Efficacy of telemedicine for thrombolytic therapy in acute ischemic stroke: a meta-analysis

Abstract

Summary

The aim of this study was to assess the benefits of telemedicine in the delivery of thrombolytic therapy for patients with acute ischemic stroke. We performed a meta-analysis using combinations of the following terms: telestroke, telemedicine, tissue plasminogen activator/t-PA, and acute ischemic stroke. The primary outcome was favorable outcome based on the modified Rankin score. Secondary outcomes were incidence of symptomatic intracranial hemorrhage and overall mortality. We found no significant difference in favorable outcome between the telemedicine and control groups, and no significant difference was found between these groups in the rate of symptomatic intracranial hemorrhage or overall mortality. Patients with acute ischemic stroke who were treated with intravenous thrombolysis had similar outcomes regardless of whether telemedicine was used or they were treated in-person at a medical facility. Telemedicine can be used to support hospitals with limited experience in administering thrombolytic therapy for stroke.

Keywords

telemedicine stroke thrombolytic therapy meta-analysis

Introduction

Stroke ranks third among the leading causes of death in Europe, Japan, the United States, and Canada, and ranks first among the leading causes of acquired long-term disability in North America and Europe.^1,2 Four-fifths of strokes are attributable to ischemic stroke and one-fifth to hemorrhagic stroke.¹

Functional outcomes of stroke patients have been shown to improve when they are treated with the thrombolytic agent alteplase within 3 hours of stroke onset.^2,3 However, only a small percentage of patients receive this treatment. For example, in the United States the percentage of stroke patients treated with thrombolytic therapy is only 3.7%.³ A particularly important reason why so few stroke patients receive thrombolytic treatment is that they live in areas where stroke unit care is not available such as in rural areas.^2,4 Also, even in urban areas with stroke units available the stroke team may be notified later than the recommended time.⁵

Telemedicine, or telestroke, has been used as a way to improve treatment for stroke patients who live in areas without adequate treatment facilities.⁶ Patients are usually examined by video, and brain scans are evaluated by teleradiology.⁷ Telestroke can be considered as a set of tools for providing improved care for stroke in areas lacking sufficient neurologic services.⁸ Although telestroke has been found to improve care, implementing a telestroke network can be challenging.³

A number of studies have found that telestroke increases the rate of administering thrombolytic therapy and improves functional outcomes in acute stroke patients, although none of these studies were randomized controlled trials.² The aim of this meta-analysis was to evaluate the benefits of telestroke for acute ischemic stroke patients versus in-person care at a medical facility with regard to the delivery of thrombolytic therapy.

Methods

Search strategy

During April 2014 the following databases were searched: PubMed, Cochrane, Embase, ISI Web of Knowledge, and ClinicalTrial.gov. There was no specific starting date. The search was conducted using combinations of the following search terms: (telestroke OR telemedicine OR telemedical) AND acute ischemic stroke. In the PubMed search the filters used were Abstract available, English, and Human; in the Embase search the filters used were Article and Human; and in the ISI Web of Knowledge search the filters used were English and Article. The search was carried out by two independent reviews and a third reviewer was consulted to resolve any disagreements.

Five studies were randomly selected in order to perform recoding by a different investigator to assess coder drift.⁹ A determination was made of per case agreement. This was achieved by division using the of the number of variables coded as the numerator and the total number of variables as the denominator. An appropriate level of reliability was deemed to be a mean agreement of 0.90. The coding process was determined to be reliable as coder drift was found to be 0.96.

Selection of studies

The studies for inclusion in the meta-analysis were selected based on the following criteria: 1) randomized controlled trials, and prospective and retrospective studies with a two-arm design; 2) patients diagnosed with acute ischemic stroke; and 3) studies compared telemedicine vs. in-person care at a medical facility for delivery of thrombolysis.

Cohort studies, letters, comments, editorials, case reports, proceedings, and personal communications were excluded. Studies were also excluded if patients had other types of ischemic stroke or had stroke related to chronic diseases, the study used a single-arm design, there was no quantitative measurement, or the study was published in a language other than English.

Study quality assessment

Prospective studies were assessed with the Cochrane “assessing risk of bias” table.¹⁰ There are six domains—random sequence generation, allocation concealment, blinding of patients and personnel, blinding of outcome assessment, incomplete outcome data, and selective reporting risk. These criteria were taken from the Cochrane Collaboration guidelines.

Retrospective studies were assessed with the Newcastle-Ottawa Scale.¹¹ This scale contains eight items categorized into three dimensions: selection, comparability, and exposure. A star system is used for a semi-quantitative assessment of study quality.

Data extraction

The following types of data were extracted: first author and year study was published, study design, comparison (telemedicine vs control), type of intervention, number of patients, mean age of patients, percentage of male patients, National Institutes of Health Stroke Scale (NIHSS) score at baseline, percentage of patients with a favorable outcome based on modified Rankin score, percentage of patients with symptomatic intracranial hemorrhage, overall mortality rate, and time from stroke onset to treatment. Two independent reviewers extracted the data from eligible studies. A third reviewer was consulted to resolve any disagreements.

Data analysis

The primary outcome in the meta-analysis was favorable outcome (modified Rankin score) and the secondary outcomes were incidence of symptomatic intracranial hemorrhage and overall mortality. The odds ratios (ORs) were calculated for the three outcomes in the telemedicine group compared with the control group. Heterogeneity among the studies was assessed by the Cochran Q statistic and the I² statistic. For the Q statistic, P < 0.10 was considered statistically significant for heterogeneity; for the I² statistic, which indicated the percentage of the observed between-study variability due to heterogeneity rather than to chance, the suggested ranges were as follows: no heterogeneity (I²= 0–25%), moderate heterogeneity (I²= 25–50%), large heterogeneity (I²= 50–75%), and extreme heterogeneity (I²= 75–100%). If either the Q statistic (P < 0.1) or I² statistic (>50%) indicated heterogeneity existed between studies, the random-effects model was preferred (DerSimonian–Laird method). Otherwise, the fixed-effect model (Mantel-Haenszel method) was recommended. Pooled OR of the outcome was calculated, and a 2-sided P-value <0.05 was considered as statistically significant. All statistical analyses were performed using the statistical software Comprehensive Meta-Analysis, version 2.0 (Biostat, Englewood, NJ, USA).

Results

Literature search

The results of the literature search are shown in a flow diagram (Figure 1). Of 230 reports initially identified by searching, 202 were found to be not relevant and 28 were found to be relevant and the full text of these latter articles were reviewed. The search results for the specific databases were as follows: in PubMed, 59 articles were found, 40 were excluded, and 19 qualified after removing duplicates; in Cochrane, 1 article was found and it was excluded; in Embase, 54 articles were found, 51 were excluded, and there were no duplicates so 3 articles qualified; in ISI Web of Knowledge, 116 articles were found, 110 were excluded, and there were no duplicates so 6 articles qualified; and in ClinicalTrial.gov 12 articles were found and all 12 were excluded, Based on the review, 20 studies were excluded and 8 were included in the meta-analysis.

Figure 1.

Flow diagram of study selection.

Study characteristics

The characteristics of the 8 studies included in the meta-analysis are presented in Table 1.^12–19 Among the 8 studies, the total number of participants ranged from 20 to 1,860 in the telemedicine groups and from 25 to 1,075 in the control groups. Mean age ranged from 62.5 to 71.9 years in the telemedicine groups and from 64.5 to 71.9 years in the control groups. Percentage of males ranged from 40% to 68.4% in the telemedicine groups and from 43.6% to 69.9% in the control groups. Favorable outcome ranged from 36.8% to 47% in the telemedicine groups and from 15.8% to 43% in the control groups. Incidence of symptomatic intracranial hemorrhage ranged from 1.2% to 8.3% in the telemedicine groups and from 0% to 13.3% in the control groups. Overall mortality ranged from 19% to 33.3% in the telemedicine groups and from 13% to 35% in the control groups. Time between stroke onset and treatment ranged from 113 to 188 minutes in the telemedicine groups and from 100 to 157 minutes in the control groups.

Table 1.

Summary of basic characteristics of studies included in meta-analysis.

Authors (Year)	Study design	Comparison	Intervention	Number of subjects	Age^†	Male (%)	NIHSS at baseline^†	Favorable outcome, observed/ total (%)	Symptomatic intracranial hemorrhage (%)	Overall mortality (%)	Onset to treatment (minutes)
Chowdhury (2012)	Retrospective	Telemedicine	Telemedicine	45	62.51 ± 11.52	62.30	13.45 ± 7.37	19/45 (42)¹	4.4	NA	125 (55, 105) ^&
Chowdhury (2012)		Control	Face-to-face	52	64.46 ± 12.00	55.80	14.81 ± 6.98	19/52 (36.5)¹	7.7	NA	100 (78, 120) ^&
Pedragosa (2012)	Prospective	Telemedicine	TMH	20	66.5	68.40	18	7/20 (36.8)¹	8.33	NA	188 ± 71
Pedragosa (2012)		Control	NonTMH	25	71.6	69.90	19	4/25 (15.8)¹	13.33	NA	143 ± 65
Johansson (2011)	Retrospective	Telemedicine	Telestroke	47	67 ± 15	66	9.9	14/30 (47)²	NA	19	113 ± 40
Johansson (2011)		Control	Specialized Stroke center	304	71 ± 16	50	10.4	76/179 (43)²	NA	13	122 ± 47
Zaidi (2011)	Prospective	Telemedicine	Telestroke	83	71.9 ± 14.4	53.10	12 (4–33)^¶	32/76 (42.1)¹	1.2	31.6	145.5 ± 42.8
Zaidi (2011)		Control	Stroke center	59	71.9 ± 14.1	43.60	10.5 (2–38)^¶	21/56 (37.5)¹	5.1	30.4	156.7 ± 31.6
Audebert (2009)	Prospective	Telemedicine	TEMPiS	1,860	NA	NA	NA	NA	NA	33.3	NA
Audebert (2009)		Control	Control hospital	1,075	NA	NA	NA	NA	NA	35	NA
Switzer (2009)	Prospective	Telemedicine	REACH	49	63	40	14.4	NA	4.1	NA	127.57 ± 36.33
Switzer (2009)		Control	MCG ED	26	NA	NA	NA	NA	0	NA	145.88 ± 46.99
Schwab (2007)	Prospective	Telemedicine	TEMPiS	170	69.4	59.4	12 (2–25)^¶	62/162 (38.2)²	NA	NA	140.6
Schwab (2007)		Control	Stroke center	132	69.6	63.6	11 (2–34)^¶	37/110 (33.7)²	NA	NA	143.6
Audebert (2006)	Retrospective	Telemedicine	TEMPiS	115	69.7 ± 10.7	56	12.4 ± 4.9	NA	7.8	NA	134 ± 30
Audebert (2006)		Control	Stroke center	110	69.8 ± 11.4	63	11.9 ± 5.3	NA	2.7	NA	135 ± 38

Data expressed as ^†mean ± standard deviation; ^&median (IQR); ^¶mean (range).

Favorable outcome: ¹modified Rankin score of 0–2 at 3 months; ²modified Rankin score of 0–1 at 3 months.

Abbreviations: NA, no data available; NIHSS, National Institutes of Health Stroke Scale; TMH, hospitals with a telemedicine system; TEMPiS, Telemedical Project for Integrative Stroke Care; REACH, Remote Evaluation of Acute IsCHemic Stroke; MCG ED, Medical College of Georgia Emergency Department.

Quality assessment

The results of quality assessment of the five included prospective studies are presented in Table 2. The results of quality assessment of the three retrospective included studies are presented in Table 3.

Table 2.

Quality assessment of the prospective included studies.

Author (year)	Random sequence generation (selection bias)	Allocation concealment (selection bias)	Blinding of participants and personnel (performance bias)	Blinding of outcome assessment (detection bias)	Incomplete outcome data (attrition bias)	Selective reporting (reporting bias)	Did the analysis include an intention-to-treat analysis?
Pedragosa (2012)	N	N	NA	NA	Y	Y	NR
Zaidi (2011)	N	N	N	N	Y	Y	NR
Audebert (2009)	N	N	NA	NA	Y	Y	Y
Switzer (2009)	N	N	NA	NA	Y	Y	NR
Schwab (2007)	N	N	N	N	Y	Y	NR

Y: yes; N: no; NA: not available.

Table 3.

Quality assessment of retrospective included studies.

Author (year)	Selection	Comparability	Exposure
Chowdhury (2012)	***	**	***
Johansson (2011)	***	**	***
Audebert (2006)	***	**	***

Outcome evaluation: Favorable outcome

Three studies^16,17,19 were excluded from the primary outcome analysis because they lacked favorable outcomes. There was no significant heterogeneity when data from the 5 included studies were pooled (heterogeneity test: Q = 1.30, df = 4, P = 0.861, I²= 0%); therefore, a fixed-effect model of analysis was used (Figure 2A). The overall analysis revealed that there was no significant difference in favorable outcome between the telemedicine and control groups (pooled OR = 1.28, 95% CI = 0.92 to 1.76, Z = 1.48, P = 0.139).

Figure 2.

Forest plots showing results for the meta-analysis of (A) favorable outcome; (B) incidence of symptomatic intracranial hemorrhage; (C) overall mortality. Abbreviations: OR, odds ratio; CI, confidence interval.

Outcome evaluation: Incidence of symptomatic intracranial hemorrhage

Three studies^14,16,18 were excluded from the analysis due to a lack of incidence of symptomatic intracranial hemorrhage. There was no significant heterogeneity when data from the 5 included studies were pooled (heterogeneity test: Q = 5.39, df = 4, P = 0.249, I²= 25.85%); therefore, a fixed-effect model of analysis was used (Figure 2B). The overall analysis revealed that there was no significant difference in incidence of symptomatic intracranial hemorrhage between the telemedicine and control groups (pooled OR = 1.08, 95% CI = 0.47 to 2.50, Z = 0.19, P = 0.849).

Outcome evaluation: Overall mortality

Five studies^{12,13,17–19} were excluded from the analysis due to a lack of overall mortality. There was no significant heterogeneity when data from the 3 included studies were pooled (heterogeneity test: Q = 1.68, df = 2, P = 0.431, I²= 0%); therefore, a fixed-effect model of analysis was used (Figure 2C). The overall analysis revealed that there was no significant difference in overall mortality between the telemedicine and control groups (pooled OR = 0.95, 95% CI = 0.82 to 1.11, Z = −0.66, P = 0.510).

Sensitivity analysis

The results of meta-analysis using the leave-one-out approach to assess sensitivity for the fixed-effect model are summarized in Figure 3. The direction and magnitude of pooled estimates did not vary considerably, which indicated that the meta-analysis had good reliability.

Figure 3.

Results of sensitivity analysis to examine the influence of individual studies on pooled estimates as determined use the leave-one-out approach of (A) favorable outcome; (B) incidence of symptomatic intracranial hemorrhage; (C) overall mortality. Abbreviations: OR, odds ratio; CI, confidence interval.

Publication bias

Publication bias was not assessed for the three outcomes because more than 5 studies are required to detect funnel plot asymmetry.²⁰

Discussion

The results of our meta-analysis showed that there was no significant difference in favorable outcome between patients in the telemedicine group and patients in the control group. Also, there was no significant difference between the groups in hemorrhage rate or overall mortality rate. Our findings provide confirmation that telemedicine can be used effectively in areas not served by hospitals that can readily provide thrombolytic therapy to stroke patients.

The 8 studies included in the meta-analysis had mostly similar findings, although there were some noteworthy differences. In the study of Pedragosa et al.¹³ the control group had a significantly poorer outcome compared with the telemedicine group. Also, the control group had a higher rate of hemorrhage. Also, in the study by Chowdhury et al.¹² the time from onset of stroke to treatment was significantly shorter than in the control group whereas in the other studies there were no significant differences between the groups, although in the study by Pedragosa et al.¹³ the time from stroke onset to treatment was about 45 minutes less than in the control group.

Our meta-analysis had several limitations. None of the 8 studies were randomized controlled trials and 3 studies were retrospective. The 8 studies differed with regard to distances from the hospital and transportation. Also, there were differences among the studies in the criteria for favorable outcome. Of the 5 studies included in the analysis of favorable outcome, 3 used modified Rankin score of 0–2 at 3 months and two used modified Rankin score of 0–1 at 3 months.

For our meta-analysis we selected studies that compared telestroke with in-person care at a medical facility but there have also been studies comparing telestroke with telephone consultation. In the STRoKe DOC trial, a randomized prospective study, that there was no significant difference in mortality rates or intracerebral hemorrhage rates after thrombolytic therapy was given but correct treatment decisions were made more often with telemedicine than with telephone consultation.²¹ Rubin and Demaerschalk²² recently reviewed these studies, which included 2 similarly designed randomized controlled trials. They noted that a pooled analysis of the randomized controlled trial confirmed the conclusions of the original trials which found that there was no difference in outcomes between groups or in hemorrhage and 90-day mortality rates, and when telestroke was used the correct decision to use thrombolytic therapy was significantly more likely. The evidence from these studies suggests that telephone consultation is not a good alternative to telemedicine.

Telemedicine has also been used for rehabilitation after stroke. The literature on telerehabilitation for stroke care was systematically reviewed by Johansson and Wild.²³ They found that the studies on telerehabilitation for stroke care suggested that this approach is promising for improving post-stroke care but cautioned that the quality of evidence in these studies should be considered to be at a low level.

In summary, the results of our meta-analysis showed that acute ischemic stroke patients treated with thrombolytic therapy via telemedicine had similar outcomes compared with patients who received thrombolytic therapy when treated in person at a medical facility. Our findings provide additional evidence that telemedicine can be successfully used to support regional hospitals with limited experience in treating stroke patients with thrombolytic therapy and can raise the standards of care and minimize inequalities in receiving treatment.

Footnotes

Conflict of interest

None.

Acknowledgement

This study is supported by the National Science & Technology Huimin Program (2013GS410101), Major Program of Science & Technology of Henan Province (121100111100), Youth Foundation of the first Affilliated Hospital, Zhengzhou University, and Innovation Scientists and Technicians Troop Construction Projects of Henan Province (144100510017). Sponsored by Program for Science?Technology Innovation Talents in Universities of Henan Province (15IRTSTHN023, 15HASTIT010).

References

Meyers

Schumacher

Connolly

Jr . Current status of endovascular stroke treatment. Circulation 2011; 123: 2591–601. doi:10.1161/CIRCULATIONAHA.110.971564.

Bodechtel

Puetz

. Why Telestroke networks? Rationale, implementation and results of the Stroke Eastern Saxony Network. J Neural Transm 2013; 120(Suppl 1): S43–37. doi: 10.1007/s00702-013-1069-y. Epub 2013 Jul 24.

French

Day

Watkins

McLoughlin

. The challenges of implementing a telestroke network: a systematic review and case study. BMC Med Inform Decis Mak 2013; 13: 125–125. doi: 10.1186/1472-6947-13-125.

Stewart

Switzer

. Perspectives on telemedicine to improve stroke treatment. Drugs Today (Barc) 2011; 47: 157–67. doi: 10.1358/dot.2011.47.2.1576694.

Tang

. Strategies to consider: telestroke and 24-hour primary stroke centres. Hong Kong Med J 2012; 18: 76–8.

Johansson

Wild

. Telemedicine in acute stroke management: systematic review. Int J Technol Assess Health Care 2010; 26: 149–55. doi: 10.1017/S0266462310000139.

Audebert

Schwamm

. Telestroke: scientific results. Cerebrovasc Dis 2009; 27(Suppl 4): 15–20. doi: 10.1159/000213054. Epub 2009 Jun 18.

Müller-Barna

Schwamm

Haberl

. Telestroke increases use of acute stroke therapy. Curr Opin Neurol 2012; 25: 5–10. doi: 10.1097/WCO.0b013e32834d5fe4.

Cooper

Hedges

Valentine

. The handbook of research synthesis and meta-analysis, 2nd ed. New York: Russell Sage Foundation, 2009.

10.

Higgins J, Green S. Cochrane handbook for systematic reviews of interventions Version 5.1. 0 [updated March 2011]. Chichester, West Sussex, UK: The Cochrane Collaboration, 2011.

11.

Stang

. Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses. Eur J Epidemiol 2010; 25: 603–5. doi: 10.1007/s10654-010-9491-z. Epub 2010 Jul 22.

12.

Chowdhury

Birns

Rudd

. Telemedicine versus face-to-face evaluation in the delivery of thrombolysis for acute ischaemic stroke: a single centre experience. Postgrad Med J 2012; 88: 134–7. doi: 10.1136/postgradmedj-2011-130060. Epub 2012 Jan 19.

13.

Pedragosa

Alvarez-Sabín

. Impact of telemedicine on acute management of stroke patients undergoing endovascular procedures. Cerebrovasc Dis 2012; 34: 436–42. doi: 10.1159/000345088. Epub 2012 Dec 1.

14.

Johansson

Mutzenbach

Ladurner

. Telemedicine in acute stroke care: the TESSA model. J Telemed Telecare 2011; 17: 268–72. doi: 10.1258/jtt.2011.101213.

15.

Zaidi

Syed

Jumma

. Telestroke-guided intravenous tissue-type plasminogen activator treatment achieves a similar clinical outcome as thrombolysis at a comprehensive stroke center. Stroke 2011; 42: 3291–3.

16.

Audebert

Schultes

Tietz

. Telemedical Project for Integrative Stroke Care (TEMPiS). Long-term effects of specialized stroke care with telemedicine support in community hospitals on behalf of the Telemedical Project for Integrative Stroke Care (TEMPiS). Stroke 2009; 40: 902–8. doi: 10.1161/STROKEAHA.108.529255. Epub 2008 Nov 20.

17.

Switzer

Hall

Gross

. A web-based telestroke system facilitates rapid treatment of acute ischemic stroke patients in rural emergency departments. J Emerg Med 2009; 36: 12–18. doi: 10.1016/j.jemermed.2007.06.041. Epub 2008 Feb 1.

18.

Schwab

Vatankhah

Kukla

. Long-term outcome after thrombolysis in telemedical stroke care. Neurology 2007; 69: 898–903.

19.

Audebert

Kukla

Vatankhah

. Comparison of tissue plasminogen activator administration management between Telestroke Network hospitals and academic stroke centers: the Telemedical Pilot Project for Integrative Stroke Care in Bavaria/Germany. Stroke 2006; 37: 1822–7. Epub 2006 Jun 8.

20.

Sutton

Duval

Tweedie

. Empirical assessment of effect of publication bias on meta-analyses. BMJ 2000; 320: 1574–7.

21.

Meyer

Raman

Hernmen

. Efficacy of site-independent telemedicine in the STRokE DOC trial: a randomized, blinded, prospective study. Lancet Neurol 2008; 7: 787–95.

22.

Rubin

Demaerschalk

. The use of telemedicine in the management of acute stroke. Neurosurg Focus 2014; 36(1): E4–E4. doi: 10.3171/2013.11.FOCUS13428.

23.

Johansson

Wild

. Telerehabilitation in stroke care–a systematic review. J Telemed Telecare 2011; 17: 1–6. doi: 10.1258/jtt.2010.100105. Epub 2010 Nov 19.