Article Text

Download PDFPDF

Original research
Discrepancy between early neurological course and mid-term outcome in older stroke patients after mechanical thrombectomy
  1. Justus F Kleine1,
  2. Tobias Boeckh-Behrens1,
  3. Sascha Prothmann1,
  4. Claus Zimmer1,
  5. Thomas Liebig2
    1. 1Department of Neuroradiology, Technische Universität München, Klinikum Rechts der Isar, München, Germany
    2. 2Department of Neuroradiology, Universitätsklinik Köln, Köln, Germany
    1. Correspondence to Dr Justus F Kleine, Abteilung für Neuroradiologie, Klinikum Rechts der Isar, Ismaningerstr 22, 81675 München, Germany; justus.kleine{at}


    Background Stroke in aged patients has a relatively poor prognosis, even after recanalizing therapy. Potential reasons include mechanisms that relate directly to the extent of brain tissue damage, but also age-dependent factors which are not, or only indirectly, stroke-related, such as pre-existing functional deficits, comorbidities, and post-stroke complications (eg, infections).

    Objective To compare early neurological course with subsequent functional outcome in older (≥80 years) and younger stroke patients in order to estimate the relative impact of these factors. Specifically, to examine if the strong age-dependency of modified Rankin Scale (mRS) outcome scores in stroke patients after mechanical thrombectomy is paralleled by a similar age dependency of early postinterventional National Institute of Health Stroke Scale (NIHSS) scores—a more specific measure of stroke-induced brain damage.

    Methods We evaluated technical results, pre- and postinterventional NIHSS scores, mid-term mRS scores and early and overall mortality and their relation to age in 125 patients, 40 of them ≥80 years, with acute middle cerebral artery occlusion, treated by mechanical thrombectomy.

    Results Technical success, pre- and postinterventional NIHSS scores and early mortality were age-independent. Early neurological improvement depended on successful recanalization, but not on age. Nevertheless, good mRS outcome (mRS 0–2) was much rarer, and overall mortality almost threefold higher in aged patients.

    Conclusions Older patients exhibit a similar early neurological course and responsiveness to mechanical thrombectomy as younger patients, but this is not reflected in mid-term functional outcome scores. This indicates that post-stroke complications and other factors that are not, or only indirectly, related to the brain tissue damage induced by the incident stroke have a dominant role in their poor prognosis.

    • Stroke
    • Thrombectomy
    • Thrombolysis
    • Brain
    • Intervention

    Statistics from

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


    Stroke in patients ≥80 years presently accounts for about one-third of all strokes in developed societies, and this proportion will continue to increase in the upcoming decades.1 Advanced age is associated with relatively poor post-stroke functional outcome, with ‘outcome’ commonly assessed by the modified Rankin Scale (mRS),2 at 3 months after the incident stroke. This negative association holds for all relevant acute treatment modalities, including intravenous thrombolysis,3–5 intra-arterial thrombolysis,6 ,7 and mechanical thrombectomy (MTE).8–12 It has important implications for clinical decision-making in individual patients, health economics, and for the design and interpretation of stroke trials. From the comparably poor mRS outcome of older patients, researchers had inferred a high likelihood of ‘futile’ recanalization,8 a relatively unfavorable risk–benefit-ratio,7 ,11 and relatively poor cost-effectiveness13 of endovascular stroke therapy in this age group. Recent subgroup analyses of the MR CLEAN and ESCAPE trials indicate that MTE with stent retrievers is an effective treatment also for older patients14 ,15 but their average outcome is still modest, and mortality remains much higher than in younger patients,15 so that further improvement is desirable.

    However, the poorer mRS outcome of older patients with acute stroke is not necessarily due to the direct negative impact of the stroke-induced brain damage alone. mRS scores are known also to be sensitive to, for example, comorbidities, pre-stroke physical and cognitive impairment, and medical complications such as infections—that is, factors that are unrelated, or only indirectly related, to the incident stroke.16 ,17 This may have negligible influence in young patients, in whom any observable post-stroke functional deficit is likely to be caused directly by the incident stroke, but become important in older patients, in whom such conditions are more prevalent and/or graver.

    The National Institute of Health Stroke Scale (NIHSS) has been designed to assess stroke-induced neurological deficits and reflects functionally relevant brain tissue damage more directly and specifically. Treatment-related effects of acute therapies that are intended to salvage brain tissue at risk—and, correspondingly, also the lack, or age-related limitations, of such effects—may therefore be recognized more clearly in early NIHSS scores than in mid-term mRS scores. Hence, if the relatively poor mRS outcome of older stroke patients after thrombectomy is mainly attributable to more extensive or more severe neurologically relevant brain damage, this should be reflected in age-related differences in neurological scores that parallel the age dependency of mRS outcome data.

    In this study, we examined this subject by analyzing and comparing technical results, pre- and post-treatment NIHSS scores, subsequent mRS outcome, and early and overall mortality in older and younger stroke patients with acute occlusion of the middle cerebral artery (MCA), treated by MTE.

    Patients and methods

    Single-center analysis was carried out of all consecutive patients (n=125; age range 18–96 years) with acute, isolated occlusion of the MCA, subjected to MTE between March 2007 and February 2012, including 40 ‘older’ patients (≥80 years). The threshold of 80 years was chosen, because most previous related studies, including recent large trials on MTE (MR CLEAN, ESCAPE), used the same cutoff value.1 ,3–7 ,10 ,14 ,15 ,18 ,19 Formalized data on living conditions and nursing requirements of the patients before the stroke were retrieved from the medical documentation system of our institution, allowing us to identify patients who had been ‘functionally independent at home’, corresponding to pre-stroke mRS scores of 0–2 (table 1). ‘Bridging’ thrombolysis with intravenous recombinant tissue plasminogen activator (rtPA) was applied in 83 patients (66%), approximately equally frequently in older and younger patients (73% and 64%, respectively, p=0.3). MTE, in most cases with stent-retrievers, was performed according to standard techniques described previuosly.20 Data for the technical results, and clinical outcome had been collected prospectively and were retrieved from the large-scale stroke database run by the departments of neurology and neuroradiology of our institution. Alberta Stroke Program Early CT scores (ASPECTS) in pretreatment CT-scans, and post-MTE recanalization status were reassessed by two experienced neuroradiologists blinded to the clinical data. Recanalization was graded according to the Thrombolysis in cerebral Infarction (TICI) scale, with grades 2b/3 defined as ‘successful’. NIHSS scores were assessed in all patients at presentation (NIHSS-pre) and after treatment before discharge (NIHSS-post) by trained neurologists. Improvements in post-treatment versus pretreatment NIHSS scores ≥4 and ≥8 were defined as ‘moderate’ and ‘substantial’, respectively. Mid-term mRS scores at ∼90 days were assessed by telephone interview or derived from the formalized documentation of the patient's rehabilitation centers, by two trained judges blinded to recanalization grading, with a high concordance rate (Cohen's unweighted κ=0.80). Discrepancies were resolved by consensus agreement. mRS scores of 0–2 were defined as ‘good outcome’. Overall, mRS scores were obtained for 103/125 patients (69/85 younger patients, 81.2%; 34/40 older patients, 85%; p=0.6), with similar proportions of scores derived from rehabilitation center documentation (39/69 vs 16/34; p=0.37). Twenty patients were known to be alive at 90 days after the incident stroke, but valid mRS scores could not be acquired. The remaining two patients were lost to follow-up after discharge, and were excluded from mid-term mRS and overall mortality analyses. All procedures were carried out in accordance with institutional guidelines and approved by the local ethics committee.

    Table 1

    Summary of baseline and technical and clinical outcome data

    Statistical analysis

    Dichotomized variables were analyzed by cross-tabulation and Fisher's exact test. Variables on an ordinal scales (NIHSS scores, mRS scores) were assessed by non-parametric tests (Mann–Whitney). All statistical analyses were carried out using SPSS V.21. p Values <0.05 were considered significant.


    Baseline data and technical and clinical results are summarized in table 1. NIHSS scores at presentation were similar, and there was an equal distribution between right-sided and left-sided strokes in older and younger stroke patients (table 1). Pre-treatment ASPECTS tended to be worse in younger patients, but this difference was not significant.

    Technical results

    Recanalization rates were almost identical in older and younger patients (TICI 2b/3 in 80% vs 82%, p=0.81). In those treated successfully, time from symptom onset to reperfusion and procedure durations were not significantly different (257±97 min in older vs 272±85 min in younger patients, p=0.5, and 58±52 min vs 66±38 min, p=0.46, respectively). There was no evidence of higher rates of procedural complications in older patients.

    Clinical outcome

    Symptomatic intracranial hemorrhage (SICH) occurred in two patients aged 72 and 76 years, respectively. There were no significant age-related differences in pre- and postinterventional NIHSS –scores (p=0.89 and 0.13, respectively, Mann–Whitney), but a highly significant difference in mid-term mRS scores (p<0.001, Mann–Whitney, figure 1A, B). As NIHSS-post was evaluated at discharge and not at a predefined time after the incident stroke, it is worth mentioning that (with cases of early, in-hospital deaths excluded) there were no age-related differences in the lengths of hospital stay (10.3 days in older vs 13.0 days in younger patients, p=0.79), and no significant correlation between lengths of hospital stay and postinterventional NIHSS scores (ρ=−0.07, p=0.45).

    Figure 1

    Summary of early neurological and mid-term modified Rankin Scale (mRS) outcome data. (A) Distributions of pre- and postinterventional National Institute of Health Stroke Scale (NIHSS) score for the entire patient sample and stratified according to recanalization success. Aged patients (≥80 years) are indicated by the black part of the stacked bars. (B) Mid-term mRS scores for all patients according to age. (C) Mid-term mRS scores according to age and recanalization success. Note reversed order of the bars in the lower panel for facilitating visual comparison between age groups for TICI 2b/3 patients. TICI, Thrombolysis in Cerebral Infarction.

    There was a stringent, quasi-linear relation between postinterventional NIHSS scores and subsequent mRS outcome in younger patients (figure 2A). For instance, none of the younger patients who had a postinterventional NIHSS score of ≤10, died, and all patients with a ‘substantial’ NIHSS score improvement (Δ NIHSS ≥8) had mid-term mRS scores of 3 or better. In the older patients, this stringent relation was lost (figure 2B). Here, the majority of patients with postinterventional NIHSS scores ≤10, including those with ‘substantial’ NIHSS score improvement, had mRS outcome scores of 4 or worse, and some even died during the 3 months’ observation period (figure 2B).

    Figure 2

    Early, postinterventional National Institute of Health Stroke Scale (NIHSS) scores plotted against mid-term modified Rankin Scale (mRS) scores for younger (A) and older patients (B). Data points from patients with identical NIHSS–mRS value pairs were slightly shifted along the mRS axis to avoid superimposition of symbols. Filled circles indicate patients with ‘substantial’ NIHSS score improvement (NIHSS pre—NIHSS post ≥8).

    Analyses restricted to patients who had been independent at home before the stroke yielded equivalent results. The mRS outcome in this subsample was somewhat better in both age groups, but still much worse in the older than in the younger patients (mRS 0–2 14% vs 60%, overall difference in mRS scores p<0.001, Mann–Whitney). In contrast, as in the entire group, there were no significant age-related differences in postinterventional NIHSS scores (p=0.13, Mann–Whitney), or postinterventional NIHSS improvement (Δ NIHSS ≥4 57% vs 75%, p=0.12, Δ NIHSS ≥8 39% vs 49%, p=0.48).


    Early, in-hospital death rates were similar (7.5% in older vs 9.4% in younger patients, p=0.73), but overall mortality was 2.7-fold higher in older patients (33% vs 12%, p<0.01). In patients who had been functionally independent, mortality was lower in both age groups, but still ∼2.6-fold higher in older than in younger patients (21% vs 8%). Eight of 10 documented case fatalities of the younger patients, but only 3 of 13 of the older patients, occurred during the acute in-hospital phase. At least five of these eight early fatalities in the younger patients were directly stroke-related: four died from large ‘malignant’ MCA infarctions, and one from intracranial hemorrhage. In contrast, the three acute phase fatalities in the older patients were caused by aspiration pneumonia, pulmonary embolism, and severe abdominal hemorrhage, respectively. At least four of the 10 delayed fatalities in the older patients were due to infections, notably pneumonia (n=3). There were insufficient data as to the precise cause of death in the remainder of the patients.

    Impact of recanalization

    NIHSS-post scores depended strongly on recanalization success in both older and younger patients. There was a prominent shift of the NIHSS-post distribution versus the NIHSS-pre distribution towards more favorable values after successful recanalization, but only a small shift after failed recanalization, in both age groups (figure 1A). Successful recanalization had a beneficial impact on mRS scores as well, but more so for younger patients (figure 1C). In particular, poor outcome scores (mRS 5–6) and overall mortality were substantially reduced in the younger, but not in the older patients. However, a clear benefit of recanalization was still retained in the older patients: good clinical outcome was achieved in 14.3% after successful recanalization, but in 0% after failed recanalization (in younger patients the figures were 62.5% vs 23.1%, respectively).


    These data indicate that older patients, with strokes of similar severity, similar vascular occlusion patterns (isolated MCA occlusion), and similar pretreatment ASPECTS, show a comparable early neurological course to that of younger patients. Moreover, with respect to their early neurological status, older patients were similarly responsive to MTE as younger patients. Technical results, procedure duration, and safety of MTE were all similar. SICH was equally rare, confirming previous studies.6 ,9–11 Post-treatment NIHSS scores, and the likelihood of NIHSS score improvement after successful recanalization, were slightly lower in older patients, but these differences were not significant. The comparable early neurological course, however, did not lead to similarly comparable mid-term mRS scores. In particular, overall mortality remained high, and good functional outcome was rare in older patients, often in spite of successful treatment and substantial early neurological recovery.

    The age dependency of mRS outcome scores is well documented,4–12 ,18 but data on early post-treatment neurological status in older patients are few and inconsistent. One study reported worse post-treatment (at day 7) than pretreatment NIHSS scores in older patients after intra-arterial thrombolysis and additional mechanical clot disruption by microwire manipulation,19 an endovascular approach which today is obsolete. A recent study reported higher NIHSS scores in older than younger patients after MTE with stent-retrievers in MCA and carotid-T occlusion.10 In that study, however, there were considerable sample asymmetries—with a preponderance of carotid-T occlusions in older (12/22), but not in younger patients (29/87), and a ∼30 min longer average symptom onset to recanalization time in the older subgroup—which may explain the observed differences in postinterventional NIHSS scores. In contrast, several other studies found no age dependency of—sometimes only imprecisely defined—postinterventional neurological improvement after IV rtPA5 and/or various modes of endovascular therapies,6 ,7 ,12 which is corroborated by our data.

    Diverse age-dependent factors have been identified that contribute to the poor functional mRS outcome of older stroke patients. These include pre-existing cognitive or physical functional impairments,18 leukoaraiosis,21 pre-existing age-related neuronal loss and diminished neuronal plasticity,22 ,23 socioeconomic factors such as restricted access to rehabilitation resources,24 and a higher rate of comorbidities and post-stroke complications (eg, infections, and injuries due to falls).25 Evidently, only some of these factors (eg, reduced vascular reserve in leukaraiotic tissue) may directly influence the extent of brain tissue damage induced by the incident stroke. Others, rather, may limit the functional restoration capacity in older stroke patients (eg, pre-existing impairment, reduced neuronal plasticity, and restricted rehabilitation resources). Still others are only indirectly, or not at all, related to stroke-induced neurological deficits (eg, post-stroke infections, comorbidities).

    The NIHSS, a dedicated neurological stroke scale, reflects stroke-induced neurological deficits more directly and specifically than a global functional score as the mRS. NIHSS scores correlate well with lesion volumes after acute ischemic stroke.26 ,27 Moreover, NIHSS score improvements after MTE have been shown to correlate well with the volume of tissue at risk salvaged by successful recanalization.28 The observed similarity in postinterventional NIHSS scores of older and younger patients thus indicates that both groups have a comparable degree of neurologically relevant brain damage. Factors that are directly related to the extent of stroke-induced brain tissue damage are therefore unlikely to explain much of the relatively poor functional mRS outcome of older patients. Conversely, it follows that the poorer outcome of older stroke patients is, to a substantive degree, due to the negative impact of factors not directly linked to the brain tissue damage induced by the incident stroke.

    This notion is supported by several observations. First, the discrepancy between similar, early neurological course and much poorer subsequent outcome in older as compared with younger patients had two main aspects: (a) a lower likelihood of achieving functional independence (mRS 0–2) in spite of similar early NIHSS scores, and (b) a high overall (delayed) mortality in contrast to similarly low early death rates. While the former (a) may be largely explained by factors such as pre-existing functional impairments, stroke-induced cognitive deficits (which may be under-rated by NIHSS scores29), and limited functional restoration capacities of the damaged ageing brain, such mechanisms are unlikely to account for (b), which is more plausibly explained by the relatively high prevalence of post-stroke complications and comorbidities in older patients. Second, and in line with this, none of even the three early fatalities in the older patients was directly linked to stroke-related brain damage (all three were due to medical complications), whereas five of eight early deaths in the younger patients were directly related to malignant infarction or SICH. Third, the stringent, quasi-linear relationship between early postinterventional neurological status and subsequent mRS outcome seen in the younger patients, was disrupted, rather than merely shifted towards less favorable mRS values, in the older patients (figure 2).

    The finding that the poorer outcome—in particular, the excess mortality—of older patients is substantially determined by factors only indirectly or unrelated to the incident stroke has two major implications. First, it calls for larger-scale systematic research to comprehensively assess the post-stroke clinical course of older patients in order to better understand these factors and their respective quantitative relevance. This will help to identify and prioritize potentially modifiable factors (such as post-stroke infections, other medical complications, and issues related to rehabilitation), and hopefully lead to improved preventive and therapeutic measures, allowing greater transference of the potential benefit of brain tissue salvaging therapies into clinically relevant, lasting functional improvement for this age group also. As acute stroke therapy is becoming more and more effective in averting or mitigating stroke-induced brain damage and neurological impairment, these issues should gain relevance. There is now strong evidence that MTE with stent-retrievers is highly effective, both technically and clinically, and significantly more effective than the current standard treatment (systemic thrombolysis with rtPA).14 ,15 ,30 Subgroup analyses of the recently published MR CLEAN and ESCAPE trials indicate that this also holds for older patients.14 ,15 MTE-treated older patients in these studies had on average significantly better mid-term mRS scores than their older controls. However, in comparison with MTE-treated younger patients, their outcome was much poorer, and their mortality rate ∼2.9-fold higher (in ESCAPE), resembling our data.15

    Remarkably, this held despite the fact that in ESCAPE, enrollment was confined to patients who had been functionally independent before the incident stroke (Barthel index ≥90). Corroborating these findings, our own analyses yielded consistent results when restricted to those older and younger patients who had been functionally independent. These observations indicate that pre-stroke morbidity alone cannot account for the relatively poor mid-term outcome and high mortality of older patients Our own limited data point to a prominent negative impact of post-stroke infections, but many other factors, as outlined above, are likely to play relevant roles, which need to be assessed by further research.

    Second, in older patients, mid-term mRS scores alone may not always be ideally suited to evaluation of acute therapies that are primarily intended to reduce brain tissue damage. Even if such therapies are reasonably effective, they cannot necessarily be expected to prevent post-stroke infections or other complications in older patients, nor to compensate for other age-dependent factors which have a potentially negative impact on mRS scores (eg, comorbidities, restrictions in rehabilitation resources). Such confounding influences may attenuate or even annihilate initial positive effects of tissue-salvaging therapies, rendering mid-term mRS scores relatively insensitive towards potential benefits that may eventually be turned into real, effective ones —for example, by future improvements in anti-infective treatments, post-stroke patient management, or geriatric medicine. In order not to miss or under-rate potential benefits that are less conspicuous than the extraordinarily strong effects of MTE and prematurely discard smaller, but potentially meaningful therapeutic improvements, trials of acute therapies in older patients should make use of specific neurological scores and, if possible, of objective surrogate markers such as lesion volumes, as secondary outcome measures supplementing (not replacing) mid-term mRS scores.

    This study has several limitations. Most importantly, a detailed analysis of the diverse factors that contribute to the poor outcome of older stroke patients is urgently needed, but such work is beyond the scope of this article. In addition, there are the common limitations of an observational study. Specifically, although most data (including TICI and NIHSS scores) were collected prospectively, some mRS scores are missing, and others were obtained from formalized medical case records rather than by standard direct or telephone interview. However, results were essentially equivalent, when analyses were restricted to patients for whom mRS scores had been obtained in the standard way. Moreover, as regards the observed age-dependency of mid-term mRS scores, this study only replicates robust findings from many previous studies, including recent well-controlled MTE trials.14 ,15 In contrast, the important finding focused on here is the discrepant age-independency of early NIHSS scores after MTE, which is not affected by this limitation. Finally, there were small, non-significant age-dependent differences in postinterventional NIHSS scores and NIHSS score improvement, which might have reached significance in a larger sample. Even then, however, the comparably minor differences in early NIHSS values would be very unlikely to fully account for the poor mRS outcome of the older patients, and could not explain their excess delayed mortality.

    In summary, we conclude that MTE in older patients may be (even) more effective in mitigating stroke-induced brain damage and neurological deficits than suggested by mid-term mRS scores alone. Further research is needed to understand, quantify, and eventually reduce the negative impact of the diverse factors that contribute to the poor mid-term outcome of older patients. This will help to transform more of the potential of tissue-salvaging therapies into effective, lasting functional benefits in this age group.


    We thank our colleagues from the Department of Neurology of our institution—notably, Dr S Wunderlich and PD Dr H Poppert, for excellent patient care and for contributing clinical data. We thank Dr D Frahm for help in the assessment of modified Rankin Scale scores. We also gratefully acknowledge the important support of PD Dr Ch Lücking, Schön Klinik München Schwabing, München, and we thank PD Dr S Waldt, Department of Radiology, Klinikum Rechts der Isar, München, for helpful comments.


    View Abstract


    • Collaborators Dr Holger Poppert, Department of Neurology, Klinikum Rechts der Isar, Technische Universität München, München, Germany.

    • Contributors JFK conceived the research, acquired and analyzed the data, and drafted and revised the paper. TB-B contributed to the research design, participated in data acquisition and analysis and revised the draft paper. SP and TL participated in data acquisition and data analysis and revised the draft paper. CZ participated in data analysis and revised the draft paper.

    • Competing interests None declared.

    • Ethics approval Ethics committee.

    • Provenance and peer review Not commissioned; externally peer reviewed.