Determinants of outcome and safety of intravenous rt-PA therapy in the very old: a clinical registry study and systematic review

SIR—The only curative treatment for acute ischemic cerebrovascular syndrome (ACVS) [ 1 ] is rt-PA administered intravenously within 3 h of symptom onset. In France, it has been licensed since 2002, but is not recommended for patients ≥80 years [ 2 ]. As most strokes occur in the elderly, and the life expectancy is increasing, the benefits and risks of thrombolytic treatment are worth evaluating in this group. However, with the exception of the second part of the NINDS rt-PA trial [ 1 ] in which 42 patients ≥80 years were included, patients in this age group have been excluded from most randomised controlled trials [ 3–6 ]. A number of recent clinical experiences have reported that 3-month outcomes are less favourable in patients ≥80 years than in those much younger, despite any evidence of an increased risk of symptomatic intracranial haemorrhage (SICH) [ 7 ]. Our aim was to evaluate the efficacy and safety of intravenous rt-PA in ACVS patients aged ≥80 years in comparison to the younger patients by examining data compiled in our prospective clinical registry and by combining our data with previously published post-licensing studies. At Bichat Hospital, a stroke team is on duty 24 h/day. Magnetic resonance imaging (MRI) is available during week-days and computed tomography (CT) is available at all times. Patients are treated by rt-PA according to the NINDS criteria [ 1 ] with the additional restrictions: early infarct signs > one third of the middle cerebral artery territory on CT scan and/or severely impaired consciousness (National Institutes of Health Stroke Scale (NIHSS) definition). The existence of microbleeds or the absence of diffusion weighted imaging (DWI)/perfusion weighted imaging (PWI) mismatch in MRI are not contraindications in our centre. Age is not a contraindication for this treatment and all patients are admitted if they are under the first 3 h of their symptoms. The severity of the ACVS was assessed using the NIHSS, with severe ACVS defined as a score ≥16. All patients had a follow-up CT/MRI scan 24 h after thrombolysis. A SICH was defined as the detection of a haemorrhage on the follow-up CT/MRI scan associated with a clinical deterioration ≥4 points in the NIHSS. Final outcome was assessed after 3 months by a neurologist, using the modified Rankin scale (mRs). We considered a mRs of 0–1 as favourable, 2–3 as moderate, and ≥4 or death as poor outcome. To identify previous studies comparing efficacy and safety of intravenous rt-PA between patients of ≥80 versus <80 years of age, we searched the MEDLINE database (January 1996 to December 2006) using OVID and the search terms (‘thrombolysis’ OR ‘tissue plasminogen activator’) AND ‘stroke’ without language restriction. Univariate comparisons were performed using Mann– Whitney U test or chi-square tests (Fisher's exact test was used when the expected cell frequency <5). Factors associated in univariate analysis ( P < 0.10) with poor outcome were included in multivariable logistic regression model; we introduced blood glucose level as a binary variable, using the median cut-off point of 6.7 mmol/l. We also tested the homogeneity of the relationship between poor outcome and age across severe and moderate ACVS. For Bichat registry and each previously published post-licensing study, we calculated the odds ratios (ORs) and 95% confidence intervals (CIs) for the association of unfavourable outcome and SICH with the older age group. Heterogeneity in ORs across studies was tested by Breslow–Day tests. Since we found no evidence of heterogeneity, we used the Mantel–Haenszel method to combine the ORs. Statistical testing was done at the two-tailed alpha level of 0.05, except for homogeneity tests where an alpha level of 0.10 was used. SAS software was used for Bichat Registry Analysis and Cochrane Collaboration's Review Manager software for the meta-analysis. Among the 131 consecutive patients with ACVS treated with intravenous rt-PA between January 2002 and March 2006, two men <64 years were lost to follow-up; neither had haemorrhagic complications and their NIHSS scores improved during the first 24 h by 6 and 3 points, respectively. Of the remaining 129 rt-PA-treated patients, 22 (17%) were ≥80 years old. There were less significantly men and current smokers in the older group (Table 1 ). Median admission NIHSS score was 18 (range: 7–23) in older and 13 (range: 4–24) in the younger ( P < 0.01). Similarly, patients classified as having severe ACVS or ACVS attributable to cardioembolic causes were more frequent among the older (Table 1 ). Baseline characteristics and outcomes of patients with ACVS treated by intravenous rt-PA at Bichat University Hospital NIHSS, National Institutes of Health Stroke Scale. The causes were blood pressure >185/110 mmHg ( n = 12), ASPECT score <7 ( n = 6), time over 3 h ( n = 11), NIHSS >22 ( n = 5). Baseline characteristics and outcomes of patients with ACVS treated by intravenous rt-PA at Bichat University Hospital NIHSS, National Institutes of Health Stroke Scale. The causes were blood pressure >185/110 mmHg ( n = 12), ASPECT score <7 ( n = 6), time over 3 h ( n = 11), NIHSS >22 ( n = 5). The rate of poor outcome was higher in the older than in the younger (64% versus 22%, P < 0.0001) whereas no difference in favourable outcome was found (27% versus 37%, P = 0.37). With regard to baseline characteristics, poor outcome was associated with high blood glucose level (median, 6.5 versus 7.4 mmol/l, P = 0.016) and severe ACVS (35% versus 78%, P < 0.0001). In multivariable analysis, the ORs (95% CIs) of poor outcome were 4.86 (1.58–14.91) for age ≥ 80 years, 5.93 (2.27–15.50) for severe ACVS and 2.36 (0.93–5.99) for high blood glucose level. Relationship between age group and poor outcome was modified by ACVS severity ( P = 0.09). If the ACVS was severe, the risk of poor outcome was higher in older than in younger [86.7% ( n = 13/15) versus 34.8% ( n = 16/46), P < 0.001], whereas if the ACVS was moderate the risk of poor outcome was similar in both groups [14.3% ( n = 1/7) versus 11.5% ( n = 7/61) P = 1.00]. Intracranial haemorrhage occurred in 35 patients (27%) including 11 SICH, without significant difference between age groups (Table 1 ). Fatal SICH resulted in four young patients and in two old ( P = 0.27). We identified ten studies relating to post-licensing experience with intravenous rt-PA for ACVS [ 8–17 ]. Of these studies, two [ 16 , 17 ] were excluded because they reported data for elderly patients only and one [ 14 ] because data were included in a national prospective cohort study [ 18 ] who also reported analysis by age group [ 9 ]. Baseline characteristics and outcomes of the seven included studies are shown in the appendix in the supplementary data on the journal website ( http://www.ageing.oupjournals.org/ ). Combining our results with those of the five studies that reported outcome at 3 months [ 8 , 9 , 11 , 12 ] the incidence of unfavourable outcome was significantly greater in older than younger patients (72.7% versus 59.3%, P < 0.0001), without heterogeneity across studies (Figure 1 ). In one study [ 10 ] the incidence of SICH was significantly higher in the older group, with a wide CI (Figure 1 ). In pooled analysis, we found no evidence that the incidence of SICH differed between the older and younger groups (6.1% versus 5.2%, P = 0.25). Odds ratios of unfavourable outcome at 3-month follow-up [ 8 , 9 , 11 , 12 ], and symptomatic intracranial haemorrhage [ 8–13 , 15 ] associated with age ≥ 80 years, in individual and combined studies The analysis of our data and published post-licensing studies has shown that the incidence of SICH in patients ≥80, treated by rt-PA for ACVS, was not significantly higher than in younger. Even though the 3-month outcome was less favourable, 27–38% of patients ≥80 years recovered completely or to a level of independence. We also found that outcome may be dependent on the initial severity of the ACVS. In the meta-analysis of data from published studies and our clinical registry, we found that the incidence of unfavourable 3-month outcome (mRs > 1) was significantly greater in the older group ( P < 0.0001). Nevertheless, 27.3% of patients ≥80 did have a favourable outcome, which is higher than the 4% of patients ≥75 of the NINDS study placebo group [ 1 ]. In most of the studies, the benefit was not attenuated by an excess risk of SICH, which varied, from 4.3–13.6% in the old to 2.0–8.4% in the young. The natural history of stroke in the very old is associated with a poor prognosis [ 19–21 ]. Given this poor natural evolution of ACVS in the elderly, our meta-analysis shows that rt-PA treatment could benefit patients in this age group, even if the benefit is not as clinically significant as in the younger. A high admission NIHSS is associated with a poor 3-month prognosis [ 22 ]; nevertheless, some patients >80 benefit from treatment with rt-PA. In our registry, in patients with a high admission NIHSS [ 7 , 16 ], younger ones were more likely to recover than older ones. These results concur with findings in the NINDS trial, in which none of the 49 patients ≥75 years with a baseline NIHSS >20 had a favourable outcome [ 23 ]. However, in patients with an admission NIHSS <16, young and old patients had similar outcomes. In our meta-analysis, it was not possible to evaluate the impact of initial NIHSS on outcome, as only one study reported the data [ 14 ]. This study also found that severe NIHSS score at admission was associated with poor outcome. As in our clinical registry, there was no difference between groups such as those with a moderate NIHSS score at baseline. One limitation of our study is that while patients≥80 years represent 30% of the stroke population, they represented only 17% of our clinical registry and 21% of the meta-analysis population. This difference suggests a selection bias; even if age was not a contraindication, it is possible that only those with good health until then were considered for rt-PA treatment. Ideally, studies should compare rt-PA treatment with placebo treatment in this age group. In conclusion, our results suggest that patients≥80 and who are treated with rt-PA for ACVS have the same incidence of SICH as younger patients. Although older patients are less likely to have a favourable outcome after rt-PA treatment, this seems to be related to other co-morbidities. Until a randomised controlled trial is conducted in this age group, our data and meta-analysis support the recommendation that patients≥80 years should not be excluded systematically from rt-PA therapy. Patients ≥80 years receiving intravenous thrombolysis had a worst outcome than patients <80. They had a similar rate of favourable outcome The likelihood of symptomatic haemorrhage did not differ between groups Overall, the weight of evidence supports the recommendation that very old patients should not be systematically excluded from rt-PA thrombolysis. This should prompt for future randomised controlled trials. We have no conflict of interest to declare