Editorials.

Two substantial Australian multi-centre clinical trials have recently provided answers to controversial questions about optimal fluid resuscitation strategies. What were the messages for critical care clinicians? Hypertonic saline (HTS) had been proposed as a better fluid for pre-hospital resuscitation of trauma patients especially those with traumatic brain injury (TBI). Potential advantages included rapid increase in intravascular volume, relatively small intravenous fluid volumes, and decreased intracranial pressure. Subgroups and meta-analysis of previous trials reported increased survival in TBI patients receiving HTS and HTS-dextran, and commercial products including HTS were approved and marketed in many countries. The HTS study was a double blind randomised trial including all 11 Victorian hospitals receiving trauma patients. Over 3 years, 229 patients with blunt injury, traumatic coma and hypotension were randomised to receive 250 mL 7.5% HTS or 250 mL Hartmann’s solution in addition to routine paramedic protocols. In both study groups hypotension was corrected on arrival to hospital, and in both groups survival was better than predicted by the usual scoring systems, and also better than patients in the meta-analysis. HTS tended to decrease the first measured ICP (P = 0.08) but did not improve long term neurological function measured 6 months after injury using the extended Glasgow Outcomes Score. The proportion of patients with favourable neurological outcomes was also the same in both groups. Therefore, despite theoretical advantages of HTS resuscitation in patients with TBI, in a well developed paramedic based trauma system, pre-hospital HTS was not beneficial and should not replace current isotonic crystalloids. However, in many Australian intensive care units, HTS is instead used as a preferred alternative to mannitol in patients with increased ICP. The Victorian trial did not address this question and indeed many of the study patients would also have received HTS in ICU as one component of therapy for intracranial hypertension. Accordingly, it would not seem appropriate for ICU clinicians to change practice concerning HTS osmotherapy on the basis of this trial. A well-constructed future trial of HTS for intracranial hypertension in ICU may be appropriate. The truly remarkable saline versus albumin fluid evaluation (SAFE) study compared the effect of 4% albumin and 0.9% saline resuscitation on 28-day mortality in nearly 7000 ICU patients. Patients admitted after cardiac surgery, liver transplant and burns were excluded. Mortality after both types of resuscitation was the same (RR 0.99; 95% CI 0.91 1.09). Clinician’s interpretation of this finding will undoubtedly be influenced by previous convictions. Those who favour theoretical advantages of colloids including albumin will likely be reassured that reports of increased mortality with albumin were not supported. Furthermore, their future choices of colloid type may shift towards albumin, being the only colloid tested in a large definitive trial and found to be safe. In contrast, crystalloid users may likely see no reason to change current practice, and also may be reassured that they use and teach a cheaper and equally effective product. It will be intriguing to see whether the recent international trend to decreased albumin usage after the Cochrane meta-analysis is reversed by the new data, or continued due to the cost considerations. In SAFE, there were also important findings in preplanned subgroups, and interestingly these were in opposite directions. In trauma patients, there was a strong trend toward increased mortality in patients treated with albumin (RR 1.36; 95%CI 0.99 1.86). This finding was in keeping with a meta-analysis, which found that trauma patients receiving colloids had a higher mortality than those receiving crystalloids. In trauma patients in SAFE, the increased mortality was almost entirely found in patients having trauma associated with brain injury (RR 1.62; 95%CI 1.12 2.34). Critical care clinicians are now likely to choose crystalloids for resuscitating future trauma patients and especially patients with associated brain injury, until more definitive data are reported. In the meanwhile, detailed re-examination of the SAFE-TBI patients is urgently required and is ongoing. A future randomised trial of albumin versus saline in TBI patients is inviting, but may be unethical if more detailed examination of the SAFE patients supports the present finding as real. On the other hand, in sepsis patients, the trend to increased survival in albumin treated patients (RR 0.87; 95% CI 0.74 1.02) is exciting but also requires confirmation, and invites a future randomised trial. This trial would be large, would have similar logistic and funding complexities to the SAFE study, and may be well supported by industry. It is intriguing to consider that the much-maligned Cochrane albumin meta-analysis may have been partly correct. We know that the quality of albumin production has improved with time, and that current Australian and