Randomized Evaluation of a Positioning Pillow for Lumbar Puncture in Paediatric Hematology-Oncology

Our study was an open, randomized, controlled trial, with two parallel groups assigned to "LP pillow" or "no intervention" conducted in patients undergoing LP.

Participants Children aged 2 to 18 years undergoing an LP were eligible for inclusion. Children who had already participated or used the LP pillow, who had a medical condition (orthopaedic anomaly) contraindicating the use of the LP pillow or whose parents refused consent were excluded.

A permuted-block algorithm was used for randomization, with participants stratified according to centre. Concealed allocation was achieved by a phone call to the coordination centre after eligibility check and baseline data collection.

Intervention The LP pillow was made of polyethylene microcellular foam, coated with rubber to facilitate decontamination. It was placed on the thighs of the child sitting with the trunk leaning forward. This position ensures maximum lumbar flexion whereas resting the trunk on the pillow allows relaxation of paravertebral muscles. The axis of the body and the spinal column were maintained perfectly symmetrical in the sagittal plane (see illustration Fig 1).

The pillow included side supports for the head; the face remained uncovered so that the child could breathe and speak comfortably; this also facilitated nitrous oxide administration. Props supporting the patient's arms maintained the cervical column in a neutral position. An opening allowed touching the child's hands and forearms. Splints at the bottom side immobilized the child's thighs to ensure effective stabilization.

Four sizes were available corresponding approximately to the following age ranges: 2-6 years, 6-10 years, 10-15 and 15-18 years. Two sets were available for the study, one per centre.

Outcomes Primary outcome was the technical quality of the LP, classified as success or failure. A successful LP is defined as one that achieves its purpose (sampling and/or treatment) at the first attempt, without visible haemorrhage and with RBC < 50/mm3 in the cerebrospinal fluid (CSF) sample (biochemical analysis).

Secondary outcomes included: child's pain, evaluated by self-administered visual analogical scales (VAS) for children over 6 years of age; parents' and caregivers' perception of the child's pain; satisfaction of the children, parents, caregivers and physician performing the LP assessed using verbal scoring; child cooperation rated with the "LeBaron Scale" (Lebaron 1984). Other outcomes included the number of attending persons (parents and/ or caregivers); duration of positioning and procedure; incidence, symptoms and duration of post-LP syndrome assessed 48 hours after LP by phone interview.

Data collection included anaesthetic, sedative and analgesic drugs used; needle size (19G, 20G, 22G); general status of the child evaluated with Karnofsky or Lansky score (Lansky 1985); platelet count; number of previous LP; date of and satisfaction with the last LP (verbal scale), practitioner experience in performing LP, presence of the parents, aim of the LP: diagnosis, therapeutic monitoring, treatment injection; amount of CSF removed (drops).

The study protocol was approved by the institutional review board of Lyon A - Hôtel-Dieu on June 8th 2004, and was conducted in accordance with the Helsinki Declaration. All parents gave written informed consent before participation of their child. Children were asked to give written consent when fully able to understand the proposed procedure.

Data were managed using CLININFO S.A. (France) software and services for clinical trial data management and quality control.

Sample size and statistical analysis The protocol initially included a total of 80 children. The sample size calculation was based on a rate of success of 70% in the control group and 95% in the pillow group, with 80%-power and 0.05 two-sided significance level (N Query® software). Based on the rate of success in the control group after inclusion of the 40th child, the sample size was re-estimated and increased to 124 children (62 per group) and the inclusion period was extended from 12 to 24 months. Comparisons between groups were performed using non parametric methods, i.e. the Wilcoxon rank test or the Fisher's exact test, depending on the nature of the variable. A logistic regression model was fitted to the platelet count and the number of LP prior to entering the study. An exploratory subgroup analysis was performed in children over 6 years of age. Comparisons with p-value less than 0.05 were considered significant. The STATA 9.2 (SatatCorp 2005. Stata Statistical Software: Release 9.0 College Station, TX: Satat Corporation) software was used to perform statistical analyses.