*

Data helps predict which drug will help specific patients

lose

weight

*

Researchers used AI to analyze effect on patients' other

medical

conditions

*

Drug benefit-risk scores need to be developed, tested

By Nancy Lapid

Nov 19 (Reuters) - U.S. researchers are beginning to

identify clinical characteristics that distinguish "super

responders" to GLP-1 weight-loss drugs like Wegovy and Zepbound

from patients who lose only moderate amounts of weight at best,

according to a report published online ahead of peer review.

The massive data analysis may eventually help personalize

treatment decisions for the drugs already being used by millions

of patients. Individual patients' health status before starting

treatment can also help guide drug selection, the analysis

suggests.

"If I am a clinician seeing patients, I need to know what

medicine will best benefit my patient," said study leader Venky

Soundararajan of Massachusetts-based data analysis company

nference. "I also need to know... what benefits and what side

effects are they likely to have" given their unique medical

history.

Researchers analyzed 14 million doctors' notes and 15

million clinical data entries from more than 135,000 patients

with and without diabetes who each took only one GLP-1 drug.

They found that roughly 12.5% were "super responders" who lost

more than 15% of their weight in the year after starting

treatment.

Another 35% were considered moderate responders, having lost

5% to 15% in the first year. The largest group - the minimal

responders, accounting for 47% - lost less than 5% of their body

weight, while an additional 5% lost roughly 5% of their weight

but gained it back again within a year.

The trajectories of weight loss were very diverse,

Soundararajan said. "But when you break it down by the brands,

you can see the medicines are getting better and better over

time in lowering the percentage of patients who still continue

to be in that minimal weight-loss group."

With Eli Lilly's ( LLY ) Zepbound and Mounjaro, for example,

23% to 28% of patients fall into the minimal weight-loss group,

compared to 30% to 43% with Novo Nordisk's Wegovy and

Ozempic. With earlier-generation GLP-1s like Lilly's Trulicity

and Novo's Saxenda and Victoza, 46% to 63% of patients fell into

the minimal weight-loss category, the study found.

AI TOOLS ANALYZE MEDICAL CONDITIONS

Using artificial intelligence tools, the researchers

analyzed not only weight-loss outcomes but also the presence and

absence of 1,300 different medical conditions before and after

treatment.

For example, the pre-treatment presence of muscle stiffness

without knee pain or osteoarthritis increased the probability a

patient prescribed Zepbound would become a super responder.

That suggests patients with obesity-related muscular

dysfunction but preserved joint health may be particularly

likely to achieve exceptional weight-loss outcomes with

tirzepatide, the active ingredient in Zepbound and Mounjaro, the

researchers said.

"If you have knee pain, osteoarthritis, chest pain, sleep

apnea, or fibromyalgia, you are less likely to be a Zepbound

super responder" in terms of weight loss, Soundararajan said.

Patients with sciatica saw improvements if they were

prescribed Wegovy, the researchers also found.

"If you have melanoma, you are very likely to respond to

Wegovy. If you have actinic keratosis, you are highly likely to

respond to Mounjaro. If you have aged osteoporosis, you are very

likely to respond to Ozempic," Soundararajan said of weight-loss

prospects.

Regardless of whether they received a Lilly or Novo drug,

patients with sinus pressure beforehand reported improvements

afterward.

Given that any individual patient is likely to have a

cluster of medical conditions, the researchers say their next

step is to develop an algorithm that yields scores to indicate

the likely benefit and risk for each drug under different

circumstances and test that in prospective studies.

"These signals will continue to get more and more and more

refined as data is collected from more and more patients,"

Soundararajan said.