In a recent publication, the results of a large Phase 3 trial comparing solithromycin with moxifloxacin in the treatment of CAP are presented . As this was a global registration trial, we should refer to the indication as CABP, the current FDA acronym for a very defined entity which has some very specific entry and evaluation criteria. The “B” stands for bacterial, of course, but in this as in so many other CABP trials, one half of patients do not have a documented bacterial infection. Nonetheless, FDA insists on making the ITT population, which includes all patients, i.e., those with “B” and those without “B” the basis of the primary efficacy determination. Not an easy logic to follow, but it has to do with randomization: FDA wants us to use both sides of the brain without prejudice, stupid!
CABP studies are full of quirks and pitfalls, in case you have not noticed yet. From case selection to disease severity classification, from efficacy timing to interpretation of results, sponsors have to follow regulatory definitions which have their own logic. It is sad to see how this drive towards more rigorous case selection and greater standardization of the study population has created so much extra work without a clear benefit for patients and clinicians. It also has resulted in lengthy articles chock-full of arcane details.
Arguably, the biggest beneficiaries of the revised FDA Guidelines are the CROs who continue to rake in million dollar contracts to enroll patients and manage study sites
The latest FDA Guidance criteria are the main reason why CABP studies have become so complex . No wonder the Lancet article is 10 pages long. It would be interesting to learn how many clinicians are interested in the strict regulatory definitions which are presented there, as it only speaks to folks with an interest in study design for regulatory submissions. There is also a 5 page appendix with additional data kindly provided by Cempra. We wish the ratio had been reversed; 5 pages for the main article would seem plenty.
So, our blog today is not so much about solithromycin and its activity vs moxifloxacin – both were shown to be equals in the CABP indication alright – but about the inordinate length of articles describing a regulatory submission study. We are using the SOLITAIRE-ORAL trial as an example but could have chosen another recent vintage CABP publications.
A CABP trial in Lancet is a 15-page affair, Nature published the discovery of the Double Helix on 2 pages. The key article on X-rays and all its properties by Roentgen fit on 9 small-sized pages.
Such data dump is overkill; why not put standard technical detail in the appendix? Transparency and attention to detail are important, no doubt, but the results could be presented in a much more understandable fashion. It is really hard to see the wood for the trees (well, in this case the trees look more like heaps of wood chips).
They taught you in school to adjust presentations with the audience in mind. The Lancet article speaks more to the regulator than the practicing clinician
Despite its length, some pieces of information are lacking. For instance, one would have liked to learn about the outcome of the valid patient population that had a valid entry pathogen and a full 5-7 day course of antibiotic therapy. In the past, this was called the CE (clinically evaluable) and ME (microbiologically evaluable) populations; they are subsets of the ITT and mITT-2 populations, respectively. Without this information it is virtually impossible to compare the SOLITAIRE data to older CAP studies which used the previous FDA efficacy criteria*.
For the SOLITAIRE-ORAL trial, it took 21 months and 114 centers to randomize 860 patients; a total of 1032 patients were screened for eligibility. In the microbiological mITT subset (all treated with a valid entry pathogen), there were 561 cases, worse even, in the stricter mITT-2 population we find only 225 cases between the 2 study arms. In summary, 225/860 or only 26% of patients were evaluable for this ITT-based efficacy analysis. The ME population would be even smaller.
The complexities introduced by non-harmonized trial design requirements between FDA and EMA are considerable and can become derailers. This was an operationally challenging study and a huge (and costly) effort. Cempra has to be congratulated for what seems to be an extremely well executed trial.
Our plea: provide clinical trial information in a more condensed way, concentrating on what is important for clinical practice. Relegate details of marginal interest of clinicians to an appendix for those readers who wants to dig deeper.
* The CE-Short Term f/u data are presented. However, the definition of CE-ST f/u is different from the prior TOC (Test-of-Cure) or EOT (End-of-Therapy) time points.
 C Barrera. Eﬃcacy and safety of oral solithromycin versus oral moxiﬂoxacin for treatment of community-acquired bacterial pneumonia: a global, double-blind, multicentre, randomised, active-controlled, non-inferiority trial (SOLITAIRE-ORAL). Lancet Infect Dis 2016; 16: 421–30