FormalPara Take-home message

A systematic review was conducted to produce guidelines on intravenous maintenance fluid therapy in acutely and critically ill children. Sixteen recommendations were produced, which suggest favouring isotonic balanced glucose-containing fluids designed for children, and to infuse them in lower amounts than Holliday and Segar’s formula.

Introduction

Intravenous maintenance fluid therapy (IV-MFT) has been defined as the water and electrolyte prescription designed to replace anticipated physiologic water and electrolyte losses over the ensuing 24-h period [1]. Maintenance fluid therapy is provided by enteral hydration and/or IV-MFT which comprises both specific IV-MFT prescription, but also additional fluids administered as vectors of various treatments, blood products, or line flush infusions. IV-MFT should be differentiated from fluid boluses and resuscitation fluids that are administered to correct a relative or true fluid deficit, or from replacement fluids that are administered to correct abnormal fluid losses, even if this might be challenging as they are often administered simultaneously. IV-MFT is a standard of care for many hospitalized children in both acute and critical care settings (i.e. emergency department, pediatric wards, surgical wards or intermediate (high dependency) care units and pediatric intensive care units (PICU)). Despite this, maintenance fluid prescribing practices vary considerably [2] and guidelines are scarce [3]. Historically, IV-MFT prescriptions (fluid volume calculation formulas and solution composition) were based on Holliday and Segar’s work [4]. Since then, the use of these guidelines has been associated with potentially severe complications such as hyponatremia, fluid overload and hyperchloremic acidosis, due to inappropriate fluid composition and/or infusion rates/volumes [3, 5,6,7]. Several established practices are currently questioned. First, the ideal fluid tonicity, i.e. the osmotic pressure gradient of the fluid which is determined by the concentration of osmoactive or non-penetrating solutes, mainly sodium [8]. Second, the chloride content and balanced nature of the fluid (i.e. balanced or buffered solutions are produced after the replacement of part of the chloride anions by organic anions to align to the plasma chloride levels, which may impact on acid–base equilibrium). Finally, the volume of fluid administered, especially in situations where patients are at risk of increased anti-diuretic hormone (ADH) secretion and free water excretion impairment [7]. Over the past decade, several pediatric societies have developed guidelines and recommendations surrounding IV-MFT, but their impact has been limited by a lack of dissemination, a lack of recent systematic reviews and because some important questions remained unanswered [3, 9, 10]. In 2020, the Metabolism, Endocrine and Nutrition (MEN) section of the European Society of Pediatric and Neonatal Intensive care (ESPNIC) formed a working group to develop European guidelines on IV-MFT. These guidelines aim to provide evidence-based recommendations around the indications for IV-MFT, the tonicity of IV-MFT, the electrolyte or glucose content and the volume of IV-MFT administered to children from term to 18 years of age.

Method

To generate these evidence-based guidelines, we followed the Scottish Intercollegiate Guidelines Network (SIGN) 50 methodology [11, 12] and results are presented in line with the EQUATOR PRISMA checklist for systematic reviews [13] and the AGREE guideline reporting checklist [14]. The systematic review was registered on PROSPERO on December 15, 2020, when the research protocol was finalized by the working group (CRD42020218847) [15].

Selection of members

In September 2020, under ESPNIC, three project leaders (DB, LT and FVV) formed a working group of members within the MEN section. This group comprised medical doctors, nurses, pharmacists and dieticians, and was created following a call for interested candidates. Selection by the project leaders was based on expertise in the methods and experience in the field of pediatric acute and intensive care, metabolism, and research. This work group also included an academic librarian specialized in systematic reviews, a systematic review methodologist, an epidemiologist and a biostatistician specialized in meta-analyses. There was no industry input into the guidelines development. No member of the working group received honoraria for any role in the process and all members have declared any potential conflicts of interest.

Question development

The content and wording of the clinical questions was agreed at the first online meeting. All questions were structured in the Population, Intervention, Control, and Outcome(s) (PICO) format as follows:

Population: acute and critically ill children aged term to 18 years. Critically ill children are those presenting with severe organ failure(s) or requiring pediatric intensive care admission; acutely ill children are those presenting with an acute non-critical condition and requiring in hospital care (e.g. admitted to the emergency department, to pediatric wards, intermediate or high dependency units, or post-surgery). Preterm babies (< 37 weeks gestational age) and the intra-operative setting were outside the scope of the guidelines.

The agreed questions were as follows:

PICO1—Indication: Does IV-MFT versus other hydration therapies (none, oral or enteral route) impact on clinical outcomes?

PICO2—Tonicity: Do isotonic solutions versus hypotonic solutions (as IV-MFT) impact on clinical outcomes?

PICO3—Balanced fluids: Do balanced solutions versus non-balanced solutions (as IV-MFT) impact on clinical outcomes?

PICO4—Composition: Does the composition of IV-MFT in terms of glucose, electrolytes (P, Mg, Ca, K), vitamins and trace elements impact on clinical outcomes?

PICO5—Amounts: Does the use of a restrictive IV-MFT volume versus the standard Holliday and Segar calculated volume impact on clinical outcomes?

Outcomes: Standard outcomes (i.e. mortality and length of hospital or PICU stay) were selected. PICO specific outcomes were further identified: hypo or hyper -natremia for PICO 2; hyper-chloremia and acidosis for PICO 3; hypo or hyper -glycemia, -kalemia, -phosphoremia and -magnesemia for PICO 4; fluid balance for PICO 5. However, the preliminary search revealed a large inconsistency in outcome reporting among studies and among PICOs, with discrepancy between outcome definition, presentation, and time of assessment. Consequently, all reported outcomes in reviewed studies were extracted and later included in meta-analysis whenever possible.

Each member of the working group was allocated to a sub-group dedicated to one of the five PICO questions. Allocation of the members to these subgroups was decided by the project leaders according to each member’s preferences and expertise, ensuring a balanced distribution by countries, between professional disciplines, between junior and senior clinicians and between researchers. In each of the five subgroups, group leaders were allocated based on their methodological experience in conducting previous systematic reviews and guideline development.

Literature search and inclusion criteria

After an initial scoping search, each group identified MESH terms for their PICO questions. Subsequently, the medical librarian undertook the literature search on five databases (Pubmed/Medline; Web of Science; Scopus; Cochrane; Embase) for each PICO question. The inclusion criteria for papers were as follows: (1) all papers published until November 2020; (2) written in English, French, Spanish or German with an English abstract; (3) inclusion of critically ill or acutely ill (in the hospital setting) children aged from 37 weeks’ gestational age (GA) to 18 years of age; (4) study designs were randomized controlled trials, cohort studies, before and after studies and case/control studies. To ensure an exhaustive search, literature reviews and editorials were sought and their reference lists checked, but they were not included in the data extraction. Animal studies, case studies, conference abstracts and letters were excluded. The search equations for each PICO question are presented in Supplementary Material 1.

After the removal of duplicates, the librarian uploaded the paper abstract for each PICO question, into the review online software Rayyan [Rayyan Systems Inc. Cambridge, Mass, USA], which access was shared with all the members of the working group.

Selection of relevant studies

Using Rayyan software, each PICO group was responsible for screening for relevance, by title and abstract, the articles for their PICO search, as per the inclusion criteria. At least, two group members, blinded to each other, performed this screening, to reduce subjectivity. In the case of disagreement, differences in decision were resolved via a discussion. If it could not be resolved, a third member was involved. Once all relevant papers were agreed, the full text papers were retrieved by the librarian and shared with each PICO group. A similar screening process was applied to full text manuscripts (double blind screening,) based on inclusion/exclusion criteria, until a final list of papers was agreed for data extraction.

Data extraction and assessment of methodological quality

Once papers were agreed for inclusion for each PICO, papers were independently analyzed by two group members (excluding any authors of the paper). Each member extracted key data and summarized the main findings in a standardized data extraction form, according to the study design. Concurrently the same two reviewers assessed the risk of bias according to the Cochrane risk-of-bias tool for randomized trials [16] and to the Newcastle–Ottawa Quality Assessment Form for cohort studies [17]. In case of disagreement, differences in assessment were resolved via discussion, involving the group leader(s) if required.

Data analysis

When appropriate, data were combined statistically in a meta-analysis. To be combined the data had to meet the following criteria: (1) more than one study; (2) the combined studies were of one study design either randomized trials or observational studies; (3) the population and the intervention were sufficiently similar; (4) the outcomes were the same, or for continuous outcome variables, data on the distribution of the variable was available; (5) the risk of bias was not considered critical according to the SIGN grading system. If two or more groups of patients in a same study were independent, we analyzed each sample as an independent study. Two biostatisticians (AB, JJP) conducted the meta-analyses, but did not participate in development of the recommendations or the voting process.

To compare experimental and control groups of randomised controlled trials (RCTs), we calculated the effect sizes and their 95% confidence intervals using the mean difference for quantitative endpoints and odds ratio for qualitative endpoints, weighted by the inverse of the variance. The analyses were performed using a random effects model. The heterogeneity of outcomes was determined using chi-squared and Higgins I2 tests: a p-value < 0.05 or an I2 ≥ 50% indicated significant heterogeneity. If the heterogeneity was significant and the conditions for the validity of the analysis were verified, a sensitivity analysis was performed to assess the robustness of the results, excluding the most influential studies. The publication bias was explored for meta-analyses with four or more included studies using the funnel plots.

A p value less than 0.05 was considered significant; all p values were two-tailed. Meta-analyses were performed using Review Manager software (RevMan, Version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014).

Consensus methodology, grading of the recommendations and voting rounds

Two online consensus meetings took place in October and November 2021. At these meetings, each group presented their conclusions based on the results from the systematic review and the meta-analyses. A first draft of recommendations was proposed to the whole study team. The wording of the recommendations was discussed with the study team and reworded if required. The classification of the grades of recommendation (A–D, Good Clinical Practice) was undertaken according to the SIGN grading system [11] (supplemental material 2). After those meetings, the recommendations, the supporting rationales, and the recording of the meetings were made available online to the whole team before the online voting. Every member of the study team attended the meetings or watched the meeting recordings and read the proposed rational before voting.

In December 2021, all the PICO recommendations were combined in an online survey (survey Monkey, Inc. (Palo Alto, CA)) for the members to vote, the intention being to gain consensus using a modified Delphi method [18]. Members were required to indicate “agreement”, “partial agreement with a suggested minor wording change” or “disagreement, with free comments” for each recommendation.

Consensus was defined as “strong consensus” (> 95%), as “consensus” (75–95%) and “no consensus” (< 75%) agreement. In the case of less than 95% agreement during the first online voting, each PICO group was asked to modify the wording of the recommendation according to members suggestions or, if they chose to make no amendment, to provide an explanation. A second and final online voting round took place in January 2022 to attempt to reach strong consensus on the set of recommendations. Members were required to indicate “agreement” or “disagreement”, with explanations and proposal.

Results

A total of 18,399 abstracts were screened (Fig. 1 PRISMA flow chart). Subsequently 56 publications from 1969 to 2021 were included [19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74] for data extraction, which included 11,689 patients in both acute and critical care, medical and surgical care settings and used in the development of 16 recommendations (Table 1).

Fig. 1
figure 1

PRISMA flow chart of literature search and study screening

Full size image
Table 1 Intravenous maintenance fluid therapy (IV-MFT) recommendations, level of evidence according to SIGN grading, and consensus within the expert group
Full size table

Overall, the level of evidence was low [11] with most studies graded 1 or lower with a serious risk of bias (Table 2 and supplemental files 3–6).

Table 2 Summary table of studies included to produce recommendations and their SIGN level of evidence
Full size table

Furthermore, the data were only suitable to combine in a meta-analysis (Figs. 2–5) for some outcomes in four of the PICO questions (PICO 4 on composition did not allow for any meta-analysis). However, outcomes varied significantly between studies and limited the scope of the meta-analysis. All forest plots are available in Supplement files 3, 4, 5 and 6 with the respective heterogeneity assessment and relative risks.

Fig. 2
figure 2

Meta-analysis of studies comparing the impact on length of stay of intravenous versus enteral hydration

Full size image

Table 2 provides the list of included studies and their main characteristics; Table 3 summarizes for each PICO the profile of the patients recruited in these studies.

Table 3 Characteristics of the patients recruited in the included studies
Full size table

For PICO 1 on indications for IV-MFT, a meta-analysis included 1668 children. It showed no significant difference in length of stay between enteral and parenteral hydration (mean difference = 9.09, 95% CI [−1.24–19.43], p = 0.08) (Fig. 2) but a trend towards a reduction in length of hospital stay in the enteral group [23, 24, 27, 28].

For PICO 2 on isotonic IV-MFT solutions, a meta-analysis of 17 RCTs (involving 3356 patients) assessing the risk of developing hyponatremia showed that isotonic solutions significantly reduced this risk compared with hypotonic fluids, with an odds ratio of 0.41 (95% CI [0.26–0.67], p = 0.0003), although the heterogeneity was high between studies [31,32,33,34, 36,37,38,39,40, 42, 45, 46] (Fig. 3).

Fig. 3
figure 3

Meta-analysis of studies comparing the impact on hyponatremia occurrence of isotonic versus hypotonic solutions

Full size image

In PICO 3 on balanced IV-MFT solutions, the length of acute care or PICU stay were slightly but significantly decreased in children receiving balanced solutions in a meta-analysis of 5 studies, including 283 patients (mean difference: −0.20 days; 95% CI [−0.33; −0.08], p = 0.001 [57, 58, 61,62,63] (Fig. 4).

Fig. 4
figure 4

Meta-analysis of studies comparing the impact on acute or critical care stay of balanced versus non-balanced solutions

Full size image

For PICO 4, the data did not allow the conduct of any meta-analysis.

In PICO 5 on the amount of IV-MFT, a restrictive strategy was significantly associated with a lower change in plasma sodium (mean difference = 1.95; 95% CI [0.29; 3.62], p = 0.02) after 12 h or more of treatment, in a meta-analysis pooling 167 patients in six subgroups in three RCTs, including patients in the PICU [71, 73, 74] (Fig. 5).

Fig. 5
figure 5

Meta-analysis of studies comparing the impact on natremia of a restrictive versus a non-restrictive fluid strategy

Full size image

Figure 6 provides a short answer to each PiCO question.

Fig. 6
figure 6

Short answers to PICO questions

Full size image

The grading of the 16 recommendations was heterogenous (ranging from good clinical practice and expert opinion to Grade A), according to the SIGN grading [11] and the quality of the meta-analysis (Table 1). After the 2-round voting process, strong consensus (> 95% agreement) was reached for 11/16 (69%) and consensus (90% < agreement < 95%) for 5/16 (31%) of the recommendations.

A long rationale supporting each recommendation is provided for each PICO question in supplemental files 7–11. This details the underlying background and pathophysiology of the PICO question, summarizes current European practice [2], discusses available recommendations in adults or other pediatric settings, provides further details of our literature search, data extraction and analysis of the data, analyzes the volume of evidence, its applicability, generalizability, consistency and clinical impact of the recommendations and finally provides the recommendation.

Discussion

Summary of findings

These ESPNIC evidence-based recommendations provide guidance for clinicians using IV-MFT in children both in acute hospital and intensive care settings. They are based on a comprehensive literature search, including literature up to November 2020.

We were able to propose 16 recommendations corresponding to our questions with a high level of consensus, and to provide an extensive rationale to support each of them, available in supplemental material. Unfortunately, the general level of evidence remains low, both in terms of quantity and quality. Few well conducted and low risk of bias RCTs were available. Furthermore, the study populations, settings and interventions were heterogeneous and, in many cases, made it difficult to pool and to compare results. Thus, even with a strong consensus, the majority of the recommendations 12/16 (75%) might be considered as weak (C or under) as they are based on a low level of evidence.

Consistency with other guidelines and recent publications

Our recommendations are consistent with the 2018 American Academic of Pediatrics clinical practice IV-MFT guidelines regarding the use of isotonic fluids [3]. Indeed, the American Academic of Pediatrics guidelines strongly recommended the use of isotonic fluids in children between 28 days to 18 years of age in acute or intensive care settings. However, our guidelines extend beyond the American guidelines as they also make recommendations on the indications for IV-MFT, on the composition of IV-MFT, and on the amount of IV-MFT that should be prescribed. In 2021, Leung and colleagues [75] published a consensus statement on IV-MFT in hospitalized children, based on a review of the literature (2000–2019). Their recommendations are similar to ours regarding prioritization of the enteral route, the use of isotonic fluids, the reduced amount of fluids administered and the type of monitoring. With the exception of isotonic fluids (strong recommendation with high quality of evidence), recommendations were all consensus based on low-quality evidence and expert opinion. The use of balanced fluids was not addressed.

Our study protocol included literature until November 2020 only, we used the same search equations and inclusion criteria to extract recent IV-MFT studies (December 2020–June 2022) to ensure the consistency of their findings with our guidelines. Twenty-one new studies (and 3 ongoing trials) were identified and are summarized in supplemental digital content 12. Their findings are globally consistent with our recommendations, especially regarding the use of isotonic fluids, reduced infusion volumes and the use of the enteral route when possible. The impact of isotonic fluids has now been studied in specific sub-groups of children such as newborns and patients with diabetic keto acidosis. Balanced solutions are currently being studied, and we have identified three ongoing RCTs comparing them to non-balanced solutions which may help revising our guidelines or adapting their level of evidence in the future.

Implications for practice

The reporting of key outcomes was inconsistent within studies, which prevented us from conducting other meta-analyses. For example, the impact of isotonic and/or balanced solutions on Na and Cl plasma levels was presented as absolute values in some papers, or changes over time, or absolute difference in others. Improving the standardization of outcome reporting in future trials is essential to allow pooling of the data. In the future, our expert group aims to conduct an international Delphi study to gain expert consensus and develop a core outcome set to be reported in future studies on IV-MFT in children. Despite this, these guidelines highlight several important points that must be considered in daily practice when prescribing IV-MFT for children. First, the intravenous route for hydration is not required in every clinical situation. Second, isotonic IV-MFT should be preferred over hypo or hypertonic solutions. Third, balanced fluids should be the standard IV-MFT solution used in children. Fourth, the IV fluid composition is important, and glucose and plasma electrolyte monitoring is essential. Finally, the harm of excessive fluids and volume overload is highlighted, and the daily calculation of fluid balance is essential for any child receiving IV-MFT.

Future implementation challenges

As per its definition, IV-MFT needs to be differentiated from IV replacement therapy, which aims to compensate for abnormal losses (e.g. skin losses in case of major wounds or burns, intestinal losses in case of severe enteropathy, losses through drains, etc.). However, IV-MFT and replacement fluids may be administered through one single IV fluid prescription which then needs to combine both IV-MFT recommendations (in terms of amounts and composition) and adapt to the rate of loss and composition of fluid losses. The implementation of these recommendations into clinical practice may be challenged by the lack of availability of ready-to-use solutions adapted for children in some European countries. A recent survey by Morice et al. [2] reported some centers having to reconstitute isotonic glucose IV-MFT solutions within clinical care units. The use of ready-to-use IV-MFT is beneficial to avoid reconstitution errors, physico-chemical stability issues and microbiological contaminations. Furthermore, products designed for adults do not usually provide glucose and, therefore, do not meet the requirements of younger children. Perioperative maintenance fluids designed as per Sümpelmann et al. recommendations [10] (Glucose 1–2.5%) may not provide sufficient amounts of glucose if used outside the perioperative setting. Isotonic balanced solutions, providing some glucose (4 to 10%) and limited amounts of potassium (+/- 4 mmol/L), would meet most children’s requirements in terms of IV-MFT. Solutions such as these can be found in certain countries in various amounts (250, 500 and 1000 mL bags) and have an osmolarity compatible with peripheral infusion. Nevertheless, such solutions are inconsistently available through Europe. Availability of ready-to-use IV-MFT solutions that are tailored to childrens needs would help facilitate the implementation of these new ESPNIC recommendations.

Limitations

The main limitation of these evidence-based recommendations is the general paucity of evidence and the low level of evidence around some of the questions. We have included all studies published until 2021 due to the paucity of studies available, but most (80.0%) were published in the last decade, which reduces the bias from studies published within a large time period. Our definition of acute and critical illness may slightly differ from the one used in some studies as PICU admission policies vary within institutions. For most of the recommendations, apart from PICO 1, sparse evidence prevented us from translating this to specific pediatric populations or specific clinical situations; sub populations should be further studied in future trials to secure extrapolation of these guidelines to specific groups of patients and for different age ranges. Apart from the trials conducted on IV-MFT during phototherapy, term neonates were rarely analyzed as a specific group or a subgroup in most of the studies, and extrapolation of the results to this specific population should be considered with caution. Cardiac patients were included in some of the study populations and recommendations for PICO 2 and 5 are likely to apply to this specific population. Some studies assessed both IV-MFT and replacement or bolus fluid therapy, as these treatments were administered simultaneously or consequently over the study period. Consequently, this may have introduced a bias in the interpretation of the results, even if the results were consistent with other studies focusing on IV-MFT. However, a consistent fluid strategy considering resuscitation, replacement and maintenance fluid therapy seems reasonable. Furthermore, due to inconsistencies among the interventions (fluids and volumes used) and outcomes, meta-analyses were not possible to conduct for many questions and heterogeneity between studies may have resulted in some bias. Finally, in these evidence-based recommendations, the PICO questions and consensus voting only reflects the view of our ESPNIC expert group. Service users (acute and critically ill children) and their parents were not involved in the design of the project. Despite these limitations, this is the most up to date and extensive review of IV-MFT in children both in acute and critical care settings.

Conclusions

These evidence-based recommendations provide a ‘best-available-evidence’ guide for both pediatric and intensive care clinicians around the prescription of IV-MFT. Due to a paucity of robust evidence, the evidence level for most of the recommendations is low, even when strong consensus was reached among the expert group. Thus, many recommendations are based on expert opinion. This review clearly identifies the urgent need and gaps for future research in this field. Each of our PICO questions deserves further robust research, and new RCTs should be conducted specifically to clarify the impact of the use of isotonic solutions or balanced solutions in various age groups (term neonates, infants and older children) and a variety of clinical conditions. The lack of evidence regarding optimal electrolyte compositions (K, P, Mg, Ca) of IV-MFT is striking and so is the need for micronutrient additions in IVMFT. Consistent reporting of relevant outcome is mandatory to enable future metanalysis. ESPNIC intends to update these guidelines every 5 years, following the same methodology. In the future, our expert group also aims to produce tools to help implement these guidelines into clinical practice and promote auditing and monitoring of their implementation.