Fluid balance is one of the most challenging aspects of critically-ill children requiring pediatric intensive care. While early and aggressive fluid resuscitation is lifesaving and widely practiced, excess fluid may be detrimental. Mechanically ventilated children may be more prone to fluid overload as they have lesser evaporative losses due to warmed, humidified gases and thermocontrolled environment [1]. Presence of non-osmotic stimuli, in addition to osmotic stimuli, for anti-diuretic hormone (ADH) release may contribute to fluid retention [2,3]. In addition, overestimation of fluid requirement occurs if traditional methods designed primarily for healthy children are used [4].

This prospective observational study was conducted in an eight-bedded PICU of All India Institute of Medical Sciences, New Delhi from July 2013 to March 2015. Children aged >1 month to £15 years requiring mechanical ventilation for at least 24 hours were enrolled and followed-up for a maximum of 14 days and for final outcome during PICU stay. Children with congenital heart disease, acute or chronic renal failure, renal tubular defects, post renal transplant, those receiving peritoneal or hemodialysis, and those with advanced malignancies were excluded. Children who were mechanically ventilated ³24 hours prior to admission in PICU were also excluded due to uncertainty of previous course and fluid received. Only the first admission was considered for every patient, and readmissions were excluded. The study was approved by the Institutional Ethics Committee and written informed consent was obtained from either of the parents or legal guardian.

Fluid overload percent (FO%) was defined as = [Total fluid input in 24 hours (mL) – total fluid output in 24 hours (mL)/weight at admission (g)] × 100. Cumulative fluid balance on any particular day was defined as the sum total of daily fluid balances till that particular day (minimum of 24 hours to maximum of 2 weeks). Significant fluid overload was defined as a cumulative fluid balance of ³15% attained at any time point during the 14 day study period and peak fluid overload was defined as the maximum percentage cumulative fluid overload on any day during the study period. Baseline demographic characteristics including weight, height, body mass index (BMI) and percentage predictive mortality at admission using Pediatric Index of Mortality-2 (PIM2) score [9] were recorded. At admission to PICU, infants were weighed using an infant weighing scale, while older children were weighed using a weighing chair scale. We were able to weigh all the intubated children taking precautions to avoid extubation. Subsequently, daily input was calculated as the sum of the total parenteral [intravenous (IV) fluids, medications, boluses, fluid replacements and blood products] and enteral fluids (oral, nasogastric or orogastric) received. Fluid output was calculated as sum of daily urine output and other body fluid output (nasogastric aspirate/ drain, drain from intercostals, or other surgical drains); these were recorded using the nursing charts on a daily basis. Insensible losses and fecal fluid losses were not included. From these data, daily fluid balance was calculated and cumulative figures generated for the total duration of mechanical ventilation.

Daily documentation of ventilator parameters [respiratory rate, inspiratory time (Ti), peak inspiratory pressure (PIP), positive end expiratory pressure (PEEP), FiO2, mean airway pressure] was done. Ventilatory settings were standardized for all children as per PICU protocol at our institute; these were influenced by the lung conditions and modified based on clinical condition and arterial blood gas reports. Arterial blood gas analysis was performed using ABL Flex (Radiometer Copen-hagen, Denmark) and required parameters (pH, PCO2, PO2) were documented. Also parameters required for determining PELOD score were recorded daily for an objective scoring of organ dysfunction .

Statistical analysis was performed using Stata 11.2 (Stata Corp, College Station, TX, USA). The relationship between the fluid overload and the outcome variables was examined using bivariate (chi-square tests for categorical data and　t-tests for continuous variables) and multivariate analyses (logistic regression) for mortality, controlling for confounders such as age, gender, nutritional status, and severity of illness at admission (PIM2 score).

Results

During the study period, 278 children required mechanical ventilation, out of whom 118 (58.5% boys) were enrolled (Fig. 1). The median (IQR) age of the study participants was 14.5 (4, 72) months. Table I show the baseline characteristics. The most common indications for mechanical ventilation in patient management were septic shock (49.1%), impending respiratory failure secondary to pulmonary causes (24.6%) and CNS causes (11.1%) followed by hepatic encephalopathy (3.4 %). At admission, children had a mean (SD) predicted mortality using PIM2 score of 15.1 (18.7).

Fig. 1 Flow of participants in the study.

 

TABLE I	Baseline Characteristics of Study Children (N=118)

Variables	Values  (n = 118)
Age (mo),  median (IQR)	14.5 (4, 72)
Male gender, n (%)	69 (58.5)
Weight Z score, median (IQR)	-2.65 (-4.85,  -1.42)
Height Z score, median (IQR)	-1.32 (-3.87,  - 0.07)
BMI Z score, median (IQR)	-2.78 (-4.66,  -1.28)
PIM2 score, mean (SD)	15.08 (18.65)
Admission diagnosis, n (%)
Sepsis	58 (49.2)
Pulmonary	29 (24.6)
CNS	13 (11.0)
Hepatic	4 (3.4)
Non pulmonary infections without sepsis	4 (3.4)
Miscellaneous	10 (8.5)

 

Cumulative fluid overload of ³15% was observed in 74 (62.7%) children. Around 50% of children reached cumulative fluid balance of ³15% within the first 5 days of PICU stay (Web Fig. 1). The median cumulative fluid overload percentage (IQR) was 19.6% (10.5%, 30.8%). The median (IQR) cumulative fluid overload at 48 hours was 7.9% (3.1%, 11.8%).

Of the enrolled children 38.1% (45/118) died. Among 74 children who had ³15% cumulative fluid overload, mortality rate was 40.5% (30/74) as compared to 34% (15/44) in those with <15% cumulative fluid overload (P=0.49) (Table II). The median (IQR) peak fluid overload in survivors was 10.2% (3.3%, 17.4%) as compared to 24.6% (13.6%, 35.5%) in non-survivors (P=0.11). This was not significant on multivariate analysis (P=0.50) (Web Table I).

TABLE II Fluid Overload and Mortality in Mechanically Ventilated Children (N=118)

At 48 hours, the median (IQR) peak fluid overload in survivors was 6.6% (3.1%, 10.3%) compared to 9% (6.1%, 12.7%) in non-survivors (P=0.04). However, on multivariate logistic regression analysis, there was no such association after adjusting for confounding factors (P=0.12) (Web Table I). Only the PIM2 score at admission, PELOD score and diagnosis of sepsis showed significant association with mortality (Web Table I).

The maximum OI values were similar in children with ³15% cumulative fluid overload [median (IQR): 10.3 (5.4, 18.0)] as compared to those with <15% cumulative fluid overload [median (IQR): 11.1 (5.9, 21.4)] (P=0.38).

Higher organ dysfunction as calculated using PELOD score was seen in those with ³15% cumulative fluid overload [median (IQR): 21 (11, 31)] as compared to children with <15% cumulative fluid overload [median (IQR): 12 (3.25, 21.75)] (P=0.03). On multivariate analysis, the association between fluid overload and PELOD score remained significant (P=0.02) (Web Table II). Cumulative fluid overload at 48 showed no association with organ dysfunction (P=0.31).

The duration of mechanical ventilation was significantly higher among children with ³15% cumulative fluid overload [median (IQR): 10 (6, 16.2) days] compared to that in children with <15% cumulative fluid overload [median (IQR): 4 (2.25, 8) days (P<0.001)]. On multivariate analysis, association remained significant (P<0.001) after adjusting for confounding factors (Web Table II). Cumulative fluid overload at 48 hours was also associated with longer duration of mechanical ventilation (P=0.001).

The duration of PICU stay was significantly higher among children with ³15% cumulative maximum fluid overload [median (IQR): 13.5 (7.75, 22) days] as compared to that in children with <15% cumulative maximum fluid overload [median (IQR): 6 (4, 10.75) days] on univariate as well as multivariate analysis (P<0.0001 and P<0.0001, respectively) (Web Table II). Cumulative fluid overload at 48 hours was also significantly associated with longer duration of PICU stay (P <0.0001).

In this prospective observational study, we observed that there was no significant association of fluid overload with mortality or peak oxygenation index. However, the total duration of mechanical ventilation and duration of PICU stay was significantly higher in those with fluid overload ³15%. These children also had significant organ dysfunction as calculated using PELOD score.

Unlike previous studies in children, we carried out a prospective study with a longer duration of follow up for outcomes. Also majority of previous studies were conducted in a restricted group of patients with acute lung injury [6,10] acute kidney injury [11,12] children on CRRT and hemodialysis [13-15]. We were able to study children commonly seen in a medical PICU with varied conditions to demonstrate the actual incidence and severity of fluid overload. We did not observe a relation between fluid overload and mortality which is similar to what has been reported previously [16,17]. However, few studies have shown association of fluid overload with mortality [18,19]. Previous studies have demonstrated significant effect of fluid overload on lung function and weaning and thereby the oxygenation index, duration of mechanical ventilation and ICU stay in both children and adults. Studies by Arikan, et al. [17] and Sinitsky, et al. [16] showed that peak fluid overload percent and fluid overload at 48 hours, respectively were associated with higher peak oxygenation index. Although we did not observe similar association of peak cumulative fluid overload percent of ³15% with OI, the duration of ventilation as well as the ICU stay was significantly longer in this group. These findings are also similar to previous studies [18,19].

Higher PELOD scores were observed in those with cumulative peak fluid overload of ³15%. However, it did not translate to a difference in mortality between the groups as other factors like increased severity of illness at presentation in both groups might have ultimately influenced the outcomes.