Diabetes Mellitus, Admission Glucose, and Outcomes After ...stroke.ahajournals.org/content/strokeaha/44/7/1915.full.pdf · Desilles et al Diabetes Mellitus and Outcomes After Thrombolysis
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1915
Up to 60% of acute ischemic stroke (AIS) patients will experience hyperglycemia, depending on its definition.1,2
Poststroke hyperglycemia (PSH) is associated with poor clinical outcome and death2–4 and is commonly considered to represent an underlying impaired glucose tolerance or unrec-ognized diabetes mellitus. More often this process is the result of stress hyperglycemia,1,2 which is usually defined as a hyper-glycemia resolving spontaneously after acute illness dissipa-tion.5 Several mechanisms have been proposed to explain the observed adverse effects of PSH on clinical outcome, includ-ing increased blood–brain barrier disruption with higher hem-orrhagic risk6,7; or increased lactic acid production in ischemic tissue leading to a greater infarct size.8
In patients with AIS treated with intravenous (IV) thrombolysis (IVT), several large studies reported an association between diabetes mellitus or PSH and unfavorable outcome, hemorrhagic transformation, and death after IVT.9,10 Other series showed that PSH or diabetes mellitus were associated with lower recanalization rates in patients with stroke treated with IVT.11,12 However, a recent study found that IVT-treated patients with PSH did not have significantly worse outcomes.13 Data on PSH or diabetes mellitus in endovascular therapy–treated patients are limited.14–19 Available studies have shown conflicting results with respect to a potential detrimental effect of diabetes mellitus and PSH on clinical outcome. In opposition to the aforementioned IVT studies,
Background and Purpose—The potential detrimental effect of diabetes mellitus and admission glucose level (AGL) on outcomes after stroke thrombolysis is unclear. We evaluated outcomes of patients treated by intravenous and/or intra-arterial therapy, according to diabetes mellitus and AGL.
Methods—We analyzed data from a patient registry (n=704) and conducted a systematic review of previous observational studies. The primary study outcome was the percentage of patients who achieved a favorable outcome (modified Rankin score ≤2 at 3 months).
Results—We identified 54 previous reports that evaluated the effect of diabetes mellitus or AGL on outcomes after thrombolysis. In an unadjusted meta-analysis that included our registry data and previous available observational data, diabetes mellitus was associated with less favorable outcome (odds ratio [OR], 0.76; 95% confidence interval [CI], 0.73–0.79) and more symptomatic intracranial hemorrhage (OR, 1.38; 95% CI, 1.21–1.56). However, in multivariable analysis, diabetes mellitus remained associated with less favorable outcome (OR, 0.77; 95% CI, 0.69–0.87) but not with symptomatic intracranial hemorrhage (OR, 1.11; 95% CI, 0.83–1.48). In unadjusted and in adjusted meta-analysis, higher AGL was associated with less favorable outcome and more symptomatic intracranial hemorrhage; the adjusted OR (95% CI) per 1 mmol/L increase in AGL was 0.92 (0.90–0.94) for favorable outcome, and 1.09 (1.04–1.14) for symptomatic intracranial hemorrhage.
Conclusions—These results confirm that AGL and history of diabetes mellitus are associated with poor clinical outcome after thrombolysis. AGL may be a surrogate marker of brain infarction severity rather than a causal factor. However, randomized controlled evidences are needed to address the significance of a tight glucose control during thrombolysis on clinical outcome. (Stroke. 2013;44:1915-1923.)
Received January 14, 2013; accepted April 9, 2013.From the Department of Neurology and Stroke Centre, Bichat University Hospital, Paris, France (J.-P.D., E.M., B.L., G.S., J.G.V., P.L., L.C., C.G., P.A.,
M.M.); INSERM U-698 and Paris-Diderot University, Paris, France (J.-P.D., E.M., J.L., B.L., P.C.L., L.C., C.G., I.K., P.A., M.M.); and the Department of Radiology, Bichat University Hospital, Paris, France (I.K.).
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA. 111.000813/-/DC1.
Correspondence to M. Mazighi, MD, PhD, Department of Neurology and Stroke Centre, Bichat University Hospital, 46, rue Henri Huchard, 75018 Paris, France. E-mail [email protected]
endovascular research reported no association between successful reperfusion therapy defined as a partial or complete recanalization and diabetes mellitus or PSH.11
We therefore evaluated whether diabetes mellitus history, or admission glucose level (AGL) in patients with AIS treated by IV, intra-arterial (IA) thrombolysis (IAT) therapy may be associated with worse outcomes, by examining data from our prospective clinical registry and by conducting a systematic review of previous observational studies.
MethodsBichat Clinical RegistryStroke ProgramWe identified patients from a prospective clinical registry of patients with AIS treated between February 2002 and February 2012 at Bichat University Hospital, Paris, France. As described previously,20 before April 2007, all patients had been treated with conventional IVT, and none with an IA approach. After April 2007, patients eligible for IVT were treated with conventional IVT in case of no documented arterial occlusion and with a systematic IV–IA approach in case of documented arterial occlusion.20 Patients not eligible for IV treat-ment, with a documented arterial occlusion, were treated by the IA approach. The IA approach was performed using an IA recombinant tissue plasminogen activator dose of 0.5 mg/kg,21 followed by ad-junctive mechanical endovascular therapy if the arterial occlusion persisted. In patients with a contraindication to recombinant tissue plasminogen activator,22 a direct mechanical endovascular therapy ap-proach was considered.
Standard Protocol Approvals, Registration, and Patient ConsentsInformed consent was obtained from the patient or their representa-tive, and the research protocol was approved by the Ethics Committee from Ambroise Paré Hospital.
Data Collection and DefinitionsInformation on patients’ demographic characteristics, medical histo-ry, current medications, laboratory and imaging findings, vital signs before treatment, National Institutes of Health Stroke Scale (NIHSS) scores (at admission, and at 1, 3, and 24 hours after the initiation of treatment),23 clinical outcomes were collected prospectively using a structured questionnaire. Times from symptom onset (or from when the patient was last seen in a normal condition) to initiation of treat-ments were also recorded. All patients had a computed tomography or MRI scan 24 hours after treatment onset to assess hemorrhagic com-plications. For patients receiving IA therapy, the recanalization status of the occluded artery was monitored with conventional angiography, and the time to recanalization was noted. Recanalization was mea-sured with the Thrombolysis In Myocardial Infarction (TIMI) score24 by 2 members of staff (E.M. and M.M.), and was used for all imaging modalities. Modified Rankin Scale at 3 months was assessed during face-to-face interviews or via telephone calls by a senior vascular neurologist (E.M. or M.M.), who was certified for modified Rankin Scale scoring.25
Outcome DefinitionsThe primary study outcome was the percentage of patients who achieved a favorable outcome, defined as an modified Rankin Scale score of 0 to 2 at 90 days. Secondary outcomes included excellent outcome (defined as an modified Rankin Scale score of 0–1 at 90 days), early neurological improvement (defined as an NIHSS score of 0–1 at 24 hours or a decrease of ≥4 points in NIHSS score at 24 hours), any recanalization (defined by TIMI 2–3), complete recanalization (defined by TIMI 3), 90-day mor-tality, hemorrhagic complications, and symptomatic intracerebral
hemorrhage (defined as a hemorrhage on the follow-up computed tomography/MRI scan associated with an increase of ≥4 points in NIHSS score).26
Systematic ReviewSearch Strategy and Study SelectionWe performed a computerized PubMed search of articles pub-lished between January 1996 and May 2012 to identify all obser-vational and interventional studies that investigated the effect of history of diabetes mellitus or AGL on efficacy or safety outcomes in patients with AIS treated by an IV or IA strategy. We used the search terms (thrombolysis OR fibrinolysis OR thrombolytic OR intravenous OR intra-arterial OR intra-arterial OR endovascular OR mechanical OR thrombectomy OR clot removal OR clot dis-ruption) AND stroke, without any language restriction. One author (J.L.) selected potentially relevant articles based on the title and abstract, and obtained the full text for detailed review. We also searched the reference lists of retrieved articles and published re-view articles for additional studies. Studies were selected using the following criteria: (1) if they involved subjects aged ≥18 years; (2) were retrospective or prospective (observational or interven-tional) studies with ≥100 patients with AIS treated by IV or IA approach; and (3) reported a statistical analysis on the association of outcomes (efficacy or safety) with history of diabetes mellitus or AGL. We did not select studies according to treatment strat-egy, time to treatment, or the reported outcome definitions. We screened duplicate publications based on the same data sets (ie, where data overlapped with data in other included studies). When multiple end points were reported in separate publications, or when impacts of diabetes mellitus history and AGL were reported in separate publications, duplicate publications were included. For other duplicate publications, only the report with the most com-plete data was included.
Data ExtractionTwo authors (J.L. and J.P.D.) independently extracted data using a standardized form, and any disagreement was resolved by con-sensus. We did not contact the authors of the studies to request incomplete or unpublished data. We collected the following data: report characteristics (first author’s name, journal, year of publica-tion); study design (country, study period, number of centers, AIS treatment [IV agents, IA chemical agents, IA mechanical therapy]); study sample (sample size, age, sex, admission NIHSS, time from symptom onset to treatment), incidence and definition of outcomes (functional outcome, neurological improvement, recanalization, mortality, intracranial hemorrhage), direction of association be-tween diabetes mellitus (or AGL) and each outcome, and, if avail-able, a measure of the strength of the association in univariate and multivariate analysis.
Statistical AnalysisStatistical testing was done at the 2-tailed α level of 0.05, except in tests for homogeneity in which an α level of 0.10 was used. Data were analyzed using the SAS software version 9.3 (SAS Institute, Cary, NC).
Bichat Clinical RegistryWe made univariate comparisons between patients with and without history of diabetes mellitus, using χ2 tests for categorical variables and Student t test for continuous variables (Mann–Whitney U test was used for non-Gaussian distributions). Associations between baseline characteristics and AGL were performed using linear regression analysis. We investigated the effects of diabetes mellitus history and AGL on clinical and recanalization outcomes using a logistic regression model adjusted for potential confounding factors selected on the basis of their significance with diabetes mellitus or glucose in univariate analyses (P<0.20). To check the linearity of the
by guest on June 21, 2018http://stroke.ahajournals.org/
Desilles et al Diabetes Mellitus and Outcomes After Thrombolysis 1917
associations of AGL and outcomes, we categorized the AGLs into quartiles and calculated odds ratios (ORs; and their 95% confidence intervals [CIs]) for the upper 3 quartiles relative to the lowest quartiles. Because ORs increased gradually with AGL quartiles, we also computed OR per 1 SD increase in AGL. Finally, among patients
who achieved recanalization after IA therapy, we investigated the association of time to recanalization with diabetes mellitus history and AGL using Mann–Whitney U test or Spearman rank correlation analysis, respectively. Adjustment for confounding factors was done using nonparametric covariance analysis.
Table 1. Bichat Clinical Registry: Baseline Characteristics According to History of Diabetes Mellitus
History of Diabetes Mellitus
P ValueNo (n=600) Yes (n=104)
Age, y, mean (±SD) 68.3 (±17.2) 71.6 (±11.5) 0.015
Men, n (%) 326 (54.3) 69 (66.4) 0.023
Medical history
Hypertension, n (%) 299 (50.1) 73 (70.2) <0.001
Hypercholesterolemia, n (%) 165 (27.8) 49 (48.0) <0.001
Current or former smokers, n (%) 221 (38.2) 41 (42.3) 0.44
Antithrombotic medication, n (%) 233 (38.8) 48 (46.2) 0.16
Clinical measures
Platelet count, 1000/µL, mean (±SD) 230 (±71) 228 (±73) 0.82
SBP, mm Hg, mean (±SD) 151 (±22) 155 (±22) 0.12
DBP, mm Hg, mean (±SD) 81 (±13) 80 (±14) 0.82
NIHSS, median (IQR) 12 (6–18) 11 (6–17) 0.72
Documented arterial occlusion,* n (%)
None 216 (37.6) 47 (48.5) 0.22
ICA isolated or tandem with MCA 90 (15.7) 14 (14.4)
MCA isolated 244 (42.4) 32 (33.0)
Posterior circulation 25 (4.4) 4 (4.1)
AIS treatment, n (%)
IV alone 390 (65.0) 74 (71.2) 0.44
Combined IV–IA 141 (23.5) 19 (18.3)
IA alone 69 (11.5) 11 (10.6)
Mechanical clot removal, n (%) 76 (12.7) 15 (14.4) 0.62
Onset to treatment time, min, median (IQR)
158 (120–195) 166 (134–194) 0.39
AIS indicates acute ischemic stroke; DBP, diastolic blood pressure; IA, intra-arterial; ICA, internal carotid artery; IQR, interquartile range; IV, intravenous; MCA, middle cerebral artery; NIHSS, National Institutes of Health Stroke Scale; and SBP, systolic blood pressure.
*Thirty-two patients had no arterial examination before AIS treatment (transcranial Doppler, magnetic resonance angiography, or computed tomography angiography).
Table 2. Bichat Clinical Registry: Outcomes According to History of Diabetes Mellitus
Outcome, n (%)
History of Diabetes Mellitus
P Value* OR (95% CI)† P Value†No (n=600) Yes (n=104)
CI indicates confidence interval; mRS, modified Rankin Scale; OR, odds ratio; and sICH, symptomatic intracranial hemorrhage.*χ2 test.†Adjusted for age, sex, admission National Institutes of Health Stroke Scale (NIHSS), hypertension, hypercholesterolemia, antithrombotic therapy, and admission
systolic blood pressure (logistic regression analysis).‡Early neurological improvement defined as NIHSS score 0–1 or a decrease of ≥4 points in NIHSS score at 24 h.§Defined following the European Cooperative Acute Stroke Study (ECASS) II criteria.
by guest on June 21, 2018http://stroke.ahajournals.org/
Systematic ReviewWe reported AGLs using the conventional unit (mg/dL). When neces-sary, we converted the standard unit (mmol/L) to the conventional unit by dividing by 0.0555. Using all available data, we calculated for each study, the unadjusted effect size for diabetes mellitus as ORs (and their 95% CIs) of functional outcome and symptomatic intracra-nial hemorrhage (sICH), using patients without history of diabetes mellitus as reference group. Because a majority of retrieved studies reported means (±SD) of AGL according to presence or absence of outcome, we calculated the unadjusted effect size as the standardized mean difference between patients with and without good outcome, or between patients with and without sICH. When means (±SD) were not available, we approximate the effect size using medians (interquartile range) or converting the OR associated with glucose cut-off level.27 We combined the available unadjusted effect sizes using the inverse-variance weighted random-effects model; hetero-geneity between studies was examined using the χ2 test for homoge-neity followed by the calculation of the I2 statistic. We performed a sensitivity analysis by excluding data from the Safe Implementation of Treatments in Stroke–International Stroke Thrombolysis Register (SITS-ISTR),9 because they contributed to >70% of combined data and because we cannot exclude the possibility that several studies in our meta-analysis may have participated in SITS-ISTR. Sensitivity
analyses restricted to the studies analyzed our primary study outcome (ie, favorable outcome), or those restricted to the studies where the effect size was not approximated were also reported. Finally, we ex-tracted and combined the available multivariate results; for AGL, we calculated the adjusted OR per 18 mg/dL (=1 mmol/L) increase in glucose values.
ResultsBichat Clinical RegistrySeven hundred thirty-three consecutive patients with AIS were treated with IV or IA therapy. Of these, 29 were excluded because no information on history of diabetes mellitus could be obtained (n=4), or missing AGL (n=18) or patients were lost to follow-up after discharge (n=7; Figure I in the online-only Data Supplement); finally, 709 patients were analyzed. In the study sample, 104 patients (15%) had history of dia-betes mellitus, and the mean (SD) AGL was 133 (45) mg/dL. Baseline characteristics of patients with and without history of diabetes mellitus are described in Table 1. Patients with dia-betes mellitus were older, more often men, and more often
Table 3. Bichat Clinical Registry: Outcomes According to Quartile of Admission Glucose Level
AGL, mg/dL P Value for Trend or
<103 (n=179) 103–121 (n=181) 122–148 (n=167) >148 (n=177) OR per SD
AGL indicates admission glucose level; CI, confidence interval; mRS, modified Rankin Scale; OR, odds ratio; and sICH, symptomatic intracranial hemorrhage.*Adjusted for age, sex, admission National Institutes of Health Stroke Scale (NIHSS), hypertension, hypercholesterolemia, antithrombotic therapy, and admission
systolic blood pressure.†Early neurological improvement defined as NIHSS score 0–1 or a decrease of ≥4 points in NIHSS score at 24 h.‡Defined following the ECASS II criteria.
by guest on June 21, 2018http://stroke.ahajournals.org/
Desilles et al Diabetes Mellitus and Outcomes After Thrombolysis 1919
had treated hypertension and hypercholesterolemia than their counterparts. The same baseline characteristics were associ-ated with AGL (Table I in the online-only Data Supplement); prior use of antithrombotic medications, admission systolic blood pressure, and NIHSS score were also associated with AGL.
Diabetes Mellitus and OutcomesIn univariate analysis as well as multivariate analysis, we found no significant differences for all study clinical outcomes regarding diabetes mellitus history (Table 2). In addition, among the 240 patients who received IA therapy, there was no significant difference in recanalization rates between the diabetes mellitus subgroups. Rate of any recanalization (TIMI 2–3) was 66.7% in patients with history of diabetes mellitus and 74.3% in the group without (P=0.38); complete recanali-zation rates were, respectively, 40.0% and 49.5% (P=0.33). In multivariate analysis, the OR associated with diabetes mellitus
was 0.89 (95% CI, 0.37–2.12; P=0.79) for any recanalization and 0.97 (95% CI, 0.42–2.22; P=0.94) for complete recana-lization. In the subset of patients who achieved any recana-lization (n=176), the median time from symptom onset to recanalization was 285 minutes (interquartile range, 255–311 minutes) in patients with history of diabetes mellitus and 245 minutes (interquartile range, 212–333 minutes) in the group without (unadjusted/adjusted P=0.080/0.052).
Admission Blood Glucose Levels and OutcomesWhatever the clinical outcomes, higher AGLs were associated with a decreased rate of good outcomes (Table 3). Conversely, higher AGLs were associated with an increased rate of adverse outcomes (90-day mortality and hemorrhagic compli-cations). In multivariate analysis, AGL remained significantly associated with all outcomes, except with early neurologi-cal improvement (adjusted OR per 1 SD increase, 0.83; 95% CI, 0.66–1.04; P=0.10). In patients who received IA therapy,
Figure 1. Forest plots of unadjusted odds ratios (ORs) for (A) good outcome and (B) symptomatic intracranial hemorrhage associated with diabetes mellitus history. Ellipses denote not available. CI indicates confidence interval. References to the studies listed are provided in the online-only Data Supplement.
by guest on June 21, 2018http://stroke.ahajournals.org/
AGL were not significantly associated with recanalization. The adjusted OR per 1 SD increase in glucose level was 1.08 (95% CI, 0.78–1.50; P=0.63) for any recanalization and 1.20 (95% CI, 0.89–1.61; P=0.23) for complete recanalization. In the subset of patients who achieved any recanalization, AGL tended to be correlated with higher time to recanalization (unadjusted/adjusted P=0.054/0.095)
Systematic ReviewThe literature search identified 15 735 citations, of which 218 full articles were read and 55 were judged eligible for inclu-sion in the systematic review (Figure II in the online-only Data Supplement). The main baseline characteristics, treatment strategies, and outcomes (definitions and incidences), with a brief description of direction of association between diabetes mellitus and outcomes or between AGL and outcomes are summarized in Table II in the online-only Data Supplement.
Diabetes Mellitus and OutcomesThe available unadjusted estimates of the effect of diabetes mellitus on functional outcome and sICH in individual
and combined studies (including our results) are shown in the Figure 1. Combining the available unadjusted results for functional outcome (19 studies, n=21 368), the incidence of good outcome was significantly lower in patients with history of diabetes mellitus than in those without (Figure 1A). A similar result was found in sensitivity analysis excluding SITS-ISTR (combined OR, 0.65; 95% CI, 0.56–0.76), or restricted to studies that analyzed favorable outcome (combined OR, 0.77; 95% CI, 0.73–0.80) or combining the available multivariate ORs (6 studies; n=19 364; Table 4). When combining the available unadjusted results for sICH outcome (13 studies; n=21 564), we found a significant increase in sICH risk for patients with history of diabetes mellitus (Figure 1B). A similar result was found when the National Institute of Neurological Disorders and Stroke (NINDS) definition was used in SITS-ISTR (combined OR, 1.38; 95% CI, 1.21–1.56) or when this study was excluded (combined OR, 1.57; 95% CI, 1.19–2.07). However, when we combined the available multivariable results (5 studies, n=18 478), the association between diabetes mellitus and sICH was no longer significant (Table 4).
Table 4. Individual and Combined ORs for Good Outcome and sICH Associated With Diabetes Mellitus History and Admission Glucose Levels in Multivariable Analysis
Combined without Ahmed (2010)9 0.90 (0.87–0.93) <0.001 1.14 (1.05–1.24) 0.002
CI indicates confidence interval; OR, odds ratio; and sICH, symptomatic intracranial hemorrhage.*Please see Table II in the online-only Data Supplement for the outcome definition and references.†ORs calculated per 18 mg/dL (=1 mmol/L) increase in admission glucose level.
by guest on June 21, 2018http://stroke.ahajournals.org/
Desilles et al Diabetes Mellitus and Outcomes After Thrombolysis 1921
AGLs and OutcomesCombining the available unadjusted effect sizes (calculated as the standardized mean differences; 21 studies, n=22 042), patients with good outcome had a significantly lower AGL than patients without good outcome (Figure 2A). A simi-lar result was found in sensitivity analysis excluding SITS-ISTR (pooled effect size, −0.26; 95% CI, −0.18 to −0.33), or restricted to studies that analyzed favorable outcome (pooled effect size, −0.26; 95% CI, −0.17 to −0.36), or excluding the approximate effect sizes (pooled effect size, −0.28; 95% CI, −0.20 to −0.36). Regarding sICH outcome, patients with sICH had a significantly higher AGL than patients without sICH (Figure 2B). A similar difference was found when the NINDS definition was used in SITS-ISTR (pooled effect size, 0.30; 95% CI, 0.19–0.40) or when this study was excluded (pooled effect size, 0.28; 95% CI, 0.14–0.42), or when the approxi-mate effect sizes were excluded (pooled effect size, 0.33; 95%
CI, 0.18–0.47). When combining the available multivariable results (calculated as ORs per 18 mg/dL of glucose), AGL remained associated with poor outcome and sICH (Table 4).
DiscussionIn the present meta-analysis, history of diabetes mellitus and AGL were associated with lower rate of good outcome after thrombolysis. In addition, AGL was significantly associ-ated with increased risk of sICH, which was not observed in patients with history of diabetes mellitus. In patients eligible for IA thrombolysis in our clinical registry, there was no docu-mented effect of AGL on recanalization rates.
Arterial recanalization grade and speed has been reported to be one of the strongest predictors of favorable outcome after thrombolysis.20 In 2 previous IVT studies, AGL11 or diabetes mellitus 12 were associated with lower recanalization rates, suggesting an impaired fibrinolytic response in the setting
Figure 2. Forest plots of unadjusted standardized mean difference (SMD) In admission glucose level between patients with and without good outcome (A) and between patients with and without symptomatic intracranial hemorrhage (sICH; B). aApproximate using median and interquartile range. bApproximate using the odds ratio of outcome associated with a cut-off of glucose level. Ellipses denote not avail-able. CI indicates confidence interval.
by guest on June 21, 2018http://stroke.ahajournals.org/
of elevated blood glucose concentration. Evidence supports the contribution of both chronic and acute hyperglycemia to coagulation activation,28,29 whereas hyperinsulinemia decreases fibrinolytic activity by increasing the production of plasminogen activator inhibitor.30,31 In contrast with these data, we did not find any association between recanalization and AGL or diabetes mellitus in IAT-treated patients. These results are in accordance with other previous IAT studies.14–19 The antifibrinolytic effect of hyperglycemia may be overcome by direct IAT and mechanical endovascular techniques explaining the difference between the IVT and IAT studies.32 The poorer clinical prognosis associated with AGL in case of IAT would therefore not be related to a recanalization rate difference, suggesting other underlying mechanisms.
Hyperglycemia and diabetes mellitus are strongly linked to increased risk of sICH after thrombolytic treatment.33,34 Experimental studies showed that diabetes mellitus and hyperglycemia are associated with blood–brain barrier and microvasculature impairments, as well as increased hemorrhagic infarct conversion after reperfusion.35 Recently, blood–brain barrier disruption after IAT was associated with AGL36 suggesting that the detelerious effect on clinical outcome of hyperglycemia may be secondary to a blood–brain barrier disruption and subsequent sICH. Hyperglycemia is also independently associated with reduced salvage of perfusion–diffusion lesion mismatch tissue, larger final infarct volume, and poor clinical outcome.4 This effect on the penumbra might also explain why hyperglycemia is not associated with worse outcome in lacunar stroke because a penumbra is usually not present in this subtype of stroke.37 Hyperglycemia can cause a worse clinical outcome despite recanalization with a markedly larger increase of the infarction volume.38 Although restoration of the blood flow to the ischemic tissue is essential for penumbral salvage, reperfusion itself can also induce injury; and hyperglycemia is associated with increased reperfusion injury.8,39 Despite all these supporting data, and pending the results of ongoing randomized trials, such as stroke Hyperglycemia Insulin Network Effort (SHINE) trial,40 to date, no clinical trial shows that glucose-lowering treatment improves clinical outcome in patients with AIS.41,42 Hyperglycemia could therefore be only a surrogate marker of larger infarct volume, explaining why interventional studies on hyperglycemia were all inconclusive.
The major limitations of our analysis were the use of obser-vational data, including: the lack of available adjusted results or adequate data to estimate the adjusted effect sizes in a sig-nificant number of studies (which was explained by the dif-ference in objectives between the present systematic review and previous studies); the absence of data on glucose levels time course, HbA1C values, diabetes mellitus’ duration, and type of diabetes mellitus treatment; and the intrinsic biases in baseline characteristics of nonrandomized studies. In patients treated by the endovascular approach, other unavailable data may influence clinical prognosis, such as infarct volumes or collateral flow. Due to multiple comparisons (particularly done in our registry analysis), we cannot exclude the pos-sibility of false-positive association. Finally, the presence of microbleeds or leukoaraiosis, conditions related to hemor-rhagic risk,43 were not documented in our registry or in the
retrieved studies, which is a potential source of heterogeneity in our findings.
In summary, AGL and history of diabetes mellitus are asso-ciated with poor clinical outcome in AIS patients treated by IV or IA thrombolysis. However, the causal relationship between AGL and worse prognosis remains undetermined and needs randomized controlled evidences, such as the ongoing SHINE trial. In fact, only data brought by a large multicenter trial may indicate the role of therapeutic intervention on glycemia in the setting of acute stroke, if any.
Sources of FundingFunding for this study was provided in part by SOS-ATTAQUE CEREBRALE and supported by the Département Hospitalo-Universitaire FIRE (Fibrosis Inflammation Remodeling) of Université Paris-Diderot, France.
DisclosuresNone.
References 1. Scott JF, Robinson GM, French JM, O’Connell JE, Alberti KG, Gray
CS. Prevalence of admission hyperglycaemia across clinical subtypes of acute stroke. Lancet. 1999;353:376–377.
2. Capes SE, Hunt D, Malmberg K, Pathak P, Gerstein HC. Stress hyperglycemia and prognosis of stroke in nondiabetic and diabetic patients: a systematic overview. Stroke. 2001;32:2426–2432.
3. Parsons MW, Barber PA, Desmond PM, Baird TA, Darby DG, Byrnes G, et al. Acute hyperglycemia adversely affects stroke outcome: a magnetic resonance imaging and spectroscopy study. Ann Neurol. 2002;52:20–28.
4. Mazighi M, Labreuche J, Amarenco P. Glucose level and brain infarc-tion: a prospective case-control study and prospective study. Int J Stroke. 2009;4:346–351.
6. Kamada H, Yu F, Nito C, Chan PH. Influence of hyperglycemia on oxida-tive stress and matrix metalloproteinase-9 activation after focal cerebral ischemia/reperfusion in rats: relation to blood-brain barrier dysfunction. Stroke. 2007;38:1044–1049.
7. Ennis SR, Keep RF. Effect of sustained-mild and transient-severe hyper-glycemia on ischemia-induced blood-brain barrier opening. J Cereb Blood Flow Metab. 2007;27:1573–1582.
8. Suh SW, Shin BS, Ma H, Van Hoecke M, Brennan AM, Yenari MA, et al. Glucose and NADPH oxidase drive neuronal superoxide formation in stroke. Ann Neurol. 2008;64:654–663.
9. Ahmed N, Dávalos A, Eriksson N, Ford GA, Glahn J, Hennerici M, et al; SITS Investigators. Association of admission blood glu-cose and outcome in patients treated with intravenous thromboly-sis: results from the Safe Implementation of Treatments in Stroke International Stroke Thrombolysis Register (SITS-ISTR). Arch Neurol. 2010;67:1123–1130.
10. Bruno A, Levine SR, Frankel MR, Brott TG, Lin Y, Tilley BC, et al; NINDS rt-PA Stroke Study Group. Admission glucose level and clinical outcomes in the NINDS rt-PA Stroke Trial. Neurology. 2002;59:669–674.
11. Ribo M, Molina C, Montaner J, Rubiera M, Delgado-Mederos R, Arenillas JF, et al. Acute hyperglycemia state is associated with lower tPA-induced recanalization rates in stroke patients. Stroke. 2005;36:1705–1709.
12. Zangerle A, Kiechl S, Spiegel M, Furtner M, Knoflach M, Werner P, et al. Recanalization after thrombolysis in stroke patients: predictors and prognostic implications. Neurology. 2007;68:39–44.
13. Meurer WJ, Scott PA, Caveney AF, Majersik JJ, Frederiksen SM, Sandretto A, et al. Lack of association between hyperglycaemia at arrival and clinical outcomes in acute stroke patients treated with tissue plas-minogen activator. Int J Stroke. 2010;5:163–166.
14. Galimanis A, Jung S, Mono ML, Fischer U, Findling O, Weck A, et al. Endovascular therapy of 623 patients with anterior circulation stroke. Stroke. 2012;43:1052–1057.
15. Christoforidis GA, Karakasis C, Mohammad Y, Caragine LP, Yang M, Slivka AP. Predictors of hemorrhage following intra-arterial thrombolysis
by guest on June 21, 2018http://stroke.ahajournals.org/
Desilles et al Diabetes Mellitus and Outcomes After Thrombolysis 1923
for acute ischemic stroke: the role of pial collateral formation. AJNR Am J Neuroradiol. 2009;30:165–170.
16. Nogueira RG, Liebeskind DS, Sung G, Duckwiler G, Smith WS; MERCI; Multi MERCI Writing Committee. Predictors of good clinical outcomes, mortality, and successful revascularization in patients with acute ischemic stroke undergoing thrombectomy: pooled analysis of the Mechanical Embolus Removal in Cerebral Ischemia (MERCI) and Multi MERCI Trials. Stroke. 2009;40:3777–3783.
17. Hallevi H, Barreto AD, Liebeskind DS, Morales MM, Martin-Schild SB, Abraham AT, et al; UCLA Intra-Arterial Therapy Investigators. Identifying patients at high risk for poor outcome after intra-arterial therapy for acute ischemic stroke. Stroke. 2009;40:1780–1785.
18. Natarajan SK, Dandona P, Karmon Y, Yoo AJ, Kalia JS, Hao Q, et al. Prediction of adverse outcomes by blood glucose level after endovascu-lar therapy for acute ischemic stroke. J Neurosurg. 2011;114:1785–1799.
19. Kase CS, Furlan AJ, Wechsler LR, Higashida RT, Rowley HA, Hart RG, et al. Cerebral hemorrhage after intra-arterial thrombolysis for ischemic stroke: the PROACT II trial. Neurology. 2001;57:1603–1610.
20. Mazighi M, Serfaty JM, Labreuche J, Laissy JP, Meseguer E, Lavallée PC, et al; RECANALISE investigators. Comparison of intravenous alteplase with a combined intravenous-endovascular approach in patients with stroke and confirmed arterial occlusion (RECANALISE study): a prospective cohort study. Lancet Neurol. 2009;8:802–809.
21. Berridge DC, Gregson RH, Hopkinson BR, Makin GS. Intra-arterial thrombolysis using recombinant tissue plasminogen activator (r-TPA): the optimal agent, at the optimal dose? Eur J Vasc Surg. 1989;3:327–332.
22. Rouchaud A, Mazighi M, Labreuche J, Meseguer E, Serfaty JM, Laissy JP, et al. Outcomes of mechanical endovascular therapy for acute ischemic stroke: a clinical registry study and systematic review. Stroke. 2011;42:1289–1294.
23. Brott T, Adams HP Jr, Olinger CP, Marler JR, Barsan WG, Biller J, et al. Measurements of acute cerebral infarction: a clinical examination scale. Stroke. 1989;20:864–870.
24. The thrombolysis in myocardial infarction (timi) trial. Phase i findings. Timi study group. N Engl J Med. 1985;312:932–936.
25. van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van Gijn J. Interobserver agreement for the assessment of handicap in stroke patients. Stroke. 1988;19:604–607.
26. Hacke W, Kaste M, Fieschi C, von Kummer R, Davalos A, Meier D, et al. Randomised double-blind placebo-controlled trial of thrombolytic therapy with intravenous alteplase in acute ischaemic stroke (ECASS II). Second European-Australasian Acute Stroke Study Investigators. Lancet. 1998;352:1245–1251.
27. Chinn S. A simple method for converting an odds ratio to effect size for use in meta-analysis. Stat Med. 2000;19:3127–3131.
28. Vaidyula VR, Rao AK, Mozzoli M, Homko C, Cheung P, Boden G. Effects of hyperglycemia and hyperinsulinemia on circulating tissue factor pro-coagulant activity and platelet CD40 ligand. Diabetes. 2006;55:202–208.
30. Pandolfi A, Giaccari A, Cilli C, Alberta MM, Morviducci L, De Filippis EA, et al. Acute hyperglycemia and acute hyperinsulinemia decrease plasma fibrinolytic activity and increase plasminogen activator inhibitor type 1 in the rat. Acta Diabetol. 2001;38:71–76.
31. Meigs JB, Mittleman MA, Nathan DM, Tofler GH, Singer DE, Murphy-Sheehy PM, et al. Hyperinsulinemia, hyperglycemia, and impaired hemostasis: the Framingham Offspring Study. JAMA. 2000;283:221–228.
32. Arenillas JF, Sandoval P, Pérez de la Ossa N, Millán M, Guerrero C, Escudero D, et al. The metabolic syndrome is associated with a higher resistance to intravenous thrombolysis for acute ischemic stroke in women than in men. Stroke. 2009;40:344–349.
33. Arnold MF, Haag AF, Capewell S, Boshoff HI, James EK, McDonald R, et al. Partial complementation of Sinorhizobium meliloti bacA mutant phenotypes by the Mycobacterium tuberculosis BacA protein. J Bacteriol. 2013;195:389–398.
34. Demchuk AM, Morgenstern LB, Krieger DW, Linda Chi T, Hu W, Wein TH, et al. Serum glucose level and diabetes predict tissue plasminogen activator-related intracerebral hemorrhage in acute ischemic stroke. Stroke. 1999;30:34–39.
35. Hawkins BT, Lundeen TF, Norwood KM, Brooks HL, Egleton RD. Increased blood-brain barrier permeability and altered tight junctions in experimental diabetes in the rat: contribution of hyperglycaemia and matrix metalloproteinases. Diabetologia. 2007;50:202–211.
36. Desilles JP, Rouchaud A, Labreuche J, Meseguer E, Laissy JP, Serfaty JM, et al. Blood-brain barrier disruption is associated with increased mortality after endovascular therapy. Neurology. 2013;80:844–851.
37. Uyttenboogaart M, Koch MW, Stewart RE, Vroomen PC, Luijckx GJ, De Keyser J. Moderate hyperglycaemia is associated with favourable outcome in acute lacunar stroke. Brain. 2007;130(pt 6):1626–1630.
38. Els T, Klisch J, Orszagh M, Hetzel A, Schulte-Mönting J, Schumacher M, et al. Hyperglycemia in patients with focal cerebral ischemia after intravenous thrombolysis: influence on clinical outcome and infarct size. Cerebrovasc Dis. 2002;13:89–94.
39. Yip PK, He YY, Hsu CY, Garg N, Marangos P, Hogan EL. Effect of plasma glucose on infarct size in focal cerebral ischemia-reperfusion. Neurology. 1991;41:899–905.
40. Southerland AM, Johnston KC. Considering hyperglycemia and throm-bolysis in the Stroke Hyperglycemia Insulin Network Effort (SHINE) trial. Ann N Y Acad Sci. 2012;1268:72–78.
41. Gray CS, Hildreth AJ, Sandercock PA, O’Connell JE, Johnston DE, Cartlidge NE, et al; GIST Trialists Collaboration. Glucose-potassium-insulin infusions in the management of post-stroke hyperglycaemia: the UK Glucose Insulin in Stroke Trial (GIST-UK). Lancet Neurol. 2007;6:397–406.
42. McCormick M, Hadley D, McLean JR, Macfarlane JA, Condon B, Muir KW. Randomized, controlled trial of insulin for acute poststroke hyper-glycemia. Ann Neurol. 2010;67:570–578.
43. Dubas F. Small vessel pathology and cerebral hemorrhage. J Neuroradiol. 2003;30:298–302.
by guest on June 21, 2018http://stroke.ahajournals.org/
is online at: Stroke Information about subscribing to Subscriptions:
http://www.lww.com/reprints Information about reprints can be found online at: Reprints:
document. Permissions and Rights Question and Answer process is available in the
Request Permissions in the middle column of the Web page under Services. Further information about thisOnce the online version of the published article for which permission is being requested is located, click
can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.Strokein Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:
by guest on June 21, 2018http://stroke.ahajournals.org/
… (…) … NI (≥ 8-points decrease in 24h NIHSS or 24h
NIHSS=0) 89 (31)
… / 0
… / 0
NINDS trial
1991-1994 48
312 68 (57)
14a
IV rtPA [90 min
a]
… (22) 149 ICH sICHNINDS
34 (11) 20 (6)
0 / … 0 / …
0 / 0 0 / 0
Australia/ Europe
SITS-ISTR, 2002-2007 49
16 049 69 (59%)
12a
IV rtPA [145 min
a] 2735 (17) 117a Favorable outcome
90-day mortality sICHNINDS
sICHSITS-MOST
7222 (53) 2038 (15) 1193 (8) 263 (2)
- / - + / + + / 0 + / 0
- / - + / + + / + + / +
North America (USA/Canada)
MERCI, Multi-MERCI, 2001-2005 50
305 72 (48%)
19a MET (+-IV,IAT) [258 min
a]
… (….) … Recanalisation (TIMI 2-3 post-treatment)
Favorable outcome 90-day mortality
197 (65) 94 (32) 114 (38)
0 / 0 - / 0
0 / …
0 / 0 - / 0
0 / …
Worldwide
SAINT I, SAINT II trials, 2003-2006 51
965 68 (57%)
14 IV rtPA [145 min]
193 (20) 133 Favorable outcome Any ICH Asymptomatic ICH sICHNINDS
… (46) 221 (23) 167 (17) 54 (6)
… / 0 0 / 0 0 / 0 0 / 0
… / 0 + / 0 + / 0 0 / 0
North America GAIN trial, 1998-1999 52
333 68 (55%)
14 IV rtPA […]
74 (22) … Excellent outcome 90-day mortality
91 (28) … (19)
… / 0 …. / 0
… / … … / …
North America PROACT II trial, 1996-1998 53
110 65 (61%)
17 IA proUK [309 min]
16 (14) 134 sICHNINDS 12 (11) 0 / … + / +
North America, Europe
Multicenter rt-PA Stroke Survey, 1994-1998 54-55
1205 67 (56%)
≤16 60%
IV rtPA [<3h, 91%]
240 (20) 138 Excellent outcome Any ICH sICHNINDS
… (33) 158 (13) 72 (6)
- / - + / + + / +
- / - + / + + / 0
Abbreviation : CASES=canadian alteplase for stroke effectiveness study; ECASS=european cooperative acute stroke study; GAIN=glycine antagonist in neuroprotection; J-ACT=japan alteplase clinical trial; IA=intraarterial; ICH=intracranial hemorrhage; IV=intravenous; MERCI=mechanical embolus removal in cerebral ischemia; MET=mechanical endovascular therapy; mRS=modified Rankin Scale; NI=neurological improvement; NIHSS=national institutes of health stroke scale; NINDS=national institute of neurological disorders and stroke; PH=parenchymal hematoma; PROACT=Prolyse in acute cerebral thrombolysis; TIBI=thrombolysis in brain ischemia; TICI=thrombolysis in cerebral infarction; TIMI=thrombolysis in myocardial infarction; rTPA=recombinant tissue plasminogen activator; SAINT=stroke acute ischemic NXY treatment; SAMURAI=stroke acute management with urgent risk-factor assessment and improvement; sICH=symptomatic intracranial hemorrhage; SITS-ISTR=safe implementation of treatments in stroke-international stroke thrombolysis register; UCLA=university of california los angeles; UK=urokinase; UTH=university of texas at Houston. Data are presented as mean (a or median) unless otherwise indicated. … indicates not done. Favorable outcome defined as a 90-day mRS ≤2 and excellent outcome as a 90-day mRS ≤1. Brief summary of systematic review of association between diabetes and outcomes : Of the seven independent studies (n=1226 patients) that evaluated the effect of diabetes on neurological improvement (with various definitions), one reported a detrimental effect of diabetes in univariate analysis but not in multivariate analysis.6 Of the 23 independent studies (n=23710) that evaluated the effect of diabetes on
functional outcome (favorable in 13 studies, excellent in 7 studies and other definition in 3 studies), six (n=18755)2,33,49,54 reported a detrimental effect of diabetes, with the statistical analysis including the highest number of covariables. Of the eleven studies (n=20861) that analyzed mortality outcome (nine at 90-day and two at hospital discharge), only the largest study (n=16049)49 reported an increased risk of death among diabetics in univariate as well as in multivariate analysis. Regarding sICH, 14 studies (n=4796) reported similar risks in the diabetic and non-diabetic groups, two (n=16654) reported an increased risk associated with diabetes in univariate analysis only,32,49 and one (n=1205)55 reported an increased risk in both univariate and multivariable analyses. When any or asymptomatic ICH was studied (12 studies, n=6066), two reported an increased risk associated with diabetes in multivariable analyses.39,55 Finally, among the eight studies (n=3198) that analyzed recanalization, one reported an unadjusted detrimental effect of diabetes in recanalization assessed at 24h after IV rtPA treatment24; all of the 4 studies restricted to IA therapy did not report significant association. Brief summary of systematic review of association between glucose and outcomes : Of the eight studies (n=1408) that evaluated the effect of AGL on neurological improvement, two reported an inverse association in univariate and multivariate analysis;1,6 all other reported no association. Of the 25 independent studies that evaluated the effect of AGL on favorable outcome, 13 reported no association (n=3028), three (n=1102) reported an inverse association in univariate analysis only18,35,50 and 9 (n=20143) reported an inverse association in univariate and multivariate analysis.4,10,14,29,35,43,,45,49,54 Of them, one reported a positive association with favorable outcome but not with excellent outcome.14 Regarding mortality (11 studies, n=19979), an increase risk of death associated with higher AGL was reported in 6 studies (n=18940)4,10,14, 24,35,45,49 which remained significant after adjustment in 5 studies. Regarding sICH (18 studies, n=23794), an increase risk associated with higher AGL was reported in 7 studies (n=19659)5,32,37,40,49,53,55 which was significant in multivariate analysis for 3 studies (n=16647).6,37,53 When any or asymptomatic ICH was analyzed (11 studies, n=5936), two studies (n=1805) reported a positive association in multivariable analyses19,55 and 3 (n=2245) reported a positive association in univariate analysis only.5,26,51 Finally, among the 5 studies (n=1285) that analyzed recanalization, two reported an inverse association between AGL and recanalization rate (one in univariate analysis).24,28
Supplemental figures :
Figure e-1. Flow chart of study sample for Bichat Registry.
29 patients with missing data on diabetes, pre-treatment glucose level, or 90-day follow-up. (mean age=64 years; men=69%)
733 patients with AIS treated by IV and/or IA therapy between February 2002 and February 2012.
704 patients included in analysis : - 464 treated by IV approach - 160 treated by IV/A approach (adjunctive MET, n=53) - 80 treated by IA approach (direct MET, n=30; adjunctive MET, n=8)
600 patients with no history of diabetes mean±SD glucose level : 124 ± 35 mg/dl
104 patients with history of diabetes mean±SD glucose level : 187 ± 57 mg/dl