What it really means when a clinical trial fails its main goal

When a large trial misses its main goal, the news often reads like a verdict: the drug failed. For people living with the condition, and especially for people who took part, that can feel like a door closing. The reality is usually more detailed. A trial that does not meet its primary outcome has still produced evidence, and knowing how to read that evidence helps you judge what comes next.

This guide uses a real example from 2026, the CARDIO-TTRansform trial of eplontersen in transthyretin amyloid cardiomyopathy (ATTR-CM), which we reported in our eplontersen trial article. It covers what "failed" means, what a neutral result can still tell you, why subgroup findings are treated with caution, and what patients can reasonably take away.

This guide is for general information and is not medical advice. Please speak to your own care team before making any decision about treatment.

What "failed" means in a trial

A trial is designed around one primary outcome and a plan for how to test it. The team decides in advance how many people to recruit, how long to follow them, and what result would count as success. Usually success means the difference between groups is large enough that it is unlikely to be down to chance, judged by a p-value below 0.05 and a confidence interval that does not cross the line of no effect.

When a trial "fails", it means the primary result did not clear that bar. That is a narrower finding than headlines suggest. The trial could not show a benefit on the main outcome with enough certainty, in the people it enrolled and over the time it ran. Whether the drug does anything at all, whether it is safe, and whether the science behind it was wrong are separate questions that the result alone does not answer.

The CARDIO-TTRansform example

ATTR-CM is a condition in which a misfolded protein, transthyretin, builds up in the heart muscle and stiffens it, leading to heart failure. Eplontersen is an antisense medicine designed to lower production of transthyretin in the liver. It is already approved for the nerve form of the disease.

CARDIO-TTRansform was a large, double-blind, placebo-controlled trial (NCT04136171). It randomised 1,438 people: 715 to eplontersen and 717 to placebo. The main outcome combined cardiovascular deaths and recurrent cardiovascular events.

The key results, published in the New England Journal of Medicine, were:

• a rate ratio of 0.89, with a 95% confidence interval of 0.73 to 1.09
• a p-value of 0.28
• at least one primary outcome event in 29.4% of people on eplontersen and 32.2% on placebo
• 74 cardiovascular deaths on eplontersen and 70 on placebo

Reading a neutral result

Start with the point estimate. A rate ratio of 0.89 means events happened at about 89% of the placebo rate in the eplontersen group. On its own, that leans towards a small benefit.

Then look at the confidence interval. It runs from 0.73, which would be a meaningful 27% reduction, to 1.09, which would be a 9% increase. Because the interval crosses 1.0, the trial cannot rule out no effect, and it cannot rule out slight harm. The p-value of 0.28 says the same thing in another way.

Now the absolute numbers. The share of people with an event was 2.8 percentage points lower on eplontersen. That gap is small and could easily be chance, given the interval. Cardiovascular deaths were close to even.

Put together, the honest reading is "no clear benefit shown". That is different from "proven to have no benefit". Showing that a drug truly has no effect needs a different kind of trial design and usually very large numbers.

If these terms are new to you, our guide on how to read clinical trial results explains confidence intervals and absolute risk using the SINGLE-AF trial.

Why trials miss their goals

There are many reasons a promising drug can come up short. Some of the common ones:

The comparison group did better than expected. In CARDIO-TTRansform, many participants were already taking a transthyretin stabiliser such as tafamidis. Better background care lowers event rates in both groups and leaves less room to show an extra benefit.

The trial was not long enough. Diseases that progress slowly may need years before a difference in hard outcomes appears.

The dose or timing was not right. A drug can work on its target without changing the outcome that matters to patients, or may need to be given earlier in the disease.

The outcome was hard to move. Combined outcomes such as death plus hospital visits are demanding, and a drug that helps quality of life may not shift them.

The drug really does not help for this use. Sometimes the simplest explanation is correct.

Each of these leads to different next steps, which is why researchers spend months picking through a neutral result.

The problem with subgroups

After a neutral main result, you will often read that a drug "worked" in one group. In CARDIO-TTRansform, the analysis hinted at a possible benefit in people who were not taking a stabiliser.

Subgroup findings deserve careful handling for three reasons.

First, chance. If you split a trial into enough slices, one of them will look striking by luck alone. With 20 independent comparisons, you would expect about one to cross the 0.05 line even when the drug has no effect.

Second, size. A subgroup is a fraction of the trial, so it has fewer events and wider confidence intervals. Estimates in small groups bounce around.

Third, timing. A subgroup that was named and planned before the trial started carries more weight than one found afterwards. Even planned subgroups are usually treated as hypotheses.

The way to confirm a subgroup finding is a new trial designed for that group. Until then, a subgroup hint is a reason to ask more questions, and should not on its own change treatment.

Safety still counts

A neutral efficacy result does not tell you a drug is harmful, and a trial can still produce useful safety data. When a trial misses its primary goal, look at the published adverse events next to the main result. In a large double-blind trial, any imbalance in serious side effects between groups is useful information for doctors using the drug for its approved uses.

What happens after a trial fails

Several things can follow a neutral phase 3 result:

• The sponsor may stop development for that use, while the drug stays on the market for other approved conditions.
• Researchers may publish further analyses, such as results by subgroup or by longer follow-up.
• A new trial may be designed to test a narrower group, a different dose, or a different outcome.
• Regulators may still be asked to consider the data, but they look first at whether the primary outcome was met.

For eplontersen, the drug's approval for ATTR polyneuropathy is unaffected by a cardiomyopathy trial. What changes is the evidence for using it in the heart form of the disease.

If you took part in a trial that failed

Taking part in a trial that misses its goal is still a contribution. The results stop other patients from being exposed to a treatment that may not help them, steer research money towards better options, and add to knowledge about the disease.

If you were a participant, you can usually:

• ask the study team for a plain-language summary of the results, which many sponsors now publish
• ask whether you were on the drug or placebo, once the trial is unblinded
• ask what your options are now, including any open-label extension or other trials

In the UK, the Health Research Authority expects sponsors to share a summary of results with participants. In the EU, the Clinical Trials Regulation requires a lay summary on the public trials database.

How to read news about a failed trial

Some signs that coverage is balanced:

• it gives the primary result first, with the confidence interval
• it reports absolute numbers alongside ratios
• it labels subgroup findings as exploratory
• it notes what the drug is already approved for, if anything
• it quotes independent experts as well as the sponsor

Coverage that leads with a subgroup, calls a neutral trial "positive", or skips the main result is worth reading with caution.

Placebo results matter too

In a double-blind trial such as CARDIO-TTRansform, nobody knew who was on the drug until the end. That design protects the result from hopes and expectations, which is one reason a neutral finding from a blinded trial is taken seriously. Our guide to placebo and blinding explains this using NOPARK, a Parkinson's trial in which the supplement group did slightly worse than placebo.

Questions to ask your doctor

If you have ATTR-CM or another condition where a trial has just missed its goal, these questions may help:

• Does this result change anything about my current treatment?
• Am I in a group that was studied in this trial?
• Is there another trial I could consider?
• What other treatments are approved for my condition?

A failed trial is part of how medicine learns. Most promising drugs never reach approval, and the ones that do have usually come through at least one disappointing result along the way. Reading a neutral result calmly, with attention to the confidence interval and to how subgroups were handled, lets you see how much was actually learned.

Related reading

For more on this, see Eplontersen in ATTR cardiomyopathy, the CARDIO-TTRansform results, How to find a clinical trial that fits you and How new medicines get approved and reach patients.