How to Tell When A Drug Company Fibs About Clinical Trial Results

By Adam Feuerstein 07/03/12 - 07:52 AM (TheStreet) -- Osiris Therapeutics(OSIR_) "disappeared" important data when the company announced results Monday from a mid-stage study of its stem cell therapy Prochymal in heart attack patients.  Naturally, Osiris didn't come out and tell investors that it was issuing a misleading press release on the Prochymal heart attack study. Instead, the company claimed the study was a success. That's not true. Figuring out Osiris' deception wasn't that difficult if you know how to parse the language of clinical trial results and look at independent sources of information for the truth.  Ride along with me as I pick apart Osiris' statements regarding the Prochymal heart attack study. Interpreting clinical trial results with a skeptical eye is a crucial tool for all biotech investors, so apply these skills universally whenever a drug or biotech company tries to convince you that its drug works. Hopefully, you'll find most companies are telling the truth, but sadly and too often, bullish pronouncements about boffo clinical trial data are just spin jobs ginned up to plaster over problems and bad data. 

Here's what Osiris issued Monday:  Osiris Therapeutics, Inc. announced today interim one-year results from its groundbreaking clinical trial evaluating Prochymal (remestemcel-L) for the treatment of patients experiencing first-time acute myocardial infarction. The trial is the largest study of allogeneic or "off-the-shelf" stem cells ever conducted in heart attack patients. A total of 220 patients were given a single infusion of either Prochymal or placebo through a standard intravenous line within seven days of an acute heart attack.  Not much to quibble with here except it's helpful to know that Osiris enrolled the first heart attack patient to this study in April 2009, so it took more than three years to complete and report "interim" results. ClinicalTrials.gov [a great, independent source of clinical trial information, by the way] lists the completion date for this study as December 2011, so Osiris appears to be reporting results six or seven months late.

Cardiac MRI assessments were conducted for six months following infarct to evaluate cardiac remodeling.  Cardiac MRI is a precise and commonly used method of measuring cardiac changes in clinical trials. MRI assessments were conducted for six months after heart attack patients were enrolled and treated with either Prochymal or a placebo, but Osiris purports to be reporting interim, one-year results. When were the MRIs conducted in the Prochymal trial? Osiris doesn't specify, which clouds the reported results, especially if MRIs were not conducted at 12 months follow up. Patients receiving Prochymal had significantly less cardiac hypertrophy, as measured by cardiac MRI, compared to patients receiving placebo (p [less than] 0.05). Patients treated with Prochymal also experienced significantly less stress-induced ventricular arrhythmia (p [less than] 0.05). Cardiac hypertrophy and ventricular arrhythmia are indicators of pathological remodeling following heart injury and provide insight into the mechanism by which mesenchymal stem cells attenuate heart injury following a myocardial infarction.  This is the paragraph where Osiris claims positive results based on Prochymal beating placebo across two efficacy endpoints. Sounds impressive except none of the Prochymal benefits disclosed by Osiris are predefined endpoints in the phase II trial.Osiris appears to have thrown out the real endpoints called for in the phase II trial and replaced them with new endpoints which just happen to show Prochymal in the best light. Why would Osiris do this? Perhaps the pre-defined endpoints in the study all failed? That's a pretty safe assumption when companies decide to swap out trial endpoints with no disclosure or explanation. Again, here is where Clinicaltrials.gov is a good fact checker. The site's listing of the Prochymal acute myocardial infarction (AMI) study shows left ventricular end systolic volume (ESV) as the primary endpoint. Secondary endpoints in the study are left ventricular ejection fraction (LVEF), infarct size and major adverse cardiovascular events (MACE).  ESV measures the volume of blood in a ventricle at the end of the heart's contraction -- an important assessment of how well the heart is able to pump blood to the rest of the body. [More blood left in the ventricle after contraction signals a weakened a heart muscle.]  Similarly, LVEF measures the percentage of blood that is pumped out of the ventricle with each heartbeat. A higher LVEF signals a stronger heart.  Osiris designed the phase II study with ESV and LVEF as two key efficacy endpoints based on the belief that the stem cells contained in Prochymal would help rebuild heart muscle, thereby lowering ESV and raising LVEF compared to placebo.  Osiris' silence on the outcomes of these two important endpoints should be deafening to investors -- and not in a good way. The mechanistic data is complemented by clinical data showing treatment with Prochymal resulted in a statistically significant reduction in heart failure. In the study, seven patients who were treated with placebo have progressed to heart failure requiring treatment with intravenous diuretics, compared to none of the Prochymal patients (p=0.01). Furthermore, patients receiving placebo tended to require re-hospitalization for cardiac issues sooner than the patients receiving Prochymal (median 27.5 days vs. 85.5 days).  These are interesting observations, but again, largely irrelevant for the purposes of this study since these weren't predefined endpoints. Importantly, Osiris doesn't disclose the time point at which these purported benefits occurred, nor does the company tell us anything about the number of patients analyzed. How was heart failure defined? Osiris doesn't say. What was the baseline incidence of heart failure in the study? Osiris doesn't say. The study only allowed for a single infusion of Prochymal or a placebo immediately after the first heart attack but patients were followed for six months or a year, so how do follow-up therapies in each arm of the study compare? Were they balanced? Again, Osiris doesn't say."This study is the largest of its kind and provides key insights into the mechanism of action of mesenchymal stem cells in the setting of acute myocardial infarction," said Lode Debrabandere, Ph.D., Senior Vice President of Therapeutics at Osiris. "These important mechanistic observations are consistent with data obtained from our preclinical models and from the first placebo-controlled human trial with Prochymal published in the Journal of the American College of Cardiology. Given the quality of the data and highly encouraging results observed thus far, we are extending the trial's duration to capture a better understanding of the long-term clinical benefits of MSCs."  The canned quote from management: Always positive and optimistic. I usually ignore this fluff but in this case, Debrabandere reminds me that Osiris conducted a previous phase I study of Prochymal in patients suffering from a first heart attack. This study enrolled 53 patients and randomized them to treatment with either Prochymal or placebo. In other words, the older phase I study is very similar to the more current phase II study.  The phase I study, "A Randomized, Double-Blind, Placebo-Controlled, Dose-Escalation Study of Intravenous Adult Human Mesenchymal Stem Cells (Prochymal) After Acute Myocardial Infarction" was published the Journal of the American College of Cardiology in December 2009. You can download the study here for free. I won't go into too much detail about the results, but suffice to say, investigators found no statistically significant difference between Prochymal and placebo in ejection fraction or a six-minute walk test.  Debrabandere's comment that Osiris is extending the current phase II study to capture the long-term clinical benefit of Prochymal is most perplexing because the phase I study was notable in that measures of cardiac improvement between Prochymal and placebo narrowed and in some cases disappeared with longer follow up. Why results would be different in the phase II study is a mystery, particularly since patients in the study are only treated with single infusion of Prochymal. 

Perhaps Osiris is extending the phase II study to delay the reporting of negative results? Again, that's a pretty safe assumption absent a better explanation. The trial also demonstrated that treatment with Prochymal was safe. There were no infusional toxicities observed in patients receiving Prochymal. Serious adverse events occurred with equal frequency in both treatment groups (31.8%). To date, there have been 5 deaths in the trial, 2 in the Prochymal group and 3 in the placebo group.  Prochymal is safe but so is placebo. That's no great comfort. Deaths in the study are essentially equivalent in both arms, a survival benefit favoring Prochymal would have been better. "For interventional cardiologists, keeping our myocardial infarction patients from progressing to heart failure is central to our mission," said Mark Vesely, M.D., Principal Investigator on the Study and Assistant Professor of Medicine (Interventional Cardiology) at the University of Maryland School of Medicine. "It is remarkable and very encouraging to see significant changes in clinically meaningful parameters this early in the study. We look forward to the additional data that will be gathered as the study progresses, which will help us to better understand both the magnitude and durability of the benefit to treatment."  Vesely's comments were remarkable but not for the reasons he states. Vesely is the researcher in charge of this phase II study and sounds encouraged but he ignores the missing data even though endpoints like EVF and LVEF are more clinically meaningful measures of cardiac health and progression to heart failure than what Osiris chose to disclose Monday.  I tried to reach Vesely to ask him about the study and the missing data but he didn't respond to my email. Bill Seiler, a spokesperson for the University of Maryland School of Medicine, reached by phone, said Vesely's comment included in the Osiris press release was not authorized or pre-approved by the school."I'm just finding out about it this morning," said Seiler adding, "I'm trying to track it all down."  This Phase 2, multi-center, randomized, double-blind, placebo-controlled study is evaluating the safety and efficacy of Prochymal (ex-vivo cultured adult human mesenchymal stem cells) intravenous infusion following acute myocardial infarction. A total of 220 patients were randomized (1:1) at 33 centers in the United States and Canada and received a single intravenous infusion of Prochymal or placebo within 7 days following first acute myocardial infarction. In addition to screening and baseline visits prior to the infusion, initially follow-up evaluations were scheduled to be conducted through 2 years. Given the encouraging results observed at the one year time-point, the trial is being extended to include 5 years of follow-up. Both male and female subjects between 21 and 85 years of age were enrolled. Patients had to have a left ventricular ejection fraction (LVEF) between 20% and 45% as determined by quantitative echocardiography or cardiac MRI at least 24 hours after successful reperfusion of the culprit vessel. In addition, troponin levels must have been greater than 4 times the upper limit of normal during the first 72 hours of hospitalization for the MI. 

Here is where Osiris describes the phase II study in detail but notice what is missing: Any discussion at all of the study's clinical endpoints!  Osiris had no such trouble describing the study's endpoints on March 2, 2011 in a press release announcing the completion of patient enrollment:  Efficacy endpoints determined from cardiac MRI include end systolic volume, LVEF and the ability of Prochymal to preserve functional heart tissue and limit scar formation following a heart attack. In addition, functional and quality of life assessments will be performed.  Osiris did not respond to an email request for the missing data from the heart attack study. The company also chose not to hold a conference call with investors following the release of the results Monday.  --Written by Adam Feuerstein in Boston.