How to Design a Clinical Trial for Aging

Current incentives push companies away from true aging drugs and toward single-indication development. This essay suggests an ‘aging’ Phase III trial design, and provides a calculator for trial sizes.

Therapies approved and labelled for slowing or reversing aging could create unprecedented patient benefit and commercial success, surpassing that of obesity drugs. Others have put thought and words to quantifying this upside. In this essay we address the challenge of proving out such upside in human trials. Measuring aging is hard, and the regulatory and reimbursement pathways for aging-related endpoints are uncertain. So far, no pivotal trials have been attempted for aging.

We propose using a composite endpoint that counts multiple age-related events, which can greatly increase the effective event rate and cut required sample sizes by half or more. We walk through why this design makes sense, what the endpoint options are, and provide a calculator for estimating trial requirements.

Best practices in clinical development disincentivize therapeutics that directly address aging mechanisms.

For the purposes of this essay, we define aging as:

• The sum of underlying biological processes that happens in every human gradually over long periods of time…

• That makes people more susceptible to diseases associated with high chronological age (e.g., neurodegeneration, cardiometabolic diseases, cancer).

Consequently, for the purposes of this essay, an aging drug is a drug that:

• Works against these underlying natural processes.

• Reduces the risk of multiple diseases with unrelated pathologies associated with high chronological age in virtually every human (vs just patient groups with specific risk factors like high LDL, obesity, etc).

To get a therapy approved and labelled for aging, you eventually need to run your pivotal clinical trial of that drug against ‘aging’, by some definition that relates to how patients feel, function, or survive. In other words, the primary endpoint of the treatment’s Phase 3 trial needs to measure aging, not be limited to any one specific indication, and connect to concrete patient outcomes¹.

So far, nobody has run such a trial. This is in part because measuring aging is hard and new: it creates uncertainty around the exact endpoints, approval criteria, and reimbursement pathways.

Instead, the majority of biotech companies that are developing aging drugs follow the conventional path of developing/validating their therapy against a specific disease. That is, they pick an age-related disease as a proxy indication, and plan to get their drug to market against that indication. Then, once the drug is on the market, they plan to expand the drug’s labelling for other indications by running additional trials against other diseases. This is a reasonable approach: we see it working today with e.g. GLP1-R agonists.

However, this approach creates an incentive misalignment: If you aim for approval against a specific indication, you will want to optimize your drug development programme for making a drug that works best against that one indication.

Chances are that the drug that works best for that one indication won’t be the exact same drug that works best against aging as a whole. Thus you will have to pick between optimizing your drug development against your target indication or optimizing your drug development against aging. Considering that making any successful drug is already very hard, and you need your drug to succeed, most will choose to optimize for making a good drug for your initial indication rather than developing a great aging drug.

We believe that developing aging drugs has a higher expected value than traditional drug development because of the much-much-higher upside (similar to what we see for incretin drugs). But the much higher risk of failure will turn many reasonable investors and entrepreneurs away from this path; as we wrote about previously, a large success with known probability is more palatable than a larger success with unknown probability. That is, for most people, having a 5% chance for a $10B upside is more attractive than a 1% chance for a $100B upside.

Being kind-of-forced to chase local maxima is frustrating. It’s therefore worth thinking about how we could align our incentives with what we, as aging researchers and biotechnologists, would want to do in the first place. Besides frustration at the individual level, this incentive-misalignment also leaves a lot of societal and patient benefits on the table. As a society, we should be pursuing the opportunities with the highest expected value.

What endpoints would we use for a trial targeting aging?

Specifically, let’s think about what the primary endpoint of our eventual Phase 3 trial can be. This is the decision that should and will drive the design of the whole trial, and will serve as the north star for all of our earlier-stage drug discovery and development processes.

We know that our primary Phase 3 endpoint needs to be:

• A generally good proxy for aging (vs for a more specific indication), to avoid misaligned incentives.

• Reimbursable.

Reimbursable means that if our Phase 3 trials succeed and a treatment shows improvement on our primary endpoint, that also means that our treatment will be:

• Approved by the FDA (and other regulatory agencies worldwide)

• Reimbursed by payers, for a very wide group of patients, at a reasonably high price tag. Payers generally only pay for therapeutics that produce very obvious and very direct benefit to the patients on their insurance plan.

The main relevant payer types we are talking about here are:

• Private commercial insurance plans (United, Cigna, etc.)

• Medicare

• Governments outside of the US

• Patients themselves (out of pocket)

Private and government plans generally care about the same treatments and evidence. Most important for aging drugs will be Medicare and other government payers, as most of the cost savings and quality-adjusted life years an aging drug will create are going to be on the 65+ year old population.

Good aging drugs will be prescribed earlier in life as well, and provide disease prevention and quality of life improvements for patients on private plans, which is needed for private plans to be incentivized to reimburse². (In this essay we’ll ignore aging drugs that need to be dosed early in life but only provide medical benefits at age 65+, since trial design for these is even less tenable).

Out of pocket payments may be a bit of a wildcard for aging drugs, as patients themselves may require less efficacy and health economic evidence as long as the upside and safety is established. There are trends in this direction, led by incretin weight loss drugs. We expect that the market-size for this is still small compared to a drug reimbursed by main payers for the foreseeable future, and that aging drugs will have less visible upside, so going forward we’ll assume the need to get reimbursed by government and private payers.

Table 1 shows endpoints used for age-related diseases:

We see that the overwhelming majority of Phase 3 primary endpoints are extremely simple. These endpoints are either:

1. Functional, i.e. measuring how well a patient functions, in the simplest terms possible, or

2. Morbidity or mortality related, i.e. measure the presence/absence/frequency of unwanted medical events (incl. death) or diseases.

The few Phase 3 endpoints in use currently that don’t fit either of the above two categories are very extensively validated surrogate endpoints. We see such endpoints in e.g. cardiovascular disease (LDL cholesterol, blood pressure), type 2 diabetes (HbA1c). The benefit of these endpoints is that we can expect to see change earlier than in functional or morbidity-related endpoints for the same disease.

Currently, surrogate biomarkers for aging are nowhere near the evidence level needed to be used as endpoints in pivotal trials. Long term, we should certainly aspire to develop such surrogate endpoints, and the way such surrogate endpoints should be developed warrant further thinking - but the first aging trials are not going to be against such endpoints.

This guides the options we have for a trial against aging. The primary endpoint of our aging Phase 3 trial will either have to be something very functional, or something morbidity/mortality related.

Option 1: Function

Functional aging endpoints would function much like today’s frailty indices: they aggregate metrics such as cognition and physical fitness into a single score. A better score indicates a higher quality of life. Importantly, a well-designed frailty index could potentially demonstrate drug efficacy earlier than other Phase 3 endpoints.

However, current frailty indices are not approved for Phase 3 reimbursement. This is partly because combining multiple domains (cognition, mobility, biomarkers) into a single metric makes it difficult to interpret exactly what an “improvement” signifies. Furthermore, these indices have not been extensively validated to predict costly medical events, such as intensive care or multi-day hospitalizations, in patients without specific diagnoses. Within individual diseases, like Parkinson’s and Alzheimer’s, composite endpoints similar to frailty indices are used. So a path is feasible, but for aging we would need to refine and validate a better frailty index than what exists today³. Such effort will not be the focus of this essay.

We would not utilize a frailty index for a first-in-class aging trial today, because using a new frailty index will carry regulatory and reimbursement risk, which means that the uncertainty we are trying to address remains. The risk is significantly higher here than with mortality or morbidity endpoints because the individual components of a frailty index are not independently reimbursable.

Option 2.1: All-cause Mortality

The simplest trial design measures all-cause mortality as the primary endpoint. This involves comparing treatment and control groups to determine if significantly fewer patients die in the treatment arm.

While this endpoint is unquestionably reimbursable, low baseline event rate means low responsiveness. Achieving statistical significance would require extremely large sample sizes or very long follow-up periods to capture enough data points.

Option 2.2: Morbidity prevention

The next logical option is to run trials targeting age-related morbidities. These trials measure the frequency of specific medical events, which can range from acute incidents (like a stroke or myocardial infarction) to new disease diagnoses. (Conversely, trials can also track positive outcomes, such as the resolution of a condition.)

Heart failure studies, such as this finerenone trial, offer a practical example. Investigators compare treatment and control groups to determine if the therapy significantly reduces adverse events, such as cardiovascular death or heart failure complications, over the trial’s duration. (Note: these trials technically measure “time-to-first-event” to allow for more nuanced analysis, but the general principle remains the same.)

Crucially, morbidity endpoints are reimbursable and yield results much faster than all-cause mortality trials. If ten years typically elapse between a diagnosis (the medical event) and death, measuring morbidity confers statistical power a decade sooner. Given the current landscape, this is our preferred trial design.

The first aging trials should focus on multimorbidity prevention

For an aging-specific trial, we expand the morbidity design to track a composite endpoint of multiple distinct medical events, rather than a single disease type. Specifically, we measure “time-to-first-event” across a broad spectrum of age-related pathologies.

A qualifying event might include:

• Cardiovascular events (e.g., heart failure, MI, stroke, or CV death)

• New diagnoses of age-related, lethal cancers

• Neurological or psychiatric disease onset

• Renal disease (e.g., late-stage CKD) onset or progression to specific stage

Note: These examples are illustrative. Actual inclusion criteria would depend on the specific drug and require precise diagnostic definitions.

In the simplest model all events are weighted equally. We count only the first qualifying event per patient. While more complex designs could weight events based on quality-of-life impact, the binary “event vs. no event” approach is standard.

This design is highly likely to be reimbursable because its individual components are already approved endpoints. Precedents already exist in standard practice, such as the 3P-MACE (death, MI, stroke) used in cardiometabolic trials. Notably, the TAME study employs a similar multi-morbidity design, with reports suggesting the FDA considers this approach reasonable.

The problem with prevention trials

It is important to recognize that this design functions fundamentally as a prevention trial. Unlike standard therapeutic trials that aim to slow progression in patients who are already diagnosed, we are recruiting healthy individuals and measuring their conversion to a diseased state.

Structurally, this mirrors vaccine development: healthy participants are treated and monitored to see if they avoid a specific outcome. The inherent challenge with this structure is statistical efficiency. In a prevention trial, participants who would never have developed the disease naturally (even without treatment) do not contribute to the efficacy signal. If the natural event rate for a specific disease is low—for example, if only 1 in 10 participants would typically get sick—the trial requires a massive sample size to achieve the same power as a standard study. This inefficiency is a primary reason why prevention trials for single chronic diseases are the method of last resort, used when biomarkers are not available.

However, the multi-morbidity endpoint helps address this. While the probability of a healthy patient developing a single specific disease in a short timeframe is low, the probability of that same patient developing any one of five or six age-related conditions is significantly higher. By aggregating these distinct pathologies into a single composite endpoint, we dramatically increase the overall event rate. This restores statistical power, allowing us to run a prevention trial with a feasible sample size and timeline.

Aging drugs may unlock prevention trials

The fundamental barrier to prevention trials is that the natural incidence of any single chronic disease in a healthy population is quite low. As shown in the table below, even prevalent age-related conditions often have annual incidence rates of 1–2% or less.

(Note: These figures are approximate estimates for illustration. They are based on published data but from different patient demographics. Precise power calculations would require rigorous adjustment for age groups and disease interdependence, but these numbers demonstrate the conceptual challenge.)

For a traditional drug targeting a single disease these low rates mandate sample sizes so large that the trial becomes economically prohibitive except for large pharma, and even then only ventured for very large potential markets.

Enter aging drugs. Because aging is the primary risk factor for multimorbidity, a true aging drug should prevent not just one, but multiple diseases simultaneously. This transforms the trial mathematics. We no longer need to wait for a specific event (like a heart attack); we can count any qualifying age-related event.

Consider a hypothetical scenario:

• Dementia incidence: 0.02/year

• Cancer incidence: 0.02/year

• Cardiovascular incidence: 0.02/year

By targeting the underlying aging process, we aggregate these risks. Suddenly, our effective incidence rate triples from 0.02 to ~0.06/year. Because sample size is inversely proportional to incidence, this roughly divides our required patient count by three. This shift in the math is what makes a preventative aging trial economically viable.

Understanding the statistics of trial design

With an understanding of endpoint selection and overall trial design, we next look at the statistical design drivers. To do this we must review the basic statistics that drive clinical trial planning. (We will keep this high-level; for a deeper dive into biostatistics, we recommend starting here.)

The primary objective of any trial is to prove that a drug successfully modulates the primary endpoint. To demonstrate this with statistical significance, the most critical decision is determining the sample size. This represents a fundamental trade-off: you need a sample size large enough to detect the drug’s effect with confidence (statistical power), but small enough to ensure the trial remains economically feasible, as cost scales directly with the number of patients.

We calculate the required scope of the trial using the following formula (an approximation of the standard Schoenfeld formula):

\(D = \left( \frac{1 + HR}{1 - HR} \right)^2 \times ( Z_{\beta} + Z_{\alpha/2} )^2\)

Where:

• D is the total number of medical events required to occur during the trial to achieve statistical significance.

• HR (Hazard Ratio) represents the expected risk reduction. It is the ratio of the hazard rates in the treatment arm versus the control arm.

• Alpha and Beta are standard statistical constants:

  • Beta (Type II error): The probability of a false negative (missing a positive result even if the drug works). This is usually set to 10%, which corresponds to 90% Power.

  • Alpha (Type I error): The probability of a false positive (detecting an effect where none exists). This is usually set to 0.05.

  • Z: A standardized normal deviate derived from the chosen alpha and beta values.

Note that D represents the number of events required, not the number of patients you need to enroll. For our proposed multimorbidity trial, we get to the number of patients needed (n) with:

\(n = \frac{D}{i \times t}\)

where:

• i is the natural incidence of the event of the patient population in our trial, measured by the number of events per year.

• t is the trial length in years.

It is important to understand the trade-off here. Statistical power depends primarily on accumulating a specific number of events (D). Therefore, sample size (n) and trial duration (t) are inversely related.

We can trade a larger sample size for a shorter trial, or vice versa. For instance, if the formula suggests a specific sample size for a one-year trial, extending the trial to five years would allow us to reduce that sample size by a factor of roughly five (assuming the event rate remains constant). We would likely prefer a longer trial, as aging drugs may require sustained dosing to demonstrate efficacy.

From this relationship, we can see that the two biggest drivers of our required sample size (n) are the hazard ratio, i.e. how good our drug is at preventing the events in question, and the incidence rate.

Estimating the trial size requirements of a drug’s multimorbidity prevention trial

In reality, an aging drug will not work equally well for every disease; it will likely show different effect sizes (Hazard Ratios) for preventing cancer versus dementia or heart failure. Determining the required sample size under these heterogeneous conditions is complex. We used multiple approaches to estimate composite hazard ratios and thus approximate trial size requirements.

Quick and dirty approach: Meta-analysis of events across types

Our first approach to modeling an aging trial with prevention of any one of several age-related events as the endpoint, was to treat it as isomorphic (structurally identical) to the problem of combining multiple trial arms in a large study.

Consider a traditional multi-center trial: a drug might work exceptionally well at one site and less so at another due to local heterogeneity. To find the true global effect size, statisticians calculate a value somewhere in between these extremes. We apply this same logic to an aging drug: if the therapy is highly effective against one morbidity but less effective against another, the overall “aging” effect size should theoretically lie in between.

Consequently, we computed an aggregate Hazard Ratio to represent the drug’s overall efficacy in preventing any contributing aging event. We borrowed statistical methods from meta-analysis, which is typically used to synthesize results from independent trials. You can think of this as a weighted average that accounts for the reliability and incidence of each specific disease endpoint when generating a global estimate.

However, we must acknowledge that this is a “quick and dirty” approximation with several limitations:

1. Dependencies: It ignores the biological reality that comorbidities are often dependent; developing one age-related disease usually increases the risk of developing another, which this model does not account for.

2. Fixed Risk: This model assumes a fixed risk for each morbidity over time, whereas in the real world, biological risk accelerates with age.

3. Underestimation: This approach averages the effects on “time-to-first-event,” which likely underestimates the drug’s true power. For example, if a drug is excellent at preventing cancer—and cancer is the event that typically happens first—the actual delay to the first medical event will be much longer than a simple average of all disease risks would suggest

Nevertheless, we captured this approach in an interactive calculator, for anyone to use in approximating the power of different multimorbidity prevention trial designs.

You can access the calculator on the link below (make a copy of the sheet):

How to Use It: The most critical inputs are the Hazard Ratios. These are your estimated effect sizes for the drug’s ability to prevent specific diseases or medical events.

Our hope is that you input the estimates for your lead aging asset and interpret the results as follows:

1. Feasible Numbers: If the estimated sample size is manageable, the trial is economically viable. Proceed to trial planning.

2. Prohibitive Numbers: If the sample size is too large to be affordable, the current asset may not be potent enough. Pivot: You likely need to find or develop a more effective molecule before warranting an aging trial.

Calculator Assumptions:

• Scope: The model assumes the drug is effective only against Cancers, Cardiovascular Diseases, Neurodegenerative Diseases (Dementia), and late-stage Chronic Kidney Disease (CKD).

• Incidence: We utilize population averages for the 65–70 age demographic.

• Independence: We assume different diseases occur independently. While this does not strictly hold in clinical reality (where one age-related disease often predicts another), we maintain this assumption here for simplicity.

Worked example of using the calculator

To illustrate how we determine trial size requirements, let us examine a specific imaginary candidate: rejuvecin.

To calculate feasibility, we first need estimates of rejuvecin’s efficacy against the specific diseases we care about. Estimating clinical effect sizes of any treatment well is very hard, how to do that is maybe the content of another essay. Coming up with the best effect size estimates specifically for rejuvecin’s effect in preventing different age-related diseases is hard again, and maybe the content of another essay.

For this illustration, let us assume that rejuvecin reduces the risk of the following events in a healthy patient:

• Cancer: 40% reduction (HR = 0.6)

• Dementia: 5% reduction (HR = 0.95)

• Cardiovascular Events (Stroke, HF, MI): 15% reduction (HR = 0.85)

• Renal Disease (Stage 4/5 CKD): 10% reduction (HR = 0.9)

(Context: For comparison, widely used drugs like statins typically demonstrate an HR of 0.7–0.8 for major cardiovascular events, while SGLT2 inhibitors often show an HR of 0.8–0.9 for heart failure hospitalizations.)

Putting these numbers into our calculator yields the following estimates:

• Composite Hazard Ratio: 0.87

• Estimated Trial Size: ~9,100 participants

These results demonstrate that a multimorbidity trial for rejuvecin would require substantially fewer patients than running independent trials for each specific arm, except cancer. Ignoring the different cost per patient per indication, launching rejuvecin as a cancer drug only may be cheaper, but offer a much narrower label that, between time versus patent life and total cost of additional trials, would not likely expand to cover the additional indications.

Drugs with ‘balanced’ effects benefit most from multimorbidity trials

The rejuvecin calculation above highlights a critical nuance: the drug’s effects are asymmetric. In our model, the bulk of its benefit comes from a strong anti-cancer signal. While a multimorbidity trial is cheaper than running all independent arms, it is likely not cheaper than running the cancer arm alone.

The economics favor the multimorbidity design when the drug is “balanced”, meaning it has pleiotropic effects across multiple domains rather than a single spike in efficacy.

Consider a drug where the disease-specific Hazard Ratios are more consistent (egalitarian) than in the rejuvecin example. Our calculator indicates that this scenario offers an even stronger economic opportunity. This occurs because the power for an asymmetric drug is effectively limited by the most affected indication, whereas pooling indications for a balanced drug optimally increases the effective incidence rate.

As shown in the table below, if the composite hazard reduction remains similar but applies to more types of events, the required sample size drops and in particular is far lower even the cheapest single-indication trial.

This confirms a key insight for longevity biotech: for an aging trial to be definitively superior to a single-disease trial, the drug must typically possess work broadly across at least several distinct aging pathologies.

Quantitative approaches

To validate the “Meta-Analysis” proxy used above, we sought a more rigorous, first-principles approach. We consulted with biostatisticians on the logic, and with Zane Koch built simulations of outcomes in control and treatment groups (scaling the incidences by the individual HRs we previously estimated) with varying trial sizes, using our previous estimates of yearly morbidity incidence rates.

The composite HR here converged to around 0.80 as the trial became better powered (achieving 90% power at ~5000 participants for a 5 year trial). This is in line with our expectation that our meta-analytic average would underestimate the effect size of an aging drug, which came out to 0.87 with the same assumptions of incidence.

To cross-check our simulation results, Mica Xu Ji proposed a third method: an analytic derivation that approximates the composite Hazard Ratio (HR) without requiring a full simulation loop. This approach calculates the expected value of the hazard ratios across all conditions, weighted by their incidence.

The approximation is defined as:

\(HR_Z \leftarrow E_c \left[ \frac{\lambda(t \mid c, Z = 1)}{\lambda(t \mid c, Z = 0)} \right] = \sum_{c} P(c) \left[ \frac{\lambda(t \mid c, Z = 1)}{\lambda(t \mid c, Z = 0)} \right]\)

Where P(c) represents the natural incidence of each condition, and λ(t | c, Z) is the Cox proportional hazards function. The ratio term inside the brackets effectively represents the specific Hazard Ratio for condition c (comparing treatment Z = 1 to control Z = 0).

When we substitute our hypothetical HR estimates into this equation, the result converges to 0.81. This is virtually identical to the 0.80 figure derived from our simulations. This convergence provides high confidence that both the granular simulation and this analytical model are accurately capturing the underlying mechanics of the trial design.

Conclusions

To summarize:

• We believe that understanding the path whereby an aging drug can be tested directly against aging (vs specific proxy-indications) will incentivize the development of true aging drugs that promise a higher upside and a higher expected value for all.

• We explore potential primary endpoints for a potential pivotal trial against aging, and claim that the type of endpoint and trial that is most likely to lead to wide reimbursement is a multimorbidity-prevention trial, measuring time to first of any age-related event.

• We suggest potential ways to calculate sample size requirements of a potential multimorbidity-prevention trial, given a drug’s estimated effect sizes against preventing specific age-related events.

• From these calculations, we conclude that:

  • In all cases, sample sizes needed to run a multimorbidity-prevention trial are substantially lower than running a prevention trial for each disease independently. That is, this trial design offers a stronger drug label for a reduced cost.

  • Aging drugs that have balanced effects across different disease groups provide even better economics: The cost of a multimorbidity-prevention trial becomes cheaper than a prevention trial for any single indication.

Self-evident disclaimer: This is an obviously imperfect analysis of how to run trials, to prompt discussion and further work. We tried to signal where we are less confident in the assumptions we’ve taken, but if that failed: please use the ideas and resources presented here thoughtfully, and use them at most as a starting point in your own careful thinking that precedes e.g. committing resources to a project. We will certainly do the same.

---

Acknowledgements

Most of the work was done by Marton Meszaros and Brennan Overhoff. Thanks to the many people who contributed and gave us guidance or feedback. Special thanks to Mica Xu Ji, Zane Koch, Frank David, and everyone in Norn Longevity Nexus.

All remaining errors in the essay are owned by Marton and Brennan only. Feel free to send feedback.

1. Brief background on drug development process and different phases of clinical trials, for those unfamiliar:

The drug development process can be broken down into four main stages. To begin, a company will conduct preclinical studies in model organisms to find a putatively efficacious compound for their indication of interest. Once they have established sufficient preclinical support they can submit an investigational new drug (IND) approval application to the FDA which, if approved, allows them to move into human studies. Phase 1 trials primarily assess safety, dosage, and pharmacokinetics in a small group of healthy volunteers or patients. Phase 2 trials expand the population size to explore efficacy in humans and further evaluate safety, often in individuals with the target condition. Phase 3 trials, also known as pivotal trials, are large-scale studies (randomized, double-blinded, and controlled) that provide the definitive evidence that the drug is efficacious (and safe). This is required by regulatory bodies like the FDA to approve the drug, as well as payers like health insurers to want to pay for the approved treatment. With exceptions, two pivotal trials precede drug approval.

Trials are run on patient populations that are defined by inclusion/exclusion criteria of the trial (such as patients aged 60 to 70 that are diagnosed with x). Trials measure one primary, and a couple of secondary endpoints. Achieving a predefined effect on the primary endpoint is by far most important goal of the trial, if you do that, your trial succeeded, and your drug is likely to get approved. Generally, upon approval your drug will be labelled (prescribed, paid for) on for the patient population you include in your Phase 3 trials. Hence the most important parts of a trial design are the patient population and the primary endpoint of the trial.

2. Even when plans are supposed to pay for certain treatments, various methods of complicating reimbursement for patients can reduce uptake.

3. We think the most promising frailty index for aging trials is the ‘intrinsic capacity’ framework developed by John Beard and others, but how to measure its different components is still being figured out.