N Engl J Med. 2025 Jun 12. doi: 10.1056/NEJMoa2416761. Online ahead of print.
A Crossover Trial of Hospital-Wide Lactated Ringer's Solution versus Normal Saline
Lauralyn McIntyre, Dean Fergusson, Tracy McArdle, Shane English, Deborah J Cook, Alison E Fox-Robichaud, Claudio Martin, John Marshall, Michael Pugliese, Kusum Menon, Kednapa Thavorn, Ian D Graham, Steven Hawken, Akshai Iyengar, Kwadwo Kyeremanteng, Raphael Saginur, Andrew J E Seely, Ian G Stiell, Daniel Bainbridge, Charles Weijer, Monica Taljaard; Canadian Critical Care Trials Group
PMID: 40503714
Why was this study done?
There have been so many trials done with balanced solutions compared to normal saline ever since Yunos and the late Rinaldo Bellomo suggested in a before/after study that balanced solutions are better for the kidneys (Yunos et al JAMA 2012). Apart from making the surgeons feel happy (Ringer's lactate is the only fluid they know), and spawning memes about ‘abnormal saline’ and bringing back ‘balance to the force', we are not really sure what this literature has achieved for changing clinical practice. We have discussed some of these trials on NephJC (SPLIT, Young et al JAMA 2015 | Summary; SMART Semler et al NEJM 2018 and SALT-ED Self et al NEJM 2018| Summary; BASICS Zampieri et al JAMA 2021 | Summary), but this was getting so repetitive that we actually skipped the last large trial (Finfer et al NEJM 2022) published in the journal. We are also skipping this one for the full NephJC treatment, but let's see what we can glean from this trial.
For those who have not been following this literature, the suggestion has been that (ab)normal saline is chloride-rich and not physiological. Balanced solutions such as Ringer's, Hartmann's, or Plasmalyte contain lower concentrations of chloride, as well as either lactate, gluconate, or acetate, which can be converted to bicarbonate. They also contain some additional cations, such as calcium or magnesium, and not just sodium. Overall, these seemingly more physiological solutions might be beneficial, especially in the setting of large volume resuscitation. The trials, however, have been decidedly mixed, with most trials being negative, except for a notable exception being the pragmatic Vanderbilt cluster trials (SMART and SALTED), which showed a benefit for the kidneys. The other exception has been BEST FLUIDS in the kidney transplant setting (Collins et al, Lancet 2023 | Summary), which we have discussed already. These patients, of course, are at high risk of AKI, in the form of delayed graft function, and they also receive high-volume resuscitation. In this scenario, it is plausible that the balanced solutions provide graft benefit.
Otherwise, for all the other patients in the hospital receiving intravenous fluids, should we be using a balanced solution and not (ab) normal saline? Even if the previous trials have been negative, what is the harm in using a balanced solution? This mostly boils down to cost, since balanced solutions are more expensive. Even if it is a few cents or a few dollars more per bag, we typically use thousands of liters of intravenous fluids in each hospital every day. Hence, the costs add up very fast. Let's see what this trial revealed in this background.
How was the trial done, and what did it show?
This was a cluster-randomized trial, done in Ontario, a Canadian province. The randomization was the order; some hospitals got saline before Ringer's and others got Ringer’s before saline. Every patient walking into the hospital for the 12-week period received this assigned fluid, as shown below in Figure S1. Being a cluster RCT, there’s no consent, there’s no eligibility (everyone gets this fluid), and outcomes are collected from the provincial administrative database at a very high level. See this #NephTrials discussion for what a cluster RCT entails and the implications.
Though the trial was powered to show a 1% difference in the outcome (primary outcome being a composite of death and hospital readmission), with 12 hospitals and 144,000 patients, they actually ended up with 7 hospitals and 43,626 patients (blame the confounded coronavirus). The actual adherence to the assigned fluid was also not perfect, as shown below, with adherence for saline being a bit better than for Ringer's. Again, this is at the hospital level, with no details available for patient-level adherence. Thus overall, these two factors of underenrolment and imperfect adherence make the trial much less powerful to demonstrate the 1% difference planned, so caveat emptor while interpreting the results.
Notably, from Table 1, ~ 58% of participants were women, and 17% were obstetric admissions. Overall, there was no difference in the primary outcome (−0.53 difference, 95% CI, −1.85 to 0.79; P=0.35) or any of the secondary outcomes, which included dialysis within 90 days (~ 0.5% in each arm). The subgroups can be seen below.
Figure 3 from McIntyre et al, NEJM 2025
What are the implications?
This is purportedly the largest trial of balanced solutions, which the authors claim increases generalizability. After all, the Vandy cluster RCTs had a puny ~ 15,000 participants compared to ~ 44,000 here. But this trial, along with SMART and SALT-ED, were cluster RCTs, and the Vandy trials had many more (smaller/shorter) clusters. Remember for a cluster RCT, the number of clusters (since they are the units of randomization) are somewhat more important than how large each individual cluster is! And the next largest trial (BASICS, from Brazil) at ~ 11,000 was a patient level trial, not cluster RCTl - so much more powerful despite a smaller patient N enrolled. So take the ‘largest trial’ claims for the present study with a pinch of saline. Secondly, the present study enrolled all comers to the hospital, compared to most of the previous trials, which enrolled some variety of critically ill, or patients with sepsis - likely a higher risk patient population in whom any benefit would be more likely to be seen. Subgroup splitters might unwisely look at the figure above from the present analysis and see some point-estimatologic differences in older/early ICU admission subgroups. Lastly, putting all the literature together, except for the SMART/SALT-ED trials (not including BEST FLUIDS, which was a specific transplant population), the rest of the trials are indeed solidly negative. And SMART/SALT-ED trials were again pragmatic cluster RCTs with small volume resuscitation in which the paradoxically positive results seem implausible. Nevertheless, a ‘Bayesian’ meta-analysis is out there suggesting a benefit with Balanced solutions (Hammond et al, NEJM Evidence 2022). My priors are that it doesn’t matter, Bayesian meta-analyses be damned.
Snark aside, let us accept that in the kidney transplant setting, we should indeed be balanced (Plasmalyte or Ringer's can be debated ad nauseam). In the modern era, where we don’t fill people up with gazillions of gallons of fluids, it is difficult to see any benefit with a bit more or less of chloride. Surely at a hospital level, it does not make sense to only stock Ringer's. A balanced approach of stocking saline and Ringer's might be the normal behaviour.
Summary by Swapnil Hiremath
Reviewed by Cristina Popa and Joel Topf