Return to site

My speculations on why the UNITY senolytic trial failed

Oliver Zolman MD

Below I discuss my speculations on the possible reasons the recent UNITY senolyic trial might have failed

WOMAC or equivalent for other OA (osteoarthritis) joints has poor correlation with hard rejuvenation outcomes (as I discovered when I was systematically reviewing all therapies and trials for osteoarthritis for my osteoarthritis rejuvenation guide);

Nor is it a placebo-resistant outcome, and in fact is affected by many things other than placebo, from mineral baths, painkillers, supplements, weight loss etc (which is not clear they adjusted for in results or covered in inclusion/exclusion criteria).

Hard rejuvenation outcomes would be more like ROM (range of motion) and quantitative imaging markers (e.g. quantitative 3 Tesla MRI multiparametric arthrography).

I have been calling this year for there to be a discussion on what set of rejuvenation core outcomes measures in effectiveness trials (COMET), as per COMET initiative -, there should be for 1) whole body rejuvenation indication (like TAME's primary outcome) and 2) rejuvenation for specific ageing caused conditions (like this unity trial).

I don't think this is the case but if FDA requires a 3 month WOMAC as a sole primary outcome in Phase 2/3 for KOA intra articular (IA) injection trials then I would consider using non-FDA regulators, or working with FDA to change this requirement.

You can analyse efficacy failure in Phase 2 trials through the PICOTS framework

P - Population

  1. Population might have been improved by changing inclusion exclusion criteria to people with lower SASP levels (I.e. score of 16+ on Longevity Level 1 (i.e. non-smokers, reasonable diet, BMI, alcohol and exercise volume), i.e. SASP levels that would be reduced by the intervention to a range that allowed optimal rejuvenation; 
  2. It may have been possible to choose a population that was less prone to an IA injection placebo response (and still pass IRB). 
  3. There's also the question of whether they have underpowered and that they overestimated their 3 month effect size.
  4. They could also have probably been more strict in their inclusion/exclusion criteria / protocol for use of all other things that affect WOMAC, as I doubt they did a systematic review on this, and its possible patients did these during the trial as they did not understand it would affect WOMAC (and even with randomisation this still have made a difference) 

I - Intervention

  1. The aging world is still fixated on monotherapies; a senomorphic adjuvent, such as Longevity Level 1 or Level 2 interventions already validated in relevant low bias RCTs to lower SASP may have aided the outcome, given the potential reliance on optimal range SASP (not just 'lower SASP' (like it lowered SASP in phase 1 but potentially irrelevant as may not have lowered to optimal range) - has to be optimal range i.e. under age 25 healthy SASP non smoker level, as this is what young mice extrapolate would be - from that Nature review showing senolytics only worked to restore knee cartilage in young healthy knee damaged mice with low SASP and from papers on SASP mediated HLA-E blocking of senescent cell clearance etc.)
  2. I normally recommend only doing Level 3 therapies (like senolytics) after implementing Level 1 and 2 therapies, as the MoA (mechanism of action) and effective value (effect size x likelihood) is often dependent on Level 1 and 2 status. 
  3. I normally recommend only doing senolytics as 'double' (senolytic + senomorphic - remember Level 1 can be a senomorphic, and remember that only 0.3% of USA score 25/25 on Level 1) or 'triple' therapy (senolytic + senomorphic + relevant cell or stem cell therapy). 
  4. If they are doing intra-articular injections they would need to specify the clinical rationale for anatomical site placement and would require ultrasound or fluoroscopy, as well as use of non-lidocaine non-bupivicaine local anaesthetic and low dose contrast as both of these are mesenchymal stem cell toxic. 
  5. PRP or auto-BM-MSC might have been a good senomorphic adjuvent. 
  6. It sounds ideal that the IA dose should have been repeated at 3 months to reduce trial entropy - and this is consistent with the MoA of senolytics - and primary outcome read first at 6 months (see below points). 
  7. This is also the relevant when calculating sample size, as they should have aimed for an effect size that was e.g. 1.5 x the largest placebo response seen in the IA literature in WOMAC, and to achieve a larger effect size you would need higher doses or more frequent/multiple doses or adjuvents and combination therapies. If their predicted effect size wasn't above this 1.5x cut off in IA placebo then they shouldn't have run the trial this way. 
  8. If you really wanted to cure KOA you'd do a much smarter umbrella nested adaptive trial adapted to include tailored combination therapies that are costed to whatever insurance incremental cost effectiveness ratio (ICER) (i.e. $ or £/QALY) is in your business plan. This adaptive trial would include things like Longevity Level 1 and 2 and other Level 3 KOA relevant interventions like PRP/auto-BM-MSC fluorscopy guided IA injections with low MSC/cartilage toxicity adjuvents. It would still be biostatistically possible to prove that your intervention is the thing that takes you to total cure or cheaper cure or provides cost effective for the increment it provided. This means you can retain financial viability whilst increasing chance of success of your trial versus traditional trial. I find it hard to see how complex aging conditions can be cured without addressing the 10 hallmarks simultaneously or serially within the same trial. The likelihood of success is so much less by using monotherapies, and monotherapy trials can be more expensive as your effect size will be smaller so you need larger power and more participants.   

C - Comparator

  1. With intra-articular injections, they would have known this group would have had a super high placebo response (as IA injections have evidence for such). 
  2. There are probably creative ways to get around this - e.g. looking at the literature to see how long IA placebo lasts for and adjusting outcomes for this. 
  3. There is the issue of whether lidocaine/bupivicaine and or contrast were NOT used in the control group, as these may cause LACK of benefit in the intervention group, hence confounding the delta between the two. 

O - Outcome

  1. As discussed WOMAC is not a rejuvenation or objective marker. 
  2. You can cure a lot of KOA if you assess it only by WOMAC if you take a load of painkillers before measuring. 
  3. Is knowing that rejuvenation markers are statistically improved but WOMAC unchanged beneficial? YES given the awful nature of WOMAC (this is in contrast to what someone was saying before in this thread that this would be useless... it would only be useless if the primary outcome was better than WOMAC). 
  4. A better primary outcome might have been knee range of motion index or quantitative 3 Tesla MRI multiparametric arthrography or index of both of these, or some more placebo resistant form of WOMAC - indexed in with the aforementioned - that's specifically designed for IA trials. 

T - Timecourse

  1. The issue with the 3-month time point for the primary outcome is that the IA injection placebo response would probably still be in effect at this time
  2. AND that, based on the dozens of KOA PRP/auto-BM-MSC RCTs and stem cell registry, its probably longer than 3 months to see full theoretical effect of KOA senolysis. 
  3. They should have set the timing for the primary outcome read to be 6 months. 

S - Setting

  1. The setting of this trial included media hype on unity. 
  2. Enrolled patients should have an inclusion or exclusion criteria requirement to have never heard of unity before. 
  3. So they are blinded to the hype of unity in the media. 
  4. The protocol should have ensured lack of media exposure throughout the trial either from clinicians talking to patients in the trial etc.  

In this case major risk of bias from incomplete followup or allocation concealment seem unlikely, but these are often things that ruin trials too and should be thought about in every trial.

All Posts

Almost done…

We just sent you an email. Please click the link in the email to confirm your subscription!