Nonsense XP

Bicarbonate adventures in critical care and the time function of marginal utility

I was at the 2022 Brazilian Intensive Care Congress, where thousands gather to renew their faith in commonplaces. All large events are the same faith revival exercise, including your national congress. However, in non-English speaking countries, the lectures have an aura of godly revelation because knowledge is produced in English and most of the audience don't speak English.

Sorry, I am already digressing from today's issue. This point deserves better treatment so I will spare it for later.

One night at the hotel, I was running through my Twitter feed, thinking how much I had spent traveling for the meeting (remember Brazil is a continent itself), and I was struck by the weakest hypothesis ever. A study hypothesized that a single bicarbonate infusion could affect acidotic patients' mortality after five years and their quality of life in three years.

I was also startled by the fact this paper was published in Critical Care Medicine. The study was a post-hoc analysis of the BICAR-ICU trial, which I think is among the most important studies in our field. Why didn't they present the follow-up data as a cohort of critical care patients? Simple descriptive statistics of all the patients would be so informative. The precise kind of study we need. Anyway, it was a tipping point. I couldn't stop thinking how five years of follow-up were wasted on such a terrible hypothesis.

Some would call it a red pill moment.¹ Maybe it was. But after a few unsettling hours, I named that feeling The Nonsense Experience (Nonsense XP)². After all, I am still in the Matrix, experiencing nonsense. At this point, I decided to start blogging about how intensivists mistake biological plausibility for clinical relevance, at least to dump the thoughts and stop thinking and get a good sleep.

The Nonsense Experience (Nonsense XP)

It is the feeling of a brief mental paralisys you get after reading something unexpectedly bizarre in a research paper. You don't feel it by simply acknowledging a Half-baked Hypothesis.

We know that survivors of critical care may undergo several long-term complications and disabilities. We instinctively associate patients' outcomes with our clinical choices during their ICU stay. It is a part of our omnipotence. I also understand it is convenient to keep milking a database for years.

However, there are dominant and marginal causes of long-term disability. To begin with, the authors should have stated the dominant drivers of long-term mortality and poor QoL. Then, they should have explained how the intervention (up to one liter of half molar sodium bicarbonate solution in 24h) affects those dominant drivers years after. Or maybe it was unnecessary because it's only magic, and we know we intensivists are naïve believers of the supernatural.

This is a vignette display of Blindness to Marginal Utility, with an interesting time component. Bicarb infusion may reduce mortality in a very short-term window, like hours. It would be reasonable to report how bicarb affects survival to the next day in a selected population of very acidotic patients. In this scenario, acidosis is a dominant driver of mortality. However, to study the same intervention in the long term is absurd.

This is another instance of a peculiar contradiction in intensivists’ minds. Everyone accepts the basics of carbonic acid chemistry, but people forget all the basics when facing a research paper. I'm about to describe this as a cognitive bias. I only need to find the best-describing name.

Infused bicarbonate (HCO3-) will find a hydrogen ion (H+) and remove it from circulation to form carbonic acid (H2CO3). Carbonic acid is a weak acid, meaning it does not fully dissociate into H+ and HCO3-in a water solution. The dissociation depends on pH, hence it is a pH buffer. Also depending on pH, carbonic acid may dissociate into H2O and CO2. Carbon dioxide (CO2) will be exhaled, and the H+ at the beginning of the paragraph is now in H2O. Of course, it can run backward from CO2 and H2O to H+. All the above reactions are timed in seconds. I'm sure there is no novelty here.

This conversation reminds me of a ghost haunting some ICUs: the idea that one can worsen acidosis by giving HCO3-. I eliminated the Chloride Kills People ghost in another post. Now let's examine the Bicarb Worsen Acidosis ghost. Of course, lung ventilation is necessary to eliminate the CO2 produced by HCO3- infusion. So there are two things to consider before pushing bicarb: (1) the actual capacity of increasing lung ventilation and (2) the extent of H2CO3 dissociation, depending on pH. I see low biological plausibility for worsening acidosis except for very unusual cases. Moreover, the BICAR-ICU trial showed no worsening acidosis. If you are that odd bicarb-hater who appears here and there ruminating about things like worsening or paradoxical acidosis, then I have a request. Please don't wait until the patient's heart stops beating to push bicarb.

Well, back to the clinical studies. I am aware that BICARICU2 is actively recruiting. It studies the effects of a single day of sodium bicarbonate infusion in patients with later-stage acute kidney injury. I look forward to seeing the results because it may change practice. I think there is a very good plausibility of a reduction in death and hemodialysis rates in the first few days, but the primary outcome is disappointing: 90 days mortality!! There is so much that can happen between day 1 and 90, that any effect of the bicarb infusion will disappear. This is surely a case of time-dependent utility. Why did they choose 90 days? Again I think of a cognitive bias. Some folks debate the proper follow-up time in critical care RCTs, implying it is possible or desirable to standardize the outcomes on 28 days, 90 days, hospital mortality, etc. It is impossible. It depends on the intervention.

Nonsense XP upsets me. Again. Now. That feeling of subtle disconnection to the world. A glimpse of the unfathomable pervasive silence that isolates each other people's minds. Why? Why will BICARBICU2 analyze the effect of bicarb on patients' quality of life 180 days after the infusion? What is the point? Desperate. Is anyone reading? I'd ask the authors: Please choose any other variable to correlate with 6-months QoL, or even better: only describe how patients are doing after 6 months.

I am leaving now, hoping to lose this feeling on my way to the hospital. I am returning to the ICU, for another night shift. I will see again the girl I admitted yesterday with DKA and oliguria. I didn't give her any bicarbonate.

Thanks for reading The Thoughtful Intensivist!

1. It refers to an iconic scene from Matrix, when Neo has to choose whether to leave his known life. I hate how this analogy is misunderstood and used out of context by a few idiots.

2. BEING a non-native speaker of English entitles me to a form of poetic license. The one thing I admire most in the English language is the liberty to use the same word as a noun, verb, or qualifier. However, it is sometimes difficult for me to figure out the exact meaning of a noum together with its preceding word. Having said that, I couldn't find a better way to express that feeling than using nonsense as an adjective, instead of nonsensical.