Multiple Risk Factor Intervention Trial Research Articles

Estimating the average treatment effect in randomized clinical trials with all-or-none compliance

Noncompliance is a common intercurrent event in randomized clinical trials that raises important questions about analytical objectives and approaches. Motivated by the Multiple Risk Factor Intervention Trial (MRFIT), we consider how to estimate the average treatment effect (ATE) in randomized trials with all-or-none compliance. Confounding is a major challenge in estimating the ATE, and conventional methods for confounding adjustment typically require the assumption of no unmeasured confounders which may be difficult to justify. Using randomized treatment assignment as an instrumental variable, the ATE can be identified in the presence of unmeasured confounders under suitable assumptions, including an assumption that limits the effect-modifying activities of unmeasured confounders. We describe and compare several estimation methods based on different modeling assumptions. Some of these methods are able to incorporate information from auxiliary covariates for improved efficiency without introducing bias. The different methods are compared in a simulation study and applied to the MRFIT.

In Memoriam: Richard H. Grimm, Jr, MD, MPH, PhD (August 22, 1946-September 8, 2022).

In Memoriam: Richard H. Grimm, Jr, MD, MPH, PhD (August 22, 1946-September 8, 2022).

The Impact of Health on Labor Market Outcomes: Evidence from a Large-Scale Health Experiment

While economists have posited that health investments increase earnings, isolating the causal effect of health is challenging due to reverse causality and unobserved heterogeneity. We examine the labor market effects of a randomized controlled trial, the Multiple Risk Factor Intervention Trial (MRFIT), which monitored nearly 13,000 men for over six years. We find that this intervention, which provided a bundle of treatments to reduce coronary heart disease mortality, increased earnings and family income. We find few differences in estimated gains by baseline health and occupation characteristics. (JEL I12, J24, J31)

ASIAN PACIFIC SOCIETY OF NEPHROLOGY CLINICAL PRACTICE GUIDELINE ON DIABETIC KIDNEY DISEASE.

ASIAN PACIFIC SOCIETY OF NEPHROLOGY CLINICAL PRACTICE GUIDELINE ON DIABETIC KIDNEY DISEASE.

Invited Commentary: Learning From Our Past to Build on Our Future- Lessons Learned From MRFIT and JUMBO.

We propose that Blackburn (Am J Epidemiol. 2020;189(6):491-498) ignores several important issues that need to be considered in the context of a historical reflection of the National Heart Institute's landmark study, the Multiple Risk Factor Intervention Trial (MRFIT), and the alternative proposal, the "JUMBO" trial, submitted to the National Heart Institute by an experienced team of extramural investigators but never funded. A key focus of this commentary is to offer an alternative perspective on both studies using our current understanding of the impact of social and structural determinants of health; evidence that policy, systems, and environmental interventions are needed to support behavior change at the individual level; and the significance of examining research from a racial/socioeconomic equity lens. While we strongly agree with Blackburn's conclusion urging the National Institutes of Health to invest in prevention research at a level consistent with its social and economic impact, we encourage the author to move beyond simply underscoring the methodological limitations and failure of the findings of MRFIT compared with the potential of the proposed JUMBO trial to consider the contribution of MRFIT to our current understanding of chronic disease prevention and treatment.

The National Cooperative Trial on Multifactorial Prevention of First Heart Attacks and Coronary Mortality in Free-Living, High-Risk, Middle-Aged American Men, Also Known as JUMBO: A Short Story.

The concept of a multiple risk-factor intervention trial (MRFIT) originated in mid-20th century efforts to determine whether modifying the "risk factors" established by cardiovascular disease epidemiology would prevent heart attacks. The term "MRFIT" probably first appeared in the 1968 report to the National Heart Institute from investigators of the National Diet-Heart Feasibility Study. Based on their pilot experience, they recommended a trial of diet alone. Aware, however, that authorities might agree with the rationale but not the implementation of such a massive and risky undertaking, they also proposed an alternative: whether coronary heart disease was preventable at all by simultaneous intervention on several risk factors; that is, a multiple risk-factor intervention trial. After some years agonizing by serial expert committees, the National Heart Institute opted against an explanatory diet trial and for a pragmatic multiple risk-factor intervention, one designed by Institute staff and operated under contract. Meanwhile, an impatient community of investigators met together in the Makarska Conference, outlined a broad cardiovascular disease prevention policy, and submitted their idealized version of a multiple risk-factor trial, called JUMBO. But the National Heart Institute, because of the plan for its own trial, had no funds left for an investigator-initiated proposal. Hence, this background and story of the MRFIT that wasn't.

Invited Commentary: Can You Do Trials of Behavioral Modification, Such as Diet or Exercise Intervention, With Hard Endpoints?

Blackburn (Am J Epidemiol. 2020;189(6):491-498) describes the evolution from the Makarska Conference in 1968 to the planning of the JUMBO trial and the decision by the National Institutes of Health (NIH) to do the Multiple Risk Factor Intervention Trial. He attributes the failure of the Multiple Risk Factor Intervention Trial to show a significant reduction in coronary heart disease or total mortality to the design of the trial, the multiple centers involved, and continued bureaucracy. The paper is important in raising 3 issues of continued relevance: 1) What is the relationship of the NIH staff in both development and monitoring of large cooperative studies, especially clinical trials, outside of the walls of the NIH? 2) Can clinical trials that utilize individual randomization to intervention and comparison be successful in evaluating a common-source epidemic, such as diet and coronary heart disease, especially in populations with readily available foods and competing advertising, etc.? 3) How much information obtained from, e.g., observational studies, animal models, and genetic studies, is necessary given limits of human population clinical trials of behavioral intervention to develop recommendations for strong public health programs to reduce morbidity and mortality, and what is the measure of successful intervention?

A win ratio approach to the re-analysis of Multiple Risk Factor Intervention Trial.

Composite outcomes, which combine multiple types of clinical events into a single outcome, are common in clinical trials. The usual analysis considers the time to first occurrence of any event in the composite. The major criticisms of such an approach are (1) this implicitly treats the outcomes as if they were of equal importance, but they often vary in terms of clinical relevance and severity, (2) study participants often experience more than one type of event, and (3) often less severe events occur before more severe ones, but the usual analysis disregards any information beyond that first event. A novel approach, referred to as the win ratio, which addresses the aforementioned criticisms of composite outcomes, is illustrated with a re-analysis of data on fatal and non-fatal cardiovascular disease time-to-event outcomes reported for the Multiple Risk Factor Intervention Trial. In this trial, 12,866 participants were randomized to a special intervention group (n = 6428) or a usual care (n = 6438) group. Non-fatal outcomes were ranked by risk of cardiovascular disease death up to 20 years after trial. In one approach, participants in the special intervention and usual care groups were first matched on coronary heart disease risk at baseline and time of enrollment. Each matched pair was categorized as a winner or loser depending on which one experienced a cardiovascular disease death first. If neither died of cardiovascular disease causes, they were evaluated on the most severe non-fatal outcome. This process continued for all the non-fatal outcomes. A second win ratio statistic, obtained from Cox partial likelihood, was also estimated. This statistic provides a valid estimate of the win ratio using multiple events if the marginal and conditional survivor functions of each outcome satisfy proportional hazards. Loss ratio statistics (inverse of win ratios) are compared to hazard ratios from the usual first event analysis. A larger 11-event composite was also considered. For the 7-event cardiovascular disease composite, the previously reported first event analysis based on 581 events in the special intervention group and 652 events in the usual care group yielded a hazard ratio (95% confidence interval) of 0.89 (0.79-0.99), compared to 0.86 (0.77-0.97) and 0.91 (0.81-1.02) for the severity ranked estimates. Results for the 11-event composite also confirmed the findings of the first event analysis. The win ratio analysis was able to leverage information collected past the first experienced event and rank events by severity. The results were similar to and confirmed previously reported traditional first event analysis. The win ratio statistic is a useful adjunct to the traditional first event analysis for trials with composite outcomes.

Liberation From the P Value's Tyranny.

> P-values, next to nuclear weapons, are the worst invention of the 20th century . > > —Herman Rubin, Purdue University, 1981 More than a half century ago, the P value was selected as the measuring tool for health care research by administrators. The difficulties generated by statistical hypothesis testing in addition to the constraints it places on cardiovascular trial researchers must now be declared unacceptable by the cardiology research community. Developed >90 years ago by statistician, Ronald Fisher, while working on a manure experiment in England,1,2 P values have risen to dominate the field of cardiovascular investigation for no other reasons than those of custom and habit. The abuse of P values has become so rampant that the American Statistical Association in 2016 was compelled to provide a consensus statement decrying their misuse.3 Counterpoint, see p 1049 Statistical hypothesis testing and P values were criticized by scientists from the beginning. They thought that its reverse logic—embracing a null hypothesis that one did not think, only to have the data reject it compelling the scientist to accept what one did think—was tortured and unnecessarily complex. P values rose to prominence in health care when grant administrators, journal editors, and Food and Drug Administration officials, overwhelmed by the post–World War II explosion in research, chose the P value instrument to help identify worthy research.4 The P value was not chosen because it had any basis in epidemiology, biology, pathophysiology, or cardiovascular medicine. It was selected because these administrators needed a metric to help them to separate the research wheat from the chaff, and they could see no alternative. The fledgling clinical trial research community accepted this P value imprimatur, not realizing that this nonmalicious selection portended pernicious consequences. Even though the P value’s sole, small role was to …

Abstract 15934: Higher Midlife Serum Levels of Kidney-injury Molecule-1 (KIM-1) are Associated With Incident Fatal CHD Over Very Long-term Follow-up Among Men

Introduction: Higher levels of kidney-injury molecule-1 (KIM-1) measured in urine are associated with presence and progression of acute renal disease. A recent study reported similar results for KIM-1 measured in blood. Hypothesis: We hypothesized that KIM-1 measured in stored serum from middle-aged men who participated in the Multiple Risk Factor Intervention Trial (MRFIT) would differentiate very long-term risk of fatal CHD vs. survival to a mean age of 80 over approximately 30 year follow-up. Methods: We conducted a nested case-control study within MRFIT, which in 1973-76 randomized 12,866 high risk but CVD free men ages 35-57 to risk factor intervention vs. usual care. Serum samples were collected at baseline and stored for future use. The trial concluded in 1982 but long-term mortality follow-up was ascertained through 2005 using the National Death Index. From MRFIT participants with stored serum from baseline, we sampled 100 men who died of CHD (mean age 47.3 at baseline and 73.9 at death), and 100 men who survived to 2005 (mean age =48.4 at baseline and 80.1 in 2005.) KIM-1 was assayed from stored serum samples using high sensitivity single-molecule counting technology (Erenna ® Immunoassay System, Singulex), with limit of detection (LoD)=0.5 pg/ml, and lower limit of quantification (LLoQ)=2.0 pg/ml. Results were compared between cases and controls using Wilcoxon rank tests and logistic regression. Results: Inter-assay %CVs were 8%. Median KIM-1 was higher for smokers vs. non-smokers and for men with vs. without hypertension, but was not associated with high cholesterol. KIM-1 was significantly higher in cases (183 pg/ml (IQR: 137-239) versus controls, (161 pg/ml (IQR:109-212), p=0.03; OR (95%CI)for Q4 versus Q1 was 2.26 (1.02 - 5.02) Adjusted for age and smoking the OR(95%CI) of fatal CHD for Q4 vs. Q1 was 2.34 (1.02- 5.37), and further adjusted for diastolic BP and serum cholesterol at baseline, was 2.0 (95% CI: 0.8-4.7). Conclusions: Higher serum KIM-1 levels at midlife were associated with a ∼2-fold increased risk of fatal CHD vs. survival over ∼30 years of follow-up. This is the first report of a longitudinal association of circulating KIM-1 levels with fatal CHD in long-term follow-up.

Hypertension Editors’ Picks

<i>Hypertension</i> Editors’ Picks

Gout in African Americans

PurposeAfrican Americans have a substantially higher prevalence of risk factors for gout than Caucasians. The aim of the present study was to compare the risk for incident gout among African Americans and Caucasians. MethodsIncidence rates of physician-diagnosed gout among 11,559 Caucasian men and 931 African American men aged 35 to 57 years and at high cardiovascular risk, observed for 7 years as a part of the Multiple Risk Factor Intervention Trial, were analyzed. Cox regression models were used to account for potential confounding by age, body mass index, diuretic use, hypertension and diabetes status, aspirin and alcohol consumption, and kidney disease. ResultsAt baseline, after accounting for risk factors, African Americans had a 14% lower prevalence of hyperuricemia than Caucasians. Incidence of gout increased with increasing prevalence of risk factors in both Caucasians and African Americans. Ethnic disparities in incidence rates were most apparent among those without other risk factors for gout. In separate Cox regression models, after accounting for risk factors, African American ethnicity was associated with a hazard ratio of 0.78 (95% confidence interval [CI], 0.66-0.93) for physician-diagnosed gout and 0.88 (95% CI, 0.85-0.90) for incident hyperuricemia. Significant interactions were observed; the association was the strongest (hazard ratio 0.47; 0.37-0.60). These associations were unaffected by addition of serum urate as a covariate or by using alternate case definitions for gout. ConclusionsAfter accounting for the higher prevalence of risk factors, African American ethnicity is associated with a significantly lower risk for gout and hyperuricemia compared with Caucasian ethnicity.

Hypertension and Obesity and the Risk of Kidney Cancer in 2 Large Cohorts of US Men and Women

Kidney cancer incidence is increasing globally. Reasons for this rise are unclear but could relate to obesity and hypertension. We analyzed longitudinal relationships between hypertension and obesity and kidney cancer incidence in 156 774 participants of the Women's Health Initiative clinical trials and observational studies over 10.8 years. In addition, we examined the effect of blood pressure (BP) on kidney cancer deaths for over 25 years among the 353 340 men screened for the Multiple Risk Factor Intervention Trial (MRFIT). In the Women's Health Initiative, systolic BP (SBP) was categorized in 6 groups from <120 to >160 mm Hg, and body mass index was categorized using standard criteria. In age-adjusted analyses, kidney cancer risk increased across SBP categories (P value for trend <0.0001) and body mass index categories (P value for trend <0.0001). In adjusted Cox proportional hazards models, both SBP levels and body mass index were predictors of kidney cancer. In the MRFIT sample, there were 906 deaths after an average of 25 years of follow-up attributed to kidney cancer among the 353 340 participants aged 35 to 57 years at screening. The risk of death from kidney cancer increased in a dose-response fashion with increasing SBP (hazard ratio, 1.87 for SBP>160 versus <120 mm Hg; 95% confidence interval, 1.38-2.53). Risk was increased among cigarette smokers. Further research is needed to determine the pathophysiologic basis of relationships between both higher BP and the risk of kidney cancer, and whether specific drug therapies for hypertension can reduce kidney cancer risk.

Chronic Kidney Disease and the Risk of Incident Gout Among Middle‐Aged Men: A Seven‐Year Prospective Observational Study

The kidney is the major organ that facilitates excretion of urate in humans. Surprisingly, few studies have assessed whether a reduced glomerular filtration rate (GFR) and/or kidney damage is associated with a higher incidence of gout, and this study was undertaken to address this question. Data from a 7-year followup of patients enrolled in the Multiple Risk Factor Intervention Trial, a primary prevention trial for cardiovascular disease among 12,866 men ages 35-57 years, were used for the present investigation. Presence of gout was determined by the study physicians from the original trial. Chronic kidney disease was defined using criteria similar to those proposed by the National Kidney Foundation. The Cox proportional hazards regression model was used to assess the association between gout and chronic kidney disease, after accounting for the effects of potential confounders. Overall, there were 722 cases of physician- diagnosed incident gout over 76,602 person-years of followup. The standardized incidence ratio of gout among those with chronic kidney disease was 1,217 (95% confidence interval [95% CI] 1,191-1,244). The adjusted hazard ratio (HR) among those with chronic kidney disease was 1.61 (95% CI 1.60-1.61). Each standard deviation decline in the estimated GFR was associated with an HR of 1.43 (95% CI 1.35-1.51). Including the serum urate level, as well as the urate-chronic kidney disease interaction term, as variables in the second analysis did not attenuate the HR. Proteinuria and hematuria, two markers of kidney damage, were associated with an elevated risk of gout independent of the estimated GFR. Chronic kidney disease manifesting as reduced glomerular function or as presence of blood or protein in the urine increases the risk of incident gout.

Chlorthalidone Versus Hydrochlorothiazide

TO THE EDITOR: Dhalla and colleagues state, As typically prescribed, chlorthalidone in older adults was not associated with fewer adverse cardiovascular events or deaths than hydrochlorothiazide. However, it was associated with a greater incidence of electrolyte abnormalities, particularly hypokalemia. This observational cohort study is contrary to a meta-analysis of 108 clinical trials, which indicated that reductions in potassium levels were considered equivalent between chlorthalidone and HCTZ when they were used in the dose range between 12.5 and 25.0 mg. In addition, data from MRFIT (Multiple Risk Factor Intervention Trial) indicate that chlorthalidone significantly reduces CVEs compared with HCTZ (P < 0.001)...

The effect of cholesterol on cancer growth and metastasis: Experimental study with LDL receptor knockout mice.

e22071 Background: Cholesterol is an essential component for normal cell growth. However, epidemiological studies show cholesterol levels may be associated with cancer incidence. The Multiple Risk Factor Intervention Trial reported a significant inverse relationship between cholesterol levels and lung cancer. The effects of cholesterol on cancer growth and metastasis have not been established. The aim of the present study was to determine if cholesterol alters cancer growth or metastasis. Methods: The study used 48 wild-type mice and 18 LDL receptor knockout mice (LDLR-/-). Mice were randomized into 3 groups: Group I (wild-type mice fed normal chow, n=23); Group II (wild-type mice fed high lipid chow, n=25); Group III (LDLR-/- mice fed high lipid chow, n=18). Human melanoma cells were injected subcutaneously for the assessment of growth and intravenously for metastasis. Mice were followed for 4 weeks. At the end of the study, mice were weighed and autopsy was performed. Melanoma weight, metastasis, and blood cholesterol levels were evaluated. Data was analyzed using One-way ANOVA and Chi-square. Results: LDL-cholesterol level was significantly higher in Group III than Group I (1103±472 vs.28±6; p&lt;0.01). There was no significant difference in melanoma weight (2.74g ±1.77, 2.98g±1.98 vs. 2.18g±1.59, P=0.56) among the three subcutaneous injection groups. Lung metastasis was observed in all intravenous injection groups (11/11 vs. 10/10 vs. 8/8). There was no difference in melanoma weight or metastasis between male and female mice (P&gt;0.05). There was no difference in net weight gain among the subcutaneous injection groups (4.12g±2.50 vs. 2.64g±2.79 vs. 3.66g±1.96, P=0.30). However, among the intravenous injection groups, weight gain was observed in Group II (2.78g±2.67) and Group III (2.36g±1.73) and weight loss in Group I (-0.32g±1.34, P&lt;0.003). Conclusions: Cholesterol has no effect on melanoma growth or metastasis in mice. A high cholesterol diet may reduce weight loss in animals with metastatic melanoma. Further study is warranted to determine whether similar results exist in other types of cancers and to evaluate a possible beneficial effect of high lipid diets on patients with metastatic cancers.

Phase II Clinical Research Design in Cardiology

Clinical trials demonstrate the best of medical expertise and epidemiological elegance. From the simple building blocks of contemporaneous control groups, randomization, and blinding, they assemble a clear picture of the nature of the treatment-effect relationship. This accomplishment has earned them the star ascendant position in cardiovascular research. Their advantage was demonstrated with Bradford Hill’s work on streptomycin, and as knowledge of the pathogenesis of atherosclerotic disease produced possibilities for new treatments, cardiovascular researchers applied this new research tool to identify effective therapies in a sequential approach (Figure 1). Figure 1. The traditional phases of clinical trials. *The study takes place before Food and Drug Administration approval. This work accelerated with clinical trials demonstrating treatment benefits for chronic diseases such as hypertension,1,2 lipid abnormalities,3,4 and heart failure,5,6 cementing their role in identifying new therapies to prevent the sequela of cardiovascular disease in vulnerable populations. However, the limitations of early clinical trial interpretation also appeared. The findings of the Multiple Risk Factor Intervention Trial (MRFIT),7 the International Verapamil SR/Trandolapril Study (INVEST),8,9 the Early Versus Late Intervention Trial With Estradiol (ELITE),7,10 the Cardiac Arrhythmia Suppression Trial (CAST),11,12 and the US Carvedilol program controversy13–19 together served to undermine the confidence of cardiologists in the interpretation of clinical trial results. Thought leaders in the field identified the interpretation problem (ie, the hyperreliance on P values to reflect positive results for any analysis in a clinical trial) and called for a fundamental change in the design and analysis of clinical trials. Subsequent work produced the following clinical trials principles: (1)There should be clear and prospective declaration in the written protocol about all important aspects of the clinical trial with particular emphasis on the end points …

Scientists seek new ways to prevent cancer

Scientists seek new ways to prevent cancer

Chlorthalidone Versus Hydrochlorothiazide

See related article, pp 1001–1007 Chlorthalidone (CTD) is the best diuretic for the treatment of hypertension, both in blood pressure–lowering efficacy and, most importantly, in prevention of hypertension-related morbidity and mortality. Despite this fact, hydrochlorothiazide (HCTZ) has been and continues to be by far the most frequently prescribed diuretic in the United States.1 Even among patients with resistant hypertension, where a more potent diuretic is especially needed, CTD is chosen in only 3% of patients.2 There are numerous possible reasons for this flagrant failure of evidence-based medicine. I will refer to the Aesop tale of “The Tortoise and the Hare” as my way of explaining the paradox. In this reference, physicians are the tortoises, whereas a pharmaceutical company is the hare; here, unlike in the original, the hare easily wins the race. The 2 diuretics were approved for the treatment of hypertension within a year of each other, HCTZ in 1959 and CTD in 1960. Soon thereafter, small trials documented the equal efficacy of CTD in much smaller doses in lowering blood pressure compared with HCTZ3 or other thiazides.4 A few years later, the Multiple Risk Factor Intervention Trial (MRFIT) was begun, and the designers offered the 15 participating clinics the choice of using either HCTZ or CTD; 9 chose HCTZ and 6 chose …

Long-Term Effects of Chlorthalidone Versus Hydrochlorothiazide on Electrocardiographic Left Ventricular Hypertrophy in the Multiple Risk Factor Intervention Trial

Chlorthalidone (CTD) reduces 24-hour blood pressure more effectively than hydrochlorothiazide (HCTZ), but whether this influences electrocardiographic left ventricular hypertrophy is uncertain. One source of comparative data is the Multiple Risk Factor Intervention Trial, which randomly assigned 8012 hypertensive men to special intervention (SI) or usual care. SI participants could use CTD or HCTZ initially; previous analyses have grouped clinics by their main diuretic used (C-clinics: CTD; H-clinics: HCTZ). After 48 months, SI participants receiving HCTZ were recommended to switch to CTD, in part because higher mortality was observed for SI compared with usual care participants in H-clinics, whereas the opposite was found in C-clinics. In this analysis, we examined change in continuous measures of electrocardiographic left ventricular hypertrophy using both an ecological analysis by previously reported C- or H-clinic groupings and an individual participant analysis where use of CTD or HCTZ by SI participants was considered and updated annually. Through 48 months, differences between SI and usual care in left ventricular hypertrophy were larger for C-clinics compared with H-clinics (Sokolow-Lyon: -93.9 versus -54.9 μV, P=0.049; Cornell voltage: -68.1 versus -35.9 μV, P=0.019; Cornell voltage product: -4.6 versus -2.2 μV/ms, P=0.071; left ventricular mass: -4.4 versus -2.8 g, P=0.002). At the individual participant level, Sokolow-Lyon and left ventricular mass were significantly lower for SI men receiving CTD compared with HCTZ through 48 months and 84 months of follow-up. Our findings on left ventricular hypertrophy support the idea that greater blood pressure reduction with CTD than HCTZ may have led to differences in mortality observed in the Multiple Risk Factor Intervention Trial.