presentation

Association between the choice of IV crystalloid and in-hospital mortality

among critically ill adults with sepsis Raghunathan K, Shaw A, Nathanson B, et al.

Critical Care Medicine, July 2014

INTRODUCTION Background

• Fluid therapy is used to restore effective circulating blood volume and to promote tissue/organ perfusion. • Guidelines recommend early resuscitation with IV crystalloids as the preferred fluid therapy and against

using hydroxyethyl starches due to increased risk of AKI. • Resuscitation with isotonic saline can induce hyperchloremia and metabolic acidosis and have been

associated with alterations in renal blood flood and effects on immune function. • Balanced fluids avoid biochemical effects and have been associated with reduced perioperative mortality

and ICU morbidity. Current Thoughts

• Crystalloids differ in chloride content and strong ion difference (SID = sodium - chloride conc): - Lactated Ringer’s solution (LR): contains electrolytes, a high SID and physiologic chloride content - Isotonic saline (IS): does not contain electrolytes, has an SID = 0 and supraphysiologic chloride content

STUDY OVERVIEW Study Type • Retrospective cohort study Objective • To examine the association between receipt of balanced fluids vs. nonbalanced fluids during initial

resuscitation and in-hospital mortality, renal morbidity, ICU and hospital lengths of stay in a large cohort of adults admitted with vasopressor-dependent sepsis

Enrollment • 53,448 critically ill adult patients who were admitted to one of at least 360 geographically and structurally diverse hospitals in the US between November 2005 and December 2010

• 6,730 patients studied (3,365 patients in each group) Inclusion Criteria

• Critically ill patients ≥18 years old • Primary or secondary diagnosis of sepsis • Receiving vasopressors and received at least 2L of crystalloids for resuscitation by day 2 • Blood cultures and 3 consecutive days of antibiotic treatment

Exclusion Criteria

• Early adverse outcomes (e.g. ARF) that could dictate fluid choice • ESRD on dialysis PTA • Surgery • Transferred out of the ICU

Outcomes • Primary outcome: in-hospital mortality after hospital day 2 • Secondary outcomes: ARF ± dialysis, hospital and ICU lengths of stay

STUDY DESIGN & METHODS Study Design

• Dose-response relationships were analyzed according to the proportion of balanced fluids received • Controlled for differences severity of illness and patient management (e.g. monitoring procedures,

utilization of diagnostic tests, administration of supportive therapies and pharmacologic treatments) • Adjusted for confounding influence of hospital-specific factors (practice variation) • Conducted secondary sensitivity analyses

Statistical Analysis

• All analyses were performed using Stat/SE 11.2 • Cohorts assembled by application of inclusion/exclusion criteria and evaluated for balance on all measured

covariates • 1:1 Propensity score matching to adjust for baseline differences between the groups • Multilevel mixed logistic regression model to predict the propensity for receipt of balanced fluids by day 2 • Sensitivity analyses using Generalized Estimating Equation models to analyze the incremental effects of

balanced fluids on all outcomes within propensity-matched quintiles of total crystalloid volume and predicted mortality risk

• Univariate comparisons using chi-square or t-tests • Categorical data: frequencies and proportions; Continuous variables: medians and interquartile ranges • P-value <0.05 was significant

RESULTS Population • At baseline prior to propensity matching:

- Patients who received balanced fluids were: younger, received larger crystalloid volumes and were MORE likely to receive colloids, steroids, invasive monitoring and mechanical ventilation; LESS likely to have a diagnosis of heart or chronic renal failure and hypertension or diabetes with complications

- Similar ARF rates, dialysis, ICU and hospital length of stay Outcomes • Primary outcome: Absolute in-hospital mortality beyond day 2: balanced fluids group 19.6% vs. no-

balanced fluids 22.8% (relative risk ratio 0.86; 95%CI, 0.78-0.94) (table 1) • Secondary outcomes: no significant differences between the groups and no significant differences in

dose-response heterogeneity for ARF ± dialysis, hospital and ICU lengths of stay (supplemental data) Additional Analyses

• Median total crystalloid volume during resuscitation: 2.5L to 10.5L

• Predominant fluid type: - No-balanced group: IS - Balanced fluid group: LR (the majority received <40% of

all fluids as balanced solutions) • NNT = 31 • Relative risk of in-hospital mortality was incrementally

LOWER by 3.4% on average per 10% increase in the 0

proportion of balance fluids (figure 2) • Risk of mortality was lowest among those receiving the greatest

proportion of balanced fluids regardless of the total crystalloid volume (figure 3)

• Those receiving larger proportions of balanced fluids were less likely to die within any given volume quintile (supplemental data)

• Mortality odds DECREASED by 7.5% on average per 10% increase in the proportion of balanced fluids received among those at the lowest risk of death (supplemental data)

CONCLUSION Author’s Conclusion

• In-hospital mortality was lower following initial resuscitation with balanced versus non-balanced crystalloids among nonoperative patients admitted with early vasopressor-dependent sepsis

• Mortality was progressively lower among patients receiving greater proportions of balanced crystalloids • IF this association represents a causal relationship, they estimate one less in-hospital death for every 31

patients treated with balanced fluids rather than saline during initial resuscitation in sepsis DISCUSSION

Strengths • Large patient population • Decent external validity • Controlled for confounders and differences at baseline

Limitations • Retrospective, observational study • Based on insurance claims and ICD-9 codes • No laboratory values (e.g. chem 7, ABG, etc.) • Very few patients exclusively receive balanced fluids • Carrier solutions and IV fluids < 500 mL were excluded from totals

CLINICAL APPLICATION Currently, crystalloid fluids are used as the standard resuscitation fluid in sepsis. Many studies have compared the usage of crystalloid fluids versus colloid fluids in sepsis, but there is little data comparing the usage of different types of crystalloid fluids, except in postoperative studies. This study showed a significant difference in mortality between the usage of balanced fluids, such as Lactated Ringer’s, versus non-balanced fluids, like isotonic saline, in patients with sepsis. There are many variables that can occur during the management of sepsis and it is difficult to determine if fluid choice solely has a significant effect on mortality. Furthermore, since this is a retrospective cohort study, there are many limitations. Though the findings of this study are statistically significant, it is not necessarily that clinically significant. This study supports the need for prospective, clinical trials to investigate the effects of “balanced” versus “nonbalanced” fluids.

Lower risk of mortality

↑ proportion of balanced fluids ↓ risk of in-hospital mortality

Fig. 2

Fig. 3