In sepsis, aggressive fluid resuscitation was harmful in randomized trial
University Teaching Hospital, Lusaka, Zambia
In the U.S., the federal government strongly encourages physicians to give most patients with sepsis aggressive crystalloid fluid boluses (~2-3 liters), without regard to a patient’s individual condition. In a randomized trial in JAMA, a similar standardized approach to aggressive fluid resuscitation in Africa appeared to cause the deaths of a significant proportion of patients with sepsis.
Authors randomized 209 patients with sepsis and hypotension presenting to the emergency department of a large national referral hospital in Zambia to receive either usual care or a 6-hour bundle containing:
A 2 liter fluid bolus, followed by 2 more liters over 4 hours (stopping for tachypnea, drop in arterial oxygen saturation, or increased jugular venous pressure)
Dopamine infusion through a peripheral IV if needed to maintain mean arterial pressure >= 65 mm Hg
Patients with hypoxemia or tachypnea were excluded, because authors had previously found that aggressive fluid resuscitation in such patients with sepsis seemed to hasten their deaths by pulmonary edema (leading to that trial stopping early). Patients with end-stage renal disease or congestive heart failure exacerbations were also excluded. Most patients in the study had HIV-AIDS, not on antiretroviral therapy.
Those randomized to usual care received whatever care their doctors advised. Only 2% in the control arm received dopamine (14% in the intervention arm). Notably, half the usual care patients received no intravenous fluids whatsoever. Consider that in the major U.S. trials since the original 2001 EGDT paper, all patients received aggressive fluid resuscitation, so its relative benefits or harms have never been tested.
Aggressive Fluid Resuscitation in Sepsis: Potentially Dangerous?
Many more septic patients receiving the aggressive fluid resuscitation protocol died in the hospital (48% vs 33%) and at 28 days (67% vs 45%) compared to those receiving usual care.
There are obvious reasons not to leap to conclusions about fluid resuscitation for sepsis in the U.S., based on this single center study in Zambia. “Usual care” is very different there than in the U.S. and other high-income countries. Most enrolled patients had sepsis as a complication of advanced HIV-AIDS, sometimes with tuberculosis or malaria -- profoundly different phenotypes than the typical patient with sepsis in the developed world. Most patients in the U.S. with sepsis do not receive 4 liters of crystalloid over 6 hours (although many do).
Most importantly, Zambia has a total of 10 ICU beds for its 13 million people. Patients in pulmonary edema as a result of fluid resuscitation could not be rescued with mechanical ventilation.
Thus, this study by no means proves that aggressive fluid resuscitation is consistently harmful in sepsis. It does indicate that it can frequently cause respiratory failure or other harms -- as these authors reported previously.
In the U.S., we can intubate people with sepsis who develop iatrogenic pulmonary edema from aggressive fluid resuscitation. But might it be better to not cause pulmonary edema at all?
With the dearth of evidence supporting standardized aggressive fluid resuscitation in sepsis, the National Quality Foundation and Centers for Medicare and Medicaid should consider modifying the fluid resuscitation component of the SEP-1 core measure for sepsis, which was adopted under questionable circumstances in the first place.
It would be ethical and appropriate to conduct a large multi-center randomized trial in sepsis in the U.S. testing current standards of fluid resuscitation against patient-centered, individualized care. The latter would include vasopressor support with or without fluid resuscitation, according to physician judgment.
A few small studies testing the hypothesis that aggressive fluid resuscitation may be unhelpful in sepsis are in process or completed, according to clinicaltrials.gov.