|Home | About | Journals | Submit | Contact Us | Français|
The Interventional Management of Stroke (IMS ) I trial suggested that combined use of reduced-dose intravenous (IV) ActivaseR (alteplase, rt-PA), followed by microcatheter-delivered intra-arterial () rt-PA, was relatively safe, and may be clinically useful, in selected acute ischemic stroke patients, as compared to similar subjects treated with full dose of IV rt-PA in the National Institute of Neurologic Disease and Stroke (NINDS) rt-PA Trial.1 It had long been anticipated that multiple interventional methods might be applicable within an IV/IA rt-PA paradigm, and that these methods might further improve revascularization and clinical outcomes2. Encouraged by safety and potential efficacy reports in a small feasibility trial of the EKOS MicroLusUS Ultrasound Catheter (hereafter referred to as EKOS® Micro-Infusion Catheter), the IMS Investigators initiated another phase II trial (IMS II) to further study its use following reduced-dose IV rt-PA therapy.3 This manuscript reports in detail the revascularization outcomes in the IMS II Trial, including general as well as some detailed comparisons to revascularization outcomes from IMS I.
The IMS II Study was a 13-center, open-labeled, single-arm pilot study, whose purpose was to determine futility or lack thereof of the combined IV/IA therapy and to obtain preliminary estimates of efficacy and safety of reduced-dose IV rt-PA (0.6 mg/kg), 15% as bolus with the remainder infused over 30 minutes, followed by additional rt-PA (up to 22 mg) and low-energy ultrasound via the EKOS® Micro-Infusion Catheter at the site of arterial occlusion in eligible acute ischemic stroke patients less than 81 years of age with large ischemic strokes (NIHSS>/=10) treated within three hours of symptoms onset. Three models of the EKOS catheter were used sequentially in the Trial. The EKOS® System consists of 2 components: 1) a single-use, 3.0 F EKOSR Micro-Infusion Catheter Primo model with an end-hole infusion lumen incorporating a high-frequency (1.7 mHZ) cylindrical ultrasound transducer delivering ultrasound at low power (<0.45W) into its 3.3F distal tip, and 2) a re-usable control unit which provides the ultrasound energy source and the user interface. The first model (SV2510), similar to that used in a Feasibility Trial, was intended for passage over a 0.010 microguidewire. Because successful access occurred in only 2 of 6 subjects between 1/26/03 and 4/1/03 caused the IMS II Steering Committee to temporarily stopped the trial to change to a 0.014” guide-wire based system (SV3014) and upgraded control unit.. The Trial was restarted 5/5/03. However, successful access in 2 of 4 subjects, again prompted study delay, in expectation of a new third model under development. The Primo model was used for the remainder of the trial between 12/17/03 and 4/12/05.
Details of study design, objectives, study protocol, EKOS and microcatheter administration of IA rt-PA, and medical management and evaluation have recently been published, and are not repeated here.1
The primary study safety measure was life-threatening bleeding, particularly intracerebral hematoma or hemorrhagic infarction with clinical deterioration. CT scans within the first 36 hours after completion of rt-PA infusion were monitored for ICH and contrast deposition, including contrast extravasation (CT density greater than 90 Hounsfield units, hyperdensity persisting at 24 hours) and contrast enhancement (high attenuation disappearing within 24 hours).5 ICH was classified according to the ECASS classification, subdivided into hemorrhagic infarction (HI) type 1 and 2, and parenchymal hematoma (PH) type 1 and 2.6 Vessel perforation, dissection, and subarachnoid hemorrhage were also recorded.
This report focuses on the secondary clinical efficacy measure of outcome of modified Rankin 0–2 outcome at 3 months. Mortality was also recorded and both these measures were further analyzed according to revascularization status. Revascularization was measured according to both recanalization of the primary arterial occlusive lesion (AOL), as well as modified Thrombolysis in Cerebral Infarction (TICI) reperfusion parameters.7,8 AOL recanalization and the modified TICI perfusion were scored grade 0–3 (Table 1). The modified TICI reperfusion score was essentially equivalent to TIMI score applied in IMS I, with grade 2 further divided into A and B for post-hoc analysis. The primary study revascularization end point was % complete recanalization of the primary occlusion at (AOL 3) at 60 minutes, compared to standard microcatheter thrombolysis in the IMS I trial.
Secondary revascularization measures included AOL 3 and TICI 2/3 perfusion of the vessels distal to the AOL at 120 minutes, or the end of the procedure. In addition, any recanalization of the AOL at each 15-minute interval was recorded. Revascularization end points were compared to Modified Rankin Score (MRS) and mortality outcomes from the IMS I study. The revascularization study end points were determined in the core at the University of Cincinnati lab by a central reader (TAT). However, operators were asked to ascribe TIMI recanalization scores for their procedures, as had been recorded in IMS I. These scores were compared to the central core lab interpretation.
Angiograms were examined for new occlusions in previously uninvolved arterial distributions. Distal emboli in the MCA beyond an MCA occlusion could not be systematically identified, because their pretreatment presence/absence was seldom confidently identified or excluded. Baseline and new emboli into the ACA distribution with MCA and internal carotid terminus, ICA-T (obstructed ICA flow into the M1 and A1 segments) occlusion were recorded.
Catheter-tip temperature was recorded for all IMS II procedures using the EKOS ultrasound microcatheter and reviewed post-procedure. Temperature graphs were compared to 15-minute control angiograms to identify if observed temperature changes correlated with recanalization.
Microcatheter contrast injections were recorded, and compared with contrast extravasation and ICH.9
Infarct, symptomatic ICH, and lesion (infarct + ICH) volumes were retrospectively measured digitally using Cheshire software via manual region of interest outline, and reported (G. Ramadas, personal communication, 2006).
A -appointed Data and Safety Monitoring Board Committee reviewed all safety and outcome results of the IMS I and II Studies. Following completion of IMS II, IMS IIB was approved by NINDS for 48 additional subjects, with the expectation of learning more about the EKOS efficacy and safety, as well as to observe temperature changes as displayed on the control box in a prospective, real-time manner. However, only 3 additional subjects were uneventfully treated by the EKOS catheter, so that goal was not achieved prior to the study termination in anticipation of initiation of the IMS III Trial.
Eighty-one subjects were entered into the IMS II trial (including 73 subjects from IMS II, and 8 subjects from IMS IIB) (Fig. 1). Twenty-six were treated with IV rt-PA-only, and 55 treated with IV and IA rt-PA (36 by EKOS microcatheter, and 19 by standard microcatheter). Among the 26 treated IV-only, 2 improved clinically following IV rt-PA, did not have arteriography, but underwent MRA instead, with no major AOL. Twelve others had no major treatable AOL at arteriography, 2 with no demonstrable AOL, and 10 with distal MCA, ACA, or PCA occlusions not treated by the interventionist. Of these 14 with no major AOL demonstrated on angiography or MRA, MRS 0–2 was achieved in 9 (64.3%).
Of the remaining 12 IV-only subjects, 1 was discovered to have a high INR and was excluded from IA treatment. In one subject with coarctation of the aorta, the ascending aorta could not be accessed in the usual retrograde transfemoral fashion for diagnostic arteriography and treatment. Three subjects were beyond the prescribed 5-hour initiation time; 2 had non-protocol IA treatments other than rt-PA, both with good outcomes. The other 7 subjects in the IV-only group had AOL eligible for IA treatment. Five had major intracranial occlusions distal to cervical ICA occlusions that could not be accesse due to inability to traverse the cervical ICA occlusions. One M2/3 occlusion was neither identified nor treated by the local treating interventionist. An early EKOS Ultrasound catheter model would not traverse the distal ICA in another due to cervical and cavernous segment ICA tortuosity. MRS 0–2 outcomes were achieved in 3 of these latter 7 (42.9 %) subjects not treated by IA therapy.
Fifty-five subjects were treated with both IV and IA rtPA.. Thirty-five subjects had the final Primo model introduced. This catheter could not be advanced beyond tortuous segments of ICA in 3 subjects (8.6%), who were then treated with standard microcatheter. Three subjects were treated without ultrasound initiated, 2 regionally (proximal to the primary AOL) in the settings of ICA and vertebral artery dissection, respectively. A catheter-control box interface cable connection failed in the third subject. Twenty-nine subjects were treated with the EKOS Primo catheter with ultrasound activation. One of the 29 subjects had both an M2 AOL treated with EKOS Ultrasound Microcatheter, and a P2 occlusion treated with standard microcatheter.
Nineteen subjects were treated by standard microcatheter, including 4 where EKOS catheter advancement had been unsuccessful. In addition, there were 4 M3, and 1 ACA occlusion, where the EKOS catheter was not indicated. One other subject had an extremely tortuous cervical ICA, and the appearance of intracranial atherosclerosis, where the EKOS catheter was not indicated. Four other subjects had incomplete occlusions and IA rtPA was administered via standard microcatheter. One subject was accessed after 5 hours, and an EKOS catheter was not introduced. The 4 other subjects had EKOS-eligible occlusions, but were treated with a standard microcatheter, according to the discretion of the operator.
The primary study revascularization end point of complete recanalization of the primary AOL at 60 minutes was met in 12 of 29 (41.4%) PRIMO catheter, and 12/33 (36.3%) of all ultrasound-catheter subjects with ultrasound activated (Fig. 2). Twenty-three IMS I subjects had reliable 60-minute images for comparison, and 7 (30.4%) had complete recanalization (p=0.38). Complete recanalization of the primary AOL was achieved at 2 hours or by procedure-end in 20/33 (60.6 %) of the entire ultrasound-catheter treatment group, and 20/29 (68.9%) EKOS Primo catheter-treated group, compared to complete recanalization by the end of the procedure in 28/55 (50.5%, p=0.11) comparable AOL in IMS I.
Recanalization of the primary AOL leads to reperfusion into distal vessel segments. Final TICI 2/3 perfusion for the 33 IMS II EKOS-treated subjects was 60.6% compared to 56.8% for comparable IMS I IA-treated group. Revascularization and clinical outcomes clinical outcomes are presented in Table 2 for each category of AOL in the IMS II EKOS Ultrasound and standard microcatheter groups, as well as for the IMS I study.
Table 3 presents MRS 0–2 outcomes for ICA T and M1 occlusion in IMS II and I subjects combined, according to the target revascularization end points. No significant differences existed between studies. These pooled data demonstrate that good recanalization (AOL 2/3) occurred in 56/75 (74.6%), and good reperfusion (TICI 2/3) occurred in 46/75 (61.3%) of ICA-T and M1 occlusions. Pooled data also demonstrate that revascularization correlates with good outcome for TICI 2/3 reperfusion (p=0.0004), TICI 2B/3 reperfusion (p=0.0002), and AOL 2/3 recanalization (p=0.03) end points, compared to failure to achieve those end points, respectively. Although only 10/23 (43.5%) M2 occlusions achieved Grade 3 AOL recanalization, 16/23 (69.5%) achieved MRS 0–2 outcome. Two subjects with M2 occlusions, both of whom recanalized and were discharged to home, were lost to follow-up, and are included in the MRS > 2 group. Comparison of the TICI 2/3 and AOL 2/3 end points as predictors of MRS 0–2 outcome shows sensitivity of 90.5% vs 95.2%, and specificity of 52% vs 34%, respectively.
Sequential 15-minute control IMS II angiograms demonstrated some degree of recanalization in 69/145 (46.7%) EKOS Ultrasound Microcatheter treatment intervals, compared to 28/65 (43.1%) comparable IMS II microcatheter treatments (p=0.09), and 39/111 (35.1%) comparable IMS I treatments in 23 subjects where reliable 15-minute angiograms were available (p=0.046). Four (7.3%) reocclusions (3 complete) occurred during the procedure in 55 IA-treated IMS II subjects.
New emboli in the ACA distribution were identified in 1/22 (4.5%) M1/M2 occlusions, and in 1/14 (7.1%) ICA-T occlusions.
Retrospective temperature analysis provided 93 IMS II data points available with both arteriographic and temperature data for retrospective analysis. Temperature decreases of as little as 0.2–0.3° C could be readily identified. Forty (41.9%) temperature decreases were identified. Forty-five (48.3%) recanalizations or microcatheter retractions were identified. Temperature analysis provided 81% sensitivity 90% specificity, 85 % positive predictive value, 87% negative predictive value, and 84.9% accuracy identifying recanalization or catheter retraction on the following arteriographic sequence.
Central core lab IMS II TIMI revascularization scores demonstrated a 41% discrepancy between the core lab and the treating interventionist, with the treating interventionist usually scoring higher (better revascularization).
Available 24-hour CT lesion volumes demonstrated no significant differences in IMS II infarct volumes compared to IMS I subjects overall, or to those treated by combined IV/IA methods.
Overall IMS II mortality was 16% subjects, similar to 16% mortality in IMS I. Of the 13 deaths, 4 were treated IV-rt-PA only: 1 died secondary to SICH following failure to identify and treat an M2-3 occlusion. Three others died from the sequelae of a large cerebral infarction; one of the three had an ICA occlusion that could not be traversed for IA therapy, a second not treated after the 5-hour limit, and the third had coarctation of the aorta that couldn’t be traversed for IA therapy. Of the 9 (16.3%) IV/IA deaths, four (44.4%) subjects did not recanalize to AOL grade 3, and 6 (66.6%) did not perfuse to TICI 2/3. Overall mortality, and mortality in revascularized and non-revascularized subjects with ICA-T and M1 occlusions, is further detailed in Table 3. Reduced mortality was associated with TICI 2/3 reperfusion (p=0.01), TICI 2B/3 reperfusion (p=0.05), and AOL 2/3 recanalization (p=0.03), compared to failure to meet those revascularization end points. Mortality in subjects failing to achieve either TIMI 2/3 reperfusion or AOL 2/3 recanalization was less than 40% for both end points.
Symptomatic ICH occurred in 8/81 (9.9 %) IMS II subjects, including 1/26 (3.8 %) treated with IV rt-PA alone, and one in an ultrasound microcatheter-treated subject without ultrasound applied. Six subjects in the ultrasound microcatheter group experienced symptomatic ICH, including:
Of these 6 subjects, 5 had NIHSS ≥ 20. Four SICH were PH1, and 2 PH2, based on digital volume measurement. Mean hematoma volumes in these 6 SICH measured 23.5 cc, representing 27.9% of lesion volume, compared to 25.2 cc, or 37.1% of lesion volume, for 6 non-ultrasound IMS II and IMS I SICH.
Three of the seven SICH had microcatheter (MCI) injections. An early concern regarding a potential link between MCI and ICH was transmitted to investigators.9 No direct vessel perforation, primary SAH, or intracranial dissection was documented in IA-treated subjects.
The IMS II study was designed to study the safety and potential efficacy of the EKOS Ultrasound Microcatheter in recanalization of acute intracranial arterial occlusions in acute ischemic stroke, following reduced-dose IV rt-PA therapy. Both clinical and technical factors contributed to the actual use of an EKOS Ultrasound Microcatheter in only 33 subjects. Clinical exclusions for AOL location and etiology (atherosclerotic occlusion, dissection) occurred. Technical problems (connector cable problem, difficulty in catheter passage, etc.) further limited ultrasound administration and evaluation in several subjects. Some subjects were excluded from use based on clinical, angiographic, or other operator-defined exclusion factors.
When used, the EKOS Ultrasound Microcatheter tended to achieve faster and more complete AOL recanalization than standard microcatheter use in IMS II, an incomplete 60-minute data set from IMS I, and the complete 120-minute data set from IMS I (Fig. 2), although these differences were not statistically significant. It should be noted that standard microcatheter use in IMS I and II allowed mechanical manipulation of thrombus and guidewire, including microcatheter passage and drug delivery distal to the occlusion initially, as well as 15-minute guidewire/microcatheter advancement. Ultrasound microcatheter use was limited to proximal bolus and drip infusion, with no guide wire manipulation beyond that required for initial and subsequent microcatheter advancement. It is hypothesized that the apparent 15 to 30-minute equivalence of microcatheter use in Fig. 2, with improved early recanalization effect compared to ultrasound, may be due to mechanical microcatheter/guidewire manipulation of favorably-predisposed occlusions that will be achieved with any mechanical manipulation applied within the IV/IA paradigm. Allowing microguidewire/ultrasound microcatheter manipulation in a similar fashion may further enhance recanalization potential.
Identification of 14/81 (17.2%) subjects without AOL qualifying for IA EKOS-Microcatheter ultrasound-assisted thrombolysis confirmed pre-IMS and IMS I study observations: IV rt-PA may recanalize some vessels early, leading to normal arteriograms (n=2), or distal occlusions (n= 10). (Fig. 1).1,10 Comparison of recanalization rates using the combined approach in IMS I and II with studies of IA-only treatment must account for this early post-IV recanalization into the formula.
It is not possible to accurately compare recanalization and reperfusion results achieved with combined IV/IA rtPA therapy paradigm to other reported IA revascularization methods in which treatment was initiated later. The 78% AOL 3 recanalization and 83.3% TICI 2/3 reperfusion of isolated M1 occlusion in IMS II, comparable to 78% for both measures in IMS I, may represent a true recanalization and reperfusion benefit compared to results obtained with r-pro-urokinase treatment in PROACT II (67% TIMI 2/3 for M1 and M2 occlusions), and with the MERCI Retriever in the MERCI and Multi-MERCI Trials (45% and 54% TIMI 2,3, respectively).11,12 However, it still remains uncertain that revascularization end point definitions and application in other revascularization studies are comparable with the IMS studies.
Importantly, TICI reperfusion rates with the IMS IV/IA paradigm again suggest 33/55 (60%) subjects with treatable clot afterIV rt-PA achieve TICI 2–3 distal perfusion with EKOS-Ultrasound-assisted or standard microcatheter IA rt-PA thrombolysis, comparable to IMS I. Differences in baseline AOL in IMS II and IMS I existed, and revascularization results differed for the various AOL treated (Table 2). Most notably, there were fewer ICA bifurcation stenoses and occlusions in IMS II than in IMS I. Yet 7/11 (63%) IMS II subjects with ICA occlusion or stenosis achieved mRankin 0–2 outcome. Overall, 35/153 (23%) IMS I and II subjects presented with acute ischemia and severe cervical ICA stenosis (n=21) or occlusion (n=14). Mean baseline NIHSS for all subjects with ICA stenosis/occlusion was 17.7. Distal AOL accompanying ICA occlusion or stenosis (tandem lesions) included 10 (28%) ICA terminus, 20 (57%) M1or M2 occlusions. There were 5 instances of no or distal M3, M4 occlusion. Mean baseline NIHSS for the 30 subjects with major tandem lesions was 17.6, less than for the entire IMS I and II IA-treatment cohort. Functional clinical outcome (MRS ≤ 2 at 90 days) was achieved in 10 (33.3%) of the 30 tandem occlusion subjects, with 20% (6/30) mortality. mRS ≤2 outcomes were achieved in 3/17 (17.6%) of the atherosclerotic ICA stenosis and in 2/8 (25%) of the occlusion groups. Of 8 subjects with atherosclerotic ICA occlusions, four were stented in non-protocol, off-label fashion, with 2 achieving mRS ≤2, and with one death. The two patients with good outcomes had individual baseline NIHSS of 10 and 12, respectively. In the non-stented occlusion group, one achieved mRS ≤2, and there were 2 deaths. There were 5 ICA dissections in the tandem group. One subject with ICA dissection was stented. MRS 0–2 outcome was more favorable for those with ICA dissection compared to atherosclerotic disease (5/5 vs 5/25, p=0.002). A trend toward better outcomes with an IV/IA thrombolysis paradigm in acute ischemia with tandem ICA and distal lesions, compared to IV tPA alone, may be suggested in the IMS studies.13
A relationship between both AOL and TICI revascularization endpoints and mRS 0–2 outcome was demonstrated for ICA-T and M1 occlusion (Table 3). TICI 2/3 reperfusion better predicted good outcome than did primary AOL 2/3 recanalization. To further analyze the TICI 2 reperfusion end point in hope of identifying an acceptable or optimal stopping point for revascularization, subjects were also categorized as TICI 2A, or 2B, based on the presence of up to 50% MCA distribution reperfusion, or TICI 2B with less than 50% (Tables 1,,3).3). TICI 2B/3 perfusion best predicted good outcome statistically (p=0.0002), more robustly than TICI 2/3 (p=0.0004). There was a trend for TICI 2B superiority to TICI 2A in predicting good outcome (p=0.08). AOL 2/3 recanalization also predicted good outcome (p=0.027), suggesting recanalization predicts good outcome nearly as well as reperfusion. Safe revascularization appears to be an appropriate surrogate end point for the IV/IA treatment paradigm. It is not as clear that a link between no-revascularization and poor-outcome is due to the no-revascularization effect itself. Procedural factors (e.g., contrast deposition, heparin use, saline deposition, distal emboli, hemorrhage into injured brain) during prolonged, failed revascularization procedures may contribute to the lack of good outcomes observed with this treatment paradigm.
M2 occlusion analysis does not lend itself to the TIMI/TICI method of grading, creating, in effect, a circular reference where the score definition is dependent on perfusion through the involved sentinel M2 distributions. Good outcomes were not dependent on revascularization, with 13/17 IMS (76%) I and II isolated M2 subjects achieving good outcome despite incomplete recanalization and reperfusion. This likely relates to the much smaller volume of ischemic brain with branch occlusions as compared to occlusions of major trunk arteries like the M-1 or ICA.
Of interest, operators scored their revascularization efforts more optimistically than the core lab. Differences between a TIMI score of 2 and 3 are of questionable practical significance, since these 2 scores are commonly viewed together as significant revascularization, correlating with increase in good outcome (Table 2). Nevertheless, it raises a legitimate question of accepting unadjudicated prospective data regarding important end points as a determinant of efficacy. In addition, this observation raises questions regarding revascularization end point terminology, definition, convention, and application. “Good” recanalization (AOL 2/3) in 43/71 (60.5%), and “good” reperfusion in 53/71 (74.5%) of ICA-T and M1 occlusions suggests that both measures are legitimate determinants of revascularization, but 15% difference indicates that they they are not equivalent.
Ultrasound microcatheter activity measurably increases catheter-tip temperature. Monitoring this effect as a safety measure provided a unique opportunity to examine the effect of recanalization or catheter retraction on catheter tip temperature. Retrospective correlation of catheter-tip temperature decrease with recanalization suggests that observing tip temperature during a procedure might prospectively allow identification of recanalization between strict prescribed angiogram intervals, and therefore might allow the revascularization procedure to be shortened. Retrospective analysis of temperature changes was not complete for all subjects. However, where available, the correlation between a decrease in temperature and recanalization or catheter retraction from the occlusion was quite strong. Whereas some recanalization was identified in 47.6% of subjects, and temperature decreases were identified in 41% of subjects, treatment time might be reduced accordingly (Fig. 3).
Identification of higher ICH and contrast deposition rates with MCI suggest technical factors may be important in clinical outcomes. Higher ICH rates with ultrasound microcatheter use were confounded by this observation. Demonstration of rupture points of distal lenticulostriate arteries legislated against local ultrasound or local temperature changes as responsible for these ICH cases in several instances. No single etiologic mechanism seems responsible for the observed ICH. Multifactorial contribution of rtPA, heparin, contrast deposition, and microcatheter injection pressure/volume is hypothesized. A correlation between MCI and ICH, as well as PH, has also been found in a thrombolysis registry analysis.14
MRS 0–2 outcomes were achieved in 45% IMS II compared to 42% IMS I subjects, and 27% in the control group, and to 39% in the treated group, of the NINDS study, despite significantly quicker time to IV thrombolytic therapy in the latter. Two IMS II subjects with M2 occlusions, with baseline NIHSS 11 and 21, were lost to follow-up but included in the MRS > 2 group. Six AOLs eligible for an IA therapy did not receive it for a variety of reasons. These occurrences suggest that a target for good outcomes of 50% or greater is not unreasonable with the IV/IA treatment paradigm applied as rapidly as possible, with limited imaging selection. It is reasonable to expect that monitoring catheter tip temperature and other further advances in revascularization with improved thrombus removal devices, balloon angioplasty, and/or stents may play a part in improving interventional outcomes. The IMS III Trial, where 900 subjects will be randomized to either full-dose IV rtPA, or to reduced-dose IV rt-PA plus IA therapy with either Cordis Microcatheter, EKOS Neurowave Catheter, or the Concentric Merci Retriever Device,2 or other devices that may be introduced as their approval status becomes defined, will have the opportunity to test this hypothesis.
The IMS II study has confirmed the relative safety of IV+IA therapy for moderate to severe acute ischemic stroke. A correlation between recanalization and reperfusion with outcome is again confirmed. IMS II provides further evidence that the EKOS PRIMO ultrasound microcatheter results in a greater probability of recanalization of the AOL when compared to standard microcatheter (68.9% vs. 50.5% after 2 hours). The role of IV/IA therapy using standard microcatheter, EKOS NeuroWave™ Catheter, or the Coincentric Merci Retriever device, will be tested against full-dose IV therapy alone in the IMS-3 Tria