Low incidence of postdural puncture headache further reduced with atraumatic spinal needle: a retrospective cohort study

Abstract

Objective:

To evaluate the incidence of postdural puncture headache (PDPH) in a predominantly pediatric sample before and after a transition from conventional to atraumatic spinal needles.

Study Design:

In this retrospective cohort study, we analyzed data from 1,059 lumbar puncture (LP) procedures in 181 individuals enrolled in NIH Clinical Center research protocols. Multivariate logistic regression was used to evaluate the association between PDPH and spinal needle type after adjusting for patient age, sex and BMI. A random effect of participant was used to accommodate repeated observations.

Results:

The median age at time of procedure was 15.3 years. The overall rate of PDPH was 5.1% (54/1,059). With conventional needles and atraumatic needles respectively, the rate of PDPH was 7.7% (43/588) and 2.3% (11/471); (odds ratio 0.32, 95% CI 0.15 - 0.68).

Conclusion

Lumbar puncture for collection of cerebrospinal fluid is an essential and common procedure in pediatric clinical care and research. PDPH is the most common adverse event of the LP procedure. Our data indicate that LP is safe in pediatrics and that use of an atraumatic spinal needle further reduces the risk of PDPH.

Introduction

Lumbar puncture (LP) for collection of cerebrospinal fluid (CSF) is an essential and common procedure in pediatric medicine¹. LP may be performed to diagnose central nervous system (CNS) infections, demyelinating processes or malignancies, to administer medication directly to the CNS, or to obtain CSF for the purposes of clinical research. CSF collection is vital component of clinical and translational research of neurodegenerative disorders, as CSF biomarker analysis contributes to characterization of neurological disease pathology and has an integral role in discovery and development of targeted therapeutics. Collection of CSF via LP is a straightforward procedure, however Institutional Review Boards (IRBs) may deny inclusion of LP for research purposes, especially in the pediatric population, due to perceived risks to the research participant, such as postdural puncture headache (PDPH).

PDPH is the most common adverse event associated with the LP procedure, with published incidences ranging from 6 to 40 percent in adults but incidence in children is largely unknown^2–5. Most cases of PDPH appear within 72 hours following the procedure⁶. PDPH is characteristically positional in nature, increasing in intensity when the patient is in the upright position and with some relief when the patient is in the supine position. In infants or non-verbal children, PDPH may be observed as irritability, preference for the supine position and vomiting. While patient-specific factors such as female sex, younger age, and low body mass index (BMI) may be associated with increased risk of PDPH⁷, studies indicate that technical elements of the procedure such as needle gauge and type may be most predictive.⁸ Unfortunately, however, these data are from adult samples and there is a dearth of published research in pediatric patients.

Conventional (cutting-type) needles result in a larger hole in the dura fibers, while atraumatic (pencil-point) spinal needles separate, rather than tear the fibers, creating a smaller hole and resulting in less CSF leakage³. Numerous randomized controlled trials, a Cochrane review, and a meta-analysis support use of atraumatic needle in adults, but limitations including heterogeneity in analysis variables led to the conclusion that further research was needed⁹.

Logistical barriers also exist to the widespread use of atraumatic needles in clinical specialties outside of Anesthesiology,¹⁰ including adequate training with new equipment and overcoming resistance to change¹¹. Prepackaged LP trays typically contain a conventional needle and extra effort is required to have a separate atraumatic needle available for the procedure. Other barriers include provider access to atraumatic needles and the perceived higher cost of stocking atraumatic needles in hospitals. A review of cost-effectiveness studies suggests that although atraumatic needles carry a higher per-unit cost, reducing PDPH results in cost savings¹². As the treatment of PDPH ranges from outpatient bedrest to prolonged hospital stays or readmission for intravenous fluids and epidural blood patch,¹³ widespread use of atraumatic needles could confer savings to the US healthcare system on the order of millions of dollars per year¹⁴.

In this study we examine the incidence of PDPH in over 1,000 procedures performed in our pediatric-focused clinical research protocols over a span of 14 years, prior to and following a transition from conventional spinal needles to the nearly exclusive use of atraumatic needles. Using this unique dataset, we were able to account for many of the uncontrolled variables mentioned in the extant literature. Here we examine whether use of an atraumatic spinal needle in LP procedures decreases the risk of PDPH in a sample of pediatric clinical research participants.

Methods

Study Design

In this retrospective observational cohort study, we extracted patient and procedure-specific data, including age, sex, BMI, LP needle type (conventional or atraumatic), complications encountered during the procedure, operator experience with the LP procedure (expert or intermediate/trainee), and number of LP procedures per individual patient.

Study Population

This study included individuals who underwent an LP in research protocols conducted by our group at the National Institutes of Health (NIH) Clinical Research Center in Bethesda, Maryland between October 2004 and August 2019. Participants or guardians provided written consent and assent was obtained when possible. The investigational LP procedures and data collection were performed with approval by the NICHD Institutional Review Board.

The majority of the sample had a diagnosis of either Niemann-Pick disease, type C1 (NPC1) or Smith-Lemli-Opitz Syndrome (SLOS). NPC1 is an autosomal recessive neurodegenerative lysosomal disease¹⁵. SLOS is an autosomal recessive multiple malformation syndrome caused by an inborn error of cholesterol synthesis¹⁶. The remainder of the sample included individuals with creatine transport deficiency (CTD)¹⁷ or juvenile neuronal ceroid lipofuscinosis (CLN3 disease)¹⁸. Each of these diseases has been well characterized and no specific contraindication for LP or known increased risk for PDPH have been reported. LP was not performed in individuals with disordered bleeding, clotting, increased intracranial pressure, or other standard contraindications to the procedure. A subset of participants with NPC1 underwent serial LPs for intrathecal medication infusion and provided longitudinal data. Prior PDPH did not preclude patients from undergoing future LPs.

LP Procedure

Procedures were performed with the patient in the lateral decubitus position. The fourth lumbar vertebral body was located by palpating the iliac crest and drawing a straight line to the intersection of this point and the midline of the spinal column¹⁹. The needle was inserted with the bevel parallel to the axis of the spine, most commonly at the L4-L5 intervertebral space, with the L3-L4 space as an alternative. Local anesthesia was achieved with local subcutaneous injection of lidocaine. Needle type was obtained from the procedure note in the patient’s medical record. Needle selection was operator dependent, however after November 2015, most LPs were done using an atraumatic needle. Opening pressure was not routinely measured. At completion of CSF collection, the stylet was replaced prior to removing the needle. Following the procedure, patients were asked to remain flat (supine or side-lying) for at least an hour, followed by activity as tolerated. Intravenous (IV) fluids were not routinely administered in association with an LP, however if the patient was sedated for the procedure, they received age and weight-appropriate maintenance IV fluids for the duration of the sedation. Prophylactic analgesics were not routinely given following the procedure.

Procedure complications such as multiple attempts, visible blood in the CSF, and agitation or excessive movement during the procedure were recorded by the operator. Operators were categorized by experience level²⁰; a “Trainee” had no prior experience performing the LP procedure or had performed fewer than ten successful independent LPs, those in the “Intermediate” category had some prior training and experience with performing the procedure but fewer than 80 successful independent LPs, and an “Expert” had specialized training in performing LP (Neurology, Anesthesiology, Emergency Medicine), or at least one year at “Intermediate” level and at least 25 successful independent LPs in a 1-year period.

Definition of PDPH

An event was considered a PDPH if the following criteria were met: symptom onset within 5 days of LP procedure and either observed by a clinician or reported by the patient or family to the medical team. Symptoms indicative of PDPH include positional headache, often accompanied by dizziness, vomiting or visual disturbances and are thought to be caused by irritation of pain-sensitive meningeal structures such as membranes, nerves and blood vessels³. PDPH events were identified via clinical and research documentation, and prescription of IV fluid bolus (not otherwise indicated during these admissions), analgesics or caffeine tablets (ordered specifically for headache during that admission and after the LP occurred) with corresponding nursing documentation of pain (headache).

Data Extraction and Analysis

Data were extracted from the Biomedical Translational Research Information System (BTRIS), an NIH-managed electronic database which houses searchable clinical research data from the Clinical Center and other NIH institutes and centers. When necessary to confirm a PDPH or to collect procedure details from clinical documentation, specific observations were obtained from the electronic medical record. Occasionally height or weight, which are both necessary to compute BMI, were not obtained in the clinic. In this case the closest measurement to the missing data point was imputed, provided that the visits were within six months of each other.

Primary Analysis:

Descriptive statistics are provided for PDPH and non-modifiable risk factors. A generalized linear model assuming a binomial distribution was used to compare the rate of headache between the two needle types, controlling for mean-centered age, mean-centered BMI and sex. A random effect of participant was used to accommodate repeated observations within person.

All analyses were performed in RStudio Version 1.1.463. We follow the current American Statistical Association guidelines regarding reporting of p-values²¹; we avoid the use of categorical “statistical significance” and instead report both raw (uncorrected) p-values and 95% confidence intervals for parameter estimates.

Results

Data from 1,059 LP procedures in 181 individuals were available. Eighty-six (48%) patients in the study had more than one LP; needle type was not constant within participant and varied by procedure. Data were analyzed at the procedure level, so descriptive statistics are provided for all 1,059 procedures (Table 1). Five hundred eighty-eight (55%) procedures in 117 participants were performed with the conventional needle and 471 (45%) procedures in 97 participants were performed with the atraumatic needle. About half (52%) of the procedures were performed in males. The mean age at procedure and BMI were 14.6 ± 8.0 years and 19.6 ± 3.8 kg/M², respectively. These values were similar between needle types. All procedures were performed using a 22 Gauge LP needle, except for two LPs done using a 20G conventional needle, both of which resulted in a PDPH. CSF volume collected was 10 mL or less in all but one procedure (19 mL), which did not result in PDPH. Eight individuals had more than one (2-3) PDPH events, seven after LP with the conventional needle, and one after LP with both needle types.

Table 1.

Patient and Procedure Characteristics

Characteristic	Overall Sample N = 1059	Conventional Needle N = 588	Atraumatic Needle N = 471
Patients, n	181	117	97
Sex, n (%)
Male	549 (52)	308 (52)	241 (51)
Female	510 (48)	280 (48)	230 (49)
Age at time of procedure, years
Mean ± SD	14.6 ± 8.0	14.9 ± 7.9	14.3 ± 8.3
Median	15.3	14.8	16.3
Range	0.1 - 68.1	0.3 - 52.9	0.1 - 68.1
BMI, kg/M2
Mean ± SD	19.6 ± 3.8	19.5 ± 3.9	19.9 ± 3.8
Range	11.9 - 38.0	11.9 - 38	12 - 36.9
Procedure complications, n (%)
Multiple attempts	46 (4.3)	30 (5.1)	16 (3.4)
Blood in CSF	17 (1.6)	10 (1.7)	7 (1.5)
Excessive movement/agitation during LP	6 (0.5)	1 (0.2)	5 (1.0)
Number of procedures per patient
Median	1	2	1
IQR	1 - 4	1 - 4	1 - 2
Range	1 - 72	1 - 35	1 - 52
Operators per type, n
Expert	6	6	4
Intermediate/Trainee	13	13	1
Procedures per operator type, n (%)
Expert	882 (84)	496 (86)	386 (83)
Intermediate/Trainee	162 (16)	83 (14)	79 (17)

A total of 54 PDPH events were documented, for an overall rate of 5.1% (Table 2). When grouped by needle type, there were 43 cases (7.3%) of PDPH in the conventional needle group and 11 cases of PDPH (2.3%) in the atraumatic needle group. Using multivariate logistic regression adjusting for age, sex, and BMI (Table 3), patients undergoing LP with the atraumatic needle had lower odds of PDPH than patients who underwent LP with a conventional needle (OR 0.32, 95% CI 0.15 - 0.68).

Table 2.

Summary of PDPH events

Characteristic	Overall Sample	Conventional Needle	Atraumatic Needle
PDPH, n (%)	54 (5.1)	43 (7.3)	11 (2.3)
Sex, n (%)
Male	35 (65)	28 (65)	7 (64)
Female	19 (35)	15 (35)	4 (36)
Age, years
Mean ± SD	13.5 ± 10.1	12.3 ± 7.2	18.4 ± 17.5
Median	11.8	11.5	14.2
BMI, kg/M2
Mean ± SD	19.2 ± 3.7	19.1 ± 3.5	19.6 ± 4.5
Range	13 - 28.1	13 - 27.4	14.4 - 28.1
PDPH per operator type, n
Expert	48 (88.9)	37 (86)	11 (100)
Intermediate/Trainee	6 (11.1)	6 (14)	0 (0)
PDPH after procedure complications, n (%)
No documented complication	50 (93)	39 (91)	11 (100)
Multiple attempts	3 (5.6)	3 (7.0)	0 (0)
Blood in CSF	1 (1.8)	1 (2.3)	0 (0)
Excessive movement/agitation during LP	1 (1.8)	1 (2.3)	0 (0)
Individuals with multiple PDPH events	8	7	11

¹One individual experienced one PDPH with each needle type

Table 3.

Maximum likelihood estimates for postdural puncture headache

	Variables	Est	SE	OR	95% CI	z value	Pr(>|z|)
Risk factor	Age	0.008	0.03	0.99	0.94, 1.05	0.27	0.78
	BMI	−0.01	0.06	0.99	0.88, 1.11	−0.22	0.83
	Sex	−0.64	0.39	0.53	0.25, 1.14	−1.62	0.11
	Needle	−1.14	0.39	0.32	0.15, 0.68	−2.94	0.003

Secondary analyses examining the effect of operator experience level and procedure complications on PDPH outcome were not performed due to the low number of PDPH events in the atraumatic needle group. The small cell size in these analyses would lead to unstable estimates with extremely wide confidence intervals. Instead, we provide descriptive summaries of this information in Table 2.

Discussion

In this study, we examined 1,059 LP procedures performed by a defined team of operators to elucidate whether the risk of PDPH was decreased when an atraumatic needle was used rather than a conventional needle. Our overall rate of PDPH (5.1%) was lower than that reported in the literature and was further reduced after we transitioned to using an atraumatic needle for the procedure. Our sample included predominately pediatric patients, but patients of all ages were included. Even when controlling for age, sex, and BMI, factors known to be associated with PDPH, needle type was the only variable meaningfully associated with the risk of PDPH in our model.

We were unable to statistically determine whether the experience level of the operator or a complication during the procedure, such as multiple attempts or blood in CSF, were associated with the occurrence of PDPH in this sample. Additionally, although the two LPs performed with the 20G needle resulted in PDPH, there were too few LPs with the 20G needle to determine the impact of needle caliber on risk of PDPH in this study. While we intended to include operator experience as a variable in this study, we did not have enough operators in the atraumatic needle group for a reliable statistical analysis, which is a limitation of this study and warrants further investigation.

Anecdotally, all four expert operators who switched to the atraumatic needle described a learning curve in the transition from the conventional to atraumatic needle, citing needle length (3.5 inch atraumatic vs. 2.5 inch conventional needle), difficulty of introducing the tip of the atraumatic needle through the skin layer, and difference in resistance as the needle passes through the tissues and into the subarachnoid space. While these differences between needles caused our operators to report a greater degree of difficulty in obtaining a successful LP, the effect was temporary and after a short transition period, was no longer an issue.

As this is a retrospective study relying on database extraction, only documented observations can be included. Variables such as needle type, gauge and procedure complications are required or typically included in the procedure note, thus did not pose a significant problem in terms of missing data. A randomized controlled study of conventional versus atraumatic needle type and the association with PDPH would have been a stronger study design. Such studies have been published recently in the adult population²² with results that agree with ours.

As mentioned previously, participants in our study population had a diagnosis of a neurologic disease, some with impaired cognition or communication skills. While this may be perceived as a barrier to identifying a PDPH in the more severely affected individuals, we do not believe it to be an issue in this study, as nonverbal signs of pain or discomfort are reliable indicators of PDPH.

The medical team met and examined patients prior to LP (often over several days) and thus had a baseline assessment to which they could compare activity and behavior following the LP. Decreased eating, increased irritability, positional nature of discomfort, altered vital signs and most importantly, parent interpretation of their child’s behavior were all used as indications of prolonged discomfort and signs of PDPH in our very young or nonverbal patients.

Our overall PDPH incidence is on the lower end of that quoted in the literature. This could be due to the predominately pediatric age of the sample, but it is possible that headache occurred after the patient was discharged and was not severe enough for the family to recognize or report it to the medical team. Every attempt to identify events was made, including database search, chart review, review of adverse event logs and provider recall, thus likely decreasing the likelihood that the true number of PDPH in this study were underestimated.

This study suggests that lumbar puncture can be done safely in a pediatric population with a low risk of PDPH following the procedure. Furthermore, the odds of PDPH can be minimized by using an atraumatic spinal needle for the LP. Our results support findings in published studies of adult patients, signifying need for a more widespread practice change in patients of all ages.

Abbreviations:

BMI

Body mass index

CSF

Cerebrospinal fluid

CNS

Central nervous system

IV

Intravenous

LP

Lumbar puncture

PDPH

Postdural puncture headache