methodological study differences in mind, such as the definition of endpoints, the variable follow-up time or age range, or the statistical adjustment for other risk factors, we found that the RAP estimates for smoking were quite consistent with a range from 6.4 to 10.5 years (median, 9.7 years). In particular, the RAPs estimated from two German occupational cohorts of incident myocardial infarction, 10 years in the Göttingen Risk, Incidence, and Prevalence Study (GRIPS) (5) and 9.3 years in the Münster Heart Study (PROCAM) (6), were remarkably similar to the estimates obtained in our study.
The magnitude of the rate advancement period is inversely related to the strength of the age effect. It is important to realize that the estimate of the RAP depends on the form in which age enters the regression model and on potential interactions with the exposure variable. In the present analysis, age was included as a simple, continuous variable with no interaction with the exposure. This assumes a monotonic increasing risk of myocardial infarction with increasing age, where the relative risk increase with each additional year of age is constant across the entire 45- to 64-year age range. Comparison of other studies that have included age in the same form has shown that the age effect is fairly similar across study populations (8), even if somewhat different inclusion criteria are used.
Another assumption to be considered in the interpretation of the concept of the rate advancement periods is the assumption of the absence of competing risks. In our case of incident myocardial infarction, potential competing risks that will remove the individual from the risk of having a myocardial infarction are essentially limited to deaths not due to myocardial infarction, that is, due to cancer, injuries, suicides, and so on, which are relatively rare in this middle-aged population. In elderly populations, however, where comorbidities and competing risks are much more frequent, rate advancement periods become increasingly theoretical.
Brenner et al. (1) suggest that rate advancement periods may be a useful tool in communicating risk factor impact on disease. Intuitively, translating the complex concepts of risk into the more easily understandable measure of time seems appealing. Literature on risk communication and perception demonstrates that many issues including the format of the information need to be considered (9, 10). For example, Weinstein et al. (11) used time intervals between expected events to convey a difference in risk probabilities in a trial on perceived threat and action intention. Participants perceived more threat and expected to take more action when the absolute risk (of cancer) of 1 in 100,000 was expressed in terms of one case in 35 years (in a city of 200,000 inhabitants) versus one case in 3,500 years (in a small town of 2,000). This effect on perceived threat and action intention was almost as large as that achieved when the absolute risk level was manipulated, that is, increased by a factor of 100. Thus, Weinstein et al. (11) could observe a substantial effect of framing risk levels on the perception of risk.
Using the RAP in risk communication settings such as counseling on risk factor modification poses conceptual challenges that are shared with all other statistical measures of probability. The concept of disease risk (be it as absolute or relative risk) is usually expressed and interpreted by the epidemiologist, physician, nurse, or health educator in an objective manner as the relative frequency of occurrence over time (12). The patient or layperson, however, is interested in the risk to him- or herself and will have a subjective interpretation of a given probability as in the degree of confidence that someone has that an event will occur (12, 13). Most individuals, however, are quite aware of the fact that chronic diseases become more frequent with increasing age. The practical appeal of the RAP may therefore lie in the ability to convey disease risk information as that of premature "aging," that is, disease occurrence, especially since it can be easily adapted to the individual consultation. For example, a RAP of 11 years can be used to explain to a 50-year-old smoker that he has the same risk of myocardial infarction as does a 61-year-old nonsmoker.
In conclusion, we have outlined several issues of both the estimation and the interpretation of the RAP. The practical usefulness of the RAP in terms of risk communication still awaits thorough evaluation. However, its simple calculation and intuitive appeal should make this a worthwhile endeavor.
|
ACKNOWLEDGMENTS |
|---|
The authors acknowledge the contribution of Andrea Schneider and Dr. Jürgen Wellmann for data management and verification, the staff of the MONICA Augsburg 1984–1985 survey including Angela Döring and Jutta Stieber, and the coronary event register team.
|
NOTES |
|---|
Reprint requests to Dr. Angela D. Liese, Department of Epidemiology and Biostatistics, School of Public Health, University of South Carolina, Columbia, SC 29208 (e-mail: aliese{at}sph.sc.edu).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
- Brenner H, Gefeller O, Greenland S. Risk and rate advancement periods as measures of exposure impact on the occurrence of chronic diseases. Epidemiology 1993;4:229–36.[Web of Science][Medline]
-
Keil U, Liese AD, Hense HW, et al. Classical risk factors and their impact on incident non-fatal and fatal myocardial infarction and all-cause mortality in southern Germany. Results of the MONICA Augsburg cohort study 1984–1992. Eur Heart J 1998;19:1197–207.
[Abstract/Free Full Text] -
Tunstall-Pedoe H, Kuulasmaa K, Amouyel P, et al. Myocardial infarction and coronary deaths in the World Health Organization MONICA Project. Registration procedures, event rates, and case-fatality rates in 38 populations from 21 countries in four continents. Circulation 1994;90:583–612.
[Abstract/Free Full Text] - Scheibling U. Darstellung eines neuen Risikomaßes anhand epidemiologischer Literatur zum Thema Risiko koronarer Herzkrankheit durch erhöhtes Cholesterin. (In German). (Dissertation). Ulm, Germany: Department of Epidemiology, University of Ulm, 1995.
- Cremer P, Nagel D, Mann H, et al. Ten-year follow-up results from the Goettingen Risk, Incidence and Prevalence Study (GRIPS). I. Risk factors for myocardial infarction in a cohort of 5790 men. Atherosclerosis 1997;129:221–30.[Web of Science][Medline]
- Assmann G, Schulte H, von Eckardstein A. Hypertri-glyceridemia and elevated lipoprotein(a) are risk factors for major coronary events in middle-aged men. Am J Cardiol 1996;77:1179–84.[Web of Science][Medline]
-
Jousilahti P, Vartiainen E, Tuomilehto J, et al. Sex, age, cardiovascular risk factors, and coronary heart disease. A prospective follow-up study of 14,786 middle-aged men and women in Finland. Circulation 1999;99:1165–72.
[Abstract/Free Full Text] - Dobson AJ, Evans A, Ferrario M, et al. Changes in estimated coronary risk in the 1980s: data from 38 populations in the WHO MONICA Project. Ann Med 1998;30:199–205.[Web of Science][Medline]
- Fischhoff B, Bostrom A, Quadrel MJ. Risk perception and communication. Annu Rev Public Health 1993;14:183–203.[Web of Science][Medline]
- Yamagishi K. When a 12.86% mortality is more dangerous than 24.14%: implications for risk communication. Appl Cognitive Psychol 1997;11:495–506.[Web of Science]
- Weinstein ND, Kolb K, Goldstein BD. Using time intervals between expected events to communicate risk management. Risk Anal 1996;16:305–8.[Web of Science][Medline]
- Walker VR. Direct inference, probability, and a conceptual gulf in risk communication. Risk Anal 1995;15:603–9.
- Jardine CG, Hrudey SE. Mixed messages in risk communication. Risk Anal 1997;17:489–98.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  D. J. Campbell Why do men and women differ in their risk of myocardial infarction? Eur. Heart J., April 1, 2008; 29(7): 835 - 836. [Full Text] [PDF]
|
|
|
|
 |
|
 A. Pfahlberg, O. Gefeller, A. D. Liese, H. Brenner, and H.-W. Hense RE: "ASSESSING THE IMPACT OF CLASSICAL RISK FACTORS ON MYOCARDIAL INFARCTION BY RATE ADVANCEMENT PERIODS" Am. J. Epidemiol., September 1, 2001; 154(5): 486 - 488. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||