Experience with a cross-study endpoint review committee for AIDS clinical trials. Terry Beirn Community Programs for Clinical Research on AIDS.

OBJECTIVES: To describe the methods and results of a standardized system for clinical endpoint determination for defining and reviewing endpoints in clinical trials for HIV-infected individuals. DESIGN: A system was developed utilizing standard definitions for the 24 diagnoses or clinical events that serve as trial endpoints and together define the combined endpoint 'progression of HIV disease. A common set of case report forms were used for all trials. Thus, an event of Pneumocystis carinii pneumonia (PCP), for example, for a subject co-enrolled in an antiretroviral trial and a PCP prophylaxis trial was only reported once. METHODS: A central committee was established to define clinical events and review endpoints across all studies. Events were classified according to established criteria for confirmed, probable and possible levels of certainty. RESULTS: This report describes the methods used to ascertain and review endpoints, and summarized 2299 clinical events for 8097 subjects enrolled in one or more of nine clinical trials. Data on the diagnostic certainty of events and agreement between site clinicians and the endpoint committee are presented. CONCLUSIONS: Uniform classification of endpoints across AIDS clinical trials can be accomplished by multicenter, multitrial organizations with standardized definitions and review of endpoint documentation. Our experience suggests that nurse coordinators reviewing all submitted endpoints for every trial are warranted and the need for external review by a clinical events committee may depend on the type of trial conducted.

Method issues in dietary data analyses in the Multiple Risk Factor Intervention Trial.

The selection process in the Multiple Risk Factor Intervention Trial caused relations between risk factors to differ between participants who were randomly assigned into the study and the screening population. Cigarette smoking, blood pressure, and serum cholesterol were moderately inversely related to each other in the randomly assigned population whereas these relations in the unselected population were direct and small in magnitude. This problem was addressed by covariate adjustment in analyses. The selection process also created an artificially high initial mean concentration of serum cholesterol; the mean plasma concentration at the second screening was 15 mg/dL lower than at the first screening. Most of this difference is attributable to regression to the mean. To account for this problem, emphasis was placed on change in plasma cholesterol over time, calculated from the second-screening measurement. Examination of the reliability of nutrition data based on one 24-h dietary recall showed that nutrient-biochemical relations are subject to considerable regression-dilution bias. The ratios of "within" to "between" components of variability were typically between one and four. Analyses in which multiple follow-up measures were averaged are emphasized in this monograph. Men assigned to the special intervention group reported considerable reductions in total energy intake, which was not consistent with observed weight loss. The most likely explanation for this is underreporting or underconsumption the day before the recall. To partially adjust for this, nutrient data are often expressed both in absolute units and as nutrient densities.

Relation of dietary carbohydrates to blood lipids in the special intervention and usual care groups in the Multiple Risk Factor Intervention Trial.

This chapter explores relations between reported intake of dietary carbohydrates and measurements of plasma lipids at baseline and during trial years 1-6 of the Multiple Risk Factor Intervention Trial. With control for dietary lipids, alcohol, and other factors, total carbohydrate intake at baseline was inversely related to baseline plasma total cholesterol and high-density-lipoprotein (HDL) cholesterol; starch and other simple carbohydrates were unrelated to plasma lipids and sucrose was inversely related to HDL cholesterol. During trial years 1-6, men assigned to the special intervention group increased their intake of starch and other simple carbohydrates as they decreased their fat intake, and lowered their intakes of refined and processed sucrose. Total carbohydrate intake of these men was inversely related to total, low-density-lipoprotein (LDL), and HDL cholesterol. Starch and sucrose intakes were also inversely related to HDL cholesterol. In contrast, intake of other simple carbohydrates was directly related to HDL, and inversely related to plasma total and LDL cholesterol. For men in the highest quintile of intake of other simple carbohydrates compared with men in the lowest quintile, plasma total cholesterol was lower by 3.6 mg/dL, LDL cholesterol was lower by 4.3 mg/dL, and HDL cholesterol was higher by 1.6 mg/dL. Findings were generally similar for men in the usual care group.

Relation of dietary fiber to blood lipids in the special intervention and usual care groups in the Multiple Risk Factor Intervention Trial.

This chapter addresses relations between intake of fiber--total, soluble, and insoluble--and blood lipids in the Multiple Risk Factor Intervention Trial through use of baseline data (single measurement), averages of four to five 24-h recalls and blood lipid determinations collected during annual follow-up examinations, and change from baseline to follow-up. No significant associations were observed at baseline. Consistent highly significant inverse associations were seen in analyses of follow-up measurements. Results from change data were of intermediate strength and consistency. These variations were in all likelihood due to the low reliability of a single 24-h recall at baseline for determination of dietary intake and change in intake for individuals. From follow-up data, plasma total and low-density-lipoprotein (LDL) cholesterol concentrations were lower by approximately 5 mg/dL for men in the special intervention group in quintile 5 of total fiber intake (25 g/d) compared with men in quintile 1 (8 g/d), after adjustment for average body mass index and intake of alcohol, saturated and polyunsaturated fatty acids, and dietary cholesterol. Results were similar for men in the usual care group. There were no adverse effects on high-density-lipoprotein cholesterol, nor any consistent associations with plasma triglycerides. Thus, increasing dietary fiber can provide additional reduction in blood total and LDL cholesterol and consequent improvement in the lipid profile, over and above the beneficial effects of a fat-modified diet.

Food group and nutrient intakes at baseline in the Multiple Risk Factor Intervention Trial.

This chapter relates food and nutrient intakes at baseline to other facets of reported dietary behavior, major risk factors, and sociodemographic characteristics of men in the Multiple Risk Factor Intervention Trial. Intakes of total fat (38.4% of energy), saturated fatty acids (14.2%), and dietary cholesterol (492 mg/d) were similar to amounts seen in the first and second National Health and Nutrition Examination Surveys in the 1970s and were generally lower than findings from studies in the 1960s. There were inverse relations between total serum cholesterol and intakes of total fat, saturated and monounsaturated fatty acids, and dietary cholesterol. These paradoxical associations were largely attributable to findings in the 21% of men who reported following a special diet, indicating that use of such a diet increases with severity of hypercholesterolemia. Fat intake was directly related to number of meals per week eaten away from home, and to cigarette smoking. Patterns of food and nutrient intake were similar for men stratified by baseline blood pressure and antihypertensive treatment. Intake of total energy and percentages from various dietary fats decreased with age, as did use of sucrose and caffeine. White men consumed more dairy products than did other ethnic groups, whereas black men consumed more eggs, sugars, and sweets. Asians had the highest intake of cereal foods. Those with more education ate less high-fat meat products, more fruit, and more polyunsaturated oils, but also more high-fat dairy products and less breads and cereals; they also drank more alcohol.

Food group and macronutrient intakes, trial years 1-6, in the special intervention and usual care groups in the Multiple Risk Factor Intervention Trial.

This chapter presents changes in dietary intake reported by men in the special intervention (SI) and usual care (UC) groups from baseline through 6 y of follow-up in the Multiple Risk Factor Intervention Trial. Changes in nutrients by SI men after 1 y of following the intensive intervention program were as follows: reduced intake of total fat (from 38.4% to 34.3% of energy), saturated fatty acids (14.2% to 10.4% of energy), and cholesterol (448 to 263 mg/d), and increased intake of polyunsaturated fatty acids (from 6.4% to 8.6% of energy). These changes were maintained and did not increase through the remaining 5 y. UC men reported small changes in similar directions. Most of the change in saturated fatty acid intake by SI participants was from high-fat meat and high- and medium-fat dairy products. Reduction in dietary cholesterol was achieved primarily by substantial decreases in intake of eggs and high-fat meats. Several baseline factors were associated with amount of dietary change in SI men. Greater changes were seen in men with higher baseline serum cholesterol concentrations, in those not consuming a special diet, in nonsmokers followed by lighter smokers, in hypertensive than in non-hypertensive men, in older participants, in white than in black men, in moderate drinkers than in nondrinkers or those consuming > or = 22 drinks/wk, and in those with no "stressful life events" than in those reporting one or more life events.

Dietary intake in the Multiple Risk Factor Intervention Trial (MRFIT): nutrient and food group changes over 6 years.

The Multiple Risk Factor Intervention Trial (MRFIT) was a randomized clinical trial in the primary prevention of coronary heart disease. Middle-aged men determined to be at high risk for coronary heart disease were randomized into either a special intervention (SI) group or a group referred to usual sources of medical care (UC). Twenty-four hour dietary recall data were used to monitor the nutrient intake of the MRFIT population and guide the nutrition education program for the SI group. The SI group of participants decreased intake of dietary cholesterol by 40% and saturated fatty acids by more than one-fourth and increased intake of polyunsaturated fatty acids by one-third. Evaluation of SI dietary intake data by food groups indicates that some dietary changes were relatively easy to implement, whereas others presented more of a challenge. Changes made with relative ease included increasing the consumption of fish and poultry, skim and low-fat milk, polyunsaturated margarines and oils, fruits, and low-fat breads and cereals and reducing the consumption of egg yolks. More difficult changes included eliminating, or even reducing, the intake of high-fat beef and pork, high-fat cheeses, high-fat crackers, snacks, and desserts, and increasing the intake of vegetarian meat alternatives.

Comparison of a computerized and a manual method of food coding for nutrient intake studies.

The Dietary Data Collection (DDC) microcomputer system is currently being developed as a tool for the standardized and detailed collection of dietary intake data for human nutrition research studies. The system operates interactively, soliciting all necessary information on menu selection screens to ensure user entry of complete food descriptions and quantity information. The descriptive data are then automatically converted to food codes and gram weights for subsequent calculation of nutrient content. At the completion of the first phase of system development, a preliminary test was performed to compare the amount of time required to enter food intake data into the DDC system with the amount of time required to accomplish the same food coding task manually. Test subjects consisted of four experienced food coders and one coder trainee. Using a crossover design, each coder manually coded 16 1-day food records and entered another 16 records into the DDC system for automatic coding. Four of the five coders took significantly less time to code and enter descriptive dietary intake information using the DDC system than they took for manual coding and data entry. Time savings ranged between 9% and 44% among the test subjects.

Impact of measurement error and temporal variability on the estimation of event probabilities for risk factor intervention trials.

The impact of measurement error and temporal variability of risk factors on estimates of disease probabilities based on the logistic function is discussed. Monte Carlo results and empirical findings from the Multiple Risk Factor Intervention Trial indicate that the degree of attenuation of logistic parameter estimates is well approximated by the reliability coefficient when the errors are assumed to be normal random variates and event probabilities are small. In the design of intervention studies, measurement error and temporal variability of risk factors do not usually influence estimates of the probability of developing the disease in the control group, but can result in mis-estimation of the probability of developing the disease in the experimental group, substantially reducing the statistical power of the clinical trial.

A comparison between carboxyhemoglobin and serum thiocyanate determinations as indicators of cigarette smoking.

Cigarette smoking histories were compared to carboxyhemoglobin and serum thiocyanate concentrations obtained from 426 smokers and 191 non-smokers. The mean levels of both carboxyhemoglobin and serum thiocyanate wefe significantly higher among cigarette smokers and correlated with number of cigarettes smoked per day. The specificity of both procedures was 81 per cent, and serum thiocyanate had a higher sensitivity (93 per cent vs. 83 per cent), making it potentially more suitable for use as an index of cigarette smoking.

The MLC assay as a test for HLA-D region compatibility between patients and unrelated donors: results of a national marrow donor program involving multiple centers.

Mixed lymphocyte culture (MLC) assays from 763 patients and prospective unrelated marrow transplant donors identified through the National Marrow Donor Program (NMDP) were analyzed for their overall utility in measuring HLA-D region compatibility. The assays were performed at 31 different transplant centers (range, 1 to 197; median, 9 assays per center). A total of 325/763 (42.6%) of the tests were judged to be uninterpretable, either due to lack of sufficient control cells included in the assay (89 tests) or to insufficient reactivity by patient and/or donor cells (236 tests). Among the 438 tests that could be interpreted, HLA-Dw phenotyping with HLA-D homozygous cells was performed for a subset of 190. The relative response (RR) values from these 190 tests, however, were not clearly separable into distinct populations; i.e., RR values corresponding to Dw identity versus nonidentity between patient and donor could not be reliably discriminated. The predictive value of a nonreactive MLC for Dw identity was calculated to be 0.91 for RRs of < or = 20%, while the predictive value of a reactive MLC for Dw nonidentity was 0.35 for RRs of > 20%. These results, based on an analysis of data submitted from multiple transplant centers testing patients who had a variety of hematologic disorders, suggest that the MLC assay is a relatively imprecise method for determining HLA-D region compatibility between patient and prospective unrelated marrow donor.

Noninvasive tracing of Krebs cycle metabolism in liver.

To quantify intrahepatic Krebs cycle metabolism, phenyl acetate, excreted in urine as a glutamine conjugate, was given to healthy subjects infused with [3-14C]lactate. They were studied after 60 h of fasting and when given glucose after an overnight fast. Distributions of 14C in glutamate from urinary phenylacetylglutamine and blood glucose were determined. Corrections to the distributions because of the fixation of 14CO2 formed from the [3-14C]lactate were determined by administering [14C]bicarbonate. Comparisons of distributions in glucose and glutamate support the assumption that the glutamate distributions reflect those in hepatic alpha-ketoglutarate. From the distributions in glutamate, the extent of exchange of labeled with unlabeled carbons and relative flow rates in the cycle in liver were estimated. Dilution of 14C by 12C in the cycle was found in the fasted but not the fed state. In the fasted state, pyruvate carboxylation was estimated to be at least twice the rate of Krebs cycle flux and the rate of pyruvate's decarboxylation less than 1/25 the rate of its carboxylation. In the fed state, the rate of decarboxylation was estimated to be between one-sixth and one-half the rate of carboxylation. The rate of conversion of oxalacetate to fumarate in both states appeared to be greater than 6 times the rate of Krebs cycle flux.

Quality control in the MRFIT local screening and clinic laboratory.

Quality control was emphasized in the screening and clinic laboratory during the early stages of the trial to cover all aspects of screening, and in the later stages to cover local testing used to monitor the participant in the clinic and to guarantee collection of valid specimens to be shipped to the Central Laboratory. Special attention throughout the trial was focused on techniques for collection of specimens, since it was recognized that analytical results could not be better than the quality of the specimens. Training courses, on-site visits, and newsletters were used to sensitize the staff of the clinic laboratory about the necessity to follow protocol. Any monitoring evidence of carelessness, use of deteriorated supplies, failure to follow safety rules, and deviation from directions in the MRFIT Manual of Operations resulted in memoranda from the Coordinating Center or the Central Laboratory.

Test of effect of lipid lowering by diet on cardiovascular risk. The Minnesota Coronary Survey.

The Minnesota Coronary Survey was a 4.5-year, open enrollment, single end-time double-blind, randomized clinical trial that was conducted in six Minnesota state mental hospitals and one nursing home. It involved 4393 institutionalized men and 4664 institutionalized women. The trial compared the effects of a 39% fat control diet (18% saturated fat, 5% polyunsaturated fat, 16% monounsaturated fat, 446 mg dietary cholesterol per day) with a 38% fat treatment diet (9% saturated fat, 15% polyunsaturated fat, 14% monounsaturated fat, 166 mg dietary cholesterol per day) on serum cholesterol levels and the incidence of myocardial infarctions, sudden deaths, and all-cause mortality. The mean duration of time on the diets was 384 days, with 1568 subjects consuming the diet for over 2 years. The mean serum cholesterol level in the pre-admission period was 207 mg/dl, falling to 175 mg/dl in the treatment group and 203 mg/dl in the control group. For the entire study population, no differences between the treatment and control groups were observed for cardiovascular events, cardiovascular deaths, or total mortality. A favorable trend for all these end-points occurred in some younger age groups.

A method for quantitating the contributions of the pathways of acetoacetate formation and its application to diabetic ketosis in vivo.

A method has been developed for estimating in the intact cell the contribution of deacylation of acetoacetyl-CoA to the formation of acetoacetate relative to acetoacetate's formation via hydroxymethylglutaryl (HMG)-CoA. Estimates depend upon the fraction of the terminal four carbons of an even carbon-containing fatty acid that are converted to acetoacetate without prior conversion to acetyl-CoA, since in the formation of acetoacetate via HMG-CoA the omega-2 and omega-3 carbons of the fatty acid are converted to acetyl-CoA. Incorporation of 14C from [16-14C]palmitic acid into carbon 2 relative to carbon 4 of acetoacetate is used as the measure of the formation of the acetoacetate from the omega and omega-1 carbons of the fatty acid without acetyl-CoA as an intermediate. Incorporation of 14C from [13-14C]palmitic acid into carbon 1 relative to carbon 3 of acetoacetate is the measure of the formation of acetoacetate from the omega-2 and omega-3 carbons without acetyl-CoA as an intermediate. Comparison of these incorporations is made with incorporation into the carbons of acetoacetate of 14C from palmitic acid labeled with 14C in any of its first 12 carbons since such incorporation must proceed via acetyl-CoA as an intermediate. In an application of this approach, the specifically 14C-labeled palmitic acids were injected into rats in diabetic ketosis. Hydroxybutyric acid that each rat excreted was isolated and degraded. From the ratios of incorporation into the carbons of the hydroxybutyrates, as a minimum, 11% of the total quantity of hydroxybutyrate excreted by the rats was formed from acetoacetyl-CoA without HMG-CoA as an intermediate.

Weight loss as a predictor of survival and disease progression in HIV infection. Terry Beirn Community Programs for Clinical Research on AIDS.

Severe weight loss in HIV is associated with decreased length of survival. It is unclear whether mild weight loss is associated with an increased risk of death or opportunistic complications of HIV. Participants in four interventional studies (n = 2382) conducted by a community-based clinical trials network were evaluated for percentage change in weight during their first 4 months in the study. Proportional hazards models were performed for the occurrence of opportunistic complications and death subsequent to the 4-month visit. The relative risk of death and opportunistic complications for those with 5% to 10% weight loss over 4 months was 2.22 (p < .001) and 1.89 (p < .001), respectively, and 1.26 (p < .01) and 1.19 (p < .01) among those who lost 0% to 5% of their body weight, respectively, when compared with those with no weight loss. Among those who lost 5% to 10% of their body weight, the relative risk of individual opportunistic complications increased significantly, including Pneumocystis carinii pneumonia (PCP) (1.61; p < .01), cytomegalovirus (CMV) (2.33; p < .001), and Mycobacterium avium complex (MAC) (1.81; p < .01). As little as 5%t weight loss over a 4-month period is associated with increased risk of death and opportunistic complications in HIV. A weight loss of 5% to 10% is also associated with an increased risk of individual opportunistic complications.

HLA class I and II antigen phenotypes of North American Indians from Minnesota: implications for marrow transplants using unrelated donors.

As a result of an appeal for a bone marrow donor for a North American Indian (Native American) patient, 261 Native Americans from our community were typed for HLA-A,B,DR antigens, and 51 were typed for HLA-A,B antigens only. The HLA antigen frequencies of the Native Americans were compared with those of 12,881 white bone marrow donors and were found to differ markedly. To investigate the implications these differences in HLA antigen frequencies would have for the location of unrelated bone marrow donors, the HLA types of 12 Native American bone marrow transplant patients from our institution were used to search among 5389 HLA-A,B,DR-typed white donors in the National Marrow Donor Program file and the file of 261 HLA-A,B,DR-typed Native American donors. In the white donor file, at least two donors were found that matched at all HLA-A,B,DR antigen loci of one Native American patient (8%). Using the Native American donor file, which was less than one-twentieth the size of the white donor file, and HLA-A,B,DR-matched donor was also found for one (8%) of the patients. These results suggest that although donors for nonwhites can be identified in a file of HLA-typed white volunteers, the probability of finding a suitably matched donor for such individuals is enhanced if donors representing racial or ethnic minorities are included in unrelated donor registries.

Pathways of acetoacetate's formation in liver and kidney.

Specifically 14C-labeled palmitic acids were perfused through livers and incubated with slices of kidneys from rats in diabetic ketosis. The distribution of 14C in the hydroxybutyric acid formed was determined. In liver, the ratio of incorporation of 14C from [13-14C]palmitic acid into carbon 1 to carbon 3 of the hydroxybutyric acid was the same as the ratio in carbon 2 to carbon 4 from [6-14C]palmitic acid. In kidney, the carbon 1-to-carbon 3 ratio was more than twice the carbon 2-to-carbon 4 ratio. In both tissues, 14C from [16-14C] palmitic acid was preferentially incorporated into carbon 4 compared to carbon 2 of the hydroxybutyric acid, but more so in liver than kidney. These results mean that in liver, the sole pathway of acetoacetate formation is via hydroxymethylglutaryl-CoA, while in kidney it is not. Rather in kidney, acetoacetyl-CoA is converted to acetoacetate to a large extent by direct deacylation, presumably via a transferase- and/or deacylase-catalyzed reaction. In liver, most of the palmitic acid utilized is converted to acetoacetate while in kidney it is not. We previously estimated that, as a minimum, 11% of the hydroxybutyric acid excreted by the rat in diabetic ketosis is formed without hydroxymethylglutaryl-CoA as an intermediate. The kidney appears to be the source of this hydroxybutyric acid if the pathways operative in these tissues in vitro are those that also operate in vivo.