Article Text
PDF
Abstract
Introduction Hypernatraemia, defined as a plasma sodium concentration >145 mmol/L, is a frequent complication in critically ill patients treated in the intensive care unit (ICU) (= ICU-acquired hypernatraemia), with reported prevalence ranging from 4% to 26%. Hypernatraemia adversely affects various physiological functions and is associated with delirium, prolonged length of stay and increased ICU and post-discharge mortality. The sodium load from intravenous drug diluents significantly contributes to ICU-acquired hypernatraemia, with drug infusions comprising about 30% of the daily fluid volume of an average ICU patient. This study aims to investigate if using glucose 5% solution as the default drug diluent, instead of sodium chloride 0.9%, can reduce the prevalence of ICU-acquired hypernatraemia and improve patient outcomes.
Methods and analysis To test the effectiveness of glucose 5% solution as the default drug diluent, we will conduct a multicentre, pragmatic, embedded, open-label, stepped-wedge, cluster-randomised trial. The study will include twelve clusters (ICUs and one intermediate care unit) across six hospitals in Germany, with a projected total sample size of 4485 patients. In line with the stepped-wedge cluster-randomised design, one ICU will transition every 4 weeks, in a randomised sequence, from using sodium chloride 0.9% as the default drug diluent to glucose 5%.
The primary endpoint is the prevalence of hypernatraemia >150 mmol/L through day 28. The number of days alive and free of the ICU through day 28 will be tested hierarchically as a key secondary endpoint. Other exploratory endpoints include ICU mortality, ICU-free days, hospital-free days and other clinical outcomes. The primary endpoint will be analysed using a logistic mixed-effects model.
Ethics and dissemination The trial was approved by the Charité—Universitätsmedizin Berlin Ethics Board and by the ethics board of each enrolled hospital. The results will be submitted for publication in a peer-reviewed journal and presented at one or more scientific conferences.
Trial registration number The trial protocol was registered with the German Clinical Trials Register on 21 June 2024 prior to initiation of patient enrolment (DRKS00033397).
- NEPHROLOGY
- INTENSIVE & CRITICAL CARE
- Adult intensive & critical care
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Strengths and limitations of this study
The pragmatic design and the embedding of the trial in routine care will allow the completion of a large multicentre randomised clinical trial without external funding, setting a precedent for future investigator initiated multicentre randomised controlled trials.
Broad inclusion criteria will increase generalisability and achieve large sample size.
The embedded nature of the trial, relying on routine clinical data, could pose the risk of a high proportion of missing data.
Adherence to the intervention will not be known, as the drug diluent used is not routinely recorded for many medications.
Follow-up is limited to in-hospital data. Effects of the intervention beyond hospital discharge will remain unknown.
Introduction
Hypernatraemia, defined as a plasma sodium concentration >145 mmol/L, is a frequent complication in critically ill patients treated in the intensive care unit (ICU) (= ICU-acquired hypernatraemia), with reported prevalence ranging from 4% to 26%.1 It represents an imbalance of total body water and electrolytes.2 ICU-acquired hypernatraemia adversely affects various physiologic functions and is associated with delirium,3 prolonged length of stay (LOS)4 and increased ICU mortality and post-discharge mortality.5–7
The aetiology of hypernatraemia is multifactorial, with two principal pathomechanisms considered to be of primary importance8: first, a deficit in free water caused by increased losses (eg, due to fever, diuretics or diarrhoea) or decreased intake (eg, due to iatrogenic restriction of fluid intake or drug-induced thirst suppression) and second, an excess of sodium in the body caused by decreased sodium excretion (eg, due to medications or impaired kidney function) and/or increased sodium intake (eg, hypertonic intravenous fluids).
Clinical management of hypernatraemia can be challenging, and evidence-based treatment options are lacking. Current guidelines recommend enteral or parenteral electrolyte-free fluid administration as the first-line treatment, although this recommendation is not backed by data from randomised trials.9 Pharmacological interventions, such as thiazide diuretics or aldosterone receptor antagonists, have failed to show an effect on sodium levels in small trials.10 11
Trial rationale
The sodium load from intravenous drug diluents contributes significantly to ICU-acquired hypernatraemia.12 Both bolus and continuous intravenous drug infusions make up about 30% of the daily fluid intake of an average ICU patient.13 Assuming patients receive a cumulative volume of 700 mL from intravenous drug infusions per day, and sodium chloride 0.9% is used as the diluent, this amounts to a sodium intake of 107.8 mmol per day just as a byproduct of intravenous medications.
Two cohort studies investigated if switching the drug diluent from sodium chloride 0.9% to glucose 5% could reduce the prevalence of hypernatraemia in the ICU.14 15 One study compared two consecutive patient groups before and after the standard drug solvent was changed from sodium chloride 0.9% to glucose 5% solution in a mixed ICU population and found a relevant reduction in the prevalence of hypernatraemia.15 Another study with a similar design investigated the opposite switch. The default diluent for intermittent drug sets in the electronic ordering system was switched from glucose 5% solution to sodium chloride 0.9%, leading to an increase in the prevalence of ICU-acquired hypernatraemia.14 In our pilot study conducted in a medical ICU, the switch to glucose 5% as the standard diluent primarily led to a reduction in severe hypernatraemia >150 mmol/L.16
While these observational studies suggest the benefits of using glucose 5% as a diluent, prospective, randomised trials are needed to conclusively determine its effectiveness and safety.
Study objectives
The goal of this study is to determine if using glucose 5% solution as the standard drug diluent compared with sodium chloride 0.9% leads to a reduction in the prevalence of hypernatraemia (>150 mmol/L) and improved clinical outcomes, such as more days alive and free of the ICU.
Methods and analysis
This manuscript was prepared by the Sweet-Water investigators in accordance with the Standard Protocol Items of the Recommendations for Interventional Trials guidelines.17 This manuscript describes key elements of the trial protocol. The full study protocol provides a more detailed description of the study design and is provided in online supplemental file 1.
Supplemental material
Aim and objectives
The goal of this study is to investigate if using glucose 5% solution as the standard diluent compared with sodium chloride 0.9% leads to
A reduction in the prevalence of ICU-acquired hypernatraemia >150 mmol/L.
Improved clinical outcomes, such as more days alive and free of the ICU.
Study design
The Sweet-Water study is a multicentre, pragmatic, embedded, open-label, cluster-randomised trial with a stepped-wedge design. The trial aims to evaluate the impact of using glucose 5% solution as a standard drug diluent on the prevalence of ICU-acquired hypernatraemia, compared with 0.9% sodium chloride. A total of six hospitals with 11 ICUs and one intermediate care unit are participating as study centres (table 1). All participating centres are academic hospitals in Germany. Randomisation will be done at the ICU level; hence, the individual clusters at each centre (the seven ICUs at Charité—Universitätsmedizin Berlin and the two units at Uniklinik Göttingen) will change their treatment standards separately. Each ‘step’ in the study lasts 4 weeks. During the first 4 week step, all ICUs will continue to adhere to the current standard of using sodium chloride 0.9% as the default drug diluent. Subsequently, every 4 weeks one cluster (ie, one ICU) will switch to the glucose 5% standard. The time points for switching will be randomised at the start of the study. The first 4 weeks after switching to glucose 5% are designated as a ‘transition step’ to allow time for the implementation of the new glucose 5% dilution standard. Patients admitted during the transition step will be excluded from the analysis. The study diagram (figure 1) presents the timing of the interventions and the steps per cluster.
Overview of stepped wedge study design.
Overview of participating study centres
This trial is fully embedded in routine care, with the intervention delivered by clinical personnel, and the study data exclusively derived from routinely collected data in the participating centres’ electronic healthcare records. Furthermore, this is an investigator-initiated trial exclusively funded by internal resources of the participating centres. The trial was approved by the Charité—Universitätsmedizin Berlin Ethics Board (EA1/206/23) and by the ethics board of each enrolled hospital/ICU. The trial protocol was registered with the German Clinical Trials Register on 21 June 2024, prior to the initiation of patient enrolment (DRKS00033397).
Study setting and population
This trial will be conducted at seven academic hospitals in Germany with 11 intensive care and one intermediate care unit (table 1).
Inclusion criteria
Admitted to a participating ICU during the study period.
Age ≥18 years.
Exclusion criteria
Admission for acute brain injury (traumatic brain injury, aneurysmal subarachnoid haemorrhage or intracerebral haemorrhage).
Hyponatraemia (sodium levels below 130 mmol/L) at ICU admission.
Randomisation
The randomisation process will involve two stages. Initially, due to logistical considerations, one of the four ICUs in the Department of Nephrology and Intensive Care Medicine at Charité—Universitätsmedizin Berlin will be selected to transition first. This selection will be made through a separate randomisation process involving only these four ICUs. Following this, the order for the remaining clusters at Charité and the other participating clusters will be determined through a second randomisation process.
Study intervention and implementation
The previous standard of care of using NaCl 0.9% as the default drug diluent will be replaced by a mandate to use glucose 5% for compatible medications, as detailed in a standard table of common intravenous medications and their compatibility with glucose 5% (online supplemental table 1).
Supplemental material
In German ICUs, commonly prescribed intravenous drugs are predominantly prepared by ICU nurses. Central pharmacies typically only prepare and deliver chemotherapeutics and monoclonal antibodies. Furthermore, it is common practice that physicians prescribe intravenous medications without specifying the drug diluent, leaving the choice of solvents and diluents to the nurses. The nurses adhere to local standard procedures for solvation and dilution, which are uniformly applied across all patients in the ICU.
The information regarding whether a drug can be diluted in glucose 5% or NaCl 0.9% was obtained from the technical information leaflet provided by the pharmaceutical companies and consulted with our local pharmacists. Some antibiotics are generally compatible with glucose 5%, but stability is not sufficient for prolonged or continuous infusion.
Physicians are free to switch patients to glucose 5% as the drug solvent to treat patients with hypernatraemia during all steps in accordance with local practices. Regular virtual meetings will be held to address any questions from ICU staff and to share the best practices among participating centres.
In patients developing hyponatraemia under the glucose 5% standard, we advise switching these patients to the 0.9% sodium chloride standard. The treatment of hyponatraemia should follow local protocols and standards of care.
Transition period
During the transition step, nurses and physicians, under the guidance of site investigators, are instructed to adopt the glucose 5% drug dilution standard (online supplemental table 1). The drug solvation and dilution standard table is distributed digitally and as a paper copy. Laminated paper copies are prominently displayed across the study ICUs. Patients admitted during the transition step will be excluded from the analysis of the study.
Monitoring adherence
In most centres the diluent used to solve an intravenous drug is not recorded in the patient data management system by nurses. In each centre, we will survey the nursing staff during the intervention phase to assess compliance.
Blinding
The study will be conducted as an open-label trial.
Data collection
Data will be exclusively derived using routinely collected data from the electronic healthcare records. No additional study data is gathered. Patient data will be extracted retrospectively after the completion of the study from the patient data management systems (PDMS) of the participating centres. Detailed information about the collected data can be found in the trial protocol (online supplemental section 3).
Outcomes
Primary outcome
The primary outcome is ICU-acquired hypernatraemia >150 mmol/L through day 28 after ICU admission, defined by a single sodium measurement >150 mmol/L after exclusion of outliers through ICU day 28. Outcome ascertainment will cease at the time of hospital discharge or 28 days after enrolment, whichever occurs first.
The choice of the primary outcome is based on the results of our pilot study, which showed a reduction in hypernatraemia >150 mmol/L.16 We also believe that this is a clinically meaningful endpoint, assessing the reduction of moderate to severe hypernatraemia.
Key secondary outcome
As a key secondary endpoint, the number of days alive and free of the ICU, referred to as ICU-free days (IFDs), through day 28 will be tested hierarchically, conditional on significant findings in the primary analysis. IFDs are defined as the number of calendar days alive and free of the ICU through day 28. For patients who die before day 28, ICU-free days are recorded as zero. If patients are admitted repeatedly to the ICU during the 28 days, IFDs are counted after the final ICU discharge. Patients discharged from the hospital are assumed to survive until day 28.
Exploratory secondary outcomes
ICU mortality: death during the index ICU stay until day 28.
Number of days alive and free of hospital through day 28 (= hospitalfree- days through day 28): defined as the number of calendar days alive and free of the hospital through day 28. For patients who die before day 28, hospital-free days are recorded as zero. Patients discharged from the hospital are assumed to survive until day 28. If patients are readmitted, the number of days alive and free of hospital is counted after the final hospital discharge.
In-hospital mortality: death during the index hospital stay.
Length of stay in the ICU: defined as the number of calendar days admitted to the ICU. LOS will be counted until the final ICU discharge.
LOS in the hospital:defined as the number of calendar days admitted to the hospital. LOS will be counted until the final hospital discharge
Number of days alive and free of mechanical ventilation, referred to as ventilator-free days (VFDs), through day 28:defined as the number of calendar days alive and free of invasive mechanical ventilation beginning the day after the final receipt of invasive mechanical ventilation through day 28. For patients who die before day 28, VFDs are recorded as zero. For patients who return to invasive mechanical ventilation and are subsequently liberated from invasive mechanical ventilation prior to day 28, VFDs will be counted from the final liberation from mechanical ventilation.
Prevalence of dialysis-dependent acute kidney injury through day 28.
Duration of kidney replacement therapy (KRT) through day 28: defined as the number of calendar days from the beginning of KRT until the last dialysis. If the patient dies or is discharged within 96 hours (about 4 days) of the last dialysis session, renal recovery is not assumed.
Peak creatinine through day 28.
Sodium concentration on day 4 post ICU admission.
Sodium concentration on day 7 post ICU admission.
Number of days without hypernatraemia >150 mmol/L through day 28: defined as the number of calendar days alive and free of hypernatraemia >150 mmol/L. Patients who died prior to day 28 received a value of zero. If serum sodium is measured multiple times a day, a day is considered hypernatraemia-free only if all measurements are below the hypernatraemia threshold. Days after hospital discharge are assumed to be hypernatremia-free. If no measurement was performed, a sodium concentration of <150 mmol/L is assumed.
ICU-acquired hyponatraemia >145 mmol/L through day 28.
ICU-acquired hypernatraemia >155 mmol/L through day 28.
ICU-acquired hypernatraemia >160 mmol/L through day 28.
Exploratory safety outcomes
ICU-acquired hyponatraemia <130 mmol/L through day 28.
Rate of positive blood cultures.
Number of positive blood cultures.
Hyperglycaemia >200 mg/dL through day 28.
Hyperglycaemia >300 mg/dL through day 28.
Hyperglycaemia >400 mg/dL through day 28.
Process of care outcomes
Cumulative volume of intravenous drug diluents on day 7.
Cumulative total fluid balance on day 7.
Statistical analysis
Sample size estimation and power calculation
There are 14 steps, each lasting 4 weeks in this stepped-wedge study with 12 clusters and a transition period of one step. To estimate the total sample size, participating ICUs were asked to estimate the expected number of ICU admissions per 4 weeks (table 2).
Expected number of intensive care unit (ICU) admissions per 4 weeks by centre and ICU
Assuming these numbers of cases n per ICU per step (n=35, 20, 25, 15, 40, 25, 30, 15, 35, 15, 45, 45), the total sample size is expected to be n=4485 over the 14 steps.
Based on a pilot study (unpublished), we assume that ICU-acquired hypernatraemia above 150 mmol/L occurs in 17.9% of cases in the control phase and can be reduced to a prevalence of 10.5% in the intervention phase.
The R package ‘swCRTdesign’ was used to calculate the power. Awithin-period intra-cluster correlation of 0.01 as well as no time effect, which corresponds to a cluster auto-correlation of 1, is assumed. Since the centres are of different sizes and the power varies slightly depending on the randomised order of the centres, the specified power is based on a simulation with 100 000 simulation runs. Averaged over these runs, a power of 98.9% can be expected. The minimum power was 97.8%; the maximum, 99.3%.
Assuming that on average 20% of the data supplied cannot be analysed or is excluded from the analysis, the number of cases would be reduced to n=3588. In this case, the average power over the 100 000 simulation runs would be 97.1%, the minimum 95.0% and the maximum 97.9%.
Statistical analysis principles
R (R Foundation for Statistical Computing, Vienna, Austria) will be used for analyses. Data will be analysed in an intention-to-treat fashion.
The following patients will be excluded from the primary analysis of the study, though these patients may have received the intervention.
Treatment crossover (admitted during the 0.9% sodium chloride step and discharged in the transition or G5% phase) or discharge or death occurred after the conclusion of the study. Patients who cross over receive a mix of interventions, which could dilute the differences between the study arms and obscure the true effect of the intervention.
Admission sodium ≥150 mmol/L. Patients with hypernatraemia at admission are excluded from the primary outcome analysis because they already exceed the threshold for the study’s primary endpoint, making it impossible to assess the intervention’s effect on preventing ICU-acquired hypernatraemia in this group.
Outcome analysis
The primary outcome will be examined in a mixed logistic regression model with the primary outcome as the dependent variable and the treatment group as fixed effect. This model will be adjusted for the admitting ICU as a random intercept, alongside fixed effects for age, sex, the baseline sodium value and the week of admission. An alpha level of 0.05 is used to reject the null hypothesis. If the primary outcome is positive (p value <0.05), the key secondary outcome ICU-free days is tested hierarchically (alpha level set at 0.05). To estimate treatment effects on this outcome, a generalised linear model with a 0-inflated beta-binomial distribution will be used. The admitting ICU will be incorporated as a random effect; treatment group, age, sex, week of admission and baseline sodium value will be considered as fixed effects.
As sensitivity analyses, we will add an interaction effect between group and week of admission for first and secondary primary endpoint models.
Descriptive statistics, including mean and SD, median, interquartile ranges, minimum and maximum, and the number and percent of subjects in specified categories will be used to summarise the demographic and baseline variables.
Handling of missing data
The primary outcomes will be analysed from complete cases. For a sensitivity analysis, the primary and key secondary outcomes will additionally be analysed using multiple imputation methods to assess the robustness of the findings against missing data.
Interim analysis
Given the stepped wedge design and the short trial duration, no interim analysis is planned.
Prespecified subgroup analysis and sensitivity analysis
Sex as documented in the PDMS.
Age≥ 65 and < 65.
KRT on admission: defined as KRT (any modality) started prior to admission and continued after admission or started de novo within 24 hours of admission to the ICU.
Invasive mechanical ventilation on admission. Invasively ventilated patients on admission are defined as those receiving mechanical ventilation either prior to or within 24 hours of ICU admission.
Illness severity on admission stratified by Sequential Organ Failure Assessment (SOFA) score:
SOFA ≤8
SOFA >8.
An exploratory subgroup analysis will be conducted among patients with an ICU LOS greater than 3 days (partial days counted as full days). Because ICU LOS was measured after randomisation and is potentially influenced by the intervention, this analysis should be interpreted cautiously due to the risk of bias. The focus on patients with longer ICU stays is based on evidence from the pilot study, suggesting that the intervention’s effect on sodium balance becomes discernible only after the initial 3 days.
Furthermore, as a sensitivity analysis, the primary endpoint will also be analysed using a risk regression model (Fine and Grey), with death considered a competing risk and hospital discharge treated as a censoring event.18 Similarly, the key secondary outcome ICU-free days (as well as other analogous outcomes such as VFDs) will be analysed using a risk regression model with death as a competing risk. In this context, ICU-free days are reinterpreted as time to ICU discharge and VFDs as time to liberation from mechanical ventilation.
Trial status
Sweet-Water is scheduled to start patient enrolment on 1 December 2024. The last study step will conclude on 27 December 2025.
Ethics and dissemination
The protocol was approved by the local Research Ethics Committee (IRB) of the coordinating study centre (Charité—Universitätsmedizin Berlin) with a waiver of informed consent and by the local IRB of all participating hospitals (Uniklinik Göttingen, Universitätsklinikum OWL, Campus Bielefeld Bethel, Uniklinik Essen, Medizinische Hochschule Hannover, Universitätsklinikum Münster). Trial results will be submitted to a peer-reviewed journal for consideration of publication and will be presented at scientific conferences. The results of the study will be disseminated to patients and the public at the completion of the trial.
Patient information
Comprehensive information about the study will be made available. This included an information sheet in lay language, detailing the study’s purpose, methodology and risks (online supplemental material). This sheet will be accessible in every participating ICU and on the Charité—Universitätsmedizin Berlin website, ensuring transparency and public accessibility of study information.
Protocol changes
Any changes to the trial protocol will be recorded on the German clinical trials register (DRKS).
Ethics statements
Patient consent for publication
Acknowledgments
GPT-4, an advanced language model developed by OpenAI, was used during the writing process of this manuscript to enhance language clarity, correct grammatical errors, and eliminate spelling mistakes. We acknowledge support from the Open Access Publication Fund of Charité – Universitätsmedizin Berlin.
Footnotes
X @schenkh2
J-HBH and RMH contributed equally.
Contributors The study concept and design were conducted by J-HBH, PE, K-UE, RMH and MP. Drafting of the manuscript was undertaken by J-HBH, PE and MP. All authors critically revised the manuscript for important intellectual content, including RMH, MP, AU, HM-R, JN, DZ, TS, MW, UT, OH, BW, SW-C, CS, BT, RB, MA-T, AZ, CS, HS, UH, DvW, FB, K-UE and PE. PE is the guarantor.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.