Acetazolamide for metabolic alkalosis in ventilated patients

Patients in the ICU with pulmonary disease primarily have respiratory acidosis and/or metabolic alkalosis. The respiratory acidosis is due to the primary disease and the metabolic alkalosis is due to our attempts to manage that disease, in this case with diuretics.

Compensation for metabolic alkalosis is suppression of respiration to allow the carbon dioxide to accumulate reducing the change in pH. Suppressing respiration is obviously a concern in patients on the vent. One option in this situation is to treat the metabolic alkalosis. Although I have read about giving hydrochloric acid infusions, that is a bit nuts. The conventional choice is to cause a drug induce proximal (Type 2) RTA with acetazolamide. 

The proximal tubule reabsorbs the vast majority of filtered bicarbonate, but it does it through a complex process, where bicarbonate is metabolized to CO2 and water (catalyzed by carbonic anhydrase). The carbon dioxide diffuses into the proximal tubule cell where bicarbonate is reconstituted (again catalyzed by carbonic anhydrase) and then secreted into the circulation. So bicarbonate is destroyed in the tubules and reformed in the tubule cells. Carbonic anhydrase inhibitors slow this and cause the patient to pee bicarbonate. The loss of bicarbonate causes metabolic acidosis, or alternatively, improvement in pre-existing metabolic alkalosis. The carbonic anhydrase inhibitor of choice is acetazolamide. Other carbonic anhydrase inhibitors include Methazolamide, Dorzolamide, Brinzolamide, Topiramate. The pomegranate contains natural carbonic anhydrase activity.

This is a nice theory but in February JAMA published a clinical trial to provide empiric data for this time honored strategy.

This is a multi center, double-blind, parallel-group, RCT.

They received consent from all patients and interestingly, if the patients were unconscious and unable to consent, their next of kin were required to provide consent. If the patients regained consciousness, a post hoc consent obtained. 

They enrolled patients with COPD and invasive mechanical ventilation, either intubation or tracheotomy tube (though only one tracheostomy patient was enrolled). Patients did not require metabolic alkalosis for inclusion. Though later they stated, "Patients with pure or mixed metabolic alkalosis received acetazolamide or placebo intravenously twice daily for a maximum of 28 days." I believe the drug was only administered if the patients had a bicarbonate higher than 26 and pH > 7.35. 

Study drug was started within 24-hours of intubation.


Acetazolamide 500 mg bid (or 1,000 mg if they were also on furosemide) twice a day.

Matching placebo.

I like that they protocoled the criteria for extubation and reintubation. If this is your end-point you should not leave it to subjective assessment of the treating physician. I tried to find this protocol in the supplement but could not. This was the closest I could find:

Study Outcomes

Primary end point: duration of invasive ventilation

Secondary outcomes:

  • Changes of serum bicarbonate
  • Partial pressure of carbon dioxide in arterial blood (PaCO2)
  • pH
  • Partial pressure of oxygen in arterial blood (PaO2):FIO2 ratio
  • Respiratory parameters (tidal volume, respiratory rate, and minute ventilation)
  • Weaning duration
  • Number of spontaneous breathing trials
  • Number of unplanned extubations
  • Number of ventilator-associated pneumonia episodes


They screened 694 patients

They randomized 382 patients

  • 187 to Acetazolamide
  • 193 to Placebo

Note that 28% of acetazolamide and 22% of placebo patients never received study drug because they did not have metabolic alkalosis. Why not make that an enrollment criteria?

Baseline characteristics

See table one. There was more LV insufficiency with acetazolamide (23 vs 16%) and more sedatives at baseline with placebo (2.1 vs 8.8%).

Primary outcome

The total durations of invasive ventilation was a median of 136.5 hours (IQR, 68.7 to 234.7 hours) in the acetazolamide group vs 163.0 hours (IQR, 86.2 to 242.9 hours) in the placebo group, for a between-group difference of −16.0 hours (95% CI, −36.5 to 4.0 hours), which did not differ significantly

Secondary outcomes

The difference between the bicarbonate on acetazolamide and placebo group was 0.8 mmol/L. Along with this change in bicarbonate, the pH was 0.01 lower with acetazolamide.


The discussion states that the 10% reduction in hours on the ventilator did not reach significance. The study was powered to detect a 15% reduction. Part of the reason for this is that the duration on the ventilators was shorter than anticipated.

The authors also pointed out that not enrolling patients with metabolic alkalosis might have been a mistake:

A limitation of this study is the presence of mixed metabolic alkalosis in most patients. Indeed, the lack of acetazolamide respiratory effect may be because so many of the patients had a degree of metabolic alkalosis too mild for the intervention. 

Though their own post hoc analysis did not show that patients with metabolic alkalosis had a more robust response.