Techniques to increase lumbar puncture success in newborn babies: the NeoCLEAR RCT

Article history

The research reported in this issue of the journal was funded by the HTA programme as project number 15/188/106. The contractual start date was in September 2017. The draft report began editorial review in July 2021 and was accepted for publication in April 2022. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Copyright statement

Copyright © 2023 Roehr et al. This work was produced by Roehr et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaptation in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – NIHR Journals Library, and the DOI of the publication must be cited.

2023 Roehr et al.

Background neonatal lumbar punctures

In the UK, between 15,000 and 30,000 newborns undergo a lumbar puncture (LP) each year to rule out meningitis. 1 Neonatal meningitis is associated with high mortality and morbidity. 1 Symptoms and signs are subtle, and the diagnosis can be confirmed only by analysis of the cerebrospinal fluid (CSF), obtained by performing a LP. Thus, LPs are an essential part of the diagnostic work-up for meningitis. 1,2 LP techniques vary in current neonatal practice, with no level 1 evidence to determine the best approach. Owing to the technical challenges of neonatal LP, thousands of infants each year undergo unsuccessful LPs, resulting in repeat procedures, causing distress to infants and parents and often necessitating prolonged courses of antibiotics and hospitalisation for the infant, and, oftentimes, rooming-in for the accompanying mother (Box 1).

The LP procedure aims to obtain CSF from the lower spine for laboratory analysis to confirm the presence/absence of meningitis and to identify the causative organism. Infants with meningitis typically require 14–21 days of inpatient intravenous antibiotics, incurring significant financial costs, and often receive hospital follow-up because of the risk of long-term neurological sequelae. 3 Prolonged antibiotic use is associated with significant complications, such as necrotising enterocolitis,4 and a potential for the development of antibiotic resistance. 5 If meningitis can be excluded, then antibiotics are usually stopped after 5 days, allowing discharge with no further follow-up.

The definition of successful LP varies, but usually encompasses the acquisition of ‘clear’ CSF (colloquial medical term: a ‘champagne clear tap’). However, in neonates, CSF samples are often pink/red due to red blood cells (RBCs) sampled unintentionally from nearby blood vessels. Significant numbers of RBCs hinder CSF interpretation, and the presence/absence of meningitis cannot be confirmed. Therefore, LP often needs to be repeated, and many infants are treated with extended courses of antibiotics because meningitis cannot be excluded. Repeated procedures and concern about meningitis understandably lead to heightened parental anxiety. 6

The success rates of LP are much lower in neonates (50–60%)7,8 than in older children (78–87%). 9,10 Modifications to ‘traditional’ LP technique have been studied, but most data thus far are observational and have a high risk of bias,11 and so no improvements have been incorporated into widespread routine practice.

In 2017, it was believed that a trial aiming at establishing the most successful LP technique would be particularly timely. The, then recent, 2016 National Institute for Health and Care Excellence (NICE) guidance,12 although aiming to avoid delays in diagnosing meningitis, had resulted in many more neonatal LPs being performed,12 and this led to an increase in antibiotic use, which came at a time of growing concerns about antimicrobial resistance caused by unnecessary use of antibiotics.

It was thought that a large-scale randomised controlled trial (RCT) of LP technique would serve the neonatal knowledge base in several ways. First, optimising LP technique will mitigate the impact of the above-mentioned change in practice on NHS resources. Second, as the dangers of antibiotic resistance become ever more pressing, technologies that have the capacity to reduce hospital antibiotic use are invaluable in preventing antibiotic-resistance problems in the future. Third, it addresses the need, identified in a recent systematic review,11 for an investigation into alternative LP techniques.

Outcomes affected by lumbar puncture technique

If a large-scale RCT could demonstrate an improved LP technique, then its incorporation into clinical practice across the UK should follow, hopefully improving neonatal LP success over the current expected event rate of 59%, which is based on local cohort data. 13 During trial planning, a 10% improvement in LP success was deemed clinically significant. Such an increase in LP success rate can be expected to translate, each year, into 1600 fewer infants having repeat procedures, 14,400 fewer doses of intravenous antibiotics (with fewer complications) and 2680 fewer bed-days for mothers and infants. Parental anxiety would be reduced, as would healthcare costs through reduced hospitalisation and reduced antibiotic use (limiting the ongoing rise of antibiotic-resistant pathogens13), and efficiency of neonatal services would be improved.

Existing evidence

Updated literature search

Brief summary of the evidence review at trial inception

No LP technique was backed by high-quality evidence from studies conducted in the neonatal period. LP techniques warranting further investigation, which can be studied most efficiently and reliably with a RCT, were (1) sitting position and (2) ESR. Both techniques are free ‘existing technologies’ that are already used by some practitioners. Observational evidence for these techniques has not led to a change of routine clinical practice. Randomised evidence was felt to be important and highly necessary to provide robust and convincing data. If either technique was shown to be beneficial, it would be free and easy to introduce nationwide.

Objective

We aimed to determine the optimal technique for LP in newborn infants by evaluating the success rate and short-term clinical, resource and safety outcomes of two modifications to traditional LP technique: a change of infant position (i.e. from lying to sitting) and a change in the timing of stylet removal (i.e. from LSR to ESR). This would be, to the best of our knowledge, the first appropriately powered RCT investigating different neonatal LP techniques, one with the potential to make a significant contribution to current knowledge and even change practice.

Design

The trial rationale and design were closely discussed with, and received input from, parent representatives from the Oxford-based charity Support for the Sick Newborn And their Parents (SSNAP) [Oxford, UK; URL: www.ssnap.org.uk (accessed 14 June 2022)]. The final trial design and its protocol are published elsewhere. 37 Briefly, the Neonatal Champagne Lumbar punctures Every time – An RCT (NeoCLEAR) trial was planned as a pragmatic non-blinded multicentre 2 × 2 factorial RCT to compare the proportion of infants in whom CSF with a RBC count < 10,000/mm³ was successfully obtained on the first LP procedure. Investigated techniques included a combination of (1) the infant’s position (i.e. sitting vs. lying) and (2) the timing of stylet removal (ESR vs. LSR) [URL: www.npeu.ox.ac.uk/neoclear (accessed 14 June 2022)]. Following written parental consent, infants requiring LP [with a working weight of > 1000 g and corrected gestational age (CGA) of between 27⁺⁰ and 44⁺⁰ weeks] were randomised by web-based allocation to LP (1) in the sitting or lying position and (2) with ESR or LSR. The trial was powered to detect a 10% absolute risk difference in the primary outcome of an interpretable CSF sample, defined as containing a RBC count < 10,000/mm³. Clinicians undertaking LPs received practical training in the different trialled techniques. The application of topical anaesthetic cream prior to LP was encouraged. A minimum set of vital signs were monitored throughout the procedure [i.e. oxygen saturation (SpO₂) and heart rate (HR)] by pulse oximetry. In addition, the duration of the procedure was timed and the clinical monitoring data were documented for the purpose of the trial. All other technical and, especially, patient management decisions were in accordance with local protocols.

In practice, parents were approached for written consent when an eligible infant needed to undergo LP. Once consent was given, computerised randomisation proceeded. Staff were advised to make one or two attempts, defined as the needle passing through the skin once, per ‘procedure’. The randomised technique was to be used for up to two procedures. The requirement for and timing of a second procedure or any further procedures were determined by the clinical team. Patient characteristics and demographic data, as well as trial-relevant clinical data, were sourced by the local study teams from hospital records and recorded on trial-specific electronic case report forms (eCRFs). All CSF samples were sent to the local laboratories, as per recruiting site procedures, and only data from laboratory reports that were immediately relevant to the trial were entered into eCRFs.

To optimise study processes around recruitment, intervention delivery, training and outcome assessments, the trial included an internal pilot, which was carried out over 8 months at centres recruiting the first 250 randomised infants. The Trial Steering Committee (TSC) reviewed the pilot data, made recommendations and approved continuation. Details of TSC members are provided in Appendix 1.

Ethics approval and research governance

The NeoCLEAR RCT received approval from the NHS Health Research Authority, South Central Hampshire B Research Ethics Committee on 12 June 2018 (reference 18/SC/0222, nrescommittee.southcentral-hampshireb@nhs.net).

Trust confirmation of capacity and capability was obtained at each site. The chief investigator or delegate submitted an annual progress report, an end of study notification and the final report to the Research Ethics Committee, Health Research Authority, host organisation and sponsor, as required by the respective organisations.

The study was sponsored by the University of Oxford (Oxford, UK). The trial was run by a designated Project Management Group (PMG) at the National Perinatal Epidemiology Unit (NPEU) Clinical Trials Unit (CTU), Nuffield Department of Population Health, University of Oxford. The core PMG met once a month, either remotely or face to face. An extended PMG (i.e. the Co-Investigator Group) met every 3 months initially and then every 4 months subsequently to troubleshoot, review progress and forward plan. The PMG reported to the TSC. Meetings were held either remotely or face to face, as permitted by COVID-19 restrictions.

The trial was overseen by the TSC, which had the ultimate responsibility for considering and, as appropriate, acting on the recommendations of the Data Monitoring Committee (DMC). The TSC included an independent chairperson, at least one clinician, statistician and patient and public involvement (PPI) representative, the chief investigator and the senior co-investigator. The TSC met annually and reviewed the progress of the trial. Contributions on documentation were also received from the baby charity Bliss (London, UK).

A DMC reviewed the study data and outcomes. The DMC ensured the safety and well-being of the trial participants and, if appropriate, made recommendations regarding continuance of the study or modification of the protocol. The DMC was, therefore, also responsible for reviewing the safety reports of serious adverse events (SAEs), but, ultimately, the TSC would have the responsibility for stopping the trial on safety grounds. Lists of the members of the DMC and the TSC are in Appendices 1 and 2, respectively.

Patient and public involvement

Patient and public involvement was integral in the trial design, with feedback and input coming from several channels, including (1) early involvement of a PPI representative and co-applicant, (2) advice from the National Institute for Health and Care Research (NIHR) PPI division and information/details for a focus group sent to the NIHR PPI online mailing list, (3) feedback from a PPI representative for the NIHR Neonatal Clinical Study Group, (4) information/details for a focus group via SSNAP and (5) parental written feedback from a survey of parents of babies who previously received care on a neonatal unit.

As a direct result, we (1) confirmed the importance of this trial in improving the clinical care provided to newborns and their families, (2) prioritised our clinical outcomes relating to LP success, number of procedures and length of stay (as those were reported to be the most important outcomes for parents), (3) developed a procedural pain relief protocol aiming to provide better analgesia than current standard practice, (4) had assurance that the timing of our consenting process and randomisation was appropriate and acceptable to parents surveyed and (5) enhanced our plans for the LP training provided to each unit.

Participants

Participants of the trial were infants who were having a LP in UK neonatal units and their associated maternity wards.

Inclusion and exclusion criteria are shown in Box 3.

Setting

The NeoCLEAR trial was set in UK neonatal units and their associated maternity wards. The following centres were included:

• recruiting sites, where parental consent was obtained, infants were enrolled by randomisation and participation in the trial was commenced (n = 21) (see Appendix 3)

• ‘continuing care sites’, that is other units to which babies were discharged from the recruiting unit and where data collection continued until discharge (see Appendix 4).

Infants were eligible to participate in other clinical trials at the same time as taking part in the NeoCLEAR trial, depending on the nature of the interventions in the other trial. Other trials running concurrently were discussed by the chief investigators or their delegated representative, who then agreed if joint recruitment was appropriate.

Screening and eligibility assessment

Infants whom the clinical teams deemed necessary to undergo LP were screened for eligibility. Anonymised screening data were recorded via the randomisation website for the co-ordinating centre to review rates of ineligibility and participant uptake rates.

Informed consent and recruitment

The clinical teams provided the parents of eligible infants with both verbal information and written information, in the form of a parent information leaflet. The teams approached parents with legal parental responsibility to discuss the trial, to answer any questions the parents may have and to request consent. Parents had as much time as they needed to consider the information provided, to discuss it with the research team or other independent parties, and to decide whether or not to participate in the NeoCLEAR trial. Written informed consent for the study was then obtained by a suitably qualified member of the study team. During the study pilot parents were given a copy of a parent anxiety score [using the State–Trait Anxiety Inventory Subscale (STAI-S)] (see Parent questionnaire). 38 Parents completing the STAI-S were also asked to provide written consent for their participation (use of the STAI-S was stopped following the review of the internal pilot because completion rates were low). Consent was given with enough time to allow randomisation and to enable the clinical team to prepare for the procedure. Where this was not possible, the LP was not delayed if the infant’s clinician deemed any delay to be clinically unsound, and, in such cases, the infant was not recruited to the NeoCLEAR trial.

Site and staff training

Before undertaking procedures as part of the trial, all ‘operators’ (i.e. practitioners inserting the needle for the LP) received face-to-face training in trial background and overview and how to perform all four LP techniques to be investigated in the trial. Training included safety monitoring and recommendations for analgesia. Training included demonstration and practice on a neonatal LP manikin (LumbarPunctureBaby, EMD Services, Uttoxeter, UK). Operators were asked to sign a training log to confirm that they were confident in performing any of the four trial techniques on completion of training. Only after this were operators entered onto the delegation log. Assistants (i.e. practitioners holding the baby for LP and/or collecting CSF) were offered the same training or a shortened positioning-focused training session. Uptake of training was variable between sites. Narrated videos of each trial technique were available on the trial website throughout the recruitment period and were signposted during all training sessions. The regular trainee rotation across sites made regular training sessions necessary, and this was facilitated by the clinical research fellow (CRF) and local principal investigators (PIs).

Training the trainers

Owing to junior doctor rotation, the number of participating centres and the resulting distances between the primary study sites, the number of training sessions required to ensure ongoing recruitment became large. Additional manikins were purchased and a ‘train the trainer’ competency document was constructed to allow PIs and sub-PIs to become trainers. Sign-off of this competency document was carried out by the CRF and local PI. LP manikins were couriered to the sites participating in ‘train the trainer’ to ensure that all operators had experience in practising on the manikins. Uptake of ‘train the trainer’ made restarting recruitment after the COVID-19 hiatus possible when face-to-face site visits were not permitted. Mechanical upkeep of the training manikin was ensured by the CRF together with the manufacturer.

Interventions

Infants requiring LP were randomly allocated to one of four combinations of interventions: (1) LP in the lying (lateral decubitus) position and LSR, (2) LP in the lying position and ESR, (3) LP in the sitting position and LSR or (4) LP in the sitting position and ESR. The trial protocol suggested that term-born infants were treated with local anaesthetic cream prior to LP. Infants were clinically monitored throughout the procedures through detailed clinical observation, aided by pulse oximetry (i.e. continuous measurement of peripheral SpO₂ and HR). The duration of procedures was timed and recorded as part of the trial documentation.

Randomisation and blinding

The NeoCLEAR trial used 1 : 1 : 1 : 1 randomisation to one of the four arms [i.e. (1) lying (lateral decubitus) position and LSR, (2) lying position and ESR, (3) sitting position and LSR or (4) sitting position and ESR] using a 24 hours a day, 7 days a week, secure web-based randomisation facility, which ensured balance between the groups. A telephone back-up system was available 24 hours a day.

Stratified block randomisation was used to ensure balance between the groups with respect to the collaborating hospital and CGA at trial entry (four groups: 27⁺⁰ to 31⁺⁶ weeks, 32⁺⁰ to 36⁺⁶ weeks, 37⁺⁰ to 40⁺⁶ weeks and ≥ 41 weeks). If repeat LPs were warranted for the same infant for the same indication after an initial unsuccessful attempt or procedure, then the infant received the same allocated technique. Infants who had more than one indication for LP during the trial recruitment period were not re-randomised. Multiple births were randomised separately, with their study identification numbers linked on the database prior to analysis.

A statistician who was independent of the trial generated the randomisation schedule, and the senior trials programmer wrote the web-based randomisation program; both were independently validated. The implementation of the randomisation procedure was monitored by the senior trials programmer throughout the trial, and reports were provided to the DMC.

The NeoCLEAR trial was an open-label trial, as blinding of the practitioner and nursing staff to the allocated technique was not possible. The assessment of the primary outcome and major secondary outcomes was based on laboratory tests (effectively blinded). Parents were not usually told which technique their infant had been allocated to and were not routinely present for the procedure; however, if they requested this information, it was shared with them, and they were able to observe the LP, at the discretion of the practitioner.

Primary and secondary outcomes

Parent questionnaire

During the pilot phase (i.e. after consent but before the first LP), parents were asked ‘How have you felt physically during the last couple of days?’. Parents used a five-point Likert scale to answer the question. In addition, parents were asked to complete the STAI-S questionnaire. 38,39 The STAI-S is a well-validated measure that consists of 20 questions that identify how stressed/anxious someone is feeling at the time of assessment. All items on the STAI-S are rated on a four-point scale, and the mean score can be used in analyses. Discontinuation of the STAI-S was recommended by the TSC based on a review of the pilot.

Sample size

The NeoCLEAR trial was designed with 90% power to detect a 10% absolute difference in the primary outcome (with an estimated comparator group event rate of 59%), with a 5% two-sided significance level. A total of 483 infants were required for each arm of each comparison (i.e. sitting position vs. lying position and ESR vs. LSR). Allowing for 5% attrition, the recruitment target was 1020 infants. Initial recruitment was planned at 10 hospitals and this was later extended to 21 centres across England (see Appendix 3).

Statistical analyses

A statistical analysis plan was agreed prior to data lock. The analysis and presentation of results followed recommendations of the CONSORT (Consolidated Standards of Reporting Trials) group.

Data collection

All trial data were collected from hospital records and recorded on trial-specific eCRFs (see Appendix 6). Trial entry data included details to confirm eligibility and confirmation of parental written consent. The completion of data entry was monitored through the CTU. Most outcome data consisted of routinely recorded clinical items obtained from the clinical notes. Non-routinely collected data included procedural details, such as time taken for LP, infant movement, SpO₂ and HR. No additional blood or tissue samples were required. Outcome data were collected until discharge home. Parents completing the STAI-S were asked to complete a second questionnaire within 48 hours of the first LP (pilot phase only).

Adverse event reporting

Box 7 lists known, but rare, complications of LP. Any occurrence of a complication following the LP was reported as a SAE.

The full protocol and guidance sheets provided to sites also prespecified expected SAEs that were foreseeable and should not have been reported unless thought to be causally related to trial procedures. All unforeseeable SAEs occurring after consent until discharge home had to be reported.

Parents had the right to withdraw consent for any aspect of the study, including their infant’s future procedures and data collection, as well as their own questionnaire completion. The treating clinician was permitted to discontinue a participant if they considered it to be in the best interests of the infant’s health and well-being.

Adverse events were recorded locally and in the eCRF, and were followed up by the trials team. The DMC reviewed the study data and outcomes, including safety reports of SAEs. SAEs were collected until the infant was discharged home, as SAEs occurring after this time point were unlikely to be related to the trial intervention. As parental participation was limited to the STAI-S questionnaire, no SAE recording was conducted for this group. For the duration of the trial, the DMC ensured the safety and well-being of the trial participants and would have made recommendations to the TSC regarding discontinuance of the study or modification of the protocol, if required; however, this was not necessary throughout the NeoCLEAR trial.

Reporting procedures

All expected SAEs (see Box 7) were recorded on the eCRF and reviewed by the DMC at regular intervals throughout the trial. Any unexpected SAEs were reported by trial sites to the NeoCLEAR Coordinating Centre as soon as possible after the event had been recognised as a SAE that was not included in the list of expected SAEs. Information on each SAE was recorded on a SAE reporting form, which was electronically transferred to the NeoCLEAR Coordinating Centre. A standard operating procedure (SOP) outlining the reporting procedure for clinicians was provided with the SAE form and in the trial handbook. The NeoCLEAR Coordinating Centre processed and reported the events as specified in the CTU’s SOPs. The chief investigator informed all investigators concerned of relevant information about unexpected SAEs that could adversely affect the safety of participants. Once a year, during the recruiting period of the trial, a safety report was submitted to the sponsor and Research Ethics Committee.

Governance and monitoring

The NPEU CTU operates on a stringent set of accredited SOPs. Within this framework, the CTU oversaw the trial governance and conduct, as well as the monitoring of the trial centres. The CTU guidance comprised structured face-to-face induction visits with trial-specific training for investigators, their support team and research staff at site initiation. Written trial-specific materials were shared in printed and electronic forms. Local study centre staff were given access to multiple online training resources, which were also made available to staff at continuing care sites, together with other supporting material [URL: www.npeu.ox.ac.uk/neoclear (accessed 14 June 2022)]. A 24-hour telephone line was available to help with randomisation issues, including the option of telephone randomisation in the case of unresolvable information technology issues at the recruiting site.

Trial monitoring continued throughout the recruitment phase with review of investigator site files, the delegation logs, certificates of good clinical practice and the research curricula vitae of the local site staff. Trial data management was performed to the standards of the NPEU CTU’s SOPs, following prespecified schedules. Data from the recruiting centres were closely monitored for quality assurance. The monitoring included regular review of consent forms and reassurance of correctly assessed participant eligibility. Additional validation checks of data were carried out in intervals. Whenever needed, queries were issued to study centres for resolution. In accordance with best data management practice, final data validation checks were carried out before the database lock. Open questions were sought to be resolved through discussion with the site PI and their local teams.

The independent DMC was continuously informed of study progress in the form of written reports and at regularly scheduled physical DMC meetings. Incidents where the study statistician might raise concerns regarding the data quality were reported to the study data management staff, who would query these incidents if deemed appropriate, or followed up by further routine data validation checks, or both. DMC meetings were also used to provide external independent review of summary data, where necessary.

Summary of changes to the study protocol

A summary of the changes made to the original protocol is presented in Appendix 7.

Recruitment and retention

The NeoCLEAR trial recruited infants from August 2018 to August 2020. COVID-19 regulations forced a pause of recruitment between March and July 2020. Trial activity restarted without difficulty thereafter until completion. The trial recruited in 21 hospitals. Infant and parent characteristics at trial entry and at first LP are detailed in Tables 1 and 2, respectively. See Table 27 for the list of participating centres.

Participants

From August 2018 to August 2020, 1082 participants from 21 centres were randomised to the principal comparisons: (1) sitting position (n = 546) compared with lying position (n = 536) and (2) ESR (n = 549) compared with LSR (n = 533).

Demographic and other baseline characteristics

The baseline patient characteristics at trial entry are presented in Table 1, and patient characteristics at first LP are presented in Table 2. Characteristics were well balanced throughout the four groups. Most patients were term-born infants and most had a working weight of ≥ 2500 g. We saw a slight male predominance, which was consistent between groups. The indications for LP were as per modified NICE guidance, i.e. predominantly exclusion of meningitis, suspected either because of patient and maternal history, or laboratory parameters [i.e. raised C-reactive protein (CRP)]. 12

Outcomes

A total of 1079 participants had a first LP, of whom 166 (15.4%) had a second (each of these LP ‘procedures’ involved one or more ‘attempts’) (Figure 2). Nine participants were withdrawn during the trial, but in only one case was consent withdrawn before data collection for the primary outcome. Three infants did not undergo LP, and for two infants the consent form was missing. Overall, six infants were excluded post randomisation, leaving 1076 infants for the final (modified intention-to-treat) analysis. All infants were followed up until discharge.

FIGURE 2.

Flow of participants by principal comparison: sitting position vs. lying position. ITT, intention to treat.

Part A: principal comparison – sitting position compared with lying position

Subgroup analyses of the primary outcome: sitting position compared with lying position

The prespecified subgroup analysis of the primary outcome overall confirms the statistically and clinically significant advantages of performing neonatal LP in the sitting position (success rate 63.7%) over the lying position (success rate 57.6%).

The effect of sitting position was consistent across all subgroups of CGA and weight. The benefit of sitting was less clear in the small subgroup of infants enrolled at ≥ 3 days old. As expected, this subgroup of infants had a lower gestational age at birth and a lower birthweight and, therefore, may have been more likely to require respiratory support at the time of LP. It is also reassuring that the results for infants with a working weight of < 2500 g and a CGA of 27⁺⁰ to 36⁺⁶ weeks were consistent with the overall findings (Figure 3).

FIGURE 3.

Subgroup analysis forest plot: principal comparison – sitting position vs. lying position. Adjusted for timing of stylet removal allocation, gestational age at randomisation and centre (as a random effect).

Part B: principal comparison – early stylet removal compared with late stylet removal

The flow chart of participants by principal comparison for the interventions with ESR compared with LSR is shown in Figure 4. The infants’ characteristics at trial entry (Table 9) and at first LP were comparable, with no relevant differences between the groups (Table 10).

FIGURE 4.

Flow of participants by principal comparison: ESR vs. LSR. ITT, intention to treat.

Part C: multiarm analysis – sitting position plus early stylet removal, sitting position plus late stylet removal, lying position plus early stylet removal and lying position plus late stylet removal

Figure 5 illustrates the multiarm analysis flow of participants by randomised group. Table 17 summarises the parent and infant characteristics at trial entry. Table 18 summarises the infants’ characteristics at first LP. The groups were well matched, and no significant interaction between infant position and timing of stylet removal was detected (p = 0.136) (Table 19). Multiarm baseline characteristics and analyses did not reveal any new obvious differences (only those previously described for sitting position vs. lying position).

FIGURE 5.

Flow of participants by randomised group: multiarm analysis. ITT, intention to treat.

Serious adverse events

Four SAEs were reported during the trial (Table 24). Three were deemed unrelated to LP. One infant, from the sitting plus LSR group, developed a scrotal haematoma 2 days after LP. The infant did not undergo further investigations to identify a cause for this, and so a relationship with the LP could not be ruled out. Therefore, this event was deemed ‘possibly related’.

Introduction

In this brief chapter, we report the economic evaluation in regard to resource consumption, as part of the NeoCLEAR trial. The objective of the economic evaluation was to compare the relative cost-effectiveness of the two interventions in terms of the duration of intravenous antibiotic therapy and length of hospital stay.

Methods

We documented the number of infants receiving antibiotics during trial and the median (IQR) duration of antibiotic course from trial entry to discharge home (in days), and calculated the number (range) of missing data. Similarly, we documented the surviving infants and their median (IQR) length of stay in hospital from trial entry until discharge home (in days).

Outcomes, results and economic analysis

We did not find statistically significant differences in resource consumption between groups (see Table 21). Likewise, we did not find statistically significant differences for the number of infants receiving antibiotics during trial, the duration of antibiotic course from trial entry to discharge home (in days), survival rates, total length of stay in hospital or the number of infants with missing data.

Summary of main findings

The NeoCLEAR trial is, to the best of our knowledge, the first adequately powered RCT comparing different LP techniques in newborns. Sitting position was superior to lying for achieving a successful first LP, with an absolute risk difference of 6.1% and a NNT of 16. CSF samples were more often interpretable, and sitting LP was better tolerated by participants, who were objectively less likely to have oxygen desaturations and dips in HR. Infants also showed reduced struggling in sitting position. The timing of stylet removal did not influence LP success.

To the best of our knowledge, this is the largest trial of neonatal LP technique worldwide. The demographic and clinical patient characteristics at baseline were well balanced across the four groups. The median age of patients was 40 weeks (term), with a median working weight of 3500 g. Eighty-seven per cent of infants were term born (i.e. with a CGA of 37–44 weeks). We observed a slight male predominance, which was equally balanced between groups. Similarly, failure to obtain any CSF was also equally distributed between groups. The sitting position resulted in significantly more successful LPs than the lying position. The advantageous effect of the sitting LP position was consistent across all subgroups of CGA and weight. The number of preterm infants was relatively small, especially the number with a CGA of below 32 weeks (n = 22, 2% of the total trial cohort). The NeoCLEAR trial showed that infants aged < 3 days significantly benefited from the sitting position, whereas this benefit was not demonstrated for the small subgroup of infants enrolled ≥ 3 days of life. As might be expected, babies in this subgroup had a lower gestational age at birth and a lower birthweight and were more likely to be on respiratory support.

Our results may, in part, be explained by the anatomical advantages of the sitting position, as described previously in neonates:14,30,31 (1) the intervertebral spaces widen in the sitting position; (2) the CSF space increases at the lowest point of the spinal canal and close to the entry site of the needle; and (3) this position is more comfortable for the baby, as evidenced by the reduced struggling we observed.

The only other RCT32 of the sitting position compared with the lying position involved 168 infants aged < 90 days in a paediatric emergency room setting. Success rates in this RCT32 did not differ significantly depending on position (lying position, 63/82, 77%; sitting position, 61/85, 72%; difference 5.1%, 95% CI −8.2% to 18.3%).

The suggestion of potentially greater LP success rates with ESR was based on studies with non-styletted needles, which have reported 100% success rates,21 and observational studies in which styletted needles were used and timing of stylet removal varied. 13,30 However, the use of unstyletted needles for LP is strongly discouraged, as it bears the risk of iatrogenic intraspinal epidermoid tumour formation. 4,22 As a safe alternative, the technique of ESR was introduced. In a prospective study that included infants aged < 12 weeks, Baxter et al. 13 found that ESR was associated with successful and non-traumatic LP. The NeoCLEAR trial did not replicate this apparent benefit of ESR. Likewise, we did not find a disadvantage of this approach. Therefore, we cannot advise for or against a specific timing of stylet removal. There was no statistically significant interaction between the position and stylet timing comparisons within our 2 × 2 factorial design.

Our safety analysis showed greater physiological stability (i.e. HR and SpO₂) for babies receiving LP in the sitting position than for babies receiving LP in the lying position, and this is similar to earlier reports by Gleason et al. 29 and Weisman et al. 28 Most other safety and secondary outcomes lacked clinical and statistical significance with respect to differences between the groups.

The very basic analysis of resource use showed no difference between techniques. Comprehensively evaluating this topic would require more in-depth study, possibly outside a pragmatic RCT like this. Although there was no difference in hospital days or antibiotic use in our study, it might, indeed, be speculated that because a sitting LP can be performed as quickly and at the same cost as a lying LP, and can result in a higher proportion of diagnostic LPs without evidence of adverse events, sitting LP might be cost-effective. More successful LP translates into less waste of professional time for repeated procedures, less opportunity for harm from additional procedures and greater parent/patient satisfaction.

The NeoCLEAR trial was designed with a 10% improvement in LP success rate as the primary outcome; however, the absolute increase was 6.1%, which is a clinically important finding. The reasons why the 10% margin was not reached should be carefully explored. This finding may, in part, be explained by a substantive level of non-adherence to the allocated procedure among infants randomised to the sitting position. We observed that non-adherence was 10-fold higher in this group than in infants allocated to the lying position. Our analysis shows that non-adherence was driven predominantly by practitioners, and not by patient characteristics. The phenomenon of non-adherence itself requires further analysis, potentially post hoc. However, contrary to a speculative perception held by practitioners (attitudes and opinions were not studied in our trial) that their defaulting to standard technique might improve the outcome of the LP, our results provide evidence that the lying position was not statistically significantly superior in any of the subgroup analyses. Moreover, the observation that sitting LP offers overall practical benefits (i.e. greater success rates, equally quick procedure, greater physiological stability), together with the final statistical analysis, cumulating in an impressive NNT of 16, convincingly demonstrates the superiority of sitting position for neonatal LP. Therefore, taking the other findings into consideration, we believe it seems reasonable to suggest adopting the sitting position as standard of care for 0- to 3-day-old term-born infants with body weights > 2.5 kg.

In conclusion, our results show that the success rate for LPs performed in the sitting position is 6.1% higher than for LPs performed in the lying position (NNT = 16). We found no difference in success rate between ESR and LSR. Compared with other modifications to LP technique (e.g. ultrasound guidance), sitting LP is cost-neutral and easy to learn. The results would be applicable in similar settings worldwide and should promote the sitting technique becoming the standard for neonatal lumbar puncture.

Strengths

The NeoCLEAR trial has several strengths. To the best of our knowledge, it is the largest RCT, to date, investigating two easily practised modifications of Quincke’s traditional LP technique. The results are clear-cut, with the LP success rate 6.1% higher for the sitting position than for the lying position. This finding is in keeping with other studies,40 using ultrasound to demonstrate that sitting position increases the lumbar interspinous distance,40 which may explain why this position is more likely to be successful.

Limitations

Sitting LP was a novel concept for many practitioners, who were introduced to it for the first time during the preparatory training for the trial. The novelty factor and lack of familiarity with the technique (both for the assistant holding the baby and for the practitioner operating the needle) might explain why experienced practitioners were less compliant with sitting allocation on repeated procedures, irrespective of the training they had received at the start of the study. It might, therefore, be speculated that success rates could have been even higher if there had been more experience among the teams with sitting position LPs. This would indicate that there would be benefit in implementing education and training in a sitting LP technique through medical education, from medical school and throughout paediatric training programmes, as well as for other healthcare providers who may be involved in performing a neonatal LP, such as advanced neonatal nurse practitioners.

The NeoCLEAR trial predominantly investigated term-born infants with suspected early-onset sepsis, and only a minority of infants were preterm or ≥ 3 days of life, which may limit the generalisability of our results for all infants. However, we did not find evidence that sitting position was less successful or less safe in infants with lower CGA or lower body weight, which is reassuring.

Blinding of practitioners was not possible, but the primary outcome was based on laboratory tests that were performed blinded to allocation.

The economic analysis was rudimentary, as cost outcomes were pragmatically measured by days spent in hospital and days of receiving antibiotic therapy. Although we found no differences in resource consumption between groups, a more detailed account of costs, including material costs and the cost in nurse and physician time, as well as the emotional expenditure on the part of the parents, seems warranted.

Implications for practice

We believe that the results of the NeoCLEAR trial should be universally applicable by paediatricians and neonatologists worldwide for the population of infants examined in our trial. The results would be applicable in similar settings worldwide and should promote the sitting technique becoming the standard for neonatal LP, especially for 0- to 3-day-old term-born infants with a body weight of > 2.5 kg.

Implications for research

As outlined in Limitations, the NeoCLEAR trial predominantly investigated term-born infants with suspected early-onset sepsis, and only a minority of infants were preterm or older than 3 days of life. Further trials should investigate the optimal position for LP in low and very low birthweight preterm infants, in ventilated patients and in cohorts of infants of all gestations with suspected late-onset sepsis. A detailed analysis of resource use and cost savings might also be desirable.

Further research could use ultrasound to investigate whether the anatomical benefits of sitting position relate to wider intervertebral spaces or to a greater lumbar CSF width. In addition, different age groups in infancy warrant further investigations, and the optimal time from sitting the infant up for LP to needle insertion requires further work. Finally, further study is required to answer the question of whether or not the advantages of sitting LP are also found in infants outside the neonatal period, including in infants presenting to emergency departments or children undergoing LP for neurological, oncologic or metabolic investigations.

Independent Trial Steering Committee

The independent members of the TSC are listed in Table 25. These were Professor Paul Heath (chairperson), Dr Chris Gale (vice chairperson), Professor Richard Elmsley, Ms Marie Hubbard, Dr Julie Nelson (PPI) and Dr Bill Yoxall. The non-independent members are Associate Professor Charles Roehr and Professor Edmund Juszczak.

The investigator ensured that this study was conducted in accordance with the principles of the Declaration of Helsinki.

Independent Data Monitoring Committee

The independent members of the DMC are listed in Table 26. These were Dr David Sweet (chairperson), Professor Kate Costeloe and Professor Siobhan Creanor.

General acknowledgements

We are grateful to the parents of participating infants and staff and carers in recruiting and continuing care sites (Tables 27 and 28, respectively). We thank the members of the independent DMC and the TSC and the administrative and support colleagues at the NPEU CTU. We thank the funder, NIHR, and the sponsor, University of Oxford, for their support.

The authors thank the charities SSNAP and Bliss for their contributions to parent-facing documents.

We would like to thank Fiona Goodgame (Trial Manager, NPEU), Mariko Nakahara (Trial Manager, NPEU) and Ann Kennedy (Manager, NPEU) for contributing to logistical aspects of the trial.

The PMG included NPEU staff members Richard Welsh (Programmer, NPEU) and Christina Cole (Senior Trail manager, NPEU). In addition, Christina Cole, Jennifer Shilton-Osborne (Personal Assistant, NPEU) and Ann Kennedy assisted with completing the final report.

The main results paper was published on 29 November, 2022 in The Lancet Child & Adolescent Health. 41

Contributions of authors

Charles C Roehr (https://orcid.org/0000-0001-7965-4637) (Consultant Neonatal Medicine and Clinical Director, NPEU) conceived the idea for the study and wrote the initial protocol. He was the chief investigator and was responsible for all aspects of the study, including preparation and submission of the grant application, securing funding, regulatory approvals, project management, data collection and preparation of the manuscript for publication, including drafting the final report.

Andrew SJ Marshall (https://orcid.org/0000-0003-0529-0425) (Consultant Paediatrician, Paediatrics) conceived the idea for the study and wrote the initial protocol. He was the trial lead clinician and drafted the final report.

Alexandra Scrivens (https://orcid.org/0000-0002-3429-8007) (CRF, Neonatal Medicine) was the trial CRF and provided valuable input in developing trial training materials and was instrumental in facilitating the clinical conduct of the trial.

Manish Sadarangani (https://orcid.org/0000-0002-9985-6452) (Consultant, Paediatric Infectious Diseases) provided expertise in study design and methodological advice, and assisted with completing the final report.

Rachel Williams (https://orcid.org/0000-0002-5872-1690) (Trial Manager, NPEU) was the main trial co-ordinator and helped to design and refine the study and to develop the protocol.

Jean Yong (https://orcid.org/0000-0001-7436-6644) (Trainee, Neonatal Medicine) provided clinical input.

Louise Linsell (https://orcid.org/0000-0003-3205-6511) (Statistician, NPEU) was a trial statistician, provided expertise in study design and methodological advice, and assisted with completing the final report.

Virginia Chiocchia (https://orcid.org/0000-0002-6196-3308) (Statistician, NPEU) provided expertise in study design and methodological advice.

Jennifer L Bell (https://orcid.org/0000-0001-9571-0715) (Statistician, NPEU) was a trial statistician, provided expertise in study design and methodological advice, and assisted with completing the final report.

Caz Stokes (https://orcid.org/0000-0003-4945-2895) (lay individual, Public and Patient Panel) provided public and patient input.

Patricia Santhanadass (https://orcid.org/0000-0002-4464-7311) (lay individual, Public and Patient Panel) provided public and patient input.

Ian Nicoll (https://orcid.org/0000-0003-2778-0660) (Advanced Neonatal Nurse Practitioner) provided input in developing trial training materials and gave valuable clinical input.

Eleri Adams (https://orcid.org/0000-0002-7034-1287) (Consultant, Neonatal Medicine) provided clinical input.

Andrew King (https://orcid.org/0000-0001-7175-2718) (Senior Programmer, NPEU) was a member of the PMG and assisted with completing the final report.

David Murray (https://orcid.org/0000-0001-9010-2905) (Programmer, NPEU) was a member of the PMG.

Ursula Bowler (https://orcid.org/0000-0002-0100-0155) (Senior Operations Manager, NPEU) provided expertise in study design and methodological advice.

Kayleigh Stanbury (https://orcid.org/0000-0002-8726-2411) (Operations Manager, NPEU) was a member of the PMG, contributed to logistical aspects of the trial and assisted with completing the final report.

Edmund Juszczak (https://orcid.org/0000-0001-5500-2247) (Statistician, Director NPEU, Clinical Trails Unit) provided expertise in study design and methodological advice.

Data-sharing statement

All data requests should be submitted to the corresponding author for consideration. Access to data may be granted following review.

This work uses data provided by patients and collected by the NHS as part of their care and support. Using patient data is vital to improve health and care for everyone. There is huge potential to make better use of information from people’s patient records, to understand more about disease, develop new treatments, monitor safety, and plan NHS services. Patient data should be kept safe and secure, to protect everyone’s privacy, and it’s important that there are safeguards to make sure that it is stored and used responsibly. Everyone should be able to find out about how patient data are used. #datasaveslives You can find out more about the background to this citation here: https://understandingpatientdata.org.uk/data-citation.

Disclaimers

This report presents independent research funded by the National Institute for Health and Care Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, the HTA programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, the HTA programme or the Department of Health and Social Care.