Routine maintenance

Surveillance proposal

The section of the guideline on routine maintenance should not be updated.

Rate of administration for routine maintenance

2019 surveillance summary

A Cochrane review (Maconochie at al. 2016) of 3 RCTs (n=420 children) compared maintenance and restricted volumes of initial fluid administration (up to 72 hours after first presentation) for acute bacterial meningitis. The maintenance and restricted volumes compared by the 3 studies were:

  • Intravenous maintenance fluids (100 ml/kg/day for the first 10 kg of body weight, plus 50 ml/kg/day for the next 10 kg, plus 20 ml/kg/day for subsequent weight) plus replacement fluids for any estimated deficit versus two thirds of maintenance fluids.

  • Intravenous maintenance fluids (110 ml/kg/day for the first 10 kg of body weight, 50 ml/kg/day for the next 10 kg and 25 ml/kg/day for subsequent weight) versus 65% of maintenance fluids.

  • Intravenous maintenance fluids (100 ml/kg/day for the first 10 kg of body weight, 50 ml/kg/day for the next 10 kg, and 20 ml/kg for subsequent weight) versus 60% of maintenance fluid equivalent as nasogastric milk-based fluids.

The review permitted evidence from adults and children but found only paediatric studies. The largest of the 3 trials was from settings with high mortality rates. There was no significant difference between maintenance and restricted fluid for the 2 primary outcomes: number of deaths (n=407 children; low-quality evidence) or acute severe neurological sequelae (n=407 children; low-quality evidence). However, when neurological sequelae were defined further, there was a significant benefit of maintenance fluid for spasticity (n=357 children); and seizures at both 72 hours (n=357 children) and 14 days (n=357 children). Maintenance fluid had significant benefits over restrictive fluid for chronic severe neurological sequelae at 3 months follow-up (n=351 children; very low-quality evidence).

An RCT (Ashok et al. 2017) (n=150 children) compared restricted (10 ml/kg/h) with liberal (30 ml/kg/h) intraoperative maintenance infusion of Ringer's lactate solution in children undergoing lower abdominal and penile surgery under general anaesthesia. All patients also received a caudal block and intravenous paracetamol for analgesia. No opioids and muscle relaxants were used. Incidence of the primary outcome of nausea and vomiting during 24 hours postoperatively was significantly less in the liberal than the restricted group. The incidence of oral fluid intake during the first 6 post-operative hours was significantly higher in the restricted group, with significantly more children complaining of thirst than in the liberal group. Parent satisfaction was significantly higher in the liberal than the restricted group. No children had complications attributed to the liberal fluid therapy.

Intelligence gathering

No topic expert feedback was relevant to this section of the guideline.

Impact statement

The Cochrane review found no difference in mortality or severe neurological sequelae between maintenance and restricted fluid in acute bacterial meningitis. The maintenance rates examined by the included studies followed exactly or very closely those recommended in the guideline, therefore the lack of difference found by the review between maintenance and restricted fluids is unlikely to affect recommendation 1.4.1 to calculate routine maintenance fluid rates using the Holliday–Segar formula.

A trial found 30 ml/kg/h more effective than 10 ml/kg/h Ringer's lactate in reducing nausea and vomiting during fluid maintenance after lower abdominal surgery. However, this outcome was not deemed to be critical or important by the guideline committee, and additionally the population is very specific and may not be generalisable to NG29 in which condition-specific recommendations are out of scope. The evidence is unlikely to affect the guideline which does not currently recommend either of the fluid rates examined in the study.

New evidence is unlikely to change guideline recommendations.

Fluid type for routine maintenance

2019 surveillance summary

Balanced crystalloids versus isotonic (0.9%) or hypotonic (0.45%) sodium chloride

An RCT (McNab et al. 2015) in Australia (n=690 children) examined Plasma-Lyte 148 plus 5% glucose versus 0.45% sodium chloride plus 5% glucose for patients in hospital needing intravenous maintenance hydration for 6 hours or longer. Patients received fluid for 72 hours or until their intravenous fluid rate decreased to lower than 50% of the standard maintenance rate. Significantly fewer patients experienced the primary outcome of hyponatraemia (serum sodium concentration <135 mmol/litre with a decrease of at least 3 mmol/litre from baseline) with Plasma-Lyte 148 plus 5% glucose versus 0.45% sodium chloride plus 5% glucose. No cerebral oedema occurred in either group. Two serious adverse events potentially related to intravenous fluid and 1 seizure occurred with isotonic fluid, versus none and 7 respectively with hypotonic fluid.

An RCT (Lima et al. 2019) (n=53 children) examined a balanced crystalloid (unspecified in the abstract) versus 0.9% sodium chloride for intravenous fluid administration during and for 24 hours after brain tumour resection. The absolute difference from the beginning of surgery to after surgery in both plasma chloride concentrations (the primary outcome), and base excess, significantly increased with 0.9% sodium chloride infusion compared with balanced crystalloids. Hyperchloraemic acidosis incidence was also significantly greater with sodium chloride, though brain relaxation score (a measure of brain oedema) was comparable between groups.

Isotonic (0.9%) versus hypotonic (<0.9%) sodium chloride

Isotonic versus hypotonic maintenance fluids were assessed in 11 paediatric RCTs (Pemde et al. 2015, Flores Robles et al. 2016, Torres et al. 2019, Bagri et al. 2019, Kumar et al. 2019, Friedman et al. 2015, Ramanathan et al. 2016, Raksha et al. 2017, Shamim et al. 2014, Valadao et al. 2015, Omoifo et al. 2018) (table 1 in the data tables appendix). Six trials compared 0.9% with 0.45% sodium chloride, 6 trials compared 0.9% with 0.18% sodium chloride, and 1 trial compared 0.9% with 0.3% sodium chloride. Almost all intervention and comparator solutions additionally contained dextrose (frequently 5%).

Results for the following outcomes were observed across the trials:

  • Hyponatraemia: significantly lower with 0.9% sodium chloride than with a hypotonic solution in 8 of 12 comparisons across 9 trials (Pemde et al. 2015; Flores Robles et al. 2016; Torres et al. 2019; Bagri et al. 2019; Kumar et al. 2019; Ramanathan et al. 2016; Raksha et al. 2017; Shamim et al. 2014; Omoifo et al. 2018).

  • Hypernatraemia: no difference between groups in 2 comparisons (Torres et al. 2019; Kumar et al. 2019).

  • Plasma sodium level: significantly higher with 0.9% sodium chloride than with a hypotonic solution in 4 of 6 comparisons across 5 trials (Flores Robles et al. 2016; Bagri et al. 2019; Friedman et al. 2015; Ramanathan et al. 2016; Valadao et al. 2015). Though in the 3 comparisons quoting numeric values in the abstract, this did not amount to hypernatraemia.

  • Other adverse events: no difference between groups in 3 comparisons across 2 trials (Flores Robles et al. 2016; Torres et al. 2019).

3.3% versus 5% dextrose

An RCT (Martinez Carapeto et al. 2018) (n=130 children) compared 0.3% sodium chloride in 3.3% dextrose with 0.9% sodium chloride in 5% dextrose as maintenance intravenous fluid in critical care after elective general surgery. There was no significant difference between groups in the incidence of hyperglycaemia at 8, 24 or 48 hours.

Intelligence gathering

Topic experts queried whether balanced crystalloids were superior to 0.9% sodium chloride, and noted they were becoming more common practice.

Impact statement

Balanced crystalloids versus isotonic (0.9%) or hypotonic (0.45%) sodium chloride

A study found a lower risk of hyponatraemia during routine maintenance with Plasma-Lyte 148 plus 5% glucose than 0.45% sodium chloride plus 5% glucose without an increase in adverse effects. This evidence is unlikely to affect recommendation 1.4.3 to use isotonic crystalloids that contain sodium in the range 131–154 mmol/litre (of which Plasma-Lyte is an example). The guideline does not specify a particular isotonic fluid, however decisions on the potential superiority of Plasma-Lyte in reducing hyponatraemia would ideally be based on a comparison of Plasma-Lyte (an isotonic fluid) with another isotonic fluid, not on a comparison with a hypotonic solution as in this study.

A second trial did compare a balanced crystalloid (unspecified) with 0.9% sodium chloride for maintenance and found that it reduced both variation in plasma chloride and hyperchloraemic acidosis (an outcome deemed important by the original guideline). However, the evidence is from a single trial in a very specific population (brain tumour resection). Further evidence from other populations to support these findings, including additional outcomes deemed critical or important by the original guideline (such as mortality or hospital stay), would be useful before considering any impact on recommendation 1.4.3 which does not currently specify a particular isotonic fluid.

New evidence is unlikely to change guideline recommendations.

Isotonic (0.9%) versus hypotonic (<0.9%) sodium chloride

The new evidence found that 0.9% sodium chloride for routine maintenance is associated with less hyponatraemia, with no impact on other adverse events. This is consistent with recommendation 1.4.3 to use isotonic crystalloids that contain sodium in the range 131–154 mmol/litre (of which 0.9% sodium chloride is an example) for routine maintenance.

New evidence is unlikely to change guideline recommendations.

5% versus 3.3% dextrose

The new evidence found no difference in hyperglycaemia with 0.3% sodium chloride in 3.3% dextrose versus 0.9% sodium chloride in 5% dextrose for post-operative fluid maintenance (though interpreting the effect of the dextrose concentration is confounded by differing sodium chloride levels in the solutions). This is consistent with recommendation 1.4.3 for fluid type for routine maintenance in children and young people which does not discuss glucose levels, nor (unlike fluid type for resuscitation) does it specify that it should be glucose-free.

New evidence is unlikely to change guideline recommendations.