Characteristics and Outcomes of Very Elderly Enrolled in a Prehospital Stroke Research Study
Background and Purpose—Greater numbers of individuals aged ≥80 years enjoy a high quality of life, yet historically stroke trials have excluded this population. We aimed to describe a population of very elderly successfully enrolled into an acute stroke trial and compare their characteristics and outcomes with the younger cohort.
Methods—We analyzed consecutive patients enrolled <2 hours of symptom onset in a prehospital stroke treatment trial, the FAST-MAG clinical trial (Field Administration of Stroke Therapy-Magnesium). We gathered demographic, treatment, and outcome data for nonelderly (<80 years old), very elderly (≥80 years old), and extreme elderly (≥90 years old). We describe key differences in the population of elderly and the impact of their inclusion on the clinical trial.
Results—Of 1700 participants in FAST-MAG, there were 1210 nonelderly, 490 very elderly, and 60 extreme elderly subjects. Very elderly stroke patients successfully enrolled in a research study were more likely to be women, white, and have an ischemic mechanism rather than an intracerebral hemorrhage. Although the very elderly had generally poorer outcomes, 4 in 10 were functionally independent at 90 days.
Conclusions—Inclusion of the very elderly population in acute stroke clinical trials would both significantly increase study participation and generalizability of future acute stroke clinical trials.
See related article, p 2679.
The US population of the very elderly (≥80 years old) is now one of the most rapidly growing age groups second to only the baby boomers (55–69 years old).1 This population is projected to nearly double by 2030.2 With improving life expectancy, the population of extreme elderly (≥90 years old) is also projected to continue to grow exponentially. This trend can be attributed to better healthcare, environmental factors, and higher standards of living in not only the United States but also worldwide. The very elderly are at an increased risk for stroke, have the highest risk for complications from stroke treatments,3 and are at significantly higher risk of mortality secondary to stroke. Despite the growing elderly population, acute stroke trials have frequently excluded their participation in studies, which has led to a limited understanding of therapeutic efficacy in a significant proportion of patients with stroke. For example, the NINDS tPA trials (National Institute of Neurological Disorders and Stroke tissue-type plasminogen activator) only included 42 patients aged ≥80 years.4,5 Potential reasons for their exclusion include a greater number of comorbidities and generally poorer outcomes.
As stroke is projected to become the leading cause of mortality in developed nations within 30 years6 and the US and worldwide population continue to grow older, there will be changes to the landscape of healthcare and a demand for a better understanding of stroke in the very elderly. It is important that we include the population of very elderly and extreme elderly patients with stroke in clinical research, but we are not certain how their inclusion would affect acute stroke trials. For example, is it possible to enroll this population and how would their demographics and outcome affect a clinical trial? We sought to describe the characteristics of the very elderly (patients aged 80–95 years) enrolled in a multicenter prehospital acute stroke treatment trial, the FAST-MAG study (Field Administration of Stroke Therapy-Magnesium).
The FAST-MAG clinical trial is a phase 3, National Institutes of Health-NINDS–sponsored, randomized, placebo-controlled, clinical trial of field-initiated magnesium sulfate versus placebo in suspected hyperacute stroke patients within the first 2 hours of last known well time (LKWT).7 The detailed methods of this study have been previously published.8–10 The purpose of the FAST-MAG trial was to identify whether magnesium sulfate, which has been shown to be a neuroprotective agent in experimental stroke models, is an effective and safe treatment for acute stroke. It also aimed to demonstrate field enrollment and treatment of suspected patients with stroke within the hyperacute time window for phase 3 stroke trials. The primary end point for this study was a difference in distribution of scores between the magnesium sulfate and placebo groups on the 3-month modified Rankin Scale with a target enrollment of 1700.
The clinical trial was active and enrolled in Los Angeles County from 2005 to 2012 and Orange County from 2010 to 2012. The racially, ethnically and socioeconomically diverse population of >13 million people in Los Angeles and Orange Counties includes >400 000 people aged >80 years.11
The study utilized the modified Los Angeles Prehospital Stroke Screen to identify likely patients with stroke en route to hospitals with a paramedic and physician-confirmed LKWT under 2 hours. Age criteria for enrollment were 40 to 95 years, and the study also required subjects to be independent in their daily activities or be able to perform most without assistance. Also having an underlying disease process such as dementia would interfere with primary outcome measures and was also included in the exclusion criteria. All subjects were enrolled using explicit informed consent by cell phone from either the potential subject or a legally authorized representative. In rare cases, subjects were enrolled through an exception to explicit informed consent.
The FAST-MAG trial gathered data including age, sex, race, comorbidities, tobacco use status, LKWT, time of paramedic arrival on scene, time of emergency department (ED) arrival, National Institutes of Health Stroke Scale, use of intravenous tissue-type plasminogen activator and final diagnosis (ischemic stroke, transient ischemic attack, intracranial hemorrhage, and other). These data were broken down into the age groups of nonelderly (age <80 years), very elderly (age ≥80 years), and extreme elderly (age ≥90 years). Rates of study enrollment, demographic characteristics, and long-term outcomes were compared among each group.
Of 1700 subjects enrolled in FAST-MAG from 2005 to 2014, 490 (29%) subjects were ≥80 years of age with 60 (4%) of those subjects were ≥90 years. The remaining 1210 (71%) subjects made up the nonelderly group were <80 years of age (Figure I in the online-only Data Supplement). Primary analysis was focused on describing very elderly group and comparing them to the nonelderly. A second analysis focused on the extreme elderly subgroup. Treatment with magnesium was not associated with any changes in 90-day poststroke outcomes in the very elderly (P=0.450) and the extreme elderly (P=0.391).
Table 1 shows a descriptive comparison of the nonelderly, very elderly, and extreme elderly populations of the trial. The mean age for the nonelderly group was 63 (SD, 11) years, for the very elderly group 85 (SD, 4) years, and for the extreme elderly 92 (SD, 2) years. The very elderly are more often white (87% versus 74%; P<0.01) and more likely to be women (53% versus 38%; P<0.01). The sex bias is even more pronounced in the extreme elderly group, which was 68% women (P<0.01). The extreme elderly group has a predominance of participants classified as white (88%). Subjects were less likely to be of Hispanic ethnicity in the very elderly (15%) and extreme elderly (12%) groups when compared with the nonelderly (27%; P<0.01 and P=0.026, respectively). More subjects were enrolled using Spanish-language consent for the study in the nonelderly group (13%; P<0.01 versus 7%; P<0.01 of very elderly). Ischemic stroke was more common in the very elderly group (82% versus 70%; P<0.01), and intracerebral hemorrhage was less common (14.5% versus 26.1%).
Comorbidities and Risk Factors
Prevalence of comorbidities differed between the very elderly and nonelderly (Table 2). The very elderly was more likely to have hypertension (85% versus 75%; P<0.01), hyperlipidemia (52% versus 45%; P<0.01), atrial fibrillation (38% versus 15%; P<0.01), and coronary artery disease (28% versus 18%; P<0.01). Conversely, prevalence of diabetes mellitus was lower in the very elderly population (19% versus 24%; P=0.024). When comparing the extreme elderly group to the rest of the population, history of atrial fibrillation (35% versus 21.2%; P=0.01) and atrial fibrillation on ECG (28% versus 18%; P=0.038) were more frequently noted.
Prevalence of behavioral risk factors such as alcohol and tobacco use was less common in the elderly (Table 2). Of note, only 4% of the very elderly and none of the extreme elderly participants used tobacco as opposed to 23% of the nonelderly (P<0.01).
Initial systolic and diastolic blood pressures (161/93 versus 158/84; P=0.025, P<0.01) and heart rate (87 versus 82 bpm; P<0.01) of nonelderly subjects were higher than those of the very elderly population when assessed in the field. Upon ED arrival, diastolic blood pressure (88 versus 80 mm Hg; P<0.01) and pulse (82 versus 78 bpm; P<0.01) remained elevated, but systolic blood pressure was equal in both groups (160 mm Hg; P=0.99). Higher blood pressure in the nonelderly may reflect the underlying cause of the cerebrovascular event, such as intracerebral hemorrhage. There was no significant difference in stroke severity by National Institutes of Health Stroke Scale at presentation to the ED and at 24 hours.
Treatment and Outcomes
Table 3 shows treatment and outcome data for the different age groups. There was no significant difference between the proportion of subjects who received reperfusion therapies including endovascular intervention or intravenous tissue-type plasminogen activator. Door-to-imaging times were longer in the very elderly (42 versus 33 minutes; P=0.048). All other measurements including LKWT to 911 call, LKWT to paramedic arrival, LKWT to scan, and arrival of paramedics to arrival to ED times were similar throughout age groups. Onset-to-treatment times were also similar in all participant groups.
Outcome measured by modified Rankin Scale at 90 days varied significantly in the different age groups. Table 4 shows the proportion of each subgroup falling in the different modified Rankin Scale categories of poor outcome and 90-day modified Rankin Scale. Although very elderly subjects had higher modified Rankin Scale scores when compared to the younger participants, there were significant numbers of very elderly (40%) and extreme elderly (25%) subjects who were functionally independent at 90 days post stroke.
We found that large numbers of very elderly and extreme elderly patients can be enrolled in acute stroke research. If we had excluded the very elderly, the FAST-MAG trial would have required an additional 2 years to complete if enrollment rates remained stable. Enrolled very elderly patients with stroke were different in many ways and results of treatment trials that exclude their participation, or under enroll this subgroup, may not reflect the actual effect in the target population. Many of the previously published National Institutes of Health–funded major stroke trials including the NINDS Tissue-Type Plasminogen Activator and TOAST12 trials (Trial of ORG 10172 in Acute Stroke Treatment) have not included or had a low representation of this population.
There were significant demographic differences between the participant age groups. The higher prevalence of female participants can likely be attributed to the fact that women have longer life expectancy: 83.2 versus 78 years for men.1 The racial breakdown throughout the groups varied slightly although whites were consistently the majority of participants. This is no surprise as whites are also the largest racial group in Los Angeles County and make up 71% of the population.13 The proportion of white participants increased in the very elderly and extreme elderly groups despite whites having a shorter life expectancy (80.3 years) than Hispanics (82.2 years) and Asian/Pacific Islanders (85 years).6 The proportion of black participants decreased in the older age groups, perhaps reflecting shorter life expectancy (74.3 years).6
The contrasting trends of decreasing prevalence of diabetes mellitus in older age groups compared with an increasing prevalence of hypertension, coronary artery disease, and hyperlipidemia is noteworthy because diabetes mellitus is associated with these comorbidities. One possible explanation is that the very elderly participants with diabetes mellitus had already developed renal failure or severe baseline disability at that point and were excluded from the study.
Very elderly participants had comparable rates of intravenous tissue-type plasminogen activator administration, and there was little difference in the expediency of care. The only notable difference in care was time from ED arrival to first scan between the very elderly and nonelderly, which was an unexpected finding. The cause of this delay is unknown but could be because of bias or preference of ED staff while triaging or that the elderly likely needed greater time for stabilization.
The very elderly had worse long-term outcomes, reflecting higher burden of comorbidities and complications after stroke. Historically, this is a key reason for their exclusion from clinical trials. We found that a considerable proportion, 4 in 10, of enrolled very elderly were independent. Demonstration of improved long-term outcomes in the very elderly would be feasible in the future.
This study has limitations. Numbers of extreme elderly patients in the study were low, limiting comparison to the much larger nonelderly group. The study design also included several exclusion criteria such as renal failure, absence of a consent provider, or patient declination on trial participation that limited the scope of the trial population.14 Prehospital process enrollment could have overcome inherent bias against enrollment of very elderly. Additionally, although we had data on place of residence before event, we do not know whether participants lived alone which may lead to delayed presentation to the ED in the very elderly and extreme elderly groups.
In conclusion, we think that enrollment of very elderly and extreme elderly subjects in clinical stroke research is feasible and desirable. The FAST-MAG trial provided valuable insight into the characteristics of the very elderly population in a stroke trial. Inherent differences among the very elderly provide further evidence of the importance of their inclusion in research.
Sources of Funding
This study was supported by NIH-NINDS Award U01 NS 44364.
Dr Liebeskind serves as a consultant for Stryker and Medtronic. The other authors report no conflicts.
Guest Editor for this article was Seemant Chaturvedi, MD.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.116.013318/-/DC1.
- Received February 29, 2016.
- Revision received July 11, 2016.
- Accepted July 19, 2016.
- © 2016 American Heart Association, Inc.
- Howden LM,
- Meyer JA
- Ortman JM,
- Velkoff VA,
- Hogan H
- Furie KL,
- Kasner SE,
- Adams RJ,
- Albers GW,
- Bush RL,
- Fagan SC,
- et al
- Saver JL,
- Starkman S,
- Eckstein M,
- Stratton S,
- Pratt F,
- Hamilton S,
- et al
- Saver JL,
- Starkman S,
- Eckstein M,
- Stratton S,
- Pratt F,
- Hamilton S,
- et al
- Sanossian N,
- Starkman S,
- Liebeskind DS,
- Ali LK,
- Restrepo L,
- Hamilton S,
- et al
- Lofquist D,
- Lugiala T,
- O’Connell M,
- Feliz S
- 13.↵US Census Bureau State and County Quickfacts 2013. US Census Bureau Website. http://www.census.gov/quickfacts/table/PST045215/06. Accessed May 15, 2016.