Safety Findings of the Phase I Study of Conivaptan for Cerebral Edema in ICH

1. Safety Findings of the Phase I Study of Conivaptan for Cerebral Edema in ICH Jesse J. Corry, MD John Nasseff Neuroscience Specialty Clinic United Hospital, part of Allina Health Division of Stroke Neurology Division of In-patient Neurology

2. Disclosures • Nothing relevant to this study.

3. Outline • Current state of ICH treatments • Perihematomal Edema and ICH outcome • Review of Study Protocol • Screening, Logistics, and Demographics • Drug Related Adverse Events • Safety Data • Lessons Learned • Next Steps

4. Current State of ICH Treatment • Intracerebral hemorrhage (ICH) represents ~10-15% of all strokes in the United States (U.S.), numbering 50-70,000 cases per year. Twice as common as subarachnoid hemorrhage (SAH), the 30 day mortality is ~40%, and greatest among the poor, elderly, black and Asian. Only 20% of ICH patients are independent at 6 months. (Adams, 2005; Thomas et al., 1996) • Only admission to a neuro-ICU, and the use of a CT scanner, have been shown to improve outcome from ICH (Thomas et al., 1996; Diringer & Edwards, 2001). • With respect to cost, ICH represents 34% of years of potential life lost to stroke (Thomas et al., 1996). Per case, the cost is ~$124,000, and the total lifetime cost for annual US cases >$4B (Thomas et al., 1996).

5. Reduce Hematoma Expansion • BP control • Surgery • Hemostasis

6. Blood Pressure Reduction • The association between BP elevation and risk of ICH expansion is not clear without a definite relationship found in prospective studies (Broderick et al. 1990, Brottet al. 1997, Ohwakiet al. 2004.) • Attempts to limit hematoma expansion via aggressive BP reduction to SBP < 140mmHg, as demonstrated by the INTERACT 2 and ATACH 2 trials failed to show significant differences in death or disability at 3 months (Anderson et al. 2013, Qureshi et al. 2016.) ATACH 2 demonstrated a higher portion of adverse renal events and INTERACT 2 failed to demonstrate reduced hematoma volume. • In patients w/ ICH similar to those in these trials, BP reduction appears safe (Boulouiset al. 2017, Zhang et al. 2017.) • Metabolic suppression of tissue surrounding the hemorrhage mitigates any potential ischemia from BP reduction(Zazuliaet al. 2001.)

7. Surgical Management • Cochrane Reviews of traditional surgical evacuation of ICH w/ craniotomy vs. medical care alone have an OR of 0.74 favoring surgery, but the data was highly heterogeneous (Prasad et al. 2008.) • STITCH I showed no benefit from early surgery c/w medical management (Mendelow et al. 2005.) • STITCH II, focusing on hemorrhages < 1 cm from cortex w/ a moderate GCS and w/o IVH failed to demonstrate decreased death or disability at 6 mo(Mendelowet al. 2013.)

8. MISTIE III • Minimally invasive catheter evacuation and ICH thrombolysis for moderate to large hemorrhages failed to improve the proportion of patients achieving mRS 0-3 at 1 year (Hanley et al. 2019.) • Of the 58% of patients who had their hematoma reduced to a volume of 15 mL or less—the surgical goal—there was a 10.5% difference in the likelihood of achieving a good functional outcome favoring the MISTIE group (P = 0.03). • Better ICH evacuation was associated with deep ICH (OR 2.08, p=0.03) and greater site and surgeon experience (P<0.0001). • Perihematomal edema also decreased among the surgical group compared with the medical group (reduction of 4.7±13.8 mL vs -11.8±15.1 mL, respectively; P <.001). Those participants who experienced less than 20% ICH clot removal saw an increase in perihematomal edema. However, removal of 20% to 65% correlated with a 2.1 mL reduction, and removal of more than 65% correlated with a 6.6 mL reduction (P <.001). More removal was associated with more reduction (P <.001). Each 10 mL increase in perihematomal volume corresponded to a 25% increase in mortality risk among both groups at 1 month and a 15% increase at 6 months.

9. Factor VIIa and Hemorrhage Growth • The FAST (Factor Seven for Acute Hemorrhagic Stroke) trial failed to demonstrate a significant difference in death or severe disability at 90 days and only minor reductions in ICH expansion (Mayer et al. 2008.) • The TICH-2 trial is currently looking at tranexamic acid to limit hematoma expansion.

10. State of Therapy Summary BP, hemorrhage volume, and surgery have failed to demonstrate a consistent effect on outcome in ICH.

11. Perihematomal Edema and ICH • Although the association between perihematomal edema and ICH outcomes has conflicting results in the literature, perihematomal edema growth appears to be a biomarker for secondary injury, and a number of clinical early phase studies have evaluated pharmacologic agents targeting inflammatory responses with edema reduction as a clinical end point. • There is currently no proven clinical therapy that both reduces perihematomal edema and improves outcomes.

12. Perihematomal Edema and ICH • Worsening cerebral edema has been implicated in delayed neurological deteriorationand worse outcome, through the elevation of intracerebral pressure (ICP) (Mayer & Sacco, 1994.) • It is associated with activation of inflammatory pathways by the toxic biochemical and metabolic effects of clot products (Urdayet al. 2015.) • Elevations in ICP reduce the ability of blood to reach the brain, exacerbating the injury and producing ischemia. The formation of peri-hematomal edema contributes to an increased volume of ~75% (Gebel et al., 2002.) • Animal models of ICH demonstrate a large peri-hematomal area that undergoes neuronal death characterized by increased water content and inflammation (Theix & Tsirka, 2007.) • Reduction of this edema may reduce the degree of neuronal death, decreasing hospital length-of-stay (LOS), and improving short-term outcome(Volbers et al., 2016.)

13. PHE, ICH- Outcome • PHE correlates with age, disease severity, ICH volume, irregular shape, and glucose (Wu et al. 2017.) Faster edema growth results in greater MLS w/ resulting herniation and mortality. • Faster PHE expansion in the first 24hr (0.74mL/h in patients who died vs. 0.17mL/h in survivors) correlate inversely with clinical outcome regardless of location (Urday et al. 2016, Grunwald et al. 2016.) • In deep hemorrhages, PHE expansion at 3 days (> 0.10mL/h) correlate with poor outcome (mRS> 3)(Grunwaldet al. 2016.)

14. PHE, ICH- Medical Management • Wide variability exists in the treatment of ICP and cerebral edema among intensivists (Hays et al., 2010.) • The use of mannitol, which elevates the osmolarity within the cerebral vasculature, promoting water movement across the blood-brain barrier and into the capillary system (Galton et al., 2011), is common, but is limited by its deleterious effects on renal function, fluctuations in intravascular volume, and pH. Most concerning is the slow elimination of mannitol from the CSF which may potentially require progressively higher doses, over time, to control intracerebral pressure, and may result in rebound cerebral edema (Nau et al., 1997; McGraw & Howard, 1983.) • Increasingly, hypertonic saline (HS), which acts similar to mannitol, is being used to abate cerebral edema. However, this therapy has not demonstrated any survival or outcome benefitdespite reductions in ICP (Bulger et al., 2010; Strandvik, 2009.) Chlorine is a potent renal vasoconstrictor potentially reducing blood flow, precipitating renal ischemia and reducing glomerular filtration rate (Wilcox, 1983; Gazitua et al., 1969). Finally, the use of hypertonic saline may be associated with increased risk of blood-stream infections (BSI), and trends to increased risk of nosocomial and urinary tract infections (Bulger et al., 2010.) HS is associated with variable changes in renal function portending renal impairment in the NCC population (Corry et al. 2014.)

15. Conivaptan for Reduction of Cerebral Edema in ICH • Neurocritical care needs a therapy that can… • Reduce 24h and 72h PHE edema expansion and has limited nephrotoxicity. • Can improve clinical outcome and/or reduce cost of care. • Is safe. • Is simple.

16. Conivaptan-Reasoning • Although not the exclusive means for water flux in the brain, AQP4 is the predominant class of water channel in the brain(Yool et al., 2010; Zeynalov et al., 2008.) For edema to occur, water must enter the astrocyte compartment. Water must pass from the capillary lumen and through the luminal, abluminal, and luminal perivascular endothelium. The perivascular pool appears, in both the influx and efflux of water, to be the rate limiting step (Zeynalov et al., 2008.) Inhibition of water permeability, via reductions in AQP4, decreases ipsilateral hemispheric water content and may have therapeutic potential for cerebral edema in the clinical setting(Migliati et al., 2010; Rosenberg et al., 1992.) • AVP V(1) receptor antagonism significantly reduces hemorrhagic brain edema in rat models (Rosenberg et al., 1992.) • Experimental models have demonstrated AVP V(1) receptor antagonism attenuates injury volume, and cerebral edema, through changes in aquaporin-4 (AQP4) expression (Liu et al., 2010.) • Animal models have demonstrated AQP4 expression increases within 6 hours from injury, peaks between 48-72 hours, and remains elevated over the first week (Li & Sun, 2003; Sun et al., 2009.) This increased expression parallels brain water content (BWC) in hemorrhagic animal models. Antagonism of AVP V(1), in animal models of TBI and both hemorrhagic and ischemic stroke, is associated with parallel reductions of AQP4 expression and cerebral edema as measured by BWC (Liu et al., 2010; Taya et al., 2008; Trabold et al., 2008; Rosenberg et al., 1992.)

17. Conivaptan-Evidence • Current clinical data in traumatic brain injury patients demonstrate conivaptan is safe and well tolerated using a single dose (20mg) to increase Na+ in a controlled fashion to reduce ICP (Galton et al., 2011; Dhar & Murphy-Human, 2011.) • Further, additional work has demonstrated the safety and tolerability of conivaptan, in doses ranging from 10-80mg/day, in the neurocritical care population (Ghali et al., 2006; Zeltser et al., 2007; Verbalis et al., 2008; Annane et al., 2009; Murphy et al., 2009; Wright et al., 2009; Naidech et al., 2010; Human et al, 2012; Marik & Rivera, 2013.) • Recently, bolusing of conivaptan (20mg) has been shown to lower ICP in hyponatremia following TBI and cerebral ischemia (Dhar & Murphy-Human, 2011; Galton et al., 2011.) The authors reported improvements in cerebral perfusion pressure (CPP) and stable blood pressure, and a prolonged reduction in ICP. They theorize the V2 antagonism of conivaptan promotes free water loss to reduce BWC, and the V1 antagonism may improve cerebral blood flow (CBF) and reduce blood brain barrier permeability(Taya et al., 2008; Trabold et al., 2008; Dhar & Murphy-Human, 2011; Vakili et al., 2005; Kleindienst et al., 2006; Fernandez et al., 2001.)

18. Study Protocol • Study Design: A single-center, open-label, safety and tolerability study • Funding: United Hospital Medical Staff Education and Research Committee Study Number 32036 • Intervention:Patients will receive 20mg IV of the study drug every 12 hours equaling 40mg/day over 2 days (4 doses total), in addition to the standardized ICH management. The conivaptan bolus (20mg), which is premixed with 100ml of 5% dextrose in water, is infused (peripherally) over 30 minutes, most commonly through an already placed central line.

19. Study Protocol- Outcome Measures • Primary: Tolerance- The number of participants with abnormal seizure activity and/or abnormal lab values and/or increase in infection rate and/or adverse events that are related to conivaptan treatment. • Secondary • In-hospital mortality • Cost • LoS NCCU, hospital • Need for EVD, surgery, ventilator • mRS at d/c, follow up • Mitigation of PHE

20. Study Protocol- Inclusion Criteria • Age >18 years old and < 80 years. • Diagnosis of primary ICH > 20 cc in volume. • Enrollment within 48 hours from initial symptoms. • Signed informed consent from the patient or obtained via their legally authorized representative (if the patient is not able to sign the informed consent themselves). The patient's decisional capacity to either provide or refuse consent will be determined using the Glasgow Coma Scale (GCS), which is being assessed at baseline and at 24 hours (+/-6hrs) after enrollment. A potential study participant with a GCS > 14 will be asked to provide their own initial study consent. A GCS ≤ 14 would indicate the need to pursue consent via legally authorized representative.

21. Study Protocol- Exclusion Criteria • Current need for renal replacement therapy (RRT). • Glomerular filtration rate (GFR) of <30 mL/minute at time of admission. • Participation in another study for ICH or intraventricular hemorrhage. • ICH related to infection, thrombolysis, subarachnoid hemorrhage, trauma or tumor. • Presence of HIV or active fungal infection that is known based on information in the electronic medical record (EMR). • Continued use of digoxin or amlodipine (as recommended by the manufacturer due to cytochrome P450 3A4 "CYP3A" inhibition). • Active hepatic failure as defined by aspartate aminotransferase (AST) >160 units/L and/or alanine transaminase (ALT) >180 units/L, or total bilirubin levels greater than four times normal levels (>4.8mg/dL). • Serum Na+> 145 mmol/L (admission labs or any time prior to recruitment/enrollment). • Unable to receive conivaptan based on contraindications indicated by the manufacturer. • Pregnant or lactating females. • Not expected to survive within 48 hours of admission, or a presumed diagnosis of brain death

22. Screening and Logistics • Dates: 1Apr17-Nov 18 • Screened: 83 (discuss TS drop here) • Age: 67.5 yo (38-88yo range) • Ethnic Demographics: • Asian= 10 • Caucasian= 59 • Latino= 4 • African= 2 • Native American= 1

23. Screening and Logistics • Reasons patients not enrolled (total >76 as one patient may have multiple reasons to be excluded): • Too small ICH= 37 • Unknown TLKW or too late= 16 • “Terminal” ICH= 10 • Too old= 12 • Did not speak English= 3 • No drug available= 4 • Missed= 3 • Family refused= 1 • Renal failure= 1 • Medication contraindication= 1 • Infection associated ICH= 1 • Surgery related ICH= 1 • Trauma related ICH=1

24. Selected Patients

25. Demographics Selected Patients • Hypertensive= 4 • Smoking= 4 • AODA= 3 • Diabetes= 3 • Prior AIS= 1 • Prior ICH= 1 • Prior TBI= 1 • Hyperlipidemia= 1

26. Scores FUNC: Likelihood of mRS 0-4 at 90 days ICH Score: Prediction of 30 day mortality. 0= No mortality 1= 13% 2=26% 3= 72% 4= 97% 5&6= 100% • Score 0-4: 0% • Score 5-7: 29% • Score 8: 48% • Score 9-10: 75% • Score 11: 95%

27. Selected Patients Disease Severity

28. Cost and Drug Related Adverse Events • As per protocol, all patients received a PICC line. • All patients were placed on Allina HS protocol. http://excellian.net/_layouts/15/WopiFrame2.aspx?sourcedoc={C7DEB7DF-7E9A-4259-94F5-DEF50F273AAE}&file=33316_ORDERSET.doc&action=default&DefaultItemOpen=1 • No patients required an EVD, bolt/ICP monitor, craniotomy. • No patients developed precipitous drop in BP, unexpected elevation of Na+, pontine demyelination, AKI.

29. Cost and Drug Related Adverse Events

30. Safety Data • Conivaptan IV/PO has a peak effect in 2hr, w/ increased urine output for up to 12 hr after, and an elimination half-life of 5hr. Data will compare two windows: On drug and off drug starting from Day one. Day one is defined as first day of drug and for most represent 12-48hr post ICH. • Data from each patient will be calculated and then compared between patients to look for inter-patient effect and by mg/Kg weight based effect (i.e. 20mg dose divided by patient admission weight. Then see if concentration had an effect.)

31. Safety Data Focuses • Na+ level change (not to high) and stability. • Changes in eGFR/renal function. • Changes in MAP/HR w/ and w/o drug and need for pressors. • Urine output w/ and w/o drug. • ADEs. Specifically need for pressors, arrhythmia, pulmonary edema, DVT. • GCS and mRS.

32. Safety Data- Na+ Overall averages (mean/SEM) for Na+at baseline, during CVT administration (On) and after the 24-hr washout period (Off). Symbols represent average for each individual patient at those timepoints. P=0.014 for Na+levels being elevated after CVT (On and Off vs Baseline) In sum, conivaptan safely elevated Na+ without overshooting targets, and helped maintain that target in the days following the end of its use. Sodium (Na+) levels shown for each patient during CVT administration. Grey lines indicate timing of CVT administration. Samples were taken at 6-hr intervals.

33. Safety Data- Renal Function Conivaptan did not adversely affect renal function as measured by eGFR. Average eGFR at baseline, during CVT administration (On), and after a 24-hr washout (Off). Symbols represent averages for each individual patient.

34. Safety Data- Urine Output Average (mean/SEM) overall urine output for time on (green) and off (white) CVT Average urine output for each patient during CVT administration (green) and after a 24-hr washout (white)

35. Safety Data- BP and HR Heart Rate (upper graphs) and mean arterial pressure (lower graphs) during CVT administration. Line graphs show data by patient. Grey lines represent CVT administration. Bar graphs show average/SEM with values for each patient represented by dots. The first bar represents the baseline measure. Subsequent bars are pre-dose, 1 hr post-dose and 6 hr post-dose. Grey lines indicate CVT administration.

36. Safety Data- HR and BP Summary Only one patient required the use of a vasopressor during the administration period of conivaptan. Most required none or IV nicardipine in the acute phase.

37. Safety Data- Outcomes Edema tended to peak later. Patients with poor outcome as predicted by clinical scores received benefit.

38. Adverse Events

39. Edema Effects Conivaptan had negligible affect on ICH volume during the study period. Data suggests that as the greatest change in ratio of ICH:PHE came with the greatest conivaptan exposure in terms of mg/Kg.

40. Lessons Learned • In a next phase, will need to include older patients, more “terminal” patients. • Keep TS network abreast of findings, progress, etc. • Imperative to keep patients euvolemic, aggressive DVT prophylaxis, and aggressive chest PT. • Larger hemorrhages, in patients at risk for more PHE effect, may need longer course, continuous infusion, or both. • Perhaps adjust medication to track PHE expansion rates.

41. Next Steps • Comparison of PHE growth with matched controls from United/Allina database in 1:3 ratio to see what affect conivaptan had on edema expansion and when. • Development of a “dose finding” study to assess best delivery mechanism (bolus vs. continuous infusion), at what time from ICH, and for how long. Previous studies safely used 80mg/day, so an IND will be necessary in this phase. • Investigate cost of PHE treatment. Once a dose of conivaptan is confirmed to mitigate PHE, develop efficacy studies aimed at significant cost reduction for ICH treatment. https://www.nytimes.com/2019/03/08/us/politics/trump-health-care-rates.htmlhttps://www.wsj.com/articles/hospitals-insurers-set-to-resist-price-transparency-proposal-11552343346

42. Thanks to Collaborators • To our patients and their families. • Co-Investigators: Ganesh Asaithambi, Arif Shaik, Jeffery Lassig. • Study Coordinators: Bridget Ho, Emily Marino. • Science Writer and Statistics: Megan Tipps. • United Research Foundation

43. Thank you for your attention. Any questions? Jesse.Corry@allina.com