ASSOCIATION OF FLUID MANAGEMENT WITH outcome in COVID–19 ARDS

Sanchit Ahuja, Frederique Paulus, Harm-Jan de Grooth, Fleur L. van der Ven, Ary Serpa Neto, Marcus J. Schultz, Pieter R. Tuinman for the PRoVENT–COVID Collaborative group

Introduction

Considerable evidence from observational studies, clinical trials and systemic reviews indicates the potential benefits of restricting fluid administration in critically ill patients [1, 2]. Previous work suggests that a higher cumulative fluid balance is independently associated with mortality, longer duration of ventilation and longer duration of intensive care unit (ICU) stay [3]. In contrast, there is also the suggestion of higher than expected occurrence of extrapulmonary organ hypoperfusion, such as acute kidney injury requiring renal replacement therapy (RRT) in patients who receive a restrictive fluid strategy [4].

Based on indirect evidences, during the early stages of the coronavirus disease 2019 (COVID–19) pandemic, consensus guidelines on the management of shock in COVID–19 patients recommended to target a neutral fluid balance strategy [5]. While a predominant subset of critically ill COVID–19 patients reported to have acute respiratory failure, the wide spectrum of disease also comprises of gastrointestinal symptoms and refractory heart failure with features of cardiogenic shock [6, 7]. These pose a special dilemma in the fluid management of critically ill COVID–19 patients, including those who need invasive ventilation. Current evidence is insufficient and constantly evolving on how to best address the optimal fluid management strategy in invasively ventilated COVID–19 patients. We, therefore sought to determine the association of cumulative fluid balance with weaning outcomes in these patients. We hypothesized that the cumulative fluid balance is often high, and that this is associated with the successful liberation of ventilation in COVID–19 ARDS patients.

Methods

Design

This is a posthoc analysis of the PRoVENT–COVID study, a multicenter, retrospective, observational cohort study in COVID–19 patients with respiratory failure requiring invasive ventilation in 22 hospitals in the Netherlands in the first 3 months of the national outbreak. PRoVENT–COVID is an investigator–initiated study, the study protocol of which was approved by local institutional Review Board of Amsterdam UMC (location ‘AMC’) and registered at clinicaltrials.gov (study identifier NCT04346342).

Patients

Consecutive invasively ventilated COVID–19 patients admitted to one of the participating ICUs were eligible for participation, if aged ≥ 18 years and when COVID–19 was confirmed by RT–PCR. For this posthoc analysis, we excluded patients who received ventilation for less than 4 calendar days, and patients who were transferred from or to another ICU that did not participate in the PRoVENT–COVID study within the first 4 days of ventilation.

Exposure

The exposure of interest is the cumulative fluid balance at successive days, calculated by total fluid input minus total fluid output on a certain day of ICU admission

Endpoints

The primary outcome is the successful liberation of ventilation at day-28, defined as the timepoint at which a patient was alive and extubated. Secondary outcomes are clinical endpoints like duration of ventilation in survivors and non–survivors, ICU and hospital length of stay, and ICU –, and 28–day and 90–day mortality. Other endpoints are acute kidney injury, according to a modified KDIGO definition, as the current measured serum creatinine from an individual patient against the baseline creatinine value defined as either the lowest in the last 7 days or a median of values from the preceding 8 to 365 days depending on availability of previous results [8], use of renal replacement therapy, and vasopressor use.

Analysis plan

Descriptive statistics will be used to describe study population and fluid management parameters. Data are presented as numbers and percentages for categorical variables and as means (+/- SD) or median and interquartile range [IQR] according to distribution. Where appropriate, statistical uncertainty will be expressed by 95% confidence intervals.

We will first examine the crude association between cumulative fluid balance and successful liberation of ventilation at day 28 using a mixed-effects model with successful liberation of ventilation at day-28 as dependent variable, fluid balance as (fixed effect) independent variable and hospital as a random intercept effect. To examine potential nonlinearity in the association, fluid balance will be entered as a restricted cubic spline function with 3 knots distributed equally along the density. The complexity of the spline function will be reduced in a stepwise fashion until minimization of the Akaike Information Criterion. Depending on the shape of the final spline, exposure will possibly be broken into quantiles to facilitate interpretation. The association will then be adjusted for possible confounding variables on the day of intubation as (fixed effects) covariates in the model. The set of predefined adjustment variables measured on the day of intubation are: sex, age, body mass index, serum creatinine, norepinephrine dose, tidal volume, arterial pH, PEEP, PaO2 /FIO2, dynamic respiratory system compliance and arterial lactate. Multivariable modeling of the association of cumulative fluid balance on calendar day 4 with successful liberation of ventilation will be performed using risk adjusted hazards ratio will be calculated at X time intervals. We will use Kaplan–Meier analysis to test the correlation between cumulative day fluid balance and time to freedom from invasive ventilation, as defined above. Multilevel modeling approach will be performed to quantify between-hospital differences and adjusted for the clustered nature of the data within the hospitals.

We will assess the robustness of findings by sensitivity analysis. Restricted cubic splines may provide a way to represent non–linear relationship for cumulative fluid balance as a continuous variable over the first 4 calendar days of invasive ventilation and days free from invasive ventilation till day 28 as outcomes, both unadjusted and adjusted for the severity of ARDS. The restricted cubic spline will have 3 equally distributed knots.

All statistical analyses were performed using SPSS version and R. We considered statistical significance to be at p 0.05.

References

1. Silversides, J.A., et al., Conservative fluid management or deresuscitation for patients with sepsis or acute respiratory distress syndrome following the resuscitation phase of critical illness: a systematic review and meta-analysis. Intensive Care Med, 2017. 43(2): p. 155-170.

2. Wiedemann, H.P., et al., Comparison of two fluid-management strategies in acute lung injury. N Engl J Med, 2006. 354(24): p. 2564-75.

3. van Mourik, N., et al., Cumulative fluid balance predicts mortality and increases time on mechanical ventilation in ARDS patients: An observational cohort study. PloS one, 2019. 14(10): p. e0224563-e0224563.

4. Finfer, S., J. Myburgh, and R. Bellomo, Intravenous fluid therapy in critically ill adults. Nature Reviews Nephrology, 2018. 14(9): p. 541-557.

5. Alhazzani, W., et al., Surviving Sepsis Campaign: guidelines on the management of critically ill adults with Coronavirus Disease 2019 (COVID-19). Intensive care medicine, 2020. 46(5): p. 854-887.

6. Madjid, M., et al., Potential Effects of Coronaviruses on the Cardiovascular System: A Review. JAMA Cardiology, 2020. 5(7): p. 831-840.

7. Cha, M.H., M. Regueiro, and D.S. Sandhu, Gastrointestinal and hepatic manifestations of COVID-19: A comprehensive review. World journal of gastroenterology, 2020. 26(19): p. 2323-2332.

8. Selby, N.M., R. Hill, and R.J. Fluck, Standardizing the Early Identification of Acute Kidney Injury: The NHS England National Patient Safety Alert. Nephron, 2015. 131(2): p. 113-7.