Exploring the Impact of Age and Pre-Stroke Modified Rankin Scale in Elderly Thrombectomy: A 15-Year Single-Center Experience

Approximately 30% of all strokes occur in patients aged ≥80 years, with an incidence rate of 20 per 1,000 person-years [1]. Given the limitations of randomized controlled trials (RCTs) on endovascular thrombectomy (EVT) in this age group, further studies are essential to aid the selection of EVT among the elderly. Although the HERMES meta-analysis showed the benefits of EVT for disability across all age groups, only 15% were aged ≥80 years. The subgroup analysis of patients over 85 years of age reported lower mortality and better functional outcomes after EVT; however, the elderly fared worse compared to younger adults [2]. Furthermore, several studies have reported lower rates of good functional outcomes in the elderly than in younger adults [3‒5]. The variability in outcomes among elderly patients poses a challenge in translating trial results into “real-world” practice.

The pre-stroke modified Rankin scale (mRS) is increasingly used to assess pre-stroke function in the acute setting as studies have shown its reliability and association with prognosis [6, 7]. However, distinguishing between patients with a pre-stroke mRS score of 1 or 2 may be challenging because of the lack of clear definitions. As elderly patients may have some degree of decreased functional ability, a general perception of poor functional status in this age group may exclude them from receiving EVT. The aim of this study was to explore clinical and safety outcomes for elderly patients (≥80 years old) with variable baseline disability scores using the pre-stroke mRS, treated with EVT for acute ischemic stroke (AIS).

Approved by the South-Eastern Norway Regional Health Authority Ethics Committee, this single-center study prospectively registered consecutive patients treated with EVT at the Oslo University Hospital (OUH) from October 2007 to November 2022. The OUH is a comprehensive stroke center that performs EVT for patients within the regional stroke care network through the “drip and ship to mothership” model. Stroke patients were screened using non-contrast CT with angiography at local stroke units, with or without perfusion imaging, and intravenous thrombolysis (IV-tPA) was administered to eligible patients before transfer to our center. Additional imaging was performed prior to EVT when the time from the last imaging was prolonged or a significant change in clinical status was observed upon admission. The indication for EVT treatment was based on the presence of accessible occlusion, penumbral imaging when available, clinical symptoms, and relevant clinical history. Since 2017, our center has increased the utilization of pre-stroke mRS (Table 1) scoring to assess the functional status of patients. Mechanical thrombectomy was performed with either a stent retriever, aspiration, or a combined approach. The modified Thrombolysis in Cerebral Infarction (mTICI) [8] was used to assess reperfusion evaluated by the treating interventional neuroradiologists, an independent neuroradiologist, or an experienced stroke neurologist. Clinical outcomes were assessed in all patients at 3 months by an mRS-certified neurologist by telephone or in-person follow-up. Missing pre-stroke mRS and mTICI scores were extracted from medical records and radiological imaging. Mortality data at 3 months and 1 year after the procedure were gathered from our regional electronic health record system, Distributed Information and Patient Data System (DIPS AS) in hospitals, used by 3 of the 4 health regions in Norway, covering 80% of patients in Norway. Additionally, data were collected through our research registry software Medinsight (OUH), which automatically updates with the public registry every 14 days, covering the whole Norwegian population. We defined fair outcome in this study as mRS ≤3 or retention of a pre-stroke mRS score of 4, while a change to mRS 4–6 was considered a worse outcome.

The study population, divided into fair and worse outcomes, 3-month mortality, and 1-year survivability, were compared using the χ² test, Fisher’s exact test, and Mann-Whitney U test as appropriate. The cohort was divided into two periods: one from 2007 to 2015 (pre-RCT era) and the other from 2016 to 2022, aiming to find variables that exhibited significant differences between these time frames. Multivariable binary logistic regression analysis was used to estimate odds ratios (OR) with corresponding 95% confidence intervals (CI) and identify predictors of fair outcome, mortality at 3 months, and 1-year survival. All p values were two-sided with a significance level of 5%. Statistical analyses were performed using IBM SPSS Statistics for Windows, version 26 (IBM Corp., Armonk, NY, USA). Ordinal logistic regression analysis was performed using STATA 17 (StataCorp, College Station, TX, USA).

Between October 2007 and November 2022, 307 patients aged ≥80 years from the EVT registry were included in the analyses. The median age was 84 years (interquartile range [IQR] 82–87), wherein 31 (10%) were nonagenarians (aged ≥90 years). Two hundred and sixty-six patients (86.9%) had a pre-stroke mRS ≤2, while 30 (9.8%) and 11 (3.6%) patients had a pre-stroke mRS of 3 and 4, respectively. Comparing the periods 2007–2015 and 2016–2022, we found significantly lower National Institutes of Health Stroke Scale (NIHSS) scores, higher pre-stroke mRS scores, and better recanalization rates in the latter period (online suppl. Table 1; for all online suppl. material, see https://doi.org/10.1159/000540992).

At 3-month follow-up, 110 (41.4%) of 266 patients with a pre-stroke mRS ≤2 achieved mRS ≤2 and 159 (53.7%) of 296 patients with pre-stroke mRS ≤3 had mRS of ≤3. Three out of 11 (27.2%) patients with a pre-stroke mRS 4 retained their premorbid functional capacity. A fair outcome was achieved in a total of 162 (53%) patients at 3-month follow-up, of whom were younger (median 84 vs. 85 years, p = 0.009), had lower median NIHSS score upon admission (13 vs. 17, p < 0.001), and had lower pre-stroke mRS (0 vs. 1, p < 0.001) compared to those with worse outcomes. Creatinine levels were significantly higher (85 vs. 77, p = 0.008) while blood glucose was lower (6.8 vs. 7.1, p = 0.033) in the fair outcome group. Time variables were comparable, while higher rates of recanalization (≥TICI 2b) were achieved in the fair outcome group (91% vs. 77%) (Table 2). The probability of achieving a fair outcome decreased by 8% for every 1-year increase in age (OR 0.92, 95% CI 0.86–0.98, p = 0.008, Fig. 1).

In multiple regression analyses, NIHSS (OR 0.89, 95% CI 0.85–0.93, p < 0.001) and pre-stroke mRS (OR 0.67, 95% CI 0.53–0.84, p < 0.001) were prognostic factors for fair outcomes when adjusting for age, glucose, and creatinine. After further adjustment for mTICI, both NIHSS (adjusted odds ratio 0.87, 95% CI 0.83–0.92, p < 0.001) and pre-stroke mRS (adjusted odds ratio 0.63, 95% CI 0.50–0.79, p < 0.001) remained significant.

At 3-month follow-up, 84 (27.4%) patients were deceased. Higher age (median age 86 vs. 84 years, p < 0.001), higher median NIHSS score (18 vs. 14, p < 0.001), pre-stroke mRS score (1 vs. 0, p = 0.002), and increased glucose level (7.1 vs. 6.8, p = 0.048) were associated with mortality (Table 3). For every year increase in age, the probability of death at 3 months increased by 16% (OR 1.16, CI 1.08–1.24, p < 0.001, Fig. 1). In multiple regression, increasing age (OR 1.13, 95% CI 1.05–1.21, p = 0.001), NIHSS (OR 1.10, 95% CI 1.05–1.15, p < 0.001), and pre-stroke mRS (OR 1.37, 95% CI 1.11–1.12, p = 0.004) were all associated with higher mortality.

At 1-year follow-up, 193 patients (62.9%) were still alive. Those who survived were younger (median age of 83 vs. 86 years), female, and had a significantly lower median NIHSS (14 vs. 17) and pre-stroke mRS (0 vs. 1). An increased rate of thrombolysis administration and recanalization rate score of mTICI ≥2b were also apparent in this group (Table 4). Twelve out of 31 nonagenarians (38.7%) were alive 1 year after EVT, and 5 of 6 patients 95 years and older were deceased within 1 year. For every year increase in age, survival probability 1 year after the stroke decreased by 13% (OR 0.87, CI 0.81–0.93, p < 0.001, Fig. 1).

In this study of elderly patients treated with EVT, 53% achieved fair outcomes with mRS ≤3 or retained a pre-stroke mRS score of 4 at the 3-month follow-up. Age, pre-stroke mRS, and NIHSS scores were all significantly associated with both outcomes at 3 months and survival 1 year after EVT. For every 1-year increase in age, the probability of a fair outcome at 3 months decreased by 8%.

The reported rate of good functional outcome after EVT for AIS in the elderly has ranged between 11 and 60% in previous studies [9, 10]. Studies with sample sizes of over 100 patients reported good outcomes (mRS ≤2) ranging from 19 to 38% [5, 11‒15]. A systematic review and meta-analysis of RCT trials on large vessel occlusions including both the anterior and posterior circulation documented higher odds for good outcomes (OR 4.43) in patients aged ≥80 years treated with EVT [16]. This study however included only 139 patients from 2 RCT trials. A more recent systematic review and meta-analysis by Zarrintan et al. [17] of 47 observational studies including 20,338 patients reported 38.7% of patients aged ≥80 achieving good functional outcomes. Our study’s 3-month follow-up results, with 41.4% of patients achieving mRS ≤2, are comparable to the aforementioned publication and consistent with previous studies reporting a significant proportion of elderly patients achieving good functional outcomes after EVT. This is also comparable to another Norwegian study reporting a 45% rate of good functional outcomes in both anterior and posterior large vessel occlusions in elderly patients [18].

As a significant proportion of elderly patients exhibit functional disabilities prior to stroke, consideration of what constitutes a good outcome for this group should be adjusted. Groot et al. [5] reported that 23.4% of patients ≥80 years had a pre-stroke mRS score of ≥3 in the MR CLEAN registry [5]. The goal of achieving an mRS ≤2 in the elderly is therefore unrealistic, and we consider using a fair mRS outcome in this age group to be a more reliable and practical metric. A recent Danish study on AIS EVT using a fair outcome as the primary endpoint reported a 41.7% fair outcome at 3 months [19].

The Catalonia registry study reported no association between pre-stroke mRS and 3-month functional outcomes after adjusting for variables such as age and ASPECTS [20]. However, the registry included younger patients, and the authors also noted that disabled patients had a higher long-term risk of mortality, sICH, and death within 3 months than patients without disability. An Australian study comparing nonagenarians with pre-existing dependency to those without did not find any significant difference in the rate of favorable outcomes [21]. A higher mortality rate among dependent nonagenarians was however reported. Goda et al. [22] reported that patients with pre-stroke mRS ≥2 had significantly worse outcomes irrespective of age [22] and Laugesen et al. [19] also found worse outcomes per level increase in pre-stroke mRS [19]. A meta-analysis on patients with AIS treated with EVT with pre-stroke dependency also reported a higher rate of unfavorable clinical outcomes and mortality irrespective of age [23]. Based on our results and the aforementioned reports, the pre-stroke mRS, despite providing a rough estimate of a patient’s functional ability, may be of value in clinical practice when considering EVT for elderly patients with AIS. The emerging use of frailty indices, shown to be associated with poor outcomes in AIS, including in EVT patients [24, 25], can be translated to equivalent mRS scores in an acute setting.

Age alone is not considered a reason for withholding EVT from the elderly as a non-negligible proportion of these patients may still benefit from EVT compared to medical treatment alone [2, 9, 26, 27]. However, multiple studies comparing younger patients, octogenarians, and nonagenarians have shown worse outcomes in the elderly. The increasing probability of worse outcomes with increasing age, notably in 5-year increments, was apparent in the STRATIS registry and the study by Rezai et al. [18, 28]. This was also the case in the recently published study by Zarrintan et al. [17], noting the higher mortality among nonagenarians compared to octogenarians [17]. In the present study, we also found a direct association between age and poor outcomes. The estimated chance of favorable outcomes in the elderly of approximately 50% leaves a considerable challenge in identifying the subgroup of patients who will benefit from EVT in clinical practice. Successfully meeting this challenge necessitates a comprehensive assessment that takes into account multiple factors.

The female predominance in survivability (55.6% vs. 44.6%) in our study reflects the fact that women generally have longer life expectancies. While no significant difference was observed in the other outcome measures between males and females, we noted a more favorable trend among females, with a male predominance in negative outcomes. In the study by Xue et al. [29], it was reported that female sex was a predictor of very poor outcomes after EVT in older patients. However, in a propensity score-matched study on the impact of sex on EVT outcomes, no difference between males and females was found [30]. This topic needs further investigation, especially in the elderly, as it can be used as a prognostic factor which can guide clinicians.

Since stroke treatment outcomes depend on the entire treatment cascade, from acute reperfusion to rehabilitation, it is important to emphasize that our findings reflect results achieved within the Norwegian health care system. The Norwegian stroke system has consistently aimed to optimize stroke care streamlining enhanced acute treatment, diagnostics, and rehabilitation. Therefore, our findings may not be applicable to other health care systems that differ from our model.

The main limitation of this study is the inclusion period of 15 years, during which EVT techniques, experience, and procedures have evolved over time. We attempted to minimize this effect by correcting for relevant variables before and after 2015. The absence of a control group and the retrospective design also limits this study’s findings. The lack of a control group is a well-known obstacle in thrombectomy studies in the elderly. Randomizing the elderly to either EVT or the best medical treatment poses ethical challenges since the treatment’s effectiveness has already been established for younger patients. Our center is a regional thrombectomy center operating as a “mothership” where patients referred for thrombectomy are selected based on presumed the effect of treatment. The patients included in our study therefore differ significantly from those excluded for treatment. The lack of a control group limits comparative analysis, and interpretations of the conclusion should therefore be done with caution.

Selection bias, particularly in the earlier years, is another limitation of this study. Healthier older adults were frequently selected for treatment during this period, whereas selection became less stringent after the RCTs proved thrombectomy to be a safe and efficient treatment for acute stroke, allowing more older adults to be treated with thrombectomy. To explore if this change significantly affected the findings in our study, we analyzed data from both pre- and post-RCT periods which did not show significant differences in outcomes (online suppl. Table). Still, the patients included in this study were selected based on the treatment evidence and guidelines existing at the time of the procedure, making it difficult to completely eliminate or mitigate this bias. Limiting patient inclusion to the time period after EVT was introduced into clinical treatment guidelines would reduce this bias. Specific post-treatment complications unrelated to EVT were not registered, nor were specific radiological parameters prior to EVT included in the analysis, limiting this study further.

EVT proves beneficial for selected elderly stroke patients, achieving rates of fair outcome reaching up to 53%. Increasing age, higher pre-stroke mRS and high NIHSS scores are associated with worse 3-month functional outcomes and mortality within 1 year of treatment. Utilizing pre-stroke mRS can aid clinicians in selecting elderly patients for EVT. We suggest the use of fair outcomes as a more appropriate outcome measure for elderly patients.

We thank all the patients who participated in our study. We would also like to thank the interventional neurologists at our center: Terje Nome, Bård Nedregaard, Øivind Gjertsen, Thor Skattør, Ruth Sletteberg, and Martin Sökjer.

The study was approved by the Ethics Committee of the South-Eastern Norway Regional Health Authority (reference No. 2010/2610a & 2015/1844). Informed consent was acquired from either the patient or its legally authorized representative when the patient was incapable of providing consent. Verbal consent was secured during the period from 2007 to 2016 (REK 2010/2610a), transitioning to written consent from 2017 to 2022 (REK 2015/1844).

Anne Hege Aamodt received unrestricted research grants from Medtronic and Boehringer Ingelheim. She has received speaker honoraria from Bayer, Teva, Novartis, Roche, AbbVie, Boehringer Ingelheim, and Pfizer and declares no conflicts of interest related to this study, which is also the case for the rest of the authors.

Oslo University Hospital and South-Eastern Norway Regional Health Authority funded this study.

Brian Anthony B. Enriquez and Karolina Skagen: conception and design, patient inclusion, data acquisition and analysis, drafting, and final review. Heidi Kristine Halling: conception, data acquisition, initial drafting, and critical review. Bjørn Tennøe: conception, radiological data acquisition, and critical review. Cathrine Brunborg: conception, data analyses, and critical review. Christian Georg Lund, Mona Elisabeth Skjelland, and Anne Hege Aamodt: conception, patient inclusion, and critical review.

Data generated during this study are included in this article and its supplementary material. Further inquiries or access to fully anonymized data can be directed to the corresponding author.