1	 Exclusion of elderly persons in randomized clinical trials of drugs for 2	 ischemic heart disease
3	 4	 Running title: Exclusion of elderly persons in trials 5	 6	 Authors: Florence T. Bourgeois MD, MPH 1, 2, 3, Liat Orenstein 3, Sarita Ballakur 4, Kenneth 7	 D. Mandl MD, MPH 1, 2, 3, and John P. A. Ioannidis MD, DSc 5, 6 8	 9	 10	 11	 12	 Corresponding Author: 13	 Florence T Bourgeois MD, MPH 14	 Computational Health Informatics Program 15	 Boston Children’s Hospital 16	 300 Longwood Avenue 17	 Boston, MA 02115 18	 Phone: 617-355-6624 19	 Email: florence.bourgeois@childrens.harvard.edu 20	 21	 Author Affiliations: 22	 ¹ Division of Emergency Medicine, Boston Children’s Hospital, Boston, MA 23	 ² Department of Pediatrics, Harvard Medical School, Boston, MA 24	 3 Computational Health Informatics Program, Boston Children’s Hospital, Boston, MA 25	 4 Faculty of Arts and Sciences, Brown University, Providence, RI 26	 5Stanford Prevention Research Center, Department of Medicine, Stanford University 27	 School of Medicine, Stanford, CA 28	 6 Department of Health Research and Policy, Stanford University School of Medicine, 29	 Stanford, CA 30	 31	 Abstract word count: 263 32	 Main text word count: 2999 33	 Tables: 2 34	 Figures: 3 35	 36	 37	 Funding/Support: Drs. Bourgeois, Ioannidis and Mandl were supported by a grant from 38	 the National Institute on Aging (1R21AG043715), National Institutes of Health. 39	 40	 41	 42	 43	
	 1	

1	 ABSTRACT 2	 Background/Objectives. Evidence-based and patient-centered care of elderly adults 3	 with ischemic heart disease requires clinical trials that study the patient population seen 4	 by clinicians in practice. We aimed to measure the exclusion of elderly persons from 5	 randomized trials studying drug interventions for ischemic heart disease and describe 6	 the characteristics of these trials. 7	 Design and setting. Interventional clinical trials studying a drug intervention for 8	 ischemic heart disease and started 2006 and after were identified in ClinicalTrials.gov. 9	 Data were extracted on study features, including age-based inclusion criteria. Data on 10	 participants and their age distribution were collected from trial publications, investigator 11	 inquiry, and result data in ClinicalTrials.gov. 12	 Measurements. Proportion of trials excluding patients based on age, mean age of trial 13	 participants, and proportion of enrolled participants who were ≥65 years and ≥75 years. 14	 Results. Of 839 identified trials, 446 (53%) explicitly excluded elderly persons. The 15	 most frequent upper age limits were 80 (n=164) and 75 (n=114) with a median upper 16	 age limit of 80 (IQR 75, 80). Trials with upper age limit exclusions tended to be smaller 17	 (median number of participants 100 vs 201, p<0.001) and were more likely to be funded 18	 primarily by non-industry sources (78.3% vs 70.0%, p=0.006). The overall mean age of 19	 trial participants was 62.7 years with a mean maximum age of 74 years. The estimated 20	 proportion of participants that was ≥65 years was 42.5% and the estimated proportion 21	 that was ≥75 years was 12.3%.
	 2	

1	 Conclusion. Despite the high burden of ischemic heart disease among elderly patients, 2	 the majority of drug trials do not enroll participants reflective of age-related prevalence 3	 of the disease. 4	 Key Words: ischemic heart disease, evidence-based medicine, research methodology 5	 6	 7	 8	 9	 10	 11	 12	 13	 14	 15	 16	 17	 18	 19	 20	 21	 22	 23	
	 3	

1	 INTRODUCTION 2	 Patients 65 years of age and older comprise only 14% of the US population, but bear a 3	 large and disproportionate amount of the healthcare burden.1,2 Greater than 60% of 4	 patients with cancer, for example, and nearly 65% of patients hospitalized with heart 5	 disease are older than 65 years.3,4 Overall, this age group consumes greater than a 6	 third of total U.S. personal health care expenses every year as well as 30% of all 7	 prescription drug costs.2 However, there is strong evidence that elderly patients are 8	 persistently excluded or underrepresented in clinical trials across a spectrum of 9	 conditions including osteoarthritis, diabetes, and various cancer types.5-7 As many as 10	 half of all clinical trials have explicit upper age limitations while others limit the 11	 participation of older patients based on indirect exclusion criteria such as the presence 12	 of comorbid conditions, cognitive impairment, or polypharmacy.5,8-10 13	 14	 Ischemic heart disease represents a leading and increasing healthcare burden among 15	 elderly patients worldwide, resulting in greater than 900,000 deaths annually in the U.S. 16	 alone.11,12 Many more adults live with chronic symptoms and disability related to angina 17	 pectoris, nonfatal myocardial infarction, and ischemic heart failure.11 Age-specific 18	 pathophysiological processes and unique features of disease management in the 19	 elderly require treatment decisions that are carefully tailored to this population.13,14 A 20	 number of trials examining cardiovascular disease over the past decade have 21	 specifically focused on the study of elderly persons, but many trials—including those 22	 influencing current treatment guidelines—study primarily younger patients.13,15,16 23	 Quantifying this gap and identifying the types of trials most likely to exclude elderly
	 4	

1	 patients may inform policies addressing research needs. Our aim was to measure the 2	 extent to which elderly persons are excluded from randomized trials studying drug 3	 interventions for ischemic heart disease and describe the characteristics of these trials. 4	 5	 METHODS 6	 Search strategy and inclusion criteria 7	 Clinical trials were selected from ClincialTrials.gov, a publicly accessible, web-based 8	 registry administered by the National Institutes of Health. As a result of federal 9	 legislations and journal editor policies mandating trial registration prior to patient 10	 enrollment, prospective trial registration has become standard practice.17,18 11	 ClinicalTrials.gov represents the most comprehensive trial registry and currently holds 12	 data on approximately 218,000 trials, making it a powerful tool for the examination of 13	 research activity and practices. 14	 15	 We searched ClinicalTrials.gov on a single day (January 15th, 2016) for interventional 16	 trials addressing ischemic heart disease using the term “ischemic heart disease” in the 17	 condition field of the registry’s advanced search form. Studies were considered eligible 18	 if they were randomized controlled trials assessing drugs with a start date on or after 19	 January 1st, 2006. Trials that were terminated or withdrawn were excluded, as were 20	 trials specific to pediatric and adolescent populations, pregnancy and reproduction, or 21	 sexually transmitted diseases.19 All trials were manually reviewed to ensure that they 22	 studied the treatment or prevention of ischemic heart disease and that they met all 23	 inclusion criteria.
	 5	

1	 2	 Collection of participant age data 3	 Publications describing trial results were sought for each of the trials in the study cohort. 4	 We first searched the “publication” section of ClinicalTrials.gov and, if a publication was 5	 not identified in the trial entry, Medline was searched via PubMed without language 6	 restrictions using a systematic protocol.20,21 We considered a trial published if we 7	 identified a peer-reviewed publication reporting on trial results. If multiple publications 8	 were identified for a single trial, we reviewed each one to identify relevant participant 9	 age data. 10	 11	 If a publication was not identified, or if it did not contain all the required data for our 12	 analysis, we sought to contact the investigators. Email addresses for investigators were 13	 identified in ClinicalTrials.gov, in prior publications, and via online searches. A standard 14	 email was sent to investigators with one reminder sent two weeks later if no response 15	 was received. For industry sponsored trials, if no investigator email address was 16	 available or if no response was received after the total of four weeks, we attempted to 17	 contact the pharmaceutical company via a company email address or online form on the 18	 company website. Finally, if we were unable to identify a publication or to obtain all 19	 relevant data for a trial, we searched for results within the ClinicalTrials.gov database. 20	 21	 Data extraction for trial characteristics 22	 Data on trial characteristics were extracted from ClinicalTrials.gov, including year of 23	 registration, study start and completion dates, primary funding source, trial phase,
	 6	

1	 primary outcome, sample size, participant age eligibility criteria, and characteristics of 2	 the study design (e.g. comparator type, number of centers; Supplementary File S1). The 3	 sample size listed in the registry entry initially represents the anticipated enrollment 4	 number and can be updated by investigators to reflect the actual enrollment figure at the 5	 end of the study. Since this field is not consistently updated and we examined studies 6	 in various stages of conduct, we use sample size to refer to anticipated or actual 7	 enrollment. 8	 9	 For each publication matched to a trial, we recorded the total number of participants 10	 enrolled and extracted data on the age distribution of participants, including minimum 11	 and maximum ages, mean (with standard deviations [SD]) or median ages (with 12	 interquartile ranges [IQR]), and number of participants above specific age thresholds. 13	 Data was collected for the entire study population as well as for each of the study arms, 14	 whenever available. 15	 16	 Statistical Analysis 17	 Descriptive statistics were used to present the proportion of trials excluding elderly 18	 persons, the distribution of upper age limits, and the characteristics of the trials with and 19	 without upper age limits. The maximum ages of patients actually enrolled were 20	 estimated, whenever this information was not available, as the mean age + 2 standard 21	 deviations (SD, Supplementary File S2)5. Univariate analysis with χ" tests and the 222	 sided t tests or Mann-Whitney tests were used for categorical and continuous variables, 23	 respectively, to examine the association of trial characteristics with the explicit exclusion
	 7	

1	 of elderly persons. We performed a logistic regression to test for trends over time in the 2	 prevalence of elderly exclusion, with upper age limit as the dependent variable and year 3	 of study registration as the independent variable. We also performed a multivariate 4	 logistic regression with all variables found to be significant in the univariate tests to 5	 further examine factors predicting exclusion of elderly persons. Both forward and 6	 backward stepwise procedures were used, and the Hosmer-Lemeshow goodness of fit 7	 test was considered. Two-sided P values of <0.05 were considered statistically 8	 significant. 9	 10	 The mean age for the entire enrolled trial population was calculated as a weighted 11	 mean based on the sample size of each trial, and a combined SD was calculated based 12	 on the mean age of the entire cohort and the sample size, standard deviation and mean 13	 age of each trial.22 Different formulas were used for these calculations according to the 14	 sample sizes (Supplementary File S2).22,23 15	 16	 The proportions of participants above specific age thresholds enrolled in the trials were 17	 calculated using age distribution information and the cumulative distribution function 18	 (Supplementary File S2). For these calculations we assumed that the mean age and SD 19	 were derived from a truncated normal distribution, as each trial has either a lower age 20	 limit for the enrolled population (whenever not mentioned, a lower age limit of 18 was 21	 assumed) or lower and upper age limits.24-26 All statistical analyses were conducted in 22	 IBM SPSS Statistics version 23.0. 23	
	 8	

1	 RESULTS 2	 Characteristics of trials and exclusion of elderly persons 3	 We identified 3777 registered interventional trials studying ischemic heart disease, of 4	 which 839 met our inclusion criteria and represented randomized controlled trials of 5	 drug interventions (Figure 1). The majority of these trials was funded primarily by non6	 industry sponsors (74%) and conducted at single centers (60%) (Table 1). Over two7	 thirds (68%) were phase 3 or 4 trials and 55% included a safety outcome. Only 23% 8	 had sample sizes of 500 or more subjects, with a median of 120 subjects per trial. The 9	 most common specific drug classes studied were anti-thrombotic agents (N=355 trials), 10	 lipid-modifying agents (N=132 trials), and antihypertensive medications (N=67 trials). 11	 12	 A total of 446 trials (53%) explicitly excluded elderly persons. The most frequent upper 13	 age limits were 80 (n=164) and 75 (n=114) with a median upper age limit of 80 (IQR 75, 14	 80) (Supplementary Figure S1). Overall, 43% (n=361) of trials had an upper age limit of 15	 80 or even lower. Among trials that did not have such an upper age limit, and for which 16	 we could identify or impute the maximum age of participants (n=200), 17% (n=34) had 17	 enrolled no patients who were over 80 years old. Logistic regression found a small but 18	 significant increase in the proportion of trials that excluded elderly persons over the 19	 study period (odds ratio 1.06 per year, 95% confidence interval [CI] 1.01 to 1.12; 20	 P=0.016) (Figure 2). When we limited the analysis to 2008 through 2015 (in order to 21	 address potential biases in early adoption of ClicialTrials.gov), there continued to be a 22	 significant trend (odds ratio 1.07 per year, 95%CI 1.003 to 1.15; P=0.04) 23	
	 9	

1	 Several trial characteristics were associated with trials that excluded elderly persons 2	 (Table 1). Trials with upper age limit exclusions were more likely to be funded primarily 3	 by non-industry sources, tended to be smaller, did not employ a multicenter design, and 4	 were shorter in duration. An additional analysis was performed focusing on trials that 5	 enrolled only patients less than 80 years of age (Supplementary Table S1). Study 6	 phase, sample size, and study duration were associated with this age restriction. 7	 8	 In multivariate analysis, study phase, sample, size, multicenter status, and study 9	 duration remained significantly associated with the use of any upper age limits (Table 10	 2). 11	 12	 Age of trial participants 13	 Publications describing trial results were identified for 308 (37%) of the trials. Based on 14	 participant age data in these publications and using additional age information obtained 15	 from investigators and from the results posted in ClinicalTrials.gov, we identified 16	 participant age data for a total of 340 trials, representing 327,672 participants. For 334 17	 of these trials, we identified data on the mean age and standard deviation (or median 18	 and interquartile range) of participants. For 46 of the trials we identified information on 19	 proportion of patients aged 65 or older available (either through a publication, a 20	 ClinicalTrials.gov entry, or provided by the investigator), and we imputed this 21	 information for the remainder, as described above. There were similar numbers of trials 22	 with and without upper age eligibility criteria that had participant age data available (161 23	 and 179 trials, respectively).
	 10	

1	 2	 The overall mean age of participants in the trials was 62.7 years with a mean maximum 3	 age of 74 years. The estimated proportion of participants that was 65 years and older 4	 was 42.5% and the estimated proportion that was 75 years and older was 12.3%. 5	 Figure 3 demonstrates the distribution of estimated proportion of patients who were 65 6	 years or 75 years and older for each of the trials. 7	 8	 DISCUSSION 9	 Our results suggest that clinical drug trials across the spectrum of ischemic heart 10	 disease are not adequately enrolling elderly patients, with greater than half the trials 11	 explicitly excluding patients based on upper age limits. Overall, there has been a slight 12	 increase in such exclusions over the past 10 years, in part due to changes in certain 13	 trial characteristics. Upper age limits were most frequently set at 75 and 80 years of 14	 age, resulting in a pronounced drop in the enrollment of participants 75 years of age 15	 and older. 16	 17	 The mean age of participants across trials for ischemic heart disease was 62.7 years, 18	 which is lower than the mean age reported for patients seeking medical care for 19	 ischemic heart disease. In a population-based surveillance study conducted in Olmsted 20	 County, Minnesota between 2005 and 2010, the mean age of 1244 patients hospitalized 21	 with acute coronary syndrome was 67.7 years.27 Another large community-based study 22	 in the U. S. reported a mean age of 69 years among 46,086 patients hospitalized with 23	 myocardial infarction between 1999 and 2008.28 In Sweden, in a national registry
	 11	

1	 comprised of almost 200,000 patients admitted to coronary care units for symptoms 2	 suggestive of acute coronary syndromes between 1996 and 2008, the mean patient age 3	 was 71 years.29 The mean age in population studies is even about a decade higher for 4	 patients with heart failure, a common complication of coronary artery disease.30-32 5	 6	 The proportion of participants enrolled in clinical trials that were 65 years and older and 7	 75 years and older was 42.5% and 12.3%, respectively. When compared to the age 8	 distribution of patients with ischemic heart disease, these rates appear low, particularly 9	 for those 75 years of age and older, and less so for those 65 years or older. For 10	 instance, based on the U.S. National Health Interview Survey conducted in 2012, 55.7% 11	 of patients with coronary heart disease were 65 years and older.33 Another study based 12	 on a national sample of greater than 260,000 patients undergoing percutaneous 13	 coronary intervention for acute myocardial infarction in the U.S. from 2001 through 14	 2009, found that 45.9% of patients were 65 years of age and older.34 These rates are 15	 not much higher than the rate we estimated for the trial participants. Conversely, the 16	 divergence from population data becomes much more prominent at the 75 year 17	 threshold. For example, in a community study in Olmsted County, among 2816 patients 18	 hospitalized for myocardial infarction between 1987 and 2006, 39% were 75 years of 19	 age or older.35 Similarly, an analysis of 1.3 million hospitalizations for acute coronary 20	 syndrome in Canada between 1994 and 2006 revealed that 40% of patients were older 21	 than 75 years.36 22	
	 12	

1	 Our results are consistent with prior studies assessing the representation of elderly in 2	 clinical trials related to ischemic heart disease. One study of 194 randomized controlled 3	 trials cited in American Heart Association guidelines for acute coronary syndromes 4	 found that the mean age of participants was 61 years, with 18% of trials attaining a 5	 mean participant age of 65 years or older and none with mean ages of 75 years or 6	 older.16 Another study examining 11 phase III clinical trials of patients with non-ST7	 segment-elevation acute coronary syndrome between 1994 and 2010, revealed that 8	 53% of participants were 65 years or older and only 20% were 75 years or older.15 Our 9	 study adds to these findings by examining a large cohort of 839 randomized controlled 10	 trials conducted over the past 10 years and addressing the wide spectrum of ischemic 11	 heart disease. We provide new data indicating that the elderly continue to be excluded 12	 from the most rigorous clinical studies—randomized controlled trials—and that this 13	 practice has the greatest impact on the clinical evidence available to guide care for 14	 patients 75 years of age and older. 15	 16	 Our study also demonstrates a registry-based approach for the ongoing assessment 17	 and monitoring of elderly patients across a large number of trials. Mandatory 18	 registration of interventional trials has led to an almost comprehensive dataset of clinical 19	 trials in ClinicalTrials.gov, which can be used to examine participant enrollment 20	 practices, including age-based exclusion. Ongoing monitoring would help inform 21	 guidance and requirements for the enrolment and study of elderly patients. 22	
	 13	

1	 Results derived from trials on younger patients cannot necessarily be extrapolated to 2	 the elderly because of age-related differences in pathophysiology and unique 3	 considerations in defining patient-centric therapeutic targets. 3,37 In addition, the 4	 pharmacologic characteristics and toxicity of certain medications are age-specific, with 5	 the tolerability and effectiveness reduced among older patients.13,38,39 On the other 6	 hand, claimed age differences in treatment effectiveness and/or harms may represent 7	 spurious findings of subgroup differences that are prone to false-positive results.40 In the 8	 absence of sufficient evidence on elderly participants one runs the risk of missing 9	 important age-treatment interactions as well as of falsely claiming such interactions. 10	 11	 Recommendations have been issued by regulatory agencies to encourage the study of 12	 pharmaceuticals in elderly persons.41,42 A guideline developed by the FDA in 1989, 13	 states that “there is no good basis for the exclusion of patients on the basis of age 14	 alone, or because of the presence of any concomitant illness or medication, unless 15	 there is reason to believe that the concomitant illness or medication will endanger the 16	 patient or lead to confusion in interpreting the results of the study”.41 The International 17	 Conference of Harmonization also published a guideline in 1993 recommending the 18	 inclusion of elderly patients in clinical drugs trials that are likely to be relevant in this 19	 population.42 By contrast, the U.S. Congress in the National Institutes of Health (NIH) 20	 Revitalization Act of 1993 included provisions for the inclusion of women and minorities 21	 in clinical trials, but unfortunately did not include requirements for older patients.43 The 22	 NIH should consider requiring that investigators outline plans to recruit elderly patients
	 14	

1	 and provide targets for the proportions to be recruited in certain age groups, in line with 2	 disease distribution in the general population.44 3	 4	 One of the limitations of this study is that we examined the exclusion of elderly patients 5	 only due to explicit age criteria and many more elderly patients may be excluded due to 6	 indirect exclusions such as the presence of co-morbid conditions or functional 7	 impairments.9,45 We could indirectly document the additional impact of such exclusions, 8	 because we found that many trials did not have any participants over 80 years old, even 9	 though they did not include an upper age eligibility criterion. Our results may therefore 10	 underestimate the overall exclusion of elderly patients. In addition, because data in 11	 ClinicalTrials.gov are provided by investigators, the information cannot be verified and 12	 include occasional incomplete data elements. However, data entered in the registry 13	 undergo quality assurance measures by ClinicalTrials.gov prior to public posting and we 14	 were able to obtain additional data as needed through direct contact with investigators. 15	 Nonetheless, we were not able to obtain all participant age information from the registry 16	 or publications, and certain imputations were required. We performed several quality 17	 control checks to examine the procedures and these indicated that the methods and 18	 results were highly reliable (see Supplementary File 2). 19	 20	 In conclusion, despite the high burden of ischemic heart disease among elderly 21	 patients, the majority of drug trials for this condition continues to exclude elderly 22	 patients, with patients 75 years of age and older particularly underrepresented. Current
	 15	

1	 trial practices and policies around the inclusion of elderly patients should be reexamined 2	 in order to ensure the applicability of trial findings to this growing population. 3	 4	 5	 ACKNOWLEDGEMENT 6	 7	 Conflicts of Interest: None of the authors has any conflicts of interest to disclose. The 8	 authors have no financial relationships relevant to this article to disclose 9	 10	 Author contributions: Bourgeois, Mandl, Ioannidis: study concept and design. 11	 Bourgeois, Orenstein, Ballakur, Mandl, Ioannidis: analysis and interpretation of data. 12	 Bourgeois, Orenstein: initial draft of manuscript. Bourgeois, Orenstein, Ballakur, Mandl, 13	 Ioannidis: critical revision and final approval of manuscript. 14	 15	 Sponsor’s role: The sponsor had no role in the design, conduct, writing, or decision to 16	 publish this study. 17	 18	 19	 SUPPLEMENTARY INFORMATION 20	 21	 Supplementary File S1. Coding and Definitions of Trial Characteristics 22	 Supplementary File S2. Statistical Analysis 23	 Supplementary Figure S1. Distribution of Upper Age Limits 24	 Supplementary Table S1. Characteristics of randomized controlled trials studying drug 25	 interventions for ischemic heart disease by presence of an upper age limit lower than 80 years 26	 27	 28	 29	 30	 31	 32	 33	 34	 35	 36	 37	 38	 39	 40	 41	 42	 43	 44	
	 16	

1	 REFERENCES 2	 3	 1.	 U.S.	Census	Bureau.	United	States	Census	Population	Estimates.	2016;	 4	 https://www.census.gov/popest/data/.	Accessed	June	7,	2016.	 5	 2.	 Lassman	D,	Hartman	M,	Washington	B,	Andrews	K,	Catlin	A.	US	health	spending	 6	 trends	by	age	and	gender:	selected	years	2002-10.	Health	Aff	(Millwood).	 7	 2014;33(5):815-822.	 8	 3.	 Hutchins	LF,	Unger	JM,	Crowley	JJ,	Coltman	CA,	Jr.,	Albain	KS.	Underrepresentation	 9	 of	patients	65	years	of	age	or	older	in	cancer-treatment	trials.	The	New	England	 10	 journal	of	medicine.	1999;341(27):2061-2067.	 11	 4.	 Hall	MJ,	DeFrances	CJ,	Williams	SN,	Golosinskiy	A,	Schwartzman	A.	National	Hospital	 12	 Discharge	Survey:	2007	summary.	Natl	Health	Stat	Report.	2010(29):1-20,	24.	 13	 5.	 Liberopoulos	G,	Trikalinos	NA,	Ioannidis	JP.	The	elderly	were	under-represented	in	 14	 osteoarthritis	clinical	trials.	J	Clin	Epidemiol.	2009;62(11):1218-1223.	 15	 6.	 Cruz-Jentoft	AJ,	Carpena-Ruiz	M,	Montero-Errasquin	B,	Sanchez-Castellano	C,	 16	 Sanchez-Garcia	E.	Exclusion	of	older	adults	from	ongoing	clinical	trials	about	type	2	 17	 diabetes	mellitus.	Journal	of	the	American	Geriatrics	Society.	2013;61(5):734-738.	 18	 7.	 Sacher	AG,	Le	LW,	Leighl	NB,	Coate	LE.	Elderly	patients	with	advanced	NSCLC	in	 19	 phase	III	clinical	trials:	are	the	elderly	excluded	from	practice-changing	trials	in	 20	 advanced	NSCLC?	J	Thorac	Oncol.	2013;8(3):366-368.	 21	 8.	 Bayer	A,	Tadd	W.	Unjustified	exclusion	of	elderly	people	from	studies	submitted	to	 22	 research	ethics	committee	for	approval:	descriptive	study.	BMJ.	 23	 2000;321(7267):992-993.	 24	 9.	 Cherubini	A,	Oristrell	J,	Pla	X,	et	al.	The	persistent	exclusion	of	older	patients	from	 25	 ongoing	clinical	trials	regarding	heart	failure.	Archives	of	internal	medicine.	 26	 2011;171(6):550-556.	 27	 10.	 Van	Spall	HG,	Toren	A,	Kiss	A,	Fowler	RA.	Eligibility	criteria	of	randomized	 28	 controlled	trials	published	in	high-impact	general	medical	journals:	a	systematic	 29	 sampling	review.	Jama.	2007;297(11):1233-1240.	 30	 11.	 Moran	AE,	Forouzanfar	MH,	Roth	GA,	et	al.	The	global	burden	of	ischemic	heart	 31	 disease	in	1990	and	2010:	the	Global	Burden	of	Disease	2010	study.	Circulation.	 32	 2014;129(14):1493-1501.	 33	 12.	 Mozaffarian	D,	Benjamin	EJ,	Go	AS,	et	al.	Heart	disease	and	stroke	statistics--2015	 34	 update:	a	report	from	the	American	Heart	Association.	Circulation.	 35	 2015;131(4):e29-322.	 36	 13.	 Lazzarini	V,	Mentz	RJ,	Fiuzat	M,	Metra	M,	O'Connor	CM.	Heart	failure	in	elderly	 37	 patients:	distinctive	features	and	unresolved	issues.	Eur	J	Heart	Fail.	 38	 2013;15(7):717-723.	 39	 14.	 Vetrano	DL,	Lattanzio	F,	Martone	AM,	et	al.	Treating	heart	failure	in	older	and	oldest	 40	 old	patients.	Curr	Pharm	Des.	2015;21(13):1659-1664.	 41	 15.	 Kragholm	K,	Goldstein	SA,	Yang	Q,	et	al.	Trends	in	Enrollment,	Clinical	 42	 Characteristics,	Treatment,	and	Outcomes	According	to	Age	in	Non-ST-Segment43	 Elevation	Acute	Coronary	Syndromes	Clinical	Trials.	Circulation.	 44	 2016;133(16):1560-1573.	
	 17	

1	 16.	 2	 3	 4	 17.	 5	 6	 7	 18.	 8	 9	 19.	 10	 11	 20.	 12	 13	 21.	 14	 15	 22.	 16	 17	 23.	 18	 19	 24.	 20	 21	 22	 25.	 23	 24	 25	 26.	 26	 27	 27.	 28	 29	 28.	 30	 31	 32	 29.	 33	 34	 35	 36	 37	 38	 30.	 39	 40	 41	 31.	 42	 43	 44	 32.	 45	 46	

Sardar	MR,	Badri	M,	Prince	CT,	Seltzer	J,	Kowey	PR.	Underrepresentation	of	women,	 elderly	patients,	and	racial	minorities	in	the	randomized	trials	used	for	 cardiovascular	guidelines.	JAMA	Intern	Med.	2014;174(11):1868-1870.	 DeAngelis	C,	Drazen	JM,	Frizelle	FA,	et	al.	Clinical	trial	registration:	a	statement	from	 the	International	Committee	of	Medical	Journal	Editors.	Med	J	Aust.	 2004;181(6):293-294.	 U.S.	Food	and	Drug	Administration	Amendments	Act	of	2007.	Public	Law	No.	105185	2007.	 Bugeja	G,	Kumar	A,	Banerjee	AK.	Exclusion	of	elderly	people	from	clinical	research:	a	 descriptive	study	of	published	reports.	BMJ.	1997;315(7115):1059.	 Bourgeois	FT,	Murthy	S,	Mandl	KD.	Outcome	reporting	among	drug	trials	registered	 in	ClinicalTrials.gov.	Ann	Intern	Med.	2010;153(3):158-166.	 Ross	JS,	Tse	T,	Zarin	DA,	Xu	H,	Zhou	L,	Krumholz	HM.	Publication	of	NIH	funded	 trials	registered	in	ClinicalTrials.gov:	cross	sectional	analysis.	BMJ.	2012;344:d7292.	 Higgins	JPT	GS,	ed	Cochrane	Handbook	for	Systematic	Reviews	of	Interventions	 Version	5.1.0	[updated	March	2011].	The	Cochrane	Collaboration;	2011.	 Hozo	SP,	Djulbegovic	B,	Hozo	I.	Estimating	the	mean	and	variance	from	the	median,	 range,	and	the	size	of	a	sample.	BMC	Med	Res	Methodol.	2005;5:13.	 Cohen	A.	Estimating	the	Mean	and	Variance	of	Normal	Populations	from	Singly	 Truncated	and	Doubly	Truncated	Samples.	The	Annals	of	Mathematical	Statistics.	 1950;21(4):557-569.	 Konrat	C,	Boutron	I,	Trinquart	L,	Auleley	GR,	Ricordeau	P,	Ravaud	P.	 Underrepresentation	of	elderly	people	in	randomised	controlled	trials.	The	example	 of	trials	of	4	widely	prescribed	drugs.	PloS	one.	2012;7(3):e33559.	 Rossi	P,	Allenby	G,	McCulloch	R.	Bayesian	Statistics	and	Marketing.	John	Wiley	&	 Sons;	2012.	 Manemann	SM,	Gerber	Y,	Chamberlain	AM,	et	al.	Acute	coronary	syndromes	in	the	 community.	Mayo	Clin	Proc.	2015;90(5):597-605.	 Yeh	RW,	Sidney	S,	Chandra	M,	Sorel	M,	Selby	JV,	Go	AS.	Population	trends	in	the	 incidence	and	outcomes	of	acute	myocardial	infarction.	The	New	England	journal	of	 medicine.	2010;362(23):2155-2165.	 Desta	L,	Jernberg	T,	Lofman	I,	et	al.	Incidence,	temporal	trends,	and	prognostic	 impact	of	heart	failure	complicating	acute	myocardial	infarction.	The	SWEDEHEART	 Registry	(Swedish	Web-System	for	Enhancement	and	Development	of	EvidenceBased	Care	in	Heart	Disease	Evaluated	According	to	Recommended	Therapies):	a	 study	of	199,851	patients	admitted	with	index	acute	myocardial	infarctions,	1996	to	 2008.	JACC	Heart	Fail.	2015;3(3):234-242.	 Gabet	A,	Juilliere	Y,	Lamarche-Vadel	A,	Vernay	M,	Olie	V.	National	trends	in	rate	of	 patients	hospitalized	for	heart	failure	and	heart	failure	mortality	in	France,	20002012.	Eur	J	Heart	Fail.	2015;17(6):583-590.	 Gerber	Y,	Weston	SA,	Berardi	C,	et	al.	Contemporary	trends	in	heart	failure	with	 reduced	and	preserved	ejection	fraction	after	myocardial	infarction:	a	community	 study.	Am	J	Epidemiol.	2013;178(8):1272-1280.	 Schmidt	M,	Ulrichsen	SP,	Pedersen	L,	Botker	HE,	Sorensen	HT.	Thirty-year	trends	in	 heart	failure	hospitalization	and	mortality	rates	and	the	prognostic	impact	of	comorbidity:	a	Danish	nationwide	cohort	study.	Eur	J	Heart	Fail.	2016;18(5):490-499.	

	 18	

1	 33.	 2	 3	 34.	 4	 5	 6	 7	 35.	 8	 9	 36.	 10	 11	 37.	 12	 13	 14	 15	 38.	 16	 17	 18	 19	 20	 39.	 21	 22	 40.	 23	 24	 41.	 25	 26	 27	 28	 42.	 29	 30	 31	 32	 33	 43.	 34	 35	 44.	 36	 37	 45.	 38	 39	 40	 41	 42	 43	 44	 45	

Blackwell	DL,	Lucas	JW,	Clarke	TC.	Summary	health	statistics	for	U.S.	adults:	national	 health	interview	survey,	2012.	Vital	Health	Stat	10.	2014(260):1-161.	 Afana	M,	Brinjikji	W,	Cloft	H,	Salka	S.	Hospitalization	costs	for	acute	myocardial	 infarction	patients	treated	with	percutaneous	coronary	intervention	in	the	United	 States	are	substantially	higher	than	Medicare	payments.	Clin	Cardiol.	2015;38(1):1319.	 Roger	VL,	Weston	SA,	Gerber	Y,	et	al.	Trends	in	incidence,	severity,	and	outcome	of	 hospitalized	myocardial	infarction.	Circulation.	2010;121(7):863-869.	 Ko	DT,	Newman	AM,	Alter	DA,	et	al.	Secular	trends	in	acute	coronary	syndrome	 hospitalization	from	1994	to	2005.	Can	J	Cardiol.	2010;26(3):129-134.	 O'Hare	AM,	Kaufman	JS,	Covinsky	KE,	Landefeld	CS,	McFarland	LV,	Larson	EB.	 Current	guidelines	for	using	angiotensin-converting	enzyme	inhibitors	and	 angiotensin	II-receptor	antagonists	in	chronic	kidney	disease:	is	the	evidence	base	 relevant	to	older	adults?	Ann	Intern	Med.	2009;150(10):717-724.	 Crivellari	D,	Bonetti	M,	Castiglione-Gertsch	M,	et	al.	Burdens	and	benefits	of	adjuvant	 cyclophosphamide,	methotrexate,	and	fluorouracil	and	tamoxifen	for	elderly	 patients	with	breast	cancer:	the	International	Breast	Cancer	Study	Group	Trial	VII.	 Journal	of	clinical	oncology	:	official	journal	of	the	American	Society	of	Clinical	 Oncology.	2000;18(7):1412-1422.	 Dieppe	P,	Bartlett	C,	Davey	P,	Doyal	L,	Ebrahim	S.	Balancing	benefits	and	harms:	the	 example	of	non-steroidal	anti-inflammatory	drugs.	BMJ.	2004;329(7456):31-34.	 Ioannidis	JP.	Why	most	published	research	findings	are	false.	PLoS	medicine.	 2005;2(8):e124.	 Food	and	Drud	Administration.	Guidelines	for	the	study	of	drugs	likely	to	be	used	in	 the	elderly.	1989;	 http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformatio n/Guidances/ucm072048.pdf.	Accessed	November	30,	2011.	 International	Conference	on	Harmonisation	of	Technical	Requirements	for	 Regulation	of	Pharmaceuticals	for	Human	Use.	Studies	in	support	of	special	 populations:	Geriatrics	E7.	1993;	 http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Efficacy /E7/Step4/E7_Guideline.pdf.	Accessed	September	1,	2015.	 National	Institutes	of	Health.	National	Institutes	of	Health	Revitalization	Act	of	1993	 Subtitle	B--Clinical	Research	Equity	Regarding	Women	and	Minorities.	1993.	 Kitzman	DW,	Rich	MW.	Age	disparities	in	heart	failure	research.	Jama.	 2010;304(17):1950-1951.	 Heiat	A,	Gross	CP,	Krumholz	HM.	Representation	of	the	elderly,	women,	and	 minorities	in	heart	failure	clinical	trials.	Archives	of	internal	medicine.	 2002;162(15):1682-1688.	

	 19	

1	 Table 1. Characteristics of randomized controlled trials studying drug interventions for 2	 ischemic heart disease by presence of an upper age limit

Variable

Primary industry funding, N (%) Study phase, N (%)
Sample size, Median (IQR)1 Active comparator, N (%) Blinding, N (%)
Multicenter design2, N (%) Registration year 3, N (%)
Trial includes any safety outcome, N (%) Study duration (Years)4, Median (IQR)

No
Yes
Phase 1 Phase 2 Phase 3 Phase 4 Unknown
No active comparator Active comparator Double Blinding Single Blinding No Blinding No Yes 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 No
Yes

Upper Age Limit

Yes (N=446)

No (N=393)

349 (78.3)

275 (70.0)

97 (21.7)

118 (30.0)

27 (6.1)

84 (18.8)

77 (17.3)

211 (47.3)

47 (10.5)

100.0

(55.75222.75)

6 (1.5)

73 (18.6)

123 (31.3)

159 (40.5)

32 (8.1)

200 .5

(82.0-910.0)

322 (72.2)

273 (69.5)

124 (27.8)

120 (30.5)

222 (49.8)

201 (51.1)

74 (16.6)

44 (11.2)

150 (33.6)
306 (69.1) 137 (30.9) 27 (6.1) 32 (7.2) 45 (10.1) 32 (7.2) 46 (10.3) 49 (11.0) 60 (13.5) 56 (12.6) 45 (10.1) 53 (11.9) 205 (46.0)
241 (54.0)

148 (37.7)
194 (49.9) 195 (50.1) 28 (7.2) 37 (9.5) 41 (10.5) 39 (10.0) 42 (10.7) 54 (13.8) 44 (11.3) 32 (8.2) 42 (10.7) 32 (8.2) 173 (44.0)
220 (56.0)

1.6 (1.0-2.4) 2.2 (1.3-3.1)

Total (N=839)
624 (74.4)
215 (25.6)
33 (3.9) 157 (18.7) 200 (23.8) 370 (44.1) 79 (9.4) 120 (60.0.0 431.75)
595 (70.9)
244 (29.1)
423 (50.4)
118 (14.1)
298 (35.5)
500 (60.1) 332 (39.9) 55 (6.6) 69 (8.3) 86 (10.3) 71 (8.5) 88 (10.5) 103 (12.3) 104 (12.4) 88 (10.5) 87 (10.4) 85 (10.2) 378 (45.1) 461 (54.9)
1.8 (1.0-2.8)

P-Value 0.006 < 0.001
< 0.001 0.385 0.069
< 0.001 0.015
0. 572 < 0.001

	 20	

Anti-thrombotic No

257 (57.6)

227 (57.8)

484 (57.7)

0.968

agents, N (%) Yes

189 (42.4)

166 (42.2)

355 (42.3)

Lipid modifying No agents, N (%) Yes

372 (83.4) 74 (16.6)

335 (85.2) 58 (14.8)

707 (84.3) 132 (15.7)

0.467

Antihypertensiv No

e drugs, N (%)

Yes

413 (92.6) 33 (7.4)

359 (91.3) 34 (8.7)

772 (92.0) 67 (8.0)

0.504

1	 2	 1 One trial with missing data on sample size 3	 2 Seven trials with missing data on multicenter design status 4	 3 Three trials registered during 2004-2005 exclude 5	 4 For 9 trials we were unable to calculate study duration, due to missing completion dates. 6	

7	 8	 Table 2. Trial characteristics associated with the exclusion of elderly persons in

9	 multivariate analysis 10	

11	

Variable

Study phase
Sample size Multicenter design Study duration (Years)

Phase 1 Phase 2 Phase 3 Phase 4 Unknown
No Yes

OR
0.38 0.25 0.40 0.43 0.9998
0.59
0.86

95% CI
(0.14, 0.98) (0.10, 0.66) (0.16, 1.01) (0.16, 1.18) (0.9997, 1.0000)
(0.43, 0.82)
(0.77, 0.97)

P-Val1u2e	 13	
0.0211345		 16	 17	 18	
0.0012690		 0.0022112		 0.0122034		
25	

26	

27	 28	

29	 30	

31	

32	

33	 34	

35	 36	

37	

38	

	 21	

1	 LEGENDS 2	 Figure 1. Flow chart for trial inclusion. Age data was collected on the mean age and 3	 standard deviation (or median and interquartile range) for the study population or for 4	 each treatment arm and on the number of participants above specific age thresholds. 5	 6	 Figure 2. Explicit exclusion of elderly persons according to study registration year. 7	 Percent of trials with an upper age limit was calculated from the 836 trials with 8	 registration date between 2006 and 2015 (3 trials registered during 2004-2005 were 9	 excluded). The proportion of trials with an upper age limit increased over time (odds 10	 ratio 1.06 per year, 95% confidence interval 1.01 to 1.12; P=0.016). 11	 12	 13	 Figure 3. Distribution of proportion of patients who are 65 years and older and 75 years 14	 and older for each trial. There were 340 trials with data on proportion of patients 65 15	 years and older and 334 trials with data for patients 75 years and older (6 trials gave 16	 proportion for age 65 years and older, but did not give proportion for age 75 and older or 17	 mean age and SD that could be used to calculate this proportion). The median 18	 proportion of patients 65 years and older was 38% (IQR 26%-48%, range 0%-100%) 19	 and the median proportion of patients 75 years and older was 9% (IQR 2%- 14%, range 20	 0%-47%). 21	 22	 23	 24	 25	 26	 27	 28	 29	 30	 31	 32	
	 22