The Food and Drug Administration said Monday that Capricor Therapeutics’ stem cell treatment for Duchenne muscular dystrophy did not meet the objectives of Phase 3 trial — contrary to the company’s claims last year.

Capricor said in December that the drug, known as deramiocel, met both the primary and secondary endpoints in a large, randomized study. It was a striking result in a fatal, childhood disease that has proven stubbornly difficult to treat, despite immense advances in genetic medicine.

The data were also notable for coming primarily in teenagers and young men who have already lost the ability to walk, a population with few options. The drug appeared to both preserve their upper-arm function and stave off the heart failure most patients eventually experience, Capricor had said.

But in documents released Monday in advance of a hearing this week where FDA advisers will weigh in on the drug, regulators said Capricor’s study had in fact fallen short. There was no statistically significant difference between deramiocel and placebo on either the pre-specified primary or secondary endpoints, namely upper-arm and cardiac function.

After the study was completed, the FDA said, Capricor made a litany of changes to its statistical analysis plan, a crucial document outlining how a study will be measured. Revising that plan after a study is done raises concerns that a company may be looking to generate a positive outcome when there wasn’t one based on its original design.

And the company’s submission to the FDA included analyses that were not included in any statistical analysis plan, regulators said.

In an interview Monday, Capricor CEO Linda Marban said the company was “completely surprised” by the agency’s accusations. She said regulators were relying on an early draft of the statistical analysis plan that still had Post-it notes and “was never fully fleshed out.”

The company submitted the new statistical analysis plan at the FDA’s request before the study was unblinded, she said.

“We have no idea why the agency is doing what they’re doing, we can’t get any feedback from them, but they are clearly on a mission to shut us down,” she said.

The agency said that its own analysis showed “small and variable changes in measures of upper limb skeletal muscle function and in several cardiac imaging parameters that did not reach statistical significance and are, thus, difficult to interpret.”

It also raised concerns about safety risks, namely infusion reactions and anaphylaxis.

“For an investigational product to receive approval, there must be convincing and substantial clinical evidence that the product is effective in the proposed population and indication, and the observed benefits must outweigh the observed risks of the product in its intended use,” the agency said. Capricor’s data “does not provide substantial evidence of effectiveness for deramiocel.”

Outside advisors will review and vote on the drug’s efficacy at a hearing on Wednesday. It is likely to be a contentious debate.

Capricor’s drug has generated tremendous excitement among some patients and doctors. A litany of investigators on the study submitted a letter to the agency last week urging the agency to give the drug a broad label. Patient advocates and families are also planning to testify.

Capricor’s drug was also one of a number of rare disease medicines that ran into obstacles at the FDA under previous leadership, drawing some criticism from lawmakers and patient advocates that the agency was stifling drug development for conditions that have few treatment options.

The FDA last year rejected the company’s marketing application for deramiocel, which was submitted based just data from a small Phase 2 trial. Capricor resubmitted the application this year with new data from the Phase 3 study.

With new leaders in place, the FDA has signaled in some cases that it’s moving back to a more flexible stance on rare disease treatments, but the documents released Monday suggest that agency officials may still scrutinize medicines’ data packages closely.