We assessed the effects of a protocol restricting resuscitation fluid vs. a standard care protocol after initial resuscitation in intensive care unit (ICU) patients with septic shock.
We randomised 151 adult patients with septic shock who had received initial fluid resuscitation in nine Scandinavian ICUs. In the fluid restriction group fluid boluses were permitted only if signs of severe hypoperfusion occurred, while in the standard care group fluid boluses were permitted as long as circulation continued to improve.
The co-primary outcome measures, resuscitation fluid volumes at day 5 and during ICU stay, were lower in the fluid restriction group than in the standard care group [mean differences -1.2 L (95 % confidence interval -2.0 to -0.4); p < 0.001 and -1.4 L (-2.4 to -0.4) respectively; p < 0.001]. Neither total fluid inputs and balances nor serious adverse reactions differed statistically significantly between the groups. Major protocol violations occurred in 27/75 patients in the fluid restriction group. Ischaemic events occurred in 3/75 in the fluid restriction group vs. 9/76 in the standard care group (odds ratio 0.32; 0.08–1.27; p = 0.11), worsening of acute kidney injury in 27/73 vs. 39/72 (0.46; 0.23–0.92; p = 0.03), and death by 90 days in 25/75 vs. 31/76 (0.71; 0.36–1.40; p = 0.32).
A protocol restricting resuscitation fluid successfully reduced volumes of resuscitation fluid compared with a standard care protocol in adult ICU patients with septic shock. The patient-centred outcomes all pointed towards benefit with fluid restriction, but our trial was not powered to show differences in these exploratory outcomes.
Although this was an exploratory feasibility study (under powered for patient centered outcomes) with a number of limitations (see below) this study clearly demonstrated that a fluid restrictive strategy reduced the risk of acute kidney failure (37% vs. 54 %, p=0.03) (see Figure below). This may be counterintuitive to many clinicians who have suggested that a fluid restrictive strategy would increase the risk of kidney injury; furthermore, they content that fluid boluses should be given for oliguria.
This study has a number of limitations which likely reduced the “benefit” of a restrictive fluid strategy. Most notably the authors followed the initial Surviving Sepsis Campaign Guidelines (initial bolus of 30mls/kg) in both groups of patients. Consequently, 4.2L and 4.7L of fluid were given respectively prior to randomization… this is a LOT OF FLUID. It is noteworthy that in the recent VANISH trial (conducted between 2013 and 2015) on average 1134ml fluid was administered in the 4 hours prior to enrollment and randomization to vasopressin or norepinephrine (JAMA 2016; 316:509-18.). In Addition, there were major protocol violations in 27 patients in the restrictive group.
Many of the indications for a fluid bolus in the restrictive group were rather “nebulous”. In the fluid restriction group, isotonic crystalloid (saline or Ringer’s solutions) fluid boluses of 250–500 mL could be given intravenously during ICU stay in the case of severe hypoperfusion defined as either (1) plasma concentration of lactate of at least 4 mmol/L, (2) MAP below 50 mmHg in spite of the infusion of norepinephrine, (3) mottling beyond the edge of the kneecap (mottling score greater than 2) , or (4) oliguria, but only in the first 2 h after randomisation, defined as urinary output at most 0.1 mL/kg IBW in the last hour. Titrating fluid according to a lactate level is absurd (and an increased lactate is NOT a sign of hypoperfusion); unless the goal is to dilute the lactate with lots of fluids, the dose of norepinephrine at which point a fluid bolus is given is not defined (just increase the dose of norepinephrine) and giving fluids to a septic patient to increase urine output is a futile exercise. These factors together accounted for the relatively small (but statistically significant) separation in amount of fluid given over 5 days [mean difference 1.2 L (95 % CI 2.0 to 0.4); P < 0.001]. Despite these limitations, this study once again demonstrates that a liberal strategy of fluid resuscitation is harmful in patients with sepsis.