Functional Outcomes of Patients ≥85 Years With Acute Ischemic Stroke Following EVT: A HERMES Substudy

See related article, p 2227

Due to improvements in health care and standards of living, the older population (aged ≥65 years) now comprises the fastest-growing age group worldwide.¹ With increasing age, however, comes an increase in comorbidities, in particular those of cardiovascular and cerebrovascular nature. Seventeen percent of all stroke patients are ≥85 years,² and patients in this age group have been observed to have longer hospital stays,³ more severe disability,⁴ and higher rates of mortality.⁴ As such, age represents an important unmodifiable risk factor for all types of stroke, including acute ischemic stroke (AIS).^5,6

Endovascular treatment (EVT) has been shown to be an effective and safe treatment for AIS due to large vessel occlusion, and its benefit is maintained across a wide range of subgroups, including patients ≥80 years of age.⁷ However, although data from the HERMES (Highly Effective Reperfusion Evaluated in Multiple Endovascular Stroke Trials) collaboration showed that the effect size of EVT on functional outcome is even larger in this patient subgroup compared with younger patients, elderly patients still experienced overall worse outcomes and higher rates of mortality.⁷ Several confirmatory observational studies and systematic reviews on the effect of EVT in octogenarians have since been published,^8–12 with few targeting nonagenarian patient populations.^13–16 The latter group accounts for ≈7% of all AIS patients⁶ and was underrepresented in the large EVT trials, partly due to upper age limits in the exclusion criteria. Furthermore, even within the subgroup of the elderly, incremental increases in age appear to play a role, with nonagenarians faring worse than octogenarians with respect to complications such as symptomatic intracerebral hemorrhage (sICH), despite similarities in baseline characteristics other than age.¹⁶

A recent survey demonstrated that, despite the limited evidence, the majority of stroke physicians would offer EVT to patients ≥80 years (including nonagenarians).¹⁷ A randomized trial involving only the very elderly undergoing EVT is unlikely to occur, however. With this study, we, therefore, aim to assess the effect of age on outcome following EVT in detail using an existing large randomized clinical trial dataset. Because the original HERMES study performed an a priori determined subgroup analysis on patients ≥80 years of age, we chose a cutoff of 85 years for the current study. The main goal of the investigation was to assess whether the EVT effect is maintained in patients ≥85 years old. As a secondary aim, we sought to compare clinical outcomes following EVT of patients according to age (≥85 years versus <85 years).

Methods

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Patient Sample and Study Design

The HERMES collaboration pooled individual patient data of 7 randomized controlled trials that established EVT as a safe and effective treatment for patients with AIS due to large vessel occlusion (n=1764).⁷ Of the trials, only ESCAPE (Endovascular Treatment for Small Core and Anterior Circulation Proximal Occlusion With Emphasis on Minimizing CT to Recanalization Times),¹⁸ EXTEND-IA (Extending the Time for Thrombolysis in Emergency Neurological Deficits–Intra-Arterial),¹⁹ and MR CLEAN (Multicenter Randomized Clinical Trial of Endovascular Treatment for Acute Ischemic Stroke in the Netherlands)²⁰ included patients ≥85 years. Detailed methodology and inclusion criteria of the individual trials, as well as the HERMES meta-analysis, are published elsewhere.^7,18–24 In brief, patients who were randomized to the intervention arm underwent EVT, whereas those randomized to the control arm received standard medical care (including, where available, stroke unit care and specialized rehabilitation under the dedicated management of health care professionals with stroke expertise). Intravenous alteplase was administered to eligible patients in both arms. Patient consent was obtained unless the local ethics boards allowed for deferral of consent.

Outcome Measures

The primary outcome of interest was ordinal 90-day modified Rankin Scale (mRS). Secondary outcomes were binarized good functional outcome, defined as an mRS score of 0 to 2 at 90 days, binarized poor outcome, defined as an mRS score of 5 to 6 at 90 days, and successful expanded Thrombolysis in Cerebral Infarction score, defined as expanded Thrombolysis in Cerebral Infarction 2b/3 (binary; patients who underwent EVT only). The safety outcomes were mortality at 90 days and sICH at 24 hours. The presence of sICH was determined using the individual definitions of the participating trials.^18–24

Statistical Analyses

Baseline characteristics of patients aged ≥85 years were compared with those <85 years using descriptive statistics, stratified by treatment allocation. Adjusted logistic regression with a multiplicative interaction term (age×treatment allocation) was performed to test for an interaction effect of age (≥85 versus <85 years) on the relationship between treatment (EVT versus control) and outcomes. Multivariable binary logistic regression models tested the association between patient age (≥85 versus <85 years) and outcomes. Analyses were adjusted for treatment arm (EVT versus control; binary), patient sex (binary), baseline National Institutes of Health Stroke Scale score (continuous), administration of intravenous alteplase (binary), baseline Alberta Stroke Program Early CT Score (continuous), occlusion location (internal carotid artery versus M1 segment of the middle cerebral artery versus M2 segment of the middle cerebral artery; categorical), and time from onset to randomization (continuous). Finally, within the subgroup of patients ≥85 years, univariable (due to the low n) ordinal logistic regression was used to measure the effect of EVT on 90-day mRS compared with conservative treatment. The results are presented in comparison to those from patients <85 years.

To properly account for clustering effects within and across studies as well as treatment groups, all statistical modeling includes random effects for study effect, as well as for the age×treatment interaction.

Missing data (including loss to follow-up) were minimal (<5%) for all outcomes and predictor variables employed in the reported analyses, and hence no imputation was employed except for covariates in statistical modeling, for which simple imputation (median or mean) was used to avoid an undesirable reduction to complete-case analysis.

Unadjusted and adjusted (common) odds ratios are reported with 95% CIs. All statistical tests were 2-sided, and a conventional significance threshold (alpha=0.05) was used for interpretation. Analyses were performed with SAS software, version 9.4 (SAS Institute, Cary, NC), and R, version 3.3 (R Foundation for Statistical Computing, Vienna, Austria).

We used the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) cohort checklist when writing our report.²⁵

Results

Patient Population

The HERMES collaboration included 1764 patients, of whom 77 (4.4%) were ≥85 years of age. Compared with their younger counterparts, patients 85 years and older had higher systolic blood pressure, higher rates of atrial fibrillation, and higher Alberta Stroke Program Early CT Scores at baseline (Table 1). Patients ≥85 years undergoing EVT also less often received intravenous alteplase (73.8% versus 88.3%, P=0.013) and were more likely to have suffered from a prior stroke (26.2% versus 10.4%, P=0.004) and diabetes (28.6% versus 14.4%, P=0.024), whereas those in the control arm had higher rates of hypertension (82.9% versus 57.9%, P=0.003; Table 1). Older adults were also more often female, although this was only significant in the control arm (EVT: 57.1% ≥85 years versus 46.8% <85 years, P=0.207, control: 80.0% ≥85 years versus 46.0% <85 years, P<0.001; Table 1).

Interaction Analysis of the Effect of Age on the Relationship Between Treatment and Outcome

There was a significant interaction effect of age on the relationship between treatment allocation and poor outcome (mRS score 5–6, P=0.020) and mortality at 90 days (P=0.031) such that the beneficial effect of EVT on mortality and poor outcome was more pronounced in patients ≥85 years of age (Tables S1 through S3). In a subsequent subgroup analysis involving only patients ≥85 years, we observed a positive effect of EVT on outcome compared with conservation management, with an unadjusted common odds ratio of 4.20 (95% CI, 1.56–11.32; Figure [A]). The common odds ratio for the effect of EVT on outcome in the younger cohort was 1.90 (95% CI, 1.60–2.26; Figure [B]).

• Download figure
• Download PowerPoint

Outcomes in Patients ≥85 Years Versus <85 Years

Older adults had worse ordinal mRS at 90 days compared with younger patients (adjusted common OR [95% CI, 0.13–0.33]; Table 2). Only 18.2% of patients ≥85 years achieved good functional outcome (mRS 0–2) compared with 40.8% in the younger patient subgroup (adjusted odds ratio, 0.28 [95% CI, 0.15–0.53]; Table 2). Although mortality at 90 days was significantly higher in the older population compared with those <85 years (41.6% versus 14.8%, adjusted odds ratio, 5.18 [95% CI, 3.11–8.66]; Table 2), patients ≥85 years who received EVT had significantly lower rates of mortality compared with those ≥85 years in the control arm (31.0% versus 54.3%, Figure [A]). No significant differences in the rates of successful reperfusion (EVT arm) or sICH were observed between the 2 age groups (Table 2).

Discussion

This study observed a nominal, quantitative interaction effect of age on the relationship between treatment (EVT) and outcomes in the higher mRS range (mortality, mRS 5–6). In other words, there was a differential effect of EVT by age, indicating that EVT is at least if not more effective in the elderly subpopulation. Although patients ≥85 years experienced substantially worse outcomes and higher rates of mortality following AIS due to large vessel occlusion compared with younger patients, the positive effect of EVT effect on outcome was maintained; within the subgroup of patients ≥85 years, those who underwent EVT had better functional outcomes and lower rates of mortality at 3 months compared with those who received standard therapy alone.

These results are in line with those from previous studies involving elderly patients undergoing EVT.^12,26–28 A key strength of the current study, however, is that the data are derived from randomized clinical trials with robust control groups, as opposed to the primarily observational studies found in the literature. The disadvantage of trial data lies in its potentially limited applicability to clinical practice. In their systematic review and meta-analysis of 17 studies involving the effect of EVT in octogenarians (n=860), however, Hilditch et al. observed rates of functional outcome (27%) and mortality (34%)⁸ within close range of those from the current study (24% and 31%, respectively).

The observed effect size of EVT on outcome for patients ≥80 years in the HERMES collaboration⁷ was smaller than that which was observed for those 85 years and older from the same cohort (common odds ratio 3.68 [95% CI, 1.95–6.92] versus 4.20 [95% CI, 1.56–11.32], respectively), the latter albeit with larger confidence intervals due to the smaller overall sample size. The seemingly paradoxical finding that elderly patients have significantly worse functional outcomes despite larger EVT effect sizes compared with younger populations is again reflected in the interaction effect. This is partly explained by the different risk factors observed for the elderly population. For example, frailty and the increased risk of (post-treatment) complications in the older population make the likelihood of survival without successful reperfusion quite low. Younger patients with generally more favorable baseline characteristics, on the other hand, are more likely to survive even in the absence of successful reperfusion.¹²

It is important to note that the rates of sICH and reperfusion did not differ according to age group, which further points toward EVT being a safe and effective treatment option for the elderly. As noted above, the relatively poor outcomes observed in elderly patient populations are likely due to other prognostic factors, such as baseline disability (although the majority of patients included in the HERMES meta-analysis were prestroke independent, MR CLEAN being the only study to include patients with a prestroke mRS score ≥2 (n=46)²⁰) or increased susceptibility in the post-treatment hospitalization phase.²⁹ Nevertheless, approximately 1 out of 4 patients ≥85 years old will achieve a good outcome following EVT, suggesting that age alone should not exclude patients from thrombectomy.

Limitations

This study has several limitations. First, patients ≥85 years were underrepresented in trials that did include them, potentially leading to selection bias. Indeed, in this study, older patients had better Alberta Stroke Program Early CT Scores, indicating potentially more stringent selection criteria for EVT for this subgroup. As a result, it is unclear whether the results of this study are applicable to the corresponding patient population seen in clinical practice. Second, due to the small sample size, we were unable to adjust for potentially confounding factors; the presented results regarding the effect of EVT in the ≥85-year-old subgroup are derived from univariable analyses. Third, we were unable to assess the effect of preexisting disability on outcome, with is presumably more prevalent with increasing age. Fourth, information on additional standards of care during the post-acute phase (eg, discharge destination) is lacking. As a result, we are unable to comment on events that occurred from the time of treatment to 90-day follow-up, which may influence outcomes. Finally, while the mRS is the standard outcome measure in stroke studies, each level is a rather broad category that may require more granularity with respect to quality of life and disability, particularly within the elderly population. Despite these limitations, we believe the results of this study will be of value to physicians faced with the often-complex treatment decision-making processes involving patients suffering from AIS.

Conclusions

Patients ≥85 years old with independent premorbid function more often achieve good functional outcomes when treated with EVT compared with conservative management, with a number needed to treat of 8 for 90-day mRS score of 0 to 2, an observed treatment effect modification of age on outcome, and no differences in sICH and reperfusion rates. EVT should therefore not be withheld in this subgroup on the basis of age alone.

Article Information

Acknowledgments

We would like to thank the HERMES collaboration (Highly Effective Reperfusion Evaluated in Multiple Endovascular Stroke Trials) investigators.

Supplemental Material

Tables S1–S3

Footnotes