TXA, The Miracle Drug With Mostly Negative Trials

Tranexamic acid is a lovely drug, and, as I always say, it’s virtually prototypical for the “floor wax AND a dessert topping” class of medications. Life-threatening bleeding AND skin whitening.

With all the hype, you’d expect it to have serious chops with respect to mitigating death – but it doesn’t. CRASH-2 led the use of TXA into prime time with a 1.5% absolute risk reduction in mortality when given to major trauma in a timely fashion. However, other such famed trials as WOMAN and CRASH-3 missed their primary endpoints – and we rely on their overall favorable “trends” and subgroup analyses to support current practice.

This, again, is no different – with its own twists and turns.

In this trial, patients were enrolled prehospital with suspected moderate or severe traumatic brain injury. This protocol is based on the very reasonable hypothesis “sooner is better” when the consideration is potential hematoma expansion into a closed space such as the calvarium. Patients received either a bolus of TXA, a half-bolus and half-infusion, and/or matching placebo to maintain the blinding. The primary outcome was favorable neurologic function at 6 months.

The trial enrolled and treated 966 participants with head injury, about three-fifths of whom ultimately demonstrated intracranial hemorrhage on advanced imaging. Consequently, a similar proportion had GCS ≤10, and many had polytrauma with a median injury severity score of 17. No obvious differences between groups at baseline should be expected to dramatically bias the true point estimate within the confidence interval.

And, the results are grossly unremarkable, to say the least. The authors take a bit of an interesting approach to presenting their negative result by explicitly stating their non-significant p-value of 0.16 for benefit alongside their non-significant p-value of 0.84 for harm. These are frequentist results, of course, and the p-value for benefit describes an adjusted effect size confidence interval between -0.9% and 10.2% with a 95% likelihood of containing the true point estimate of mortality benefit to TXA. The opposing p-value is a one-sided estimate of describing the likelihood (or lack thereof) the TXA is actually responsible for harm. Similar non-significant trends are seen with 28-day mortality and progression of intracranial hemorrhage, showing similar 95% CIs are a true point estimate more likely to favor TXA. Suffice to say, the TXA is providing benefit than harm, but these data are not profoundly conclusive.

So, again, this trial joins the parade of trials requiring some exploration and faith to declare TXA the victor. These trials are, at least, remarkably consistent in their mild blessing of TXA, and certainly the effect size is greater than the apparent collection of adverse effects. There are oddities here, as in the others, such as 16% of the trial cohort being lost to follow-up, and the minimal differences in clot lysed at 30 minutes following maximum amplitude – but these are not enough to stay your hand.

The effect size on mortality and neurologic function is almost certainly small, but even large number-needed-to-treats are acceptable for safe, inexpensive drugs for and important, life-altering outcomes. We should be incorporating TXA into various treatment protocols, educated by our best interpretation of the evidence, but mindful it only has a lesser influence in patient outcomes as part of a comprehensive approach to the injured trauma patient.

“Effect of Out-of-Hospital Tranexamic Acid vs Placebo on 6-Month Functional Neurologic Outcomes in Patients With Moderate or Severe Traumatic Brain Injury”
https://jamanetwork.com/journals/jama/article-abstract/2770409

More Anti-Antibiotics in Diverticulitis

It’s hard to believe I’ve covered this topic – the evidence for reducing antibiotic exposure in diverticulitis – as long ago as 2013. While it might not be the prudent thing to be the first adopter of a new medical practice, I’d have thought this idea could have had more traction, sooner.

This latest piece of evidence is a “pragmatic” randomized, placebo-controlled, double-blind trial of patients admitted for “non-complicated” diverticulitis. “Non-complicated” in this context means, effectively, non-perforated, and explicitly defined as Hinchey 1a grade. Patients were also excluded if they had multiple systemic inflammatory response criteria, immunosuppression, or other comorbid physical status. Those who received antibiotics were given a regimen including cefuroxime, metronidazole, and amoxicillin/clavulanic acid, while the opposing side received matching placebo.

The trial was rolled out across four hospitals between 2015 and 2019, including three in Auckland, New Zealand, and one in Sydney, Australia and ultimately included 180 participants. The primary outcome was length of hospital stay, the difference for which was not statistically significant at 40 hours for the antibiotic cohort and 46 hours for placebo. The authors tracked many adverse events, discontinuation reasons, and protocol terminations, and there was no clear pattern or apparent trend favoring either cohort, within the scope of the small sample.

After so many years, it would be lovely to finally see better guideline uptake in support of antibiotic stewardship for mild diverticulitis. It certainly seems consistent across all the various trials and cohorts by multiple groups across the world, now, there is minimal, if any, additive benefit – or, at the least, the harms from antibiotic use are similar to those whose diverticulitis progresses left untreated.

“Antibiotics Do Not Reduce Length of Hospital Stay for Uncomplicated Diverticulitis in a Pragmatic Double-Blind Randomized Trial”
https://reader.elsevier.com/reader/sd/pii/S1542356520304262