|Home | About | Journals | Submit | Contact Us | Français|
Due to a shortage of stroke specialists, many outlying or “spoke” hospitals initiate intravenous (IV) thrombolysis using telemedicine or telephone consultation prior to transferring patients to a regional stroke center (RSC) hub. We analyzed complications and outcomes of patients treated with IV tPA using the “drip and ship” approach compared to those treated directly at the RSC.
A retrospective review of our Get With the Guidelines Stroke (GWTG-Stroke) database from 01/2003-03/2008 identified 296 patients who received IV tPA within 3 hours of symptom onset without catheter-based reperfusion. GWTG-Stroke definitions for symptomatic intracranial (sICH), systemic hemorrhage, discharge functional status and destination were applied. Follow-up modified Rankin Score (mRS) was recorded when available.
Of 296 patients, 181 (61.1%) had tPA infusion started at an outside spoke hospital (OSH) and 115 (38.9%) at the RSC hub. OSH patients were younger with fewer severe strokes than RSC patients. Patients treated based on telestroke were more frequently octogenarians than patients treated based on a telephone consult Mortality, sICH and functional outcomes were not different between OSH vs. RSC, and telephone vs. telestroke patients. Among survivors, mean length of stay was shorter for OSH patients but discharge status was similar and 75% of patients walked independently at discharge.
Outcomes in OSH “drip and ship” patients treated in a hub-and-spoke network were comparable to those treated directly at an RSC. These data suggest that “drip and ship” is a safe and effective method to shorten time to treatment with IV tPA.
Intravenous tissue plasminogen activator (IV tPA) is FDA approved for acute ischemic stroke (AIS) within the first 3 hours of symptom onset1.Previous studies have shown that IV tPA is underutilized and that only 3-4% of AIS patients receive thrombolytic treatment 2. A majority of acute stroke patients present to hospitals that lack 24/7 availability of a stroke specialist 3, 4, are managed by generalists5, and many do not receive tPA at the presenting facility6, 7. If practitioners opt for transfer to a regional stroke center (RSC) for initiation of tPA, delays may preclude delivery of tPA.
Hub-and-spoke networks have been developed to allow IV tPA to be initiated at an outlying spoke hospital (OSH) under the supervision of a vascular neurologist at the designated RSC hub who provides assistance in the diagnosis and management of AIS via telephone8 or telemedicine-enabled (“telestroke”)9 consultation prior to transfer. This practice has been referred to as “drip and ship” 10. The evidence supporting telemedicine for acute stroke care within a hub and spoke network has recently been reviewed, and guideline recommendations published 11, 12.
Limited data exist comparing the safety and feasibility of patients treated by “drip and ship” 13 versus patients who arrive directly at the RSC for tPA, or the methods of consultation.
All AIS patients at our RSC are prospectively identified and entered in our Get With the Guidelines-Stroke (GWTG-Stroke) quality improvement database during their admission14. Because de-identified data is collected for quality improvement purposes, only data available in the medical record is abstracted and our institutional review board (IRB) has waived the need for informed consent and approved the analysis of our GWTG-Stroke database. We identified all patients admitted to our RSC hospital from 01/03-03/08 who received IV tPA at our hospital or at an OSH under our direct supervision prior to transfer, and without subsequent catheter-based reperfusion therapy.
Twelve of the 33 referring OSHs in this report had an established telemedicine connection with our hospital during the study period; the remainder used telephone. All consultations are captured in our online acute stroke log. Our telemedicine program uses full motion, interactive videoconferencing with teleradiology, training and education, system maintenance and case feedback, and has been previously described9, 15. Hospitals requesting acute stroke consultation from our RSC hub by telephone receive the same instructions regarding eligibility as hospitals that use telemedicine (see table 1 for inclusion and exclusion criteria of tPA use in our hospital). The instructions to OSHs include post tPA care management advice. After all intravenous tPA administration using telephone or telemedicine consult, the treating physicians at our RSC hub reviews with the emergency physician at the OSH the importance of post tPA care, and basic immediate recommendations such as blood pressure parameters, and avoidance of invasive procedures, antiplatelet and anticoagulant use. Clinical informed consent for tPA was obtained in all cases.
Post-tPA symptomatic intracranial hemorrhages (sICH) were classified according to ECASS16 and GWTG-Stroke criteria (clinical worsening within 36 hours of tPA due to hemorrhage) by our institutional acute stroke quality taskforce. When available, followup modified Rankin Scores (mRS) were abstracted. mRS were extracted from patient records by a trained physician abstractor 17. No patients were contacted for the purpose of this study but the social security death index was searched. Death between discharge and 3 months was assigned to the 3 month timepoint, between 3 to 6 months to the 6 month timepoint and between 6 to 12 months to the 12 month timepoint.
Univariate analyses of baseline patient characteristics and discharge outcomes were performed in STATA 9.2 (STATA, Inc) using chi-square test, Wilcoxon rank sum test, Fisher's exact test, as appropriate. Because follow-up mRS scores were abstracted from the medical record, they were available at either 3, 6 or 12 months post discharge. A nonparametric rank test was used to compare the follow-up mRS levels between RSC patients and OSH patients; and within OSH patients, between the telestroke and telephone consultation methods. This test stratified subjects according to their length of follow-up. We implemented the test using a Cox proportional hazard model, with mRS at follow-up as the “event time,” with no censoring, and with follow-up time as the strata. We performed the analysis both with and without patients who died in hospital during their index stroke admission. Multivariable logistic regression was used to correct for imbalances between the groups. Baseline variables significant to P <0.1 in univariate analysis (age) or previously shown to strongly influence outcome (onset to TPA time and NIHSS) were entered in the model. Because of the variability in follow-up time to mRS, we stratified the rank test by follow-up time (0, 3, 6, or 12 months), but assumed a common effect for hospital, or method, across the strata. For subjects without any follow-up, we fixed their follow-up time to be zero and used their mRS at discharge. Subjects who died in-hospital were assigned a mRS of 6 and follow-up time of 0.
Among 2,526 AIS patients admitted from January 2003 to March 2008 to our RSC hub, there were 353 (14.0%) AIS patients who received IV tPA treatment within 3 hours of symptom onset at either an OSH or our RSC hub. Among these patients, 296 (83.9%) received IV tPA without any additional catheter-based reperfusion therapy and were selected for analysis. Forty-one patients received tPA under telephone/telemedicine supervision but were not transferred to our RSC hub. Those patients were not included in our study. Overall, patients were 47.6% male, 94.6% white and ranged in age from 20-102 years (mean 72.3±13.9), with 35.5% of the patients ≥ 80 years of age (Table 2). In-hospital mortality in the ≥ 80 year old subgroup was 28/102 (27.5%), with only 6 deaths in patients with symptomatic ICH. There were no apparent protocol violations in the determination of eligibility criteria for treatment with intravenous tPA. Patients were transferred from hospitals in Massachusetts (n=27 hospitals), New Hampshire (n=5 hospitals) and Maine (n=1 hospital). Among 33 OSH, 12 had telestroke capability. The median [interquartile range] distance from the OSH to our RSC was 43.3 miles [17.4, 57.6], with telestroke hospitals farther away than telephone hospitals (52.2 [42.9, 59.6] vs. 40 [17.4, 54.8], p=0.005).
Baseline patient demographics, treatment complications and discharge outcomes were compared between patients who started tPA at OSH vs. RSC (Table 2). OSH patients were younger (71.5+14.7 vs. 73.6+12.4, p = 0.05) and less likely to have NIHSS > 20 (11.6% vs. 22.6%, p =.01). There were no significant baseline differences in gender, race, vascular risk factors, baseline ambulatory status, initial mean NIHSS and onset to tPA time. In unadjusted analyses, there were no differences in treatment complications and discharge outcomes. In OSH vs. RSC patients, the rates of sICH (3.87% vs. 5.22%, p = 0.58), serious systemic hemorrhage (0.55% vs. 2.61%, p = 0.14) and in-hospital mortality (14.92% vs.17.39%, p= 0.57) were not different but mean hospital length of stay was shorter (5.91±3.7 vs.7.63±6.5, p <0.001). Among survivors, ambulatory status at discharge, and discharge disposition did not differ between groups.
Follow-up mRS scores were available in 237/296 (80.1%) patients. These scores were obtained once for each patient, and were acquired at different timepoints of either discharge (59 patients-19.8%, including all in-hospital deaths), 3 months (116 patients-39.2%), 6 months (101-34.2%) or 12 months (20-6.8%), in a similar distribution between the OSH vs. RSC hub patients”. Lack of follow-up mRS was more common in OSH than RSC patients (32.1% vs. 15%, p=0.004). RSC patients without follow-up had similar age, pre-stroke ambulation status, presenting NIHSS, use of telestroke, tPA complications, length of stay, in-hospital mortality and discharge destination as RSC patients with followup. OSH patients without follow-up had similar rates of pre-stroke ambulation, use of telestroke, symptomatic ICH, in-hospital mortality and discharge destination as OSH patients with followup. There was a trend toward OSH patients without follow-up to be older, have lower NIHSS and more systemic hemorrhage after IV tPA.
Overall, 34.8% of IV tPA patients achieved a mRS of 0-1, and 32.5% were dead. In multivariate analysis, there was no difference in the likelihood of achieving a lower mRS score at follow-up between those who started IV tPA at the RSC vs. the OSH (Table 3), when stratified by the time to followup, adjusted for differences in age, time to tPA treatment and NIHSS, with or without in-hospital deaths. Only age (0.97 per one year) and initial NIHSS (0.93 per one point) were independently associated with achieving a lower mRS.
Among the 181 OSH cases, 31.0% achieved a mRS of 0-1 and 34% died. Supervision of tPA was by telestroke in 84 (46.4%) and telephone in 97 (53.6 %) patients. In the telestroke group, there were significantly more octogenarians (45.2% vs. 26.8%, p=0.01) and a trend toward more smoking. The unadjusted sICH rate (4.76% vs. 3.09%, p=0.56), overall mortality (10.71 vs. 18.56, p=0.14) and discharge outcomes among survivors were not different between telestroke vs. telephone, however mortality among the octogenarians was significantly lower in the telestroke group (18.4% vs. 42.3%, p= 0.05) (Table 4). In the multivariate analysis, there was no difference in the likelihood of achieving a lower followup mRS score for telestroke vs. telephone, adjusted for differences in age, time to tPA treatment and NIHSS, with or without in-hospital deaths (Table 5). Only age (0.98 per one year) was independently associated with achieving a lower mRS.
This study suggests that the current “drip and ship” practice of starting IV tPA at outlying hospitals followed by rapid transfer to a stroke center is safe and feasible compared to initiation of tPA at a regional stroke center. In our study, treatment complications such as sICH, mortality, discharge outcomes and long-term functional outcomes were no different between OSH and RSC patients. The mortality rates observed in our small retrospective series of patients were greater than seen in randomized clinical trials of intravenous tPA use. This may be due to patients with more severe deficits at baseline (median NIHSS of 12) and no upper age limit for tPA use, both important predictors of outcome in acute ischemic stroke. Over 30% of the patients in this series were greater than 80 years of age. Because the available data suggest a dismal outcome in older patients with severe stroke who are untreated, it is our policy to offer all patients the option of IV tPA if they qualify for treatment regardless of age and NIHSS. We could not control for the factors that may have led families to request a withdrawal of life-sustaining treatments due to poor neurologic prognosis. There were important baseline imbalances with older and more severe strokes treated at the RSC, but after adjustment, complication rates and mortality did not differ between OSH and RSC. Taken together, these findings are consistent with prior reports that suggest that “drip and ship” approaches can be effectively implemented in a variety of environments 9, 11, 12, 18 10
There were more octogenarians treated via telestroke than telephone, perhaps reflecting the increased willingness by emergency physicians to offer IV tPA to older patients based on telestroke support. Telestroke has recently been shown to be superior to telephone in accurate decision-making for IV tPA 19, 20 and may provide greater reassurance to referring physicians who treat relatively few patients. It is also readily accepted by patients, who express confidence in the remote management of their conditions via telestroke21. Despite the significantly higher number of octogenarians in the telestroke vs. telephone group, their mortality was significantly lower and this difference was not due to eligibility protocol violations. Taking into account the older age and higher stroke severity in the telestroke group, our outcomes are in line with those of previously published series of conventionally treated 22 and telemedically-treated23 patients.
It has been shown that participation in a national stroke quality improvement program increases adherence to evidence-based guidelines14, and it is likely that increased adherence to these measures will reduce in-hospital complications and perhaps improve long term functional outcomes for patients with acute ischemic stroke. These include recommended interventions such as deep vein thrombosis prevention, early antithrombotic therapy, lipid management, dysphagia screening, smoking cessation counseling, anticoagulation for atrial fibrillation, early rehabilitation assessment and services, and patient education 24. Therefore, many stroke patients may benefit from transfer from a spoke hospital to an RSC that is engaged in a structured national stroke quality improvement program.
Initial studies validated the reliability of the telemedicine-enabled (“telestroke”) assessment of the National Institutes of Health Stroke Scale (NIHSS) 15, 25 26, 27 in both subacute and acute ischemic stroke patients. A single prior study suggested that IV tPA supervised via telephone may be feasible and time-efficient, although there are limited data on long term outcomes comparing these patients to those who arrive directly to the RSC 8. Recently, there has been increased attention on telestroke as a means to facilitate increased access to acute stroke care for rural or neurologically underserved areas28, 29. The Center for Medicare and Medicaid Services recently issued a new administrative V code (V45.88 “Status post administration of tPA in a different facility within the last 24 hours prior to admission to current facility”) to permit tracking of this “drip and ship” practice, and potentially address the reimbursement inequity created in this scenario in which neither the spoke nor hub hospital receives the appropriate higher diagnosis-related group payment (DRG 61-63). Our study provides additional scientific support for this “drip and ship” approach.
Because tPA's benefit diminishes rapidly with increased time to treatment 30-32, “drip and ship” should improve treatment rates and outcomes compared to transfer to RSC's for tPA 6. Many smaller hospitals have used telestroke “drip and ship” approaches to overcome physician reluctance 5, 13, 33 34and become “tPA-capable” centers 35. The results from our study suggest complication rates are comparable to that of the NINDS IV tPA trial 1 despite a substantially older population.
Our study has some potential limitations. It is a retrospective series and therefore may not be adequately powered to detect significant differences in complications or outcomes between the groups at spoke vs. hub hospitals or between telephone and telestroke consultation. To control for any difference associated with hospital inpatient care processes, we only included those patients treated with IV tPA at an outside hospital if they were subsequently transferred to our RSC hub. In this manner, all patients had the same post-tPA care and secondary prevention opportunities. Therefore, these results may not apply to patients who receive IV tPA under remote supervision but remain at an OSH (“drip and keep”) that is not a fully equipped stroke center. The data from the TeMPIS project in Bavaria23 strongly supports the “drip and keep” model if the use of telemedicine can fill in the missing pieces to make a functioning OSH stroke unit, but we have no data on the outcomes of patients who remain at an OSH that functions below the level of a stroke unit or primary stroke center. We do not have data on the number of patients who presented to referring hospitals with acute stroke suspect who were not discussed using telemedicine. Nevertheless, based on our rates of treatment with intravenous thrombolysis (13% of all telemedicine and telephone consults and 24% of acute ischemic stroke diagnoses from outside referring hospitals within the time period of this study) which is at or above the range of the reported rates of t-PA use in community hospitals 36, we believe that outside hospital emergency physicians are contacting our system for most of the acute stroke suspects evaluated in their emergency room. Previous studies from our group also suggest that referring emergency physicians involved in our program endorse the quality of our Telestroke system and felt confident managing patients with telemedicine support, therefore we do think that our hospital is contacted for the majority of the acute stroke suspects that present to our referring hospitals emergency departments 9. Also, since our study design deliberately excluded those patients who underwent rescue reperfusion therapy for failed IV tPA, additional studies would need to be performed to address that population. Lastly, our follow-up mRS data was limited due to the different times that follow-up was available, and the lack of follow-up data in 20% of the patients. Six and 12 month outcomes incorporates some variability due to comorbidities and rehabilitation interventions. We do not have data on rehabilitation duration and quality provided to our patients. Reassuringly, the group without followup was similar in most ways to those patients with follow-up assessment, there were no differences in the follow-up data distribution nor in inpatient rehabilitation use between the OSH and RSC hub patients, and we used a statistical approach to try to maximize the follow-up information available.
In summary, these results suggest that the “drip and ship” model of IV tPA treatment is safe and can be readily applied to a wide variety of hospital settings for acute stroke patients who arrive within 3 hours of stroke symptom onset. With only 4 neurologists per 100,000 persons in the US, and recent evidence suggesting IV tPA is effective up to 4.5 hours after symptom onset37, 38, remote support of intravenous thrombolysis will play a crucial role in meeting an ever-growing demand.
This work was supported by AHRQ RO1HS011392, NIH P50 NS051343, the American Heart Association-Bugher Foundation, Harvard NeuroDiscovery Center and the Deane Institute for Integrative Research in Atrial Fibrillation and Stroke. We recognize the generous support of the Esther U Sharp Fund, the Conway Fellowship Fund, the Lakeside Fund, and Levitt Fund. .The authors gratefully acknowledge the shared trust of the participating spoke hospitals in our network, and the clinical contributions of the MGH Acute Stroke Team staff and fellows who work tirelessly in support of the telestroke program at MGH, providing advice and recommendations to emergency physicians across the network.
Conflict of Interest The authors report no financial conflicts of interest