Introduction

Type 2 diabetes mellitus (T2DM) is a serious public health concern with a considerable impact on human life and health expenditures. About 462 million people suffer from T2DM, corresponding to 6.28% of the world’s population. The USA and Western Europe show a higher prevalence rate that continues to rise despite public health interventions [1]. The number of people living with diabetes mellitus quadrupled between 1980 and 2014 [2]. The global prevalence could increase from 6000 to about 8000 per 100,000 by 2040 [1]. T2DM is a major risk factor for cerebrovascular disease; a diabetic patient has more than doubled risk to develop an ischemic stroke (IS) [3]. Diabetes prevalence in IS was estimated to be 33% [4].

Reperfusion therapies (RT) represent a fundamental treatment in acute IS, and diabetes has a considerable influence on the outcome of treated patients [5,6,7,8]. Although there are many studies that examined the differences between diabetic and no diabetic patients, only a few evaluated factors determining prognosis among diabetic patients with IS and underwent RT.

The aims of the present study are to identify the main prognostic factors in diabetics treated with RT and to evaluate the reciprocal influence of intravenous thrombolysis (IVT) and mechanical thrombectomy (MT) among these patients.

Methods

Patients

In this retrospective study, we reviewed 962 acute IS patients treated with IVT and/or MT in our hospital between February 2014 and December 2019, and we selected 170 patients with a previous T2DM diagnosis.

Patients underwent IVT within 4.5 h after IS onset. In case of unknown symptom onset, the IVT was also performed if magnetic resonance imaging (MRI) showed an ischemic lesion on diffusion-weighted imaging (DWI) that was not visible on fluid-attenuated inversion recovery (FLAIR).

Patients underwent MT within a time frame from symptom onset to treatment ≤ 6 h for anterior circulation and ≤ 24 h for posterior circulation. Regarding stroke with unknown onset, patients were selected using MRI in order to discern between salvageable and terminally infarcted tissue with the application of DWI, perfusion scanning, and FLAIR.

Main exclusion criteria for RT were large territorial infarction defined as Alberta Stroke Program Early CT Score (ASPECTS) < 5, hospital arrival beyond time window, and elevated bleeding risk for IVT.

Patients suffering from type 1 diabetes mellitus were excluded from the study.

Data collection and clinical assessment

The following baseline information and risk factors were assessed in all participants: age, gender, weight, smoke, arterial hypertension, dyslipidemia, coronary heart disease, prior stroke or transitory ischemic attack, chronic kidney disease. Collected data regarding T2DM were antidiabetic drugs (metformin/no metformin) and duration of T2DM. Laboratory test data were admission glucose, creatinine, leucocytes, neutrophils, lymphocytes, and platelets. Admission systolic and diastolic pressures were included. Collected stroke data were etiology according to the Trial of Org 10,172 in Acute Stroke Treatment classification [9], IVT, recombinant tissue plasminogen activator (rtPA) dosage, MT, stroke-to-treatment time (interval between first symptom onset and the beginning of IVT or femoral artery puncture), duration of MT treatment (interval between femoral artery puncture and last contrast bolus injection), neurological severity on admission and discharge using the National Institutes of Health Stroke Scale (NIHSS), intracranial hemorrhage (ICH), symptomatic intracranial hemorrhage (SICH) according to the European-Australian Cooperative Acute Stroke Study 2 (ECASS2) [10], 3-month functional impairment using modified Rankin scale (mRs), and 3-month mortality.

Outcome measures

The main outcome is the mRs at 3 months.

The secondary outcomes are:

  • Neurological severity on discharge

  • 3-month mortality

  • ICH

  • SICH

All data were retrospectively obtained using our center database and medical charts. We performed a main analysis on all RT and two subanalyses focusing on patients treated with IVT alone and patients underwent MT, respectively.

Ethics approval and study consent

The research was conducted ethically in accordance with the World Medical Association Declaration of Helsinki. The study protocol had been approved by the research institute’s committee on human research. Every patient or a legal representative provided a written informed consent to data collection for scientific purposes.

Statistical analysis

All statistical analyses were performed using R software. Continuous variables were expressed as mean ± standard deviation or median and interquartile range (IQR); categorical variables were expressed as absolute frequencies and percentages. Data were analyzed by Shapiro–Wilk test to evaluate normal distribution. Mann–Whitney U or Student’s t test for independent samples was used for comparison between categorical variables with two levels and continuous variables as appropriate. ANOVA test or Kruskal–Wallis test was used for comparison between categorical variables with more than 2 levels and continuous variables on the basis of normal distribution. The chi-square test was used for comparison between categorical variables. The method of partitioning the degrees of freedom was applied to refuse H0 hypothesis as appropriate. Spearman’s Rank or Pearson’s correlation coefficient was used for comparison between continuous variables as appropriate. In multivariate analyses (MA), we used binomial logistic regression model or multiple linear regression model. A binomial logistic regression was used for binary-dependent variables (3-month mortality, ICH and SICH) whereas a multiple linear regression was used for quantitative outcomes (mRs at 3 months and neurological severity on discharge). The goodness of fit tests used were Hosmer–Lemeshow for C and H statistic, Osius-Rojek’s, Stukel’s, and Le Cessie-Van Houwelingen-Copas-Hosmer [11]. A value of P ≤ 0.05 was considered significant.

Results

Descriptive analysis

The study cohort comprised 170 diabetic patients. The mean age was 76.72 years, eighty-four (49.4%) were males, and the median admission NIHSS was 14 (IQR 8–18). The median functional outcome at 3 months measured by mRs was 4 (IQR 2–6). The median discharge NIHSS was 5 (IQR 2–10). Forty-six patients (27.1%) died within 3 months. Fifty-nine patients (34.7%) had an ICH, and fifteen (8.8%) a SICH.

Patients who underwent IVT alone were 62. The mean age was 76.89 years, thirty-three (53.2%) were males, and the median admission NIHSS was 10 (IQR 6–15). The mRs at 3 months was 3 (IQR 1–5). The median discharge NIHSS was 3 (IQR 1–5). Fourteen patients (24.6%) died within 3 months. Twenty-two patients (35.5%) had an ICH, and seven (11.5%) a SICH.

MT was performed in 108 patients. Among these, only 29 patients underwent IVT, whereas the others were not treated in most cases because of delay in hospital arrival and the unknown onset with a no permissive MRI. They had a mean age of 76.62 years, the males were 51 (47.2%), and the median admission NIHSS was 14 (IQR 10.75–18.25). The mRs at 3 months was 4 (IQR 2–6). The median discharge NIHSS was 6 (IQR 3–12). Thirty-two patients (29.6%) died within 3 months; thirty-seven patients (34.3%) had an ICH, and eight (7.4%) a SICH.

The present data, remaining baseline characteristics, and treatment information are summarized in Table 1. Table S1 shows a descriptive analysis of MT patients divided according to treatment with IVT.

Table 1 Baseline characteristics, treatments, and outcomes
Full size table

Outcome analysis in all patients undergoing reperfusion therapies

A higher 3-month mRs was associated with no metformin use (4 [2–6] vs 3 [1–5]; p-value = 0.020), MT (4 [2–6] vs 3 [1–5]; p-value = 0.034), older age (correlation coefficient, r = 0.211; p-value = 0.009), higher admission blood glucose (r = 0.202; p-value = 0.013), and elevated admission NIHSS (r = 0.407; p-value = 0.000) (see Table 2). Only admission blood glucose and admission NIHSS maintained statistical significance on MA.

Table 2 Modified Rankin scale (mRs) at 3 months
Full size table

Lower NIHSS on discharge was associated with IVT (5.11 ± 5.50 vs 8.63 ± 6.11; p-value = 0.001), no MT (5.19 ± 5.62 vs 7.71 ± 6.10; p-value = 0.013), lower neutrophil levels (r = 0.200; p-value = 0.018), higher lymphocyte levels (r =  − 0.199; p-value = 0.018), and lower admission NIHSS (r = 0.521; p-value = 0.000) (see Table 3). On MA, lower NIHSS on discharge was related only with IVT, no MT, and admission NIHSS.

Table 3 NIHSS on discharge
Full size table

The 3-month mortality was associated with older age (78.98 ± 8.31 vs 75.41 ± 8.87; p-value = 0.021), no smoke (32.0% vs 12.5%; p-value = 0.053), elevated admission blood glucose (210.00 ± 81.84 vs 175.67 ± 60.98; p-value = 0.005), and a higher admission NIHSS (17.5 [11–20] vs 13 [8–17]; p-value = 0.003) (see Table S2). Age, blood glucose, and admission NIHSS were related with mortality on MA.

Patients with ICH were older (78.97 ± 7.90 vs 75.61 ± 8.93; p-value = 0.017) and more frequently non-smoker (38.3% vs 16%; p-value < 0.031) and they had lower platelet levels (210,574.55 ± 203,000.00 vs 251,046.73 ± 102,239.28; p-value = 0.008) and a higher admission NIHSS (16 [13–19.5] vs 12 [8–17]; p-value = 0.001) (see Table 4). MA confirmed age, platelets, and admission NIHSS.

Table 4 Intracranial hemorrhage (ICH)
Full size table

The predictors of SICH were leucocyte levels (7807.86 ± 2981.04 vs 9405.96 ± 2828.66; p-value = 0.046), stroke etiology (2.6% of large artery atherosclerosis, 15.5% of cardio embolism, 33.3% of small vessel occlusion, 6.0% of undetermined subtype; p-value < 0.051), and duration of MT treatment (151.25 ± 105.11 vs 75.55 ± 38.87; p-value = 0.001) (see Table 5). Only MT duration was confirmed on MA.

Table 5 Symptomatic intracranial hemorrhage (SICH)
Full size table

Outcome analysis in patients undergoing IVT

Analyzing patients undergoing IVT, a lower 3-month mRs was associated metformin (2 [0–4] vs 4 [3–6]; p-value = 0.004), younger age (r = 0.562; p-value = 0.000), short diabetes duration (r = 0.328; p-value = 0.013), and lower admission NIHSS (r = 0.538; p-value = 0.000) (see Table 2). Age and admission NIHSS were also confirmed on MA.

A lower NIHSS on discharge was related with younger age (r = 0.332; p-value = 0.012); metformin use (2 [1–4.75] vs 5.5 [1–12]; p-value = 0.031); lower neutrophil levels (r = 0.330; p-value = 0.013); higher lymphocyte levels (r =  − 0.319; p-value = 0.016); lower admission NIHSS (r = 0.502; − value = 0,000) (see Table 3). MA confirmed only admission NIHSS.

The 3-month mortality was associated with older age (84.71 ± 3.93 vs 74.49 ± 9.82; p-value = 0.001), female sex (37.5% vs 15.2%; p-value < 0.053), longer diabetes duration (21.14 ± 10.13 vs 14.27 ± 9.53; p-value = 0.025), no metformin use (40.0% vs 10.7%; p-value < 0.013), higher admission NIHSS (16.5 [8.25–20] vs 8 [5.5–15]; p-value = 0.004) (see Table S2). MA identified older age and higher admission NIHSS as the mortality main predictors.

The ICH was associated with older age (80.73 ± 7.23 vs 74.78 ± 10.11; p-value = 0.018), no metformin use (52.0% vs 20.0%; p-value < 0.013), higher admission systolic pressure (169.10 ± 32.57 vs 154.03 ± 21.00; p-value = 0.032), lower platelet levels (210,714.29 ± 44,976.82 vs 277,375.00 ± 127,498.96; p-value = 0.024), stroke etiology (7.1% of large artery atherosclerosis, 60% of cardio embolism, 33.3% of small vessel occlusion, and 36.7% of undetermined subtype; p-value < 0.031), unknown onset time (18.2% vs 2.5%; p-value = 0.033) (see Table 4). Only age and platelets remained significant on MA.

The only predictor of SICH was metformin use (0.0% vs 20.8%; p-value < 0.009) (see Table 5).

Outcome analysis in patients undergoing MT

Analyzing patients undergoing MT, the 3-month mRs score was associated with admission blood glucose (r = 0.297; p-value = 0.004) and admission NIHSS (r = 0.267; p-value = 0.008) (see Table 2). This significance persisted on MA.

The discharge NIHSS was associated with IVT (3.5 [1.25–5] vs 7 [4–13]; p-value = 0.004; see Fig. 1) and admission NIHSS (r = 0.437; p-value = 0.000) (see Table 3). MA confirmed these associations.

Fig. 1
figure 1

NIHSS on discharge in MT patients according to IVT treatment

Full size image

The 3-month mortality was also linked with admission systolic pressure (160.34 ± 24.91 vs 148.20 ± 25.05; p-value = 0.027) and admission glucose levels (215.94 ± 92.34 vs 171.86 ± 63.83; p-value = 0.008) (see Table S2). MA confirmed these associations.

The ICH was related only with higher NIHSS (17 [14–20] vs 13 [9–18]; p-value = 0.001) (see Table 4).

The predictors of SICH were elevated weight (89.25 ± 11.06 vs 74.06 ± 10.73; p-value = 0.008), rtPA dosage (73.33 ± 14.74 vs 56.54 ± 11.18; p-value = 0.024), and prolonged MT duration (151.25 ± 105.11 vs 75.55 ± 38.87; p-value = 0.001) (see Table 5). MA identified only weight as SICH-independent predictor.

Discussion

Although T2DM is present in one-third of IS, only few studies had evaluated factors determining prognosis among diabetic patients with IS and undergoing RT.

Regarding functional outcome and mortality, we observed an association with admission glycemia in all patients and MT patients but not in the IVT-alone group. Patients undergoing MT, therefore, are probably more influenced by hyperglycemia than IVT patients and this should be considered. In our patients treated only with IVT, mRs score increased according to the glycemia levels but this trend was not significant: hyperglycemia had an influence on functional outcome but it was not a main outcome predictor. Literature data are conflicting about glycemia and functional outcome in diabetics. In IVT patients, two studies did not find an association [12, 13] whereas other three observed it [14,15,16]. In MT patients, two studies correlated hyperglycemia with a worst functional outcome [17, 18] whereas other two failed to find it [7, 19]. These discrepant responses to hyperglycemia could be due to the difference in chronic glycemic control, diabetic drugs taken, diabetes duration, and other metabolic conditions: further studies should investigate these elements. Another important element of our analysis is that mortality in MT patients was linked to high levels of systolic pressure on the admission. No other studies in literature evaluated pressure influence in diabetic patients treated with MT. A high systolic pressure was associated with a worse functional outcome in general patients undergoing MT in several studies which explained it through reperfusion injury, cerebral edema, and hemorrhagic transformation [20]. Another hypothesis is that hypertension may be a sign of stroke severity rather than a determinant: the organisms increase blood pressure in order to maintain cerebral perfusion [20]. It is interesting to observe that blood pressure was not related with discharge neurological severity, hemorrhagic transformation, and functional outcome in our patients: we could also hypothesize that the higher mortality was due to a more severe T2DM rather than stroke damage in itself. T2DM was associated with artery stiffness, autonomic dysfunction, endothelium dysfunction, and impaired nitric oxide (NO) synthesis [21] that can favor an alteration of blood pressure control. IS could be considered a stress test that could induce higher pressure peaks in patients with a more severe neurovascular dysfunction due to T2DM. Conditions such as autonomic dysfunction were indeed associated with an increased mortality in T2DM [22, 23].

Regarding neurological severity on discharge, in all RT group, we observed that MT was associated with higher NIHSS, whereas IVT with lower score probably because most of IVT patients had no large vessel occlusion. It is important to highlight that IVT in patients undergoing MT led to a reduced (almost halved) neurological severity, supporting (with the limits of our small population size) the efficacy of combined treatment in these patients.

Evaluating ICH, platelet levels and age appeared as the main predictors in all RT group and IVT patients which did not influence bleedings in MT. Our study also showed that duration of MT was associated with SICH in all RT but not in MT. We could explain this result as an increased SICH risk in patients undergoing longer MT compared with patients treated only with IVT. In patients undergoing MT, the high weight was the only predictor of SICH. This association was not evaluated in other MT study on T2DM. Regarding hemorrhagic transformation and obesity in general population treated with MT, some studies described no association [24,25,26], whereas Chen et al. reported a reduction in symptomatic hemorrhage [27]. Another study found no relationship between SICH and metabolic syndrome [28]. In our patients, the combination between T2DM and high weight probably favored an increased vessel fragility that could not be present in obesity alone and in metabolic syndrome, condition in which patients are not always diabetics.

We furthermore found that metformin had a protective role for SICH in IVT patients. Several preclinical study [29,30,31,32,33] and three clinical study showed the positive impact of metformin in ischemic stroke [34,35,36]. Indeed, metformin plays an anti-oxidant and anti-inflammatory action, favoring the blood–brain barrier integrity and a correct endothelial function [37]. The metformin beneficial effect was not observed in our MT patients, a larger population study is probably necessary in order to find it.

There are several limitations in our study. The small number of patients may have underpowered our analysis. The study in a single institution may have affected the selection of patients but it allowed us to obtain data homogeneity. The retrospective design represents another limit. The glycated hemoglobin would have been a useful data but it was missing in a significant part of patients.

Conclusion

The present study evaluated RT in patients affected by T2DM that constitutes one-third of IS victims and deserve attention in regard to their complexity and fragility. We observed a prognostic role of admission glycemia in MT but not in IVT. The study showed that neurological severity on discharge was reduced in patients undergoing both treatments compared with MT alone. A protective role of metformin for SICH was found in patients treated with IVT, whereas the high weight was a predictor of symptomatic hemorrhage in MT. Our results give several insights in regard to T2DM and stroke RT that need to be confirmed in larger studies, but they represent a starting point in order to ameliorate medical management of these patients.