Chloride administration in the intensive care unit, an independent predictor of mortality

Introduction

Chloride is the body's most important extracellular anion, important in many physiologic processes including acid base balance and osmotic pressure [1]. Hyperchloremic metabolic acidosis is a common finding in critically ill patients and may be associated with renal failure and even increased mortality [2].

Objectives

The objective of our study was to determine if chloride administration in a critically ill population is associated with hospital mortality.

Methods

We performed a retrospective observational study in a 24 bed tertiary mixed medical surgical ICU in a teaching hospital in the Netherlands. Patients admitted to the ICU in the period between January 1th 2008 and November 1th 2014 were screened for eligibility. Inclusion criteria were: 1. ≥ 18 years. 2. Length of stay (LOS) of ≥ 72 hours. Readmissions to the ICU were excluded. The primary end points for our analysis was hospital mortality. Univariate analysis was performed with Wilcoxon rank sum test for nonparametric data. Multivariate analysis was performed with predictors of ICU and hospital mortality (age, Acute Physiology and Chronic Health Evaluation IV predicted mortality) as well as factors associated with (hyper)chloremia, metabolic acidosis and fluid resuscitation. Chloride administration was defined as the total amount of chloride in mmol/l administered to a patient in a given time period.

Results

We studied 1471 patients. Mean age at admission was 67 years, APACHE II score 24, the average stay 10,5 days (table 1, baseline characteristics). In univariate analysis chloride levels at 72 hours were predictors of hospital mortality as were pH, chloride administration and total fluid administration (table 2). A multivariate analysis was performed to determine whether chloride, chloride administration, fluid administration and pH are independent predictors of mortality. This models shows that chloride levels are not an independent predictor of hospital mortality, chloride administration however is an independent predictor (OR 1.005; 95% CI 1.003-1.007 at 72 hours) as is pH, total fluid administration and age (table 3).

Table 1 Baseline characteristics.

Table 2 Univariate analysis

Table 3 Multivariate analysis

Conclusions

In conclusion, chloride administration in a critically ill population is an independent risk factor for mortality, even when corrected for total fluid administration, pH and chloride levels. Further studies the optimal resuscitation fluid with regard to chloride levels are needed.