Influenza and Cardiovascular Disease

Abstract

There is mounting evidence in support of a significant role for influenza infection in the development of atherosclerosis and the triggering of its complications. Here we review the biologic basis of this relationship, with special emphasis on the pro-inflammatory and pro-thrombotic effects of influenza infection. We also discuss the related epidemiologic findings and discuss in detail the possible causal relationship between influenza and cardiovascular disease.

We appraise the relationship between influenza and coronary heart disease, on the basis of Bradford Hill's criteria of causality. We show that our proposed relationship meets the following criteria: strength of association, consistency, temporal sequence, coherence, biologic plausibility, experimental evidence, and analogy. Further studies are needed to assess whether it meets the criterion of biologic gradient. Specificity is not met, but meeting that criterion is of least importance in the study of multifactorial chronic diseases such as coronary heart disease. These criteria do not yield indisputable evidence for or against cause-and-effect, but they can help researchers appraise available evidence and determine the areas that need further research. The case for expanding the research on the effect of influenza on cardiovascular disease is a strong one, for most of Hill's criteria are met.

Although rarely a cause of death until the early 20th century, coronary heart disease (CHD) assumed epidemic proportions by the mid 1950s. Concurrently, a decrease in deaths related to communicable diseases established CHD as the chief cause of death in the Western world. Today, coronary disease is also on the rise in developing countries. ¹ This transition in the pattern of CHD has led to a shift in epidemiologic approaches to studying the causes of disease in general. Whereas the 19th century was marked by recognition of the role of infection (a single causative factor), scientists were faced by the beginning of the 20th century with noninfectious diseases, which had multiple predisposing factors. In 1961, the Framingham Heart Study ² introduced the term “risk factor.” Extensive research is now under way to identify and reduce CHD risk factors, thereby decreasing the risk of cardiovascular events.

Infection, Inflammation, and Atherosclerosis

The role of infectious agents in atherosclerosis has been recognized for more than a century. William Osler ³ was one of the first to propose a major role for acute infection in the pathogenesis of atherosclerosis. In the early 20th century, a few pioneer scientists used several infectious agents (Salmonella typhi, streptococci, etc.) to induce atherosclerosis in animal models. By the late 1970s, scientists began to study the role of herpesviruses and Chlamydia pneumoniae and, later, of Helicobacter pylori, Mycoplasma pneumoniae, Porphyromonas gingivalis, enterovirus, and a growing list of other agents in atherogenesis (Table I). ^4–11 This effort coincided with the emergence of new evidence pointing to atherosclerosis as an inflammatory disease. ¹² The role of infection in endothelial injury and vascular wall inflammation came under scrutiny. ¹³ Many infectious agents have been investigated in this regard, but none of them has proved to play a causative and specific role. Chlamydia pneumoniae has been studied the most extensively, but the results of large clinical trials of antibiotics against this disease have been largely disappointing. ^14,15

TABLE I. Infectious Agents Implicated in Atherosclerosis

Atherosclerosis, Vulnerable Plaques, and Triggers

Atherosclerosis is a highly prevalent disease, having been present from antiquity and having been observed in ancient mummies. ¹⁶ In 1909, Osler ¹⁷ observed that “it is exceptional to find no patches of arterial degeneration in any body postmortem, and even children might show some slight foci of fatty degeneration.” More systematic data emerged from studies conducted by Enos, ¹⁸ MacNamara, ¹⁹ and colleagues, who observed CHD in young soldiers killed in action in Korea and Vietnam. Major studies have been done to elucidate the factors that convert established, stable atherosclerotic plaques into unstable, life-threatening plaques. ²⁰ Acute coronary syndromes involve the rupture of vulnerable plaques, which are usually nonobstructive and which have inflammatory cell infiltrates, a thin fibrous cap, and a large lipid core. ²¹ When the surfaces of these plaques rupture or erode, a thrombotic event will ensue. The factors that lead to plaque inflammation followed by a myocardial infarction (MI) are not fully understood. Our group has hypothesized that influenza may play a role in some patients. Whereas most other infectious agents result in a chronic, indolent infection, which presumably increases chronic inflammation of the arterial walls, influenza induces notable acute arterial-wall inflammation and may trigger plaque destabilization that leads to acute coronary syndrome.

Epidemiologic Studies

Influenza epidemics are associated with a significant increase in cardiovascular deaths. Selwyn Collins, in 1932, ²² was one of the first researchers to report that for nearly every influenza epidemic there is a peak in deaths related to organic heart disease, which corresponds chronologically to the influenza peak.

Acute MI shows a seasonal variation, having its highest incidence in the winter months. ^23–25 Influenza activity has been suggested as a reason for this winter peak in the MI rate. ^26–28 Glezen and colleagues showed that peaks in influenza activity are followed, 2 weeks later, by a peak in deaths related to ischemic heart disease, hypertension, and cerebrovascular disease. ²⁹ This finding is supported by clinical studies showing that many acute MIs are preceded by an upper respiratory tract infection. ^30–36

Effect on Excess Death

Influenza is a major cause of morbidity and death. Each year, in the United States alone, “the flu” accounts for 110,000 hospitalizations, 1 to 3 billion dollars in direct costs, and 10 to 15 billion dollars in indirect costs. ³⁷ Although earlier estimates cited 22,000 excess influenza-related deaths each year, ³⁸ newer estimates—derived from epidemiologic data—cite 36,000 deaths involving the respiratory and circulatory systems and 51,000 all-cause deaths related to influenza each year in the United States. ³⁹ This increased death is partly due to the aging of the population and to the advent of more virulent viral strains. However, the influenza-related death toll may be even higher: because influenza is not a recognized trigger of MI, it is very unlikely to be recorded on the death certificates of patients who die of MI, stroke, heart failure, or cardiac arrest. Consequently, the number of deaths triggered by influenza is under-recorded. In fact, our estimates—derived from clinical trials and case-control studies—show that, by triggering cardiovascular events, influenza may cause up to 90,000 deaths per year.

Influenza Vaccination Prevents Cardiovascular Events

In 2000, Naghavi and associates ⁴⁰ reported that influenza vaccination has a protective effect in a high-risk population (patients with known CHD). In this case-control study, vaccination against influenza was associated with a 67% reduction in the risk of MI during the subsequent influenza season (odds ratio [OR], 0.33; 95% confidence interval [CI], 0.13–0.82; P = 0.017), but it did not reduce deaths. In a case-control study, Siscovick and coworkers ⁴¹ found that influenza vaccination was associated with a 49% reduction (OR, 0.51; 95% CI, 0.33–0.79) in the risk of an out-of-hospital primary cardiac arrest. Lavallée and colleagues ⁴² found a 50% decrease (OR, 0.50; 95% CI, 0.26–0.94) in the stroke rate for patients vaccinated against influenza, including a 48% risk reduction (OR, 0.42; 95% CI, 0.21–0.81) in those vaccinated during the preceding 5 years. Conversely, in studying survivors of a 1st MI, Jackson and coworkers ⁴³ failed to show that influenza vaccination had any protective effect against recurrent coronary events (adjusted hazard ratio, 1.18; 95% CI, 0.79–1.75). However, this study may have been hampered by a misclassification of the subjects, and exclusion of the early (90-day) period after an acute MI (when complications are most likely to develop) may have limited the researchers' ability to detect a potential benefit.

During 2 influenza seasons, Nichol and associates ⁴⁴ studied community-dwelling elderly subjects (288,238 persons ≥65 years old), in whom vaccination against influenza was associated with a reduced risk of hospitalization for heart disease (reduction, 19%; P < 0.001) and cerebrovascular disease (reduction, 16%–23%; P < 0.018) and a reduced all-cause mortality rate (reduction, 48%–50%; P < 0.001) during influenza seasons. ⁴⁴

In a small, randomized, pilot clinical trial conducted by Gurfinkel and colleagues, ⁴⁵ the mortality rate was 2% for the vaccinated group versus 8% for the control group (relative risk [RR], 0.25; 95% CI, 0.07–0.86; P = 0.01). ⁴⁵ The triple composite endpoint occurred in 11% of the vaccinated patients versus 23% of the control patients (P = 0.009). At 1 year, the incidence of cardiovascular death was significantly lower in the vaccine recipients (6%) than in the control group (17%) (RR, 0.34; 95% CI, 0.17–0.71). ⁴⁶ The triple composite endpoint occurred in 22% of the vaccinated group versus 37% of the control group (P = 0.004). ⁴⁶ The two-year follow-up evaluation of patients who were revaccinated showed a decreased incidence of the combined endpoint of death and MI after revaccination (3.4% vs 9.7%; P = 0.05). ⁴⁷

Mechanisms

Influenza affects the vascular system in multiple ways. We have shown that inoculation of atherosclerotic apolipoprotein-E (apoE)–deficient mice with influenza A results in heavy infiltration of atherosclerotic plaques by inflammatory cells (macrophages and T cells), as well as smooth muscle cell proliferation, fibrin deposition, platelet aggregation, and thrombosis. ⁴⁸ These profound inflammatory and prothrombotic changes mimic those seen in coronary plaques after a fatal MI. ⁴⁹

Influenza is associated with a greatly increased number of proinflammatory, prothrombotic cytokines, and it causes endothelial dysfunction, increased plasma viscosity, tachycardia, and release of endogenous catecholamines. Clinical flu is also associated with psychological distress, dehydration leading to hypotension, and hemoconcentration, hypoxemia, and demand ischemia. Van Lenten and coworkers have shown that influenza decreases the anti-inflammatory properties of high-density lipoprotein cholesterol particles and increases entry of macrophages into the arterial wall. Influenza infection also has extensive and profound procoagulant effects (Table II). ^67,68

TABLE II. Effects of Influenza on the Coagulation System

Viremia after influenza and extrapulmonary seeding of the virus are considered to be rare; ⁶⁹ however, a direct infection of the atherosclerotic plaques and chronic plaque infection cannot be ruled out. Whether or not the virus infects the plaque, influenza may induce an antigenic cross-reactivity, which may result in enhanced immune response against plaque antigens.

In our mice experiments, only the plaques, not the normal arterial segments, were inflamed. In an ecological study, Azambuja and Duncan demonstrated an association between the distribution of the ages at which patients died of influenza and pneumonia during the 1918–19 United States influenza pandemic and the distribution of CHD death from 1920 to 1985 in survivors from the corresponding birth cohorts. ⁷⁰ These researchers suggest that the 1918 influenza pandemic (and probably subsequent flu epidemics) played a role in the CHD epidemic of the 20th century, probably by initiating immune responses related to potential antigenic mimicry.

Is There a Causal Relationship?

Three case-control studies, ^40–42 1 large cohort study, ⁴⁴ and 1 pilot, randomized clinical trial ⁴⁵ have suggested that influenza vaccination has a protective role against CHD (Fig. 1). It is important to clarify, however, whether a true cause-and-effect relationship exists or whether these results were due to chance, a bias, or a simple noncausal association.

Fig. 1 Studies of influenza vaccination and cardiovascular outcomes.

Treacherous philosophical issues regarding the nature of causality have long been recognized and were detailed by Hume in the 18th century. ⁷¹ Various “criteria” have been proposed to help assess possible causal relationships in epidemiologic studies. Although these criteria cannot be considered decisive in differentiating between causal and noncausal associations, they can help support the evidence for a causal relationship instead of one based on chance, bias, or confounding factors. ⁷² These criteria were published in the United States Surgeon General's 1st report on smoking and health (in 1964) and were later improved by Sir Bradford Hill (Table III) and Mervyn Susser. ^73–75 In the following paragraphs, we briefly appraise the relationship between influenza and CHD, on the basis of Hill's criteria of causality.

TABLE III. Hill's Criteria for Causality ⁷⁴

Strength of Association. Strong associations provide better evidence of causality than do weak ones. Influenza is associated with a 2- to 3-fold increase in the risk of an MI. Influenza vaccination is associated with an unusually marked reduction in the cardiovascular outcomes, which ranges from 20% to 70% (50% on average, Fig. 1).

Consistency. Consistency refers to the finding of similar or consistent results by different investigators, using a variety of methods, in diverse studies involving different populations and circumstances. The greater the consistency among the diverse findings, the stronger the evidence for causality. Influenza has been associated with an increased risk of CHD events in multiple previous studies, 23-25,30–36 in addition to 3 case-control studies, ^40–42 1 cohort study, ⁴⁴ and 1 clinical trial, ⁴⁵ which showed that influenza vaccination had a beneficial effect in primary and secondary prevention settings with different outcomes.

Temporal Sequence. This criterion requires that exposure to a disease precede the onset of the disease by a reasonable amount of time. Epidemiologic studies have shown a 7- to 10-day lag in the development of CHD outcomes after influenza in individual patients and a similar pattern in the peaking of CHD after influenza seasonal activity. Similarly, our animal studies have shown that the most significant pathologic changes happen 7 to 10 days after influenza infection in apoE–deficient mice.

Coherence. Coherence suggests that the available information regarding the disease forms a cohesive entity and that the proposed relationship does not contradict or conflict with existing theory and knowledge. Reports concerning an increased incidence of CHD events tend to parallel reports concerning an increase in all-cause deaths associated with influenza. In studies of the effect of vaccination on all-cause mortality, the magnitude of the observed effect fully supports the protection seen in cardiovascular studies. In a meta-analysis of 20 cohort studies, influenza vaccination reduced deaths by 68%; in case-control studies, it reduced deaths from all causes by 30%. ^76,77

Biologic Plausibility. This criterion requires the association in question to be plausible and explainable on the basis of known biologic facts. We have previously described the growing body of evidence of a procoagulant effect of influenza infection.

Biologic Gradient (Dose-Response Relationship). An increase in the level, intensity, duration, or total amount of exposure is associated with a progressive increase in the risk of disease. So far, there has been no evidence to support the theory that multiple bouts or more severe cases of influenza produce more deleterious cardiovascular effects. This question has not been adequately answered, and it merits further research. Indirect supporting evidence may be inferred from the studies of excess death in milder epidemics of influenza. During such epidemics, Simonsen and colleagues documented a lower excess death: an average of 3,200 deaths in 9 seasons of influenza B and A (H1N1). In contrast, during more severe epidemics, the excess death was higher: an average of 7,600 deaths in 11 seasons of influenza A (H3N2). ³⁸ Re-evaluation of the existing data would show whether there was a similar difference in the cardiovascular deaths during those same years. Researchers also need to determine whether multiple infections or more virulent viral strains result in worse cardiovascular outcomes in experimental and clinical cases.

Specificity. This criterion maintains that exposure leads to only 1 disease and that the disease results from only 1 exposure. However, in chronic diseases (especially coronary artery disease, which is multifactorial), this is rarely true. Therefore, an absence of specificity does not rule out a causal relationship.

Experimental Evidence. This criterion holds that a disease may be altered (prevented or ameliorated) by an appropriate experimental regimen (in vitro tests, animal experiments, or clinical trials). Satisfaction of this criterion is not a necessity; in most cases, such proof is elusive and is not even sought. However, in multiple studies showing that influenza vaccination protects against CHD events, this criterion has been met (Fig. 1).

Analogy. Analogy implies a similarity in some respects among things that are otherwise different. According to Hill, this is among the least important of the criteria. The influenza-atherosclerosis relationship is analogous to some other relationships between CHD risk factors and atherosclerosis.

Although these criteria can help researchers better appraise the available evidence, none of them, per se, can yield indisputable evidence for or against the cause-and-effect hypothesis, and none of them (except perhaps temporal sequence) is a sine qua non. Causation should be thought of as a multifactorial web. Interaction among several factors (obesity, smoking, hypertension, dyslipidemia, male sex, etc.) causes coronary artery disease, and each of these factors has its own causal web. Also, many such webs may exist for a given disease. ⁷²

Public Health Aspects

Recognition of the effects of influenza on CHD provides the medical community with a valuable opportunity to further reduce cardiovascular death and morbidity. Influenza vaccination is a safe, inexpensive, and effective method for reducing morbidity in high-risk patients with cardiovascular disease. However, this method is extremely under-utilized.

The influenza vaccine is very safe and well tolerated in all age groups. ⁷⁸ Randomized, double-blind clinical trials have documented a scarcity of systemic symptoms after vaccination. ^79,80 In 1976, after a universal vaccination program was implemented in response to a swine flu scare, a controversial increase in the incidence of Guillain-Barré syndrome (GBS) was reported. ^81,82 However, methodologic questions arose concerning these reports, and GBS was not observed in subsequent studies of other vaccination programs, including the United States Army's mass influenza vaccination program (1980–1988). ^83–85 The Advisory Committee on Immunization Practices ⁷⁸ found no indication that current influenza vaccines are associated with a substantial increase in GBS and stated that, even if the influenza vaccine does pose a risk, it probably involves slightly more than 1 additional case per million persons vaccinated. The Committee concluded that the potential benefits of influenza vaccination outweigh the possible risk of developing vaccine-associated GBS.

Multiple studies have shown that vaccination can reduce both direct medical costs and indirect costs related to absenteeism from work. ^86–88 Influenza vaccination is one of the most cost-effective interventions available and, in selected groups, it may even be cost saving. In fact, vaccination has a far more favorable cost–benefit profile than most cardiovascular prevention methods, such as statins. For this reason, the under-use of influenza vaccination is unacceptable.

Current national guidelines advise that high-risk CHD patients need to be vaccinated against influenza. Also included are children and household contacts who live with patients who have heart disease. However, the main best-practice cardiology guidelines from the American Heart Association, European Society of Cardiology, and American College of Cardiology fail to specify the need for vaccinating CHD patients. ^89–93 In general, cardiologists ignore the impact of influenza on the cardiovascular system and consider immunization a responsibility of primary care physicians. This lack of knowledge and failure to recommend or practice vaccination has led to a very low rate of immunization in cardiovascular patients. The national objectives for influenza vaccination, as stated by “Healthy People 2000” and “Healthy People 2010,” are 60% for adults 18 to 64 years old with high-risk conditions and 90% for persons 65 or older. ⁹⁴ However, according to the 1997–2001 National Health Interview Survey, the prevalence of self-reported influenza vaccination was only 22.7%, 49.2%, and 76.7% in CHD patients aged 18 to 49, 50 to 64, and ≥65 years, respectively. ⁹⁵ This significant shortcoming needs urgent and focused corrective action.

Although extremely important, education of healthcare professionals at the primary and specialty levels is not enough. Community leaders, teachers, churches, and the mass media have an equal responsibility to spread the right message, correct myths, and motivate people to get vaccinated. New research is needed to identify the obstacles to vaccination of CHD patients and to develop programs that are designed to overcome those obstacles. Public health officials must recognize prevailing myths and misinformation and must develop programs to counteract them. The prophylactic use of new neuraminidase inhibitors appears promising for preventing influenza; if this benefit is verified in clinical trials, these agents may find a place in preventing CHD. Besides offering additional protection to elderly persons and those with a weak immune response to vaccine, neuraminidase inhibitors can be especially important in vaccine-virus mis-matches. ⁹⁶

Future Research

The effect of influenza on CHD is just beginning to be recognized, and it warrants extensive research, which may pave the way for new, more efficient methods for preventing and controlling CHD. We have recommended a series of actions that can improve influenza control in CHD patients (Table IV). Basic studies are needed to elucidate the mechanism by which influenza affects the vascular system. It is essential to determine whether influenza directly affects the vascular walls and directly damages the arterial walls. Different anti-inflammatory, anticoagulant, antioxidant, lipid-lowering, immunomodulating, antimacrophage, and antiviral drugs must be tested, in animal studies, for their impact on influenza-induced vascular effects. If effective, these agents must be tested later in human clinical trials. Because of the antigenic similarity of influenza antigens and oxidized low-density-lipoprotein cholesterol epitopes, studies are needed to evaluate potential cross-activation of the immune system against plaques after influenza infections. Various viral strains may behave differently in this regard.

TABLE IV. Recommendations for Improving Influenza Control in Cardiovascular Patients

Clinical trials will determine whether vaccination against influenza can protect against cardiovascular events in high-risk subjects. Ethical issues may preclude the use of a placebo in patients with known CHD and in persons older than 50 years, because these groups are urged to receive the vaccine. However, they may be given prophylactic oseltamivir in randomized, placebo-controlled trials, since this approach would provide sufficient and additional protection against influenza. Younger subjects (< 50 years of age) with a high risk of CHD, on the basis of multiple risk factors or of subclinical CHD, are a large portion of the population and should be studied in clinical trials of influenza vaccine and oseltamivir to reduce their risk of CHD events.

Another interesting approach is to vaccinate the children and household contacts of persons who are at high risk for CHD. Reichert and associates showed that vaccination of Japanese schoolchildren against influenza decreased the incidence and mortality rate of that disease among older persons; moreover, this method attenuated the seasonal variability in mortality rate related to cardiovascular and respiratory disease and also to all causes. ^97,98

Conclusion

Growing evidence suggests that influenza may play a causal role in the development of atherosclerosis and its complications. Further basic, animal, and clinical studies are needed to elaborate the mechanism by which influenza increases the risk of CHD and to determine how vaccines and antiviral agents may protect CHD patients. To increase influenza vaccination of patients at high risk for CHD, an intense public health effort is needed. Moreover, clinical trials are necessary to identify which groups may benefit most from influenza prevention in terms of cardiovascular prevention.

Acknowledgments

The authors wish to thank Drs. Deborah Vela and Navid Sadeghi for their assistance in preparing this manuscript.