Mild hyponatremia

Isotonic fluid significantly decreased the risk of mild hyponatremia at both ≤ 24 h and > 24 h (RR = 0.38, 95% CI [0.30, 0.48], P < 0.00001; RR = 0.47, 95% CI [0.37, 0.62], P < 0.00001, respectively). The pooled results were homogeneous at both points of time (I2 = 24%, P = 0.15; I2 = 0%, P = 0.63, respectively) (Fig. 2). Our subgroup analysis at 2, 8, 12, 18, 24, 36, 48, and 72 h showed that isotonic fluid decreased the risk of mild hyponatremia at different points of time; the lowest risk was at 2 h followed by 18 h and 36 h (RR = 0.07, 95% CI [0.00, 1.19], P = 0.07; RR = 0.21, 95% CI [0.07, 0.59], P = 0.003; RR = 0.21, 95% CI [0.07, 0.67], P = 0.008, respectively). Our subgroup analysis on isotonic versus moderately hypotonic fluid and isotonic versus severe hypotonic fluid revealed comparable relative risks of mild hyponatremia (RR = 0.39, 95% CI [0.30, 0.50], P < 0.00001; RR = 0.37, 95% CI [0.27, 0.51], P < 0.00001). This comparable relative risk was also maintained in our subgroup analysis on surgical patients versus medical patients (RR = 0.53, 95% CI [0.36, 0.78], P = 0.001; RR = 0.30, 95% CI [0.22, 0.41], P < 0.00001, respectively), standard rate versus restricted rate (RR = 0.36, 95% CI [0.26, 0.49], P < 0.00001; RR = 0.60, 95% CI [0.34, 1.05], P = 0.07, respectively), neonates versus other children (RR = 0.18, 95% CI [0.06, 0.51], P = 0.001; RR = 0.39, 95% CI [0.33, 0.46], P < 0.00001, respectively), balanced versus 0.9% saline (RR = 0.24, 95% CI [0.15, 0.39], P < 0.00001; RR = 0.41, 95% CI [0.34, 0.49], P < 0.00001, respectively), and open label studies versus blinded studies (RR = 0.40, 95% CI [0.31, 0.52], P < 0.00001; RR = 0.37, 95% CI [0.25, 0.54], P < 0.00001, respectively). In addition, our subgroup analysis based on different regions of the included studies showed that isotonic saline significantly decreased the risk of mild hyponatremia in studies conducted in Asia, Australia and Oceania, and Europe (RR = 0.35, 95% CI [0.26, 0.46], P < 0.00001; RR = 0.39, 95% CI [0.24, 0.65], P = 0.0002; RR = 0.41, 95% CI [0.26, 0.64], P = 0.0001, respectively). However, there was no significant difference between the two fluids in terms of mild hyponatremia in studies conducted in both North and South America (RR = 0.39, 95% CI [0.14, 1.10], P = 0.07; RR = 0.53, 95% CI [0.23, 1.19], P = 0.12, respectively) (Supplementary Fig. 2).

Fig. 2

Pooled results for mild hyponatremia at ≤ 24 and > 24 h

Moderate hyponatremia

Isotonic fluids significantly decreased the risk of moderate hyponatremia at ≤ 24 and > 24 h, compared to hypotonic fluids (RR = 0.40, 95% CI [0.25, 0.65], P = 0.0002; RR = 0.40, 95% CI [0.20, 0.79], P = 0.008, respectively). The pooled results were homogeneous at both time points (I2 = 0%, P = 0.89; I2 = 0%, P = 0.4, respectively) (Fig. 3).

Fig. 3

Pooled results for moderate hyponatremia at ≤ 24 and > 24 h

Our subgroup analysis at 24, 36, and 72 h showed that isotonic fluid reduced the risk of moderate hyponatremia, with the lowest risk observed at 72 h (RR = 0.17, 95% CI [0.03, 0.88], P = 0.03) (Supplementary Fig. 3A).

Severe hyponatremia

Isotonic fluids significantly reduced the risk of severe hyponatremia at > 24 h, compared to hypotonic fluids (RR = 0.22, 95% CI [0.06, 0.85], P = 0.03). The pooled results were homogeneous at all-time points (I2 = 0%, P = 0.7). However, there was no significant difference between the two fluids at ≤ 24 h (RR = 0.32, 95% CI [0.06, 1.54], P = 0.15). The pooled results were homogeneous at all time points (I2 = 0%, P = 0.81) (Fig. 4).

Fig. 4

Pooled results for severe hyponatremia at ≤ 24 and > 24 h

Our subgroup analysis at 48 h showed that isotonic fluid reduced the risk of severe hyponatremia (RR = 0.18, 95% CI [0.04, 0.78], P = 0.02) (Supplementary Fig. 3B).

Hypernatremia

Isotonic fluid significantly increased the risk of hypernatremia at ≤ 24 h compared to hypotonic fluid (RR = 2.44, 95% CI [1.34, 4.45], P = 0.003). The pooled results were homogeneous (I2 = 0%, P = 0.68). However, there was no significant difference between the two groups at > 24 h (RR = 1.42, 95% CI [0.56, 3.59], P = 0.47). The pooled results were heterogeneous (I2 = 13%, P = 0.33) (Supplementary Fig. 4). Interestingly, our sensitivity analysis excluding trials conducted on neonates showed that the risk of hypernatremia became insignificant at both ≤ 24 and > 24 h (RR = 2.03, 95% CI [0.97, 4.25], P = 0.06; RR = 1.20, 95% CI [0.47, 3.09], P = 0.71, respectively) (Fig. 5). Furthermore, our subgroup analysis based on the age category showed that isotonic fluid significantly increased the risk of hypernatremia in studies conducted on neonates (RR = 3.74, 95% CI [1.42, 9.85], P = 0.008), while there was no significant difference between the two fluids in studies conducted on other children (RR = 1.58, 95% CI [0.90, 2.76], P = 0.11) (Fig. 6).

Fig. 5

Pooled results for hypernatremia at ≤ 24 and > 24 h. Sensitivity analysis excluding trials conducted on neonates

Fig. 6

Pooled results for hypernatremia with subgrouping based on the age group

Serum sodium

The hypotonic group had significantly lower serum sodium level compared to the isotonic one at ≤ 24 h (MD = –2.36 95% CI [–2.84, –1.88], P < 0.00001). The pooled results were heterogeneous (I2 = 59%, P < 0.0001). However, there was no significant difference between the two groups at > 24 h (MD = –0,92 95% CI [–1.87, 0.03], P = 0.06). The pooled results were heterogeneous (I2 = 68%, P = 0.0009) (Supplementary Fig. 5A).

Serum osmolarity

The hypotonic group had significantly lower serum osmolarity compared to the isotonic group at ≤ 24 h (MD = –4.85, 95% CI [–6.95, –2.74], P < 0.00001). The pooled results were homogeneous (I2 = 12%, P = 0.33). However, there was no significant difference between the two groups at > 24 h (MD = –8.20, 95% CI [–17.91, 1.52], P = 0.10). The pooled results were heterogeneous (I2 = 90%, P < 0.0001) (Supplementary Fig. 5B).

Serum chloride

There was a significant difference between the hypotonic and isotonic groups (favoring the hypotonic group) in terms of serum chloride level at ≤ 24 h (MD = –1.68, 95% CI [–2.94, –0.42], P = 0.009). The pooled results were heterogeneous (I2 = 64%, P = 0.004). However, there was no significant difference between the two groups at > 24 h (MD = –0.66, 95% CI [–3.23, 1.90], P = 0.61). The pooled results were heterogeneous (I2 = 65%, P = 0.06) (Supplementary Fig. 5C).

Serum potassium

There was no significant difference between the isotonic and hypotonic groups in terms of serum potassium level at both the time points, ≤ 24 and > 24 h (MD = 0.00, 95% CI [–0.18, 0.18], P = 0.94; MD = 0.01, 95% CI [–0.20, 0.22], P = 0.95, respectively). The pooled results were heterogeneous at ≤ 24 and > 24 h (I2 = 66%, P = 0.001; I2 = 74%, P = 0.004, respectively) (Supplementary Fig. 5D).

Serum creatinine

Serum creatinine level was significantly higher in the isotonic group at ≤ 24 h (MD = 0.89, 95% CI [0.84, 0.94], P < 0.00001). However, there was no significant difference between the isotonic and hypotonic groups in terms of serum creatinine level at > 24 h (MD = 0.85, 95% CI [–0.02, 1.71], P = 0.05). The pooled results were homogeneous  at both ≤ 24 and > 24 h (I2 = 0%, P = 0.79; I2 = 7%, P = 0.36) (Supplementary Fig. 5E). Interestingly, our subgroup analysis based on the composition of isotonic fluids revealed that 0.9% saline was associated with significant increase in serum creatinine levels (MD = 0.90, 95% CI [0.84, 0.96], P < 0.00001), while there was no significant difference between isotonic and hypotonic groups in studies which used balanced isotonic solutions (MD = 0.99, 95% CI [–1.91, 3.90], P = 0.50) (Supplementary Fig. 5F).

Blood pH

Isotonic fluid had significantly lower blood pH than hypotonic fluid (MD = –0.05, 95% CI [–0.08 to –0.02], P = 0.0006). The pooled results were heterogeneous (I2 = 68%, P = 0.04) (Supplementary Fig. 6A).

Blood sugar

There was no significant difference between isotonic and hypotonic fluids in terms of blood sugar (MD = 3.06, 95% CI [–0.45, 6.56], P = 0.09). The pooled results were homogeneous (I2 = 0%, P = 0.78) (Supplementary Fig. 6B).

Urinary sodium

The hypotonic group had significantly lower urinary sodium than the isotonic group (MD = –37.07, 95% CI [–47.53, –26.61], P < 0.00001). The pooled results were homogeneous (I2 = 0%, P = 0.43) (Supplementary Fig. 6C).

Length of hospital stay

There was no significant difference between isotonic and hypotonic fluids in terms of length of hospital stay (MD = –0.07, 95% CI [–0.66, 0.51], P = 0.8). The pooled results were homogeneous at all time points (I2 = 39%, P = 0.14) (Supplementary Fig. 6D).