Conference Coverage

TauRx still rooting for its methylene blue AD drug, despite controversial study results

View on the News

LMTM cohort study is fatally flawed

There are several reasons why this kind of analysis defies scientific credibility.

The term “monotherapy” is really a euphemism for substandard care before the study. These patients on monotherapy were not selected to be so. These were people with mild to moderate AD dementia who should have been on memantine or a cholinesterase inhibitor and were not, for unknown reasons. They represent a health care bias, and on that fact alone this comparison should not have even been mentioned. Calling it monotherapy is an attempt to distract from the fact that this was an indication bias that defines this group.

Dr. David Knopman Michele G. Sullivan/Frontline Medical News

Dr. David Knopman

This group of patients constituted only 20% of the entire sample, and using these small subsets in a clinical trial is usually deemed inferentially meaningless because of the high chance for random effects.

The claim that this analysis was done before the database lock is true in principal. But because the investigators had already seen this result in their previous study, which was identically designed, they cannot really claim this was truly an ad hoc analysis. They already knew what they were going to see.

In July, the investigators claimed that the 100-mg dose was effective in monotherapy. They have been convinced over the entire course of development that the 4-mg dose was ineffective. Now they are retracting that. To me, this apparent wild goose chase for any kind of effect trivializes the entire process of a clinical trial.

What I believe we are observing here is a profound placebo effect that can occur when people who have been getting substandard care are put in a clinical trial and exposed to good care.

David Knopman, MD, is a clinical neurologist at the Mayo Clinic, Rochester, Minn., and a member of the Alzheimer’s Association Medical and Scientific Advisory Council.



– An anti-tau compound that has stirred scientific controversy for 8 years will continue along its developmental pathway at a much lower dose, despite yet another study that has Alzheimer’s researchers scratching their heads.

The drug, dubbed LMTM, is a derivative of the dye methylene blue. Its most recent phase III study, reported at the Clinical Trials on Alzheimer’s Disease conference, found that 100 mg twice a day conferred no cognitive or functional benefit upon patients with mild AD, compared with a control dose of 4 mg.

Dr. Claude Wischik Michele G. Sullivan/Frontline Medical News

Dr. Claude Wischik

Some significant differences, however, did emerge in two prespecified subanalyses of the 4-mg control group. Patients who took the low dose, intended to be a placebo comparator, did better than those on the high dose – but only if they were not taking any standard symptomatic AD medications.

Based on these findings, TauRx, which is developing LMTM, will abandon the 100-mg dose and refocus on the 4-mg dose, said Claude Wischik, MD, chairman and chief executive officer of the Singapore-based company.

“I think it looks effective and there’s no advantage to going to a higher dose,” Dr. Wischik said in an interview. “The 100-mg dose doesn’t offer anything above the 4-mg dose, and we saw more dropouts in the higher-dose group. We will go forward with a new trial using 4 mg.”

The new commitment to 4 mg turns LMTM’s prior development trajectory on its head, as nothing lower than 75 mg has been investigated in a phase III study. The 4-mg control dose was used as a placebo stand-in, since LMTM colors urine blue or green. The low dose was considered biologically inactive and used to maintain the study blind.

Dr. Wischik has been investigating LMTM as a tau anti-aggregant for 10 years, first publicly reporting clinical data in 2008. LMTM has never posted significant cognitive or functional benefits in any primary analysis. Instead, it has moved forward based on a series of subanalyses that showed significant or near-significant benefits in smaller, meticulously constructed subgroups – conclusions that critics have called questionable at best. The most recent of these examined the drug’s effect in patients with mild to moderate disease and was presented last July at the Alzheimer’s Association’s International Conference (AAIC).

That study also didn’t meet its primary endpoints in the overall cohort of 891 patients, but TauRx promoted it as “promising,” based on a subgroup analysis of the 15% of patients who were not taking memantine or cholinesterase inhibitors.

Among these patients, those taking 75 mg twice daily declined 6 points less on the Alzheimer’s Disease Assessment Scale-cognitive subscale (ADAS-cog) than those taking 4 mg. Those taking 125 mg twice daily declined 6 points less than the 4-mg group. On the Alzheimer’s Disease Cooperative Study-Activities of Daily Living (ADCS-ADL), patients taking 75 mg twice daily scored 6.5 points higher than did the placebo group, indicating better function, and those taking 125 mg twice daily scored 7 points higher than did the placebo group.

At AAIC, researchers suggested that the monotherapy groups could have had a less aggressive disease course, or might not have had Alzheimer’s disease at all. Others complained about the unorthodox grouping of control patients in the subgroup analysis.

It was after digesting these data that TauRx investigators changed the statistical analysis of the current study, then in its final months, from a randomized trial to a cohort analysis. This was done before data lockdown, but it was still a dramatic shift from the original study design.

“The primary analysis was changed to essentially analyze this as a cohort study,” said Lon Schneider, MD, who presented the results at CTAD. “The comparisons of interest were patients taking 100 mg twice a day who were not on [symptomatic treatment], compared to the original control group of 4 mg. The other comparison was the 4-mg group not on cholinesterase inhibitors to the 4-mg group that was on them.”

The 18-month trial randomized 800 patients with mild AD to 100 mg LMTM twice daily or to 4 mg twice daily. Patients were drawn from two global regions: Canada and the United States, and eastern Europe and Australia.

The study group was a typical one, with a mean age of 70 years and a mean Mini Mental State Exam score of 22. The mean ADAS-cog score was 17. Most (80%) were taking a cholinesterase inhibitor, memantine, or both; 20% were naive to these medications.

Primary endpoints were the ADAS-cog11, ADCS-ADL, and left ventricular volume. Secondary outcomes included the Mini Mental State Exam and Neuropsychiatric Index.

In the primary analysis as the trial was created and conducted, LMTM 100 mg/twice daily did not confer any benefit, compared with the control 4-mg dose. The decline curves were virtually superimposable in the ADAS-cog score, ADCS-ADL score, and in loss of left ventricular volume.
This same nonsignificant pattern occurred in all the secondary endpoints, which Dr. Schneider did not show.

The cohort analyses stratified patients according to whether they were taking any cholinesterase inhibitor or memantine, or both, at baseline. That was where some differences did emerge.

The first compared the entire 4-mg cohort to the subset of patients taking 100 mg as monotherapy (absent any symptomatic medications). Both the 100-mg and 4-mg groups declined linearly on all measures, but compared to the 4 mg group the 100 mg group experienced about a 3-point benefit on both the ADAS-cog and ADS-ADL measures.* The 100-mg group also experienced significantly more dropouts (45% vs. 23%), with 16% of those being due to adverse events.

The second analysis compared the two 4-mg groups: those taking LMTM as monotherapy and those taking it in combination with standard AD medications. Again, both groups declined, but that decline was attenuated in the monotherapy group, with a 4-point advantage in the ADAS-cog and nearly a 5-point advantage in the ADS-ADL. The 3-cc ventricular volume advantage was seen as well.

Again, Dr. Schneider said, these results were recapitulated in the secondary endpoints, which he did not show.

The trial seems to upend TauRx’s earlier firm contention that the previously tested higher doses slow cognitive and functional decline – a view Dr. Wischik clung to after the July data were released. Dr. Schneider attempted to address this by suggesting that “the 4-mg dose may not have been as inactive as the developer thought.”

However, he noted, another possibility is that the patients who took the 4-mg dose but not the symptomatic drugs “may have had a more benign course of disease, compared to those taking cholinesterase inhibitors and memantine.”

This new study has now aroused the same criticism levied last summer. Maria Carrillo, PhD, chief scientific officer of the Alzheimer’s Association, was blunt in her assessment.

“The results of post hoc analyses, even when preplanned, are not valid and could be spurious,” she said. “They may mean nothing. As a field, we have been lured into rabbit holes in the past due to post hoc analyses and wasted too much time and way too much money. Of course, companies can do what they want as next steps in trials if they have the financial backing to do so.”

Dr. Wischik, however, said both the July data and the new data clearly justify taking the 4-mg dose forward in a randomized, placebo-controlled study.

“It’s really very exciting that we got exactly the same results now as we did in the post-hoc analysis [of the July data],” he said in an interview. “We predicted these results based on what we saw, we changed the statistical analysis, and we got the predicted results. This has nothing to do with data scouring.”

The question of whether the monotherapy patients are fundamentally different from those taking standard AD drugs is a valid one, he admitted. “We can’t avoid that criticism until we do another study where people who are not on any AD treatment are randomized.”

Dr. Wischik was then asked whether it would be difficult to recruit an entire cohort of patients with mild Alzheimer’s who are willing to forego approved symptomatic medications while in such a study. He did not think that would be problematic.

“Twenty percent of our cohort was already in that slot,” he said. “This practice pattern is determined somewhat by geography and somewhat by the type of clinician treating the patient. People also go on the drugs and then come off for various reasons. But even in the U.S., only 55% of Alzheimer’s patients are taking them.”

Dr. Wischik didn’t mention the problem of finding an appropriate placebo for such a study. If indeed the 4-mg dose is biologically active, such a placebo would have to be demonstrably inert, as well as provide the appropriate urine color to keep the blinding unbroken.

“That’s a challenge,” Dr. Schneider said.

Dr. Schneider was a coinvestigator on the LMTM phase III program. He has disclosed financial relationships with numerous pharmaceutical companies.

Correction, 12/12/16: An earlier version of this article misstated the results of this study.

On Twitter @Alz_Gal

Next Article: