Update: XBiotech made an important change to the 514G3 study design that raises further questions about the drug's efficacy. More details below.

An antibody designed by XBiotech (XBIT) to treat serious staph infections failed to reduce to risk of adverse events or reduce the length of hospitalization stays compared to a placebo.

Four patients with staph infection treated with the XBiotech drug 514G3 died during the phase II clinical trial. There were no deaths among the placebo patients in the study.

XBiotech did its best in a Monday morning press release to explain away the negative study results. The study was only exploratory and enrolled just a handful of patients, therefore picking up statistically significant differences between 514G3 and placebo was not expected, XBiotech said.

About those four patient deaths. XBiotech says a panel of outside experts reviewed the cases and determined that three of the patient deaths were unrelated to 514G3. Causes of death for these patients was not disclosed.

The fourth patient died from a stroke one day after receiving treatment with 514G3, which was deemed to be "possibly related" to the XBiotech drug by the same panel of outside experts.

XBiotech chalked up the death imbalance in the study to bad randomization. The staph infection patients in the 514G3 arm of the study were sicker than those slotted into the placebo arm. Four 514G3 patients entered the XBiotech study suffering from strokes, another four were diagnosed with sepsis. None of the placebo patients had strokes or sepsis.

XBiotech did not explain why patients with life-threatening co-morbidities were allowed to enter the 514G3 study.

But XBiotech also admits the 514G3-treated patients had a milder form of staph infection than placebo patients, undercutting the company's efficacy claims about its antibody.

Twenty-two percent of staph patients treated with 514G3 reported serious adverse events compared to 44% of placebo patients. The difference was not statistically significant, failing to meet the phase II study's primary endpoint.

The median duration of hospitalization stays was shorter for 514G3 patients than placebo but again, the difference failed to reach statistical significance.

The four 514G3 patients who died, including one death 24 hours after starting the study, could explain away the difference in time spent in the hospital.

"The study provided the signals we were hoping to see," said XBiotech CEO John Simard.

After all the excuse making, of course.

Update: Does 514G3 cure staph? For obvious reasons, that's an important question. Unfortunately, XBiotech Monday's press release doesn't address the question. Even more curious, the company made a recent change to the phase II study that de-emphasized the endpoint designed to answer the question.

As late as the middle of February, the co-primary endpoint of the XBiotech 514G3 study was "Time to sterile culture from date of randomization." The definition:

"The time to sterile culture is the interval in days from the first dose of study drug until 2 consecutive days of negative blood cultures has occurred. The difference in this interval will be compared between patients randomized to placebo and those who received the highest dose of 514G3."

In other words, does 514G3 eliminate staph in patients' blood?

On Feb. 16, XBiotech removed "Time to sterile culture from date of randomization" as a co-primary endpoint of the 514G3 study, relegating it to the end of a long list of secondary endpoints.

In its place, "safety and tolerability" became the co-primary endpoint, defined as "The incidence of adverse events, serious adverse events, and laboratory abnormalities will be compared between the 514G3 arm and the placebo arm."

XBiotech could have provided data on "time to sterile culture" in Monday's press release but did not.

The timing of the study-design change is important and potentially worrisome. XBiotech finished patient enrollment in December. Treatment lasted 28 days. The study design was changed in the middle of February, after all patients had been treated. Did XBiotech have data from the study in hand when it swapped out the primary endpoints?

Adam Feuerstein writes regularly for TheStreet. In keeping with company editorial policy, he doesn't own or short individual stocks, although he owns stock in TheStreet. He also doesn't invest in hedge funds or other private investment partnerships. Feuerstein appreciates your feedback; click here to send him an email.