IVT does not reduce length of stay in alcohol intoxication

I started writing this post when this study came out, and was reminded of it today when a student mentioned that they'd been told by a nurse that young intoxicated patients were a good opportunity to practice large-bore cannulae. I pointed out that most of these patients don't actually need a cannula, as well as the slight ethical greyzone of the advice they'd received.

In January's edition of EMA was this interesting little study on the use of normal saline IVT for alcohol intoxication:

Perez, S. R., Keijzers, G., Steele, M., Byrnes, J. and Scuffham, P. A. (2013), Intravenous 0.9% sodium chloride therapy does not reduce length of stay of alcohol-intoxicated patients in the emergency department: A randomised controlled trial. Emergency Medicine Australasia, 25: 527–534.

The conclusion

I'll put the main outcomes of the trial first, then outline the methodology a little bit more.

There is no evidence that in alcohol-intoxicated patients a bolus of i.v. normal saline administration, together with observation, delivers better patient-oriented outcomes than observation alone.

This study of 144 patients presenting to ED with acute, uncomplicated alcohol intoxication demonstrated that a 20ml/kg bolus of normal saline in patients presenting to ED needing observation for simple alcohol intoxication did not make any difference to the length of stay in ED (primary outcome). In addition there was no difference in secondary outcomes:

  • Treatment time (from first seen by clinician to discharge)
  • Change in intoxication score using the OAAI tool used in the World Health Organization Collaborative Study on Alcohol and Injuries
  • Change in subjective level of intoxication according to the treating nurse
  • Change in blood or breath alcohol levels

In addition, the treatment group used A$31.92 more per presentation. This difference is attributed to the cost of giving saline and the difference in staff time. The authors estimate that if the routine laboratory tests used in the trial were also avoided, then they could have saved an additional A$62.94 per patient.


Design and population

This was a single-blinded (data analysis), two-centre (tertiary and urban hospitals), randomised trial of patients aged 18 - 50 presenting with uncomplicated alcohol intoxication. Pregnant, aggressive and intellectually impaired patients and complicated presentations (head injury, concomitant condition requiring further investigation or treatment, airway compromise) were excluded. Ethics approval was given for waiver of consent with the option to withdraw consent after treatment.

The primary endpoint was ED length of stay (LOS) from triage to discharge. Discharge time data were collected on a data collection sheet. Discharge was at the discretion of the treating doctor per local protocols (GCS 15, ambulant, no ongoing concerns, responsible escort, or left before treatment complete).

Secondary endpoints were:

  1. Treatment time from first seen by treating doctor or nurse to "ready for discharge"
  2. Absolute, percentage and rate of change in an objective alcohol intoxication scale at 2 hours and discharge (adapted from the OAAI tool used in the World Health Organization Collaborative Study on Alcohol and Injuries)
  3. Change in general state of intoxication assessed by the treating RN on a 5-point scale from very severe to not intoxicated
  4. Absolute, percentage and rate of change in breath alcohol level at 2 hours and discharge

Economic evaluation was performed taking into account direct resource use (IV sets, bags of saline, laboratory tests) and self-reported staff time. ED bedspace and security were not included.


The study was largely a convenience sample, based on triage nurse identification. This means that patients were missed due to clinical priorities, lack of recognition, etc. Only 144 of 425 potentially eligible patients were included, based on primary diagnosis, out of 3752 alcohol-related presentations in total.

The observation in both groups consisted of a monitored bay with oxygen saturation, blood pressure and heart rate. Hourly Glasgow Coma Scale measurements and respiratory rate observations were performed according to hospital protocol. This is more than we do as basic observation.

There was also some incomplete data (breath and blood alcohol, intoxication scores), and no reporting of adverse events or follow up (eg IV line assoc infections, etc).

So What?

Do we still give routine IVT to intoxicated patients? Does it reduce hangover symptoms, and therefore represent a better standard of care? Does it encourage earlier wakefulness and mobilisation by filling the bladder quicker?

Share your thoughts