The authors reply

We were pleased to see this secondary analysis by Rourke et al (1), recently published in Critical Care Medicine, of preexisting heart failure (HF) from their previous study (2) addressing fluid-resuscitation predictors in sepsis and septic shock. The authors evaluated 229 patients with a HF history from their original cohort. Clinical documentation of fluid overload (FO) conferred the lowest likelihood for achieving 30 × 3 (7%), lower than ejection fraction (EF) less than 50% (30% likelihood). Additionally, documented FO led to approximately half the fluid volume over 3 hours compared with an EF less than 50% (7 vs 13 mL/kg) and a drastically lower odds of reaching 30 × 3 (odds ratio = 0.17; 95% CI, 0.07–0.43). Even though HF results universally in less fluid (2), the reduced EF phenotype does not confer additive restrictions, yet HF combined with documented FO reduced initial resuscitation fluid volume compared with all other subgroups of HF.

Despite no differences in group mortality, achievement of 30 × 3 was associated with reduced ICU duration by about 2 days in the entire HF cohort (HR = 3.16; 95% CI, 1.81–5.53), driven primarily by reductions in ICU duration among those with EFs less than 50%. This finding deserves attention, as over 40% of critical care clinicians believe 30 mL/kg boluses are inappropriate with HF reduced EF (3). A recent review highlights the dangers of underresuscitation based on EF assessments, as techniques such as point-of-care-ultrasound have extensive limitations (4). Furthermore, another recent study showed elderly patients with EFs less than 35% had an association with improved mortality with 30 × 3 achievement (5). Together, the current evidence supports either neutral or beneficial effects with prompt conventional fluid resuscitation through 30 × 3 in HF.

Even as HF raises legitimate concerns for FO, it lacks a clear definition, resulting in unwarranted under resuscitation. In their original analysis, Kuttab et al (2) defined FO as physician documentation with justification of their rationale for limiting fluids (e.g., bedside echocardiogram, ultrasound, and jugular venous distension). Hawkins et al (6) posed a less subjective definition of FO, specifying a total body weight increase by at least 10% from baseline secondary to fluid administration. However, even this simple definition is not expediently applied in the acute setting and better guides fluid management after the initial bolus. Rapid sepsis interventions are well established to improve outcomes, further supporting the generalized delivery 30 × 3 absent unequivocable contraindications, and the most actionable FO mitigation strategies are best suited for after 30 × 3 during the fluid optimization phase, when FO most likely occurs (4).

Evaluating fluid resuscitation in HF phenotypes is a long overdue addition to the sepsis literature. Reason suggests underpowering of HF subgroups precludes detecting mortality reductions, but this gap has not impeded the adoption of most sepsis interventions to this population. Though prospective controlled trials are still lacking, nearly all data support conventional 30-mL/kg boluses in HF (and HF with reduced EF); however, extreme HF (i.e., EF < 10%) presents ungeneralizable situations. A remaining question centers around whether more conservative fluid management with more liberal vasopressors would benefit HF patients in the initial hours of sepsis, and recent data show early vasopressors reduce fluid balances from 8 to 24hours (7), inherently appealing with HF-FO concerns.