European Journal of Neurology 2014, 21: 305–311
doi:10.1111/ene.12307
Change practice now! Using atraumatic needles to prevent post lumbar puncture headache A. Davisa,b, R. Dobsona,b, S. Kaniniaa,b, M. Espasandina,b, A. Berga,b, G. Giovannonia,b and K. Schmierera,b a
Blizard Institute, Barts and the London School of Medicine and Dentistry, Queen Mary University of London, London; and bBarts Health NHS Trust, The Royal London Hospital, London, UK
Keywords:
atraumatic needle, needle design, post lumbar puncture headache Received 7 August 2013 Accepted 17 October 2013
Background and purpose: Lumbar puncture (LP) is a key diagnostic procedure in medicine. Post lumbar puncture headache (PLPHA) is a well recognized complication of LP. Evidence suggests that using atraumatic needles for diagnostic LP (ATNLP) reduces risk of PLPHA. However, clinicians in Europe and the USA routinely use traumatic needles for diagnostic LP (TNLP). The occurrence of PLPHA following ATNLP and TNLP was compared in a clinical setting. Further, a survey was performed exploring use of ATNLP amongst UK neurologists. Methods: Service development study. Patients were followed up 2 and 7 days after LP using blinded telephone assessment. A questionnaire was developed to assess use of ATNLP amongst UK neurologists. Frequency, onset, duration and severity of PLPHA were recorded as were use of analgesia, general practitioner consultations, hospital readmissions, days off work due to PLPHA and cost. Neurologists were asked about their familiarity with, and use of, ATNLP. Results: One hundred and nine participants attending the Royal London Hospital were included, and 74 attendees of the Association of British Neurologists 2012 conference completed an on-site questionnaire. ATNLP reduced the rate of PLPHA (27.1% vs. 60.4%; P < 0.01). In those participants who developed PLPHA symptoms were short lived (mean 50 h vs. 94 h, P = 0.02) and less severe after ATNLP. Use of ATNLP led to significant cost savings. Only one in five UK neurologists regularly use ATNLP stating lack of training and availability of atraumatic needles as main reasons. Conclusions: ATNLP significantly reduces the risk of PLPHA. Training is required 3 to facilitate a change from TNLP to ATNLP amongst clinicians.
Introduction Cerebrospinal fluid (CSF) analysis is a key diagnostic procedure in medicine. CSF is usually obtained through lumbar puncture (LP). LP may lead to a number of complications, most commonly to post LP headache (PLPHA) [1]. According to the 3rd edition of the International Headache Classification (ICHD3), PLPHA is defined as a postural headache occurring within 5 days of LP. It presents as a burning, dull or throbbing headache and may radiate to the neck. Associated features include nausea, tinnitus, hyperacusis, photophobia and diplopia (Table 1). PLPHA has also been described following dural Correspondence: Dr Angharad Davis, Blizard Institute, 4 Newark Street, London E1 2AT, UK (tel: +44 20 7882 8605; fax: +44 20 78822180; e-mail [email protected]).
© 2013 The Author(s) European Journal of Neurology © 2013 EFNS
puncture in the context of spinal anaesthesia, myelography and intrathecal administration of drugs [2–4]. The link between damage to the dura, CSF leakage and PLPHA was recognized decades ago [5]. Substantial evidence supports the use of atraumatic needles (ATNs) for diagnostic LP (ATNLP) instead of traumatic needles (TNLP) [1,4,6,7] (Fig. 1). However, in 2001 Birnbach et al. [8] discovered that only 2% of US neurologists used ATNs in clinical practice. The body of evidence that preceded and followed this survey prompted the American Academy of Neurologists (AAN) to take action, and in 2006 the AAN recommended that ATNs were used as standard for diagnostic LP [9]. Subsequent evidence, however, suggests that ATNs remain second choice for neurologists in the USA [6] as well as in Europe [10]. In the UK the interest in PLPHA is so low that no guidance
305
306
A. Davis et al.
Table 1 Diagnostic criteria for post lumbar puncture headache A
B C D
Headache that worsens within 15 min after standing and improves within 15 min after lying, with at least one of the following and fulfilling criteria B, C and D Neck stiffness Tinnitus Hyperacusis Photophobia Nausea Dural puncture has been performed Headache develops within 5 days of dural puncture Headache resolves either* Spontaneously within 1 week Within 48 h after effective treatment of the spinal fluid leak (usually by epidural blood patch)
*In 95% of cases this is so. When headache persists, causation is in doubt. Source: International Classification of Headache Disorders, 3rd edn, 2013.
is currently available that would highlight the importance of needle choice for diagnostic LP. This is not only a failure of the medical community to address an evidently iatrogenic and largely preventable cause of morbidity, but it also hampers efforts to develop new therapies for common chronic and disabling conditions such as multiple sclerosis, where serial investigation of biomarkers in the CSF to assess treatment effects in clinical trials can only be done under the premise that the risk of PLPHA is minimized. A service development audit was therefore performed to address four questions: 1 What is the frequency, duration and severity of PLPHA in patients undergoing diagnostic LP in a routine clinical setting? 2 How does the occurrence of PLPHA following TNLP and ATNLP compare when performed in the clinical setting? 3 Is there an economic benefit to the healthcare system of using ATNLP instead of TNLP?
4 Are British neurologists aware of ATNs and – if yes – do they use them?
Methods This service development study was approved by the Clinical Effectiveness Department at Barts Health NHS Trust (project registration number 827). Patients attending the daycase unit at the Royal London Hospital and short stay inpatients undergoing LP as part of their routine clinical care were approached for inclusion in the study. Patients with idiopathic intracranial hypertension were excluded. LPs were performed by doctors with a minimum of 2 months’ experience in LP. Needles used were either 22G 9 90 mm Quincke (TNLP) or 90 mm 9 22G Sprotte (ATNLP). Initially LP was carried out using TNs, with a mid-point switch to ATNs. Needle bevel direction was standardized for the TNLP group and duration of recumbancy after LP was recorded. According to the ICHD-3 criteria PLPHA develops within 5 days and resolves within 1 week in 95% of cases. Over and above a shortterm follow-up after 2 days patients were therefore interviewed 7 days post LP. Interviews were conducted by an assessor who was unaware of the needle type used (ME or AB). The presence, severity and duration of PLPHA were documented using a standardized questionnaire. Analgesia use, general practitioner (GP) and A&E visits as well as readmission to hospital for inpatient treatment were also recorded. Data analysis was done in two stages: (i) including all patients whose headache fulfilled ICHD-3 criteria as outlined in Table 1 (broad), and (ii) excluding patients who had a potential alternative explanation for headache, such as chronic daily headache (conservative). Given that there was a normal distribution of data, the frequency of PLPHA
Figure 1 Needle design: (a) traumatic (Quincke©) and (b) atraumatic (Sprotte©) needle with (c) introducer used in LP.
© 2013 The Author(s) European Journal of Neurology © 2013 EFNS
Preventing post lumbar puncture headache
between the two groups was compared using Fisher’s exact test, and the severity and duration using an unpaired t-test. Our cost analysis took into account the price of the needle used, the cost of GP consultations, A&E assessments, readmission to hospital and inpatient treatment. A survey was conducted on ATNLP knowledge and use amongst UK neurologists attending the 2012 meeting of the Association of British Neurologists. Delegates were encouraged to fill in an in-house developed questionnaire on a tablet computer. The questionnaire consisted of five questions establishing the frequency of use (or lack thereof) of ATNs in their clinical practice, testing delegates’ knowledge of basic facts regarding the impact of ATNLP versus TNLP on PLPHA, and exploring potential reasons for avoiding use of ATNs (Data S1). Chi-squared tests were applied to test the null hypotheses regarding (i) awareness of ATNs and (ii) frequency of use of ATNs in the delegates’ clinical practice.
Results Post lumbar puncture headache
One hundred and nine patients were enrolled, 52 of whom underwent TNLP and 57 ATNLP. In three patients ATNLP was not successful and TNLP was performed. These three patients were excluded from the analysis. One patient in the TNLP group and four patients in the ATNLP group were followed up by a non-blinded assessor and were therefore also excluded from the analysis. A total of 10 patients had followup outside the 7-day window (six ATNLP, four TNLP). All 10 patients, however, were able to give a reliable headache history and were included in the analysis. Five patients in the ATNLP group were lost to follow-up. A total of 96 patients were therefore included in the final analysis, 48 in each arm. There was no significant difference between groups with respect to age [mean age: TNLP 42 years (range 19– 74 years), ATNLP 41 years (range 17–74 years)] or gender (male:female TN 23:25, ATN 29:19). Sixty-one per cent of patients included in this study were under investigation for suspected multiple sclerosis. Other indications for LP included peripheral neuropathies (7%), cognitive decline (6%), central nervous system infection (5%) and other neurological conditions (21%). Broad analysis including all 96 subjects showed a reduction of any PLPHA incidence (i.e. at either day 2 or 7) following ATNLP compared with TNLP (27.1% vs. 60.4%; P < 0.01). In the conservative © 2013 The Author(s) European Journal of Neurology © 2013 EFNS
307
analysis 15 patients were excluded, leaving 81 patients (38 TNLP, 43 ATNLP). A reduction in the rate of PLPHA in both groups was seen; however, the difference in PLPHA between ATNLP and TNLP remained significant (ATNLP 20.9% vs. TNLP 50.0%, P = 0.01) (Fig. 2). In both broad and conservative analysis, PLPHA duration in patients undergoing TNLP was significantly longer than in patients undergoing ATNLP. In the broad analysis, TNLP mean duration was 94 h, SD 60 h; ATNLP mean duration was 50 h, SD 47 h; P = 0.02. In the conservative analysis, TNLP mean duration was 89 h, SD 53 h; ATNLP mean duration was 41 h, SD 37 h; P = 0.02. In the conservative analysis, six patients who had undergone TNLP had persistent headache on day 7, whereas only one patient complained of PLPHA at day 7 following ATNLP (Table 2). There was a non-significant difference of PLPHA severity (pain scale 1–5) with mean severity in TNLP of 3.1 compared with 2.6 in ATNLP (P = 0.9). The time to onset of PLPHA was significantly shorter in the ATNLP group (mean 5 h, SD 7) compared with the TNLP group (mean 18 h, SD 17 h; P = 0.04) in the conservative analysis. Nineteen patients in the TNLP group required analgesia for headache compared with six patients in the ATNLP group
Figure 2 Incidence of post lumbar puncture headache when using traumatic and atraumatic needles.
Table 2 Duration of post lumbar puncture headache in those participants who developed the syndrome Headache at specified time points At 1h
At day 2
Broad TN 6 28 ATN 1 12 Conservative TN 4 19 ATN 0 9
At 1 h and day 2
At day 7
At day 2 and day 7
At 1 h, day 2, day 7
6 1
13 3
12 2
3 1
4 0
6 1
6 1
1 0
308
A. Davis et al.
Figure 3 Attempt rate over time in ATNLP for operators AD and SK.
Table 3 Cost analysis comparing ATN with TN accounting for cost of PLPHA and needle type
Cost per unit
Figure 4 Effect of PLPHA on work, and contact with healthcare services due to PLPHA.
(P = 0.03). There was no significant difference in the number of attempts needed for successful LP across the timeline of the study, suggesting increased familiarity with the technique over time did not influence the overall outcome of the study (Fig. 3). In the broad analysis, three patients in the TNLP group visited their GP due to PLPHA. A total of six patients required A&E or day unit assessment of their PLPHA following TNLP. Of these, three patients were readmitted for inpatient management of PLPHA, one of them requiring an epidural blood patch. By comparison, only one patient undergoing ATNLP visited their GP due to PLPHA, and there were no A&E attendances or readmissions in the ATNLP group. Using conservative analysis the number of patients attending A&E was significantly greater following TNLP than ATNLP (4 vs. 0, P = 0.04), and whilst six patients had to take time off work following TNLP this was the case for only one patient following ATNLP (P = 0.05). A total of 29.5 days were taken off work following TNLP compared with 4 following ATNLP (Fig. 4). Cost analysis
Whilst the price for ATN (sum total £454.56) was higher than for TN (sum total £164.16), this difference
Needle system GP visit A&E visit Day as inpatient Epidural blood patch Total cost
£9.47 (atraumatic) £3.42 (traumatic) £25 £68 £265 £88
Atraumatic needle group; sum total (n)
Traumatic needle group; sum total (n)
£454.56 (48)
£164.16 (48)
£25 (1) £0 (0) £0 (0)
£75 (3) £408 (6) £2650 (10)
£0 (0)
£88 (1)
£479.56
£3383.16
was far outweighed by the overall cost of PLPHA management in patients developing PLPHA after TNLP (ATN £479.56; TN £3383.16) (Table 3). Survey
Seventy-four neurologists (consultants and trainees from the first year of specialty training onwards) completed our survey. Fifty-eight (78%) were aware of ATNLP (v2 = 23.8, P < 0.001). Of the 74 delegates, 40 (54%) said they never perform ATNLP in their clinical practice (v2 = 53, P < 0.001). Only 12/74 (16%) used ATNLP routinely (Fig. 5a). The main reasons for not using ATNLP were given as lack of training (35%) and availability/higher cost of ATNs (22%). Three respondents said they were not convinced by the evidence in favour of ATNLP over TNLP (Fig. 5b).
Discussion For most people, having an LP is in itself a frightening experience. However, PLPHA is an additional and, in most cases, entirely unnecessary and avoidable iatrogenic pain syndrome. Our study provides robust © 2013 The Author(s) European Journal of Neurology © 2013 EFNS
Preventing post lumbar puncture headache
(a)
(b)
Figure 5 Neurologists survey analysis. (a) How often do you or your team use ATNs for diagnostic LP? (b) What reasons would prompt you not to use ATNs routinely?
evidence that diagnostic LP leads to significant morbidity irrespective of the needle type used for LP. Twenty per cent of people undergoing ATNLP had a, however mild, form of PLPHA. Nevertheless, our data confirm earlier studies indicating that the incidence of PLPHA can be significantly reduced in a routine clinical setting by nearly 60% [1,4]. This reduction in morbidity due to PLPHA can be achieved by a simple switch from TNLP to ATNLP. Even though the rate of PLPHA in patients who had ATNLP in our study was higher than reported previously [1,4,7], our data suggest that PLPHA in ATNLP is of shorter duration and usually does not require any input from healthcare providers. The higher than previously reported rate of PLPHA in both TNLP and ATNLP may be explained by our study design: although unaware of needle type used, patients were clearly aware of the purpose of the study. Patients consented to receive a telephone call ‘where you will be asked if you have developed a headache amongst other questions’. This may have led to greater awareness of headache, and indeed in some patients may have resulted in over-reporting of headache symptoms. Moreover, patients were specifically asked for the presence of a headache, rather than to volunteer any symptoms they may have noted following their LP. Again this may have resulted in over-reporting. Patients were unaware of needle type, although it is possible that in patients with a history of previous TNLP they were able to differentiate a difference in technique. © 2013 The Author(s) European Journal of Neurology © 2013 EFNS
309
Two main hypotheses have been proposed about the pathophysiology of PLPHA [11]. The first suggests a reduction in CSF pressure leading to a downward shift of the brain and subsequent tension on the meninges and other pain-sensitive structures when the individual sits upright. The second hypothesis proposes that the CSF leak leads to compensatory cerebral vasodilation, resulting in PLPHA. It is possible that both mechanisms play a role. The link between damage to the dura, CSF leak and PLPHA was recognized many decades ago, leading to the invention of the ATN [5]. ATNs were designed to separate, rather than cut through, the longitudinal fibres of the dura. By 1951 the needle design had been refined, and a first successful reduction in PLPHA incidence using ATNLP was reported [12]. The needle type is not the only recognized risk factor for PLPHA. Previous studies have shown an increased incidence of PLPHA associated with young age, female gender and low body mass index [13]. Procedure technique may also play a role, with lower rates of PLPHA observed when (i) the needle bevel is inserted parallel to longitudinal dural fibres [14] and (ii) the needle stylet is replaced prior to withdrawing the needle [15]. Bed rest has traditionally been thought to have a preventive role in the development of PLPHA; however, evidence supporting this hypothesis is lacking [16,17]. Given that dural leak is one of the proposed mechanisms underlying PLPHA it is not surprising that a larger needle diameter has been linked to increased risk of PLPHA [18]. In the majority of cases, PLPHA resolves spontaneously with an estimated remission rate of 72% at 5 days [18,19]. However, such a seemingly high resolution rate of an iatrogenic condition should not lead to complacency. Given the importance of LP, and therefore the frequency at which it is being performed on a daily basis in any hospital across the UK, the socioeconomic impact of more than one in four people still suffering from PLPHA 5 days after their LP is massive (Table 3). Whilst simple analgesia and intravenous caffeine have been shown sometimes to reduce the need for more aggressive intervention [18,20], evidence-free measures such as increased fluid intake and bed rest continue to be encouraged [10,16,17]. Only recently a study suggested that clinicians would prefer such ‘established practices’ rather than switching from TNLP to ATNLP [10]. By far the most effective treatment once PLPHA has developed is epidural blood patching with a success rate of up to 98% [18,21]. Epidural patches are not routinely performed by neurologists (who often
310
A. Davis et al.
do diagnostic LP), however, at least in the UK, but rather by anaesthetists, which often results in delays due to the required coordination between medical disciplines. How can a culture of avoidance be turned into a culture of adoption of ATNLP in line with the evidence? The results of our survey suggest that neurologists in the UK are generally aware of ATNs; however, they either avoid using them or are not in the position to use them as their Trust does not provide them. Of the neurologists and trainees completing our survey 35% cited lack of availability and cost as main reasons for not using ATNLP. However, in one study a decision analytical model predicted $10.4 million savings to the US healthcare system should ATNLP replace TNLP in clinical practice [22]. In our study of nearly 100 patients, savings of £2903 were calculated for using ATNLP instead of TNLP provided not only the price of the needle is taken into account but also the cost for managing PLPHA. The reluctance to use ATNLP is not unique to the UK. A recent study from Denmark detected a similarly low rate of ATNLP amongst neurologists [10]. However, the reasons for this reluctance may vary, for example due to different national training and funding streams. Further research exploring this area is warranted. A further reason not to adopt ATNLP was the lack of training. Whilst this could easily be taken care of, the lack of a significant learning curve in our study suggests that ATNLP can be learned very quickly and does not pose any significant difficulty. Of note, a meta-analysis by Halpern and Preston [23] demonstrated a higher incidence of operator difficulty when using TNs compared with ATNs, suggesting that ATNLP should be easy to adopt. Nevertheless, this problem should be addressed and training provided. At our institution an online demonstration video is being used to help train junior doctors in ATNLP (http://www.youtube.com/watch?v=cfGKRb WvloA). Whilst LP is an essential diagnostic procedure in clinical practice, it is also emerging as a means to collect CSF biomarkers to predict outcome in clinical trials of neuro-inflammatory and neurodegenerative diseases [24]. However, studies requiring patients to undergo serial LP may be threatened by high dropout rates or refused institutional review board approval should modifiable risk factors such as PLPHA not be minimized. The debate about causes and modifiable risk factors of PLPHA has lasted for over a century. Surely it is now time to finally make ATNLP standard clinical practice.
Ethical approval This service development study was approved by the Clinical Effectiveness Department at Barts Health NHS Trust (project registration number 827).
Disclosure of conflicts of interest AD, RD, SK, ME and AB have no conflicts of interest to declare. GG has received research grant support from Bayer-Schering Healthcare, Biogen-Idec, GW Pharma, Merck Serono, Merz, Novartis, Teva and Sanofi-Aventis. GG has received personal compensation for participating on Advisory Boards in relation to clinical trial design, trial steering committees and data and safety monitoring committees from Bayer-Schering Healthcare, Biogen-Idec, Eisai, Elan, Fiveprime, Genzyme, Genentech, GSK, Ironwood, Merck-Serono, Novartis, Pfizer, Roche, SanofiAventis, Synthon BV, Teva, UCB Pharma and Vertex Pharmaceuticals. KS is a PI on trials sponsored by Novartis and Roche. He has received speaking honoraria from, and served on advisory boards for, Novartis, Merck-Serono and Merck Inc.
Acknowledgements RD is funded by an Association of British Neurologists/MS Society of Great Britain Clinical Research Fellowship. GG receives grant support from the MRC, National MS Society, MS Society of Great Britain and Northern Ireland, AIMS2CURE and the Roan Charitable Trust. KS is in receipt of an HEFCE Clinical Senior Lectureship and grant support from Barts Charity, the US National MS Society, Novartis and Biogen.
Supporting Information Additional Supporting Information may be found in the online version of this article: Data S1. Association of British Neurologists annual conference, Brighton 2012 Delegates Survey.
References 1. Lavi R, Yarnitsky D, Rowe JM, et al. Standard vs atraumatic Whitacre needle for diagnostic lumbar puncture: a randomized trial. Neurology 2006; 67: 1492–1494. 2. Barker P. Headache after dural puncture. Anaesthesia 1989; 44: 696–697. 3. Flaatten H, Krakene J, Vedeler C. Post-dural puncture related complications after diagnostic lumbar puncture, myelography and spinal anaesthesia. Acta Neurol Scand 1998; 98: 445–451.
© 2013 The Author(s) European Journal of Neurology © 2013 EFNS
Preventing post lumbar puncture headache
4. Lavi R, Rowe JM, Avivi I. Traumatic vs. atraumatic 22 G needle for therapeutic and diagnostic lumbar puncture in the hematologic patient: a prospective clinical trial. Haematologica 2007; 92: 1007–1008. 5. Greene HM. A technique to reduce the incidence of headache following lumbar puncture in ambulatory patients with a plea for more frequent examination of cerebrospinal fluids. Northwest Med J 1923; 22: 240–245. 6. Adrendt K, Demaerschalk BM, Wingerchuk DM, et al. Atraumatic lumbar puncture needles: after all these years, are we still missing the point? Neurologist 2009; 15: 17–20. 7. De Almeida S, Shumaker S, LeBlanc S. Incidence of post-dural headache in research volunteers. Headache 2011; 51: 1503–1510. 8. Birnbach DJ, Kuroda MM, Sternman D, et al. Use of atraumatic needles among neurologists in the United States. Headache 2001; 41: 385–390. 9. Armon C, Evans RW. Addendum to assessment: prevention of post-lumbar puncture headaches. Neurology 2005; 65: 510–512. 10. Stendell L, Fomsgaard JS, Olsen KS. There is room for improvement in the prevention and treatment of headache after lumbar puncture. Dan Med J 2012; 59: A4483. 11. Bezov D, Lipton RB, Ashina S. Post-dural puncture headache: Part I. Diagnosis, epidemiology, etiology and pathophysiology. Headache 2010; 50: 1144–1152. 12. Hart JR, Whitacre RJ. Pencil point needle in prevention of postspinal headache. J Am Med Assoc 1951; 147: 657–658. 13. Lybecker H, Møller JT, May O, et al. Incidence and prediction of postdural puncture headache. A prospective study of 1021 spinal anesthesias. Anesth Analg 1990; 70: 389–394.
© 2013 The Author(s) European Journal of Neurology © 2013 EFNS
311
14. Richman JM, Joe EM, Cohen SR, et al. Bevel direction and postdural puncture headache: a meta-analysis. Neurologist 2006; 12: 224–228. 15. Strupp M, Brandt T. Should one reinsert the stylet during lumbar puncture? N Engl J Med 1997; 336: 1190–1198. 16. Sudlow C, Warlow CP. Posture and fluids for preventing post-dural puncture headache. Cochrane Database Syst Rev 2002; 2: CD001790. 17. Thoennissen J, Herkner H, Lang W, et al. Does residual bed after cervical or lumbar puncture prevent headache? CMAJ 2001; 165: 1311–1316. 18. Turnbull DK, Shepherd DB. Post-dural puncture headache: pathogenesis, prevention and treatment. Br J Anaesth 2003; 91: 718–729. 19. Lybecker H, Djernes M, Schmidt JE. Post dural puncture headache onset, duration, severity and associated symptoms: an analysis of 75 consecutive patients with PDPH. Acta Anaesthesiol Scand 1995; 39: 605–612. 20. Basurto OX, Martinez GL, Sola I, et al. Drug therapy for treating post-dural puncture headache. Cochrane Database Syst Rev 2011; 8: CD007887. 21. Boonmak P, Boonmak S. Epidural blood patching for preventing and treating post-dural puncture headache. Cochrane Database Syst Rev 2010; 1: CD001791. 22. Tung CE, So YT, Lansberg MG. Cost comparison between the atraumatic and cutting lumbar puncture needles. Neurology 2012; 78: 109–113. 23. Halpern S, Preston R. Postdural puncture headache and spinal needle design. Metaanalyses. Anaesthesiology 1994; 81: 1376–1383. 24. Gunnarsson M, Malmestr€ om C, Axelsson M, et al. Axonal damage in relapsing multiple sclerosis is markedly reduced by natalizumab. Ann Neurol 2011; 69: 83–89.
doi:10.1111/ene.12307
Change practice now! Using atraumatic needles to prevent post lumbar puncture headache A. Davisa,b, R. Dobsona,b, S. Kaniniaa,b, M. Espasandina,b, A. Berga,b, G. Giovannonia,b and K. Schmierera,b a
Blizard Institute, Barts and the London School of Medicine and Dentistry, Queen Mary University of London, London; and bBarts Health NHS Trust, The Royal London Hospital, London, UK
Keywords:
atraumatic needle, needle design, post lumbar puncture headache Received 7 August 2013 Accepted 17 October 2013
Background and purpose: Lumbar puncture (LP) is a key diagnostic procedure in medicine. Post lumbar puncture headache (PLPHA) is a well recognized complication of LP. Evidence suggests that using atraumatic needles for diagnostic LP (ATNLP) reduces risk of PLPHA. However, clinicians in Europe and the USA routinely use traumatic needles for diagnostic LP (TNLP). The occurrence of PLPHA following ATNLP and TNLP was compared in a clinical setting. Further, a survey was performed exploring use of ATNLP amongst UK neurologists. Methods: Service development study. Patients were followed up 2 and 7 days after LP using blinded telephone assessment. A questionnaire was developed to assess use of ATNLP amongst UK neurologists. Frequency, onset, duration and severity of PLPHA were recorded as were use of analgesia, general practitioner consultations, hospital readmissions, days off work due to PLPHA and cost. Neurologists were asked about their familiarity with, and use of, ATNLP. Results: One hundred and nine participants attending the Royal London Hospital were included, and 74 attendees of the Association of British Neurologists 2012 conference completed an on-site questionnaire. ATNLP reduced the rate of PLPHA (27.1% vs. 60.4%; P < 0.01). In those participants who developed PLPHA symptoms were short lived (mean 50 h vs. 94 h, P = 0.02) and less severe after ATNLP. Use of ATNLP led to significant cost savings. Only one in five UK neurologists regularly use ATNLP stating lack of training and availability of atraumatic needles as main reasons. Conclusions: ATNLP significantly reduces the risk of PLPHA. Training is required 3 to facilitate a change from TNLP to ATNLP amongst clinicians.
Introduction Cerebrospinal fluid (CSF) analysis is a key diagnostic procedure in medicine. CSF is usually obtained through lumbar puncture (LP). LP may lead to a number of complications, most commonly to post LP headache (PLPHA) [1]. According to the 3rd edition of the International Headache Classification (ICHD3), PLPHA is defined as a postural headache occurring within 5 days of LP. It presents as a burning, dull or throbbing headache and may radiate to the neck. Associated features include nausea, tinnitus, hyperacusis, photophobia and diplopia (Table 1). PLPHA has also been described following dural Correspondence: Dr Angharad Davis, Blizard Institute, 4 Newark Street, London E1 2AT, UK (tel: +44 20 7882 8605; fax: +44 20 78822180; e-mail [email protected]).
© 2013 The Author(s) European Journal of Neurology © 2013 EFNS
puncture in the context of spinal anaesthesia, myelography and intrathecal administration of drugs [2–4]. The link between damage to the dura, CSF leakage and PLPHA was recognized decades ago [5]. Substantial evidence supports the use of atraumatic needles (ATNs) for diagnostic LP (ATNLP) instead of traumatic needles (TNLP) [1,4,6,7] (Fig. 1). However, in 2001 Birnbach et al. [8] discovered that only 2% of US neurologists used ATNs in clinical practice. The body of evidence that preceded and followed this survey prompted the American Academy of Neurologists (AAN) to take action, and in 2006 the AAN recommended that ATNs were used as standard for diagnostic LP [9]. Subsequent evidence, however, suggests that ATNs remain second choice for neurologists in the USA [6] as well as in Europe [10]. In the UK the interest in PLPHA is so low that no guidance
305
306
A. Davis et al.
Table 1 Diagnostic criteria for post lumbar puncture headache A
B C D
Headache that worsens within 15 min after standing and improves within 15 min after lying, with at least one of the following and fulfilling criteria B, C and D Neck stiffness Tinnitus Hyperacusis Photophobia Nausea Dural puncture has been performed Headache develops within 5 days of dural puncture Headache resolves either* Spontaneously within 1 week Within 48 h after effective treatment of the spinal fluid leak (usually by epidural blood patch)
*In 95% of cases this is so. When headache persists, causation is in doubt. Source: International Classification of Headache Disorders, 3rd edn, 2013.
is currently available that would highlight the importance of needle choice for diagnostic LP. This is not only a failure of the medical community to address an evidently iatrogenic and largely preventable cause of morbidity, but it also hampers efforts to develop new therapies for common chronic and disabling conditions such as multiple sclerosis, where serial investigation of biomarkers in the CSF to assess treatment effects in clinical trials can only be done under the premise that the risk of PLPHA is minimized. A service development audit was therefore performed to address four questions: 1 What is the frequency, duration and severity of PLPHA in patients undergoing diagnostic LP in a routine clinical setting? 2 How does the occurrence of PLPHA following TNLP and ATNLP compare when performed in the clinical setting? 3 Is there an economic benefit to the healthcare system of using ATNLP instead of TNLP?
4 Are British neurologists aware of ATNs and – if yes – do they use them?
Methods This service development study was approved by the Clinical Effectiveness Department at Barts Health NHS Trust (project registration number 827). Patients attending the daycase unit at the Royal London Hospital and short stay inpatients undergoing LP as part of their routine clinical care were approached for inclusion in the study. Patients with idiopathic intracranial hypertension were excluded. LPs were performed by doctors with a minimum of 2 months’ experience in LP. Needles used were either 22G 9 90 mm Quincke (TNLP) or 90 mm 9 22G Sprotte (ATNLP). Initially LP was carried out using TNs, with a mid-point switch to ATNs. Needle bevel direction was standardized for the TNLP group and duration of recumbancy after LP was recorded. According to the ICHD-3 criteria PLPHA develops within 5 days and resolves within 1 week in 95% of cases. Over and above a shortterm follow-up after 2 days patients were therefore interviewed 7 days post LP. Interviews were conducted by an assessor who was unaware of the needle type used (ME or AB). The presence, severity and duration of PLPHA were documented using a standardized questionnaire. Analgesia use, general practitioner (GP) and A&E visits as well as readmission to hospital for inpatient treatment were also recorded. Data analysis was done in two stages: (i) including all patients whose headache fulfilled ICHD-3 criteria as outlined in Table 1 (broad), and (ii) excluding patients who had a potential alternative explanation for headache, such as chronic daily headache (conservative). Given that there was a normal distribution of data, the frequency of PLPHA
Figure 1 Needle design: (a) traumatic (Quincke©) and (b) atraumatic (Sprotte©) needle with (c) introducer used in LP.
© 2013 The Author(s) European Journal of Neurology © 2013 EFNS
Preventing post lumbar puncture headache
between the two groups was compared using Fisher’s exact test, and the severity and duration using an unpaired t-test. Our cost analysis took into account the price of the needle used, the cost of GP consultations, A&E assessments, readmission to hospital and inpatient treatment. A survey was conducted on ATNLP knowledge and use amongst UK neurologists attending the 2012 meeting of the Association of British Neurologists. Delegates were encouraged to fill in an in-house developed questionnaire on a tablet computer. The questionnaire consisted of five questions establishing the frequency of use (or lack thereof) of ATNs in their clinical practice, testing delegates’ knowledge of basic facts regarding the impact of ATNLP versus TNLP on PLPHA, and exploring potential reasons for avoiding use of ATNs (Data S1). Chi-squared tests were applied to test the null hypotheses regarding (i) awareness of ATNs and (ii) frequency of use of ATNs in the delegates’ clinical practice.
Results Post lumbar puncture headache
One hundred and nine patients were enrolled, 52 of whom underwent TNLP and 57 ATNLP. In three patients ATNLP was not successful and TNLP was performed. These three patients were excluded from the analysis. One patient in the TNLP group and four patients in the ATNLP group were followed up by a non-blinded assessor and were therefore also excluded from the analysis. A total of 10 patients had followup outside the 7-day window (six ATNLP, four TNLP). All 10 patients, however, were able to give a reliable headache history and were included in the analysis. Five patients in the ATNLP group were lost to follow-up. A total of 96 patients were therefore included in the final analysis, 48 in each arm. There was no significant difference between groups with respect to age [mean age: TNLP 42 years (range 19– 74 years), ATNLP 41 years (range 17–74 years)] or gender (male:female TN 23:25, ATN 29:19). Sixty-one per cent of patients included in this study were under investigation for suspected multiple sclerosis. Other indications for LP included peripheral neuropathies (7%), cognitive decline (6%), central nervous system infection (5%) and other neurological conditions (21%). Broad analysis including all 96 subjects showed a reduction of any PLPHA incidence (i.e. at either day 2 or 7) following ATNLP compared with TNLP (27.1% vs. 60.4%; P < 0.01). In the conservative © 2013 The Author(s) European Journal of Neurology © 2013 EFNS
307
analysis 15 patients were excluded, leaving 81 patients (38 TNLP, 43 ATNLP). A reduction in the rate of PLPHA in both groups was seen; however, the difference in PLPHA between ATNLP and TNLP remained significant (ATNLP 20.9% vs. TNLP 50.0%, P = 0.01) (Fig. 2). In both broad and conservative analysis, PLPHA duration in patients undergoing TNLP was significantly longer than in patients undergoing ATNLP. In the broad analysis, TNLP mean duration was 94 h, SD 60 h; ATNLP mean duration was 50 h, SD 47 h; P = 0.02. In the conservative analysis, TNLP mean duration was 89 h, SD 53 h; ATNLP mean duration was 41 h, SD 37 h; P = 0.02. In the conservative analysis, six patients who had undergone TNLP had persistent headache on day 7, whereas only one patient complained of PLPHA at day 7 following ATNLP (Table 2). There was a non-significant difference of PLPHA severity (pain scale 1–5) with mean severity in TNLP of 3.1 compared with 2.6 in ATNLP (P = 0.9). The time to onset of PLPHA was significantly shorter in the ATNLP group (mean 5 h, SD 7) compared with the TNLP group (mean 18 h, SD 17 h; P = 0.04) in the conservative analysis. Nineteen patients in the TNLP group required analgesia for headache compared with six patients in the ATNLP group
Figure 2 Incidence of post lumbar puncture headache when using traumatic and atraumatic needles.
Table 2 Duration of post lumbar puncture headache in those participants who developed the syndrome Headache at specified time points At 1h
At day 2
Broad TN 6 28 ATN 1 12 Conservative TN 4 19 ATN 0 9
At 1 h and day 2
At day 7
At day 2 and day 7
At 1 h, day 2, day 7
6 1
13 3
12 2
3 1
4 0
6 1
6 1
1 0
308
A. Davis et al.
Figure 3 Attempt rate over time in ATNLP for operators AD and SK.
Table 3 Cost analysis comparing ATN with TN accounting for cost of PLPHA and needle type
Cost per unit
Figure 4 Effect of PLPHA on work, and contact with healthcare services due to PLPHA.
(P = 0.03). There was no significant difference in the number of attempts needed for successful LP across the timeline of the study, suggesting increased familiarity with the technique over time did not influence the overall outcome of the study (Fig. 3). In the broad analysis, three patients in the TNLP group visited their GP due to PLPHA. A total of six patients required A&E or day unit assessment of their PLPHA following TNLP. Of these, three patients were readmitted for inpatient management of PLPHA, one of them requiring an epidural blood patch. By comparison, only one patient undergoing ATNLP visited their GP due to PLPHA, and there were no A&E attendances or readmissions in the ATNLP group. Using conservative analysis the number of patients attending A&E was significantly greater following TNLP than ATNLP (4 vs. 0, P = 0.04), and whilst six patients had to take time off work following TNLP this was the case for only one patient following ATNLP (P = 0.05). A total of 29.5 days were taken off work following TNLP compared with 4 following ATNLP (Fig. 4). Cost analysis
Whilst the price for ATN (sum total £454.56) was higher than for TN (sum total £164.16), this difference
Needle system GP visit A&E visit Day as inpatient Epidural blood patch Total cost
£9.47 (atraumatic) £3.42 (traumatic) £25 £68 £265 £88
Atraumatic needle group; sum total (n)
Traumatic needle group; sum total (n)
£454.56 (48)
£164.16 (48)
£25 (1) £0 (0) £0 (0)
£75 (3) £408 (6) £2650 (10)
£0 (0)
£88 (1)
£479.56
£3383.16
was far outweighed by the overall cost of PLPHA management in patients developing PLPHA after TNLP (ATN £479.56; TN £3383.16) (Table 3). Survey
Seventy-four neurologists (consultants and trainees from the first year of specialty training onwards) completed our survey. Fifty-eight (78%) were aware of ATNLP (v2 = 23.8, P < 0.001). Of the 74 delegates, 40 (54%) said they never perform ATNLP in their clinical practice (v2 = 53, P < 0.001). Only 12/74 (16%) used ATNLP routinely (Fig. 5a). The main reasons for not using ATNLP were given as lack of training (35%) and availability/higher cost of ATNs (22%). Three respondents said they were not convinced by the evidence in favour of ATNLP over TNLP (Fig. 5b).
Discussion For most people, having an LP is in itself a frightening experience. However, PLPHA is an additional and, in most cases, entirely unnecessary and avoidable iatrogenic pain syndrome. Our study provides robust © 2013 The Author(s) European Journal of Neurology © 2013 EFNS
Preventing post lumbar puncture headache
(a)
(b)
Figure 5 Neurologists survey analysis. (a) How often do you or your team use ATNs for diagnostic LP? (b) What reasons would prompt you not to use ATNs routinely?
evidence that diagnostic LP leads to significant morbidity irrespective of the needle type used for LP. Twenty per cent of people undergoing ATNLP had a, however mild, form of PLPHA. Nevertheless, our data confirm earlier studies indicating that the incidence of PLPHA can be significantly reduced in a routine clinical setting by nearly 60% [1,4]. This reduction in morbidity due to PLPHA can be achieved by a simple switch from TNLP to ATNLP. Even though the rate of PLPHA in patients who had ATNLP in our study was higher than reported previously [1,4,7], our data suggest that PLPHA in ATNLP is of shorter duration and usually does not require any input from healthcare providers. The higher than previously reported rate of PLPHA in both TNLP and ATNLP may be explained by our study design: although unaware of needle type used, patients were clearly aware of the purpose of the study. Patients consented to receive a telephone call ‘where you will be asked if you have developed a headache amongst other questions’. This may have led to greater awareness of headache, and indeed in some patients may have resulted in over-reporting of headache symptoms. Moreover, patients were specifically asked for the presence of a headache, rather than to volunteer any symptoms they may have noted following their LP. Again this may have resulted in over-reporting. Patients were unaware of needle type, although it is possible that in patients with a history of previous TNLP they were able to differentiate a difference in technique. © 2013 The Author(s) European Journal of Neurology © 2013 EFNS
309
Two main hypotheses have been proposed about the pathophysiology of PLPHA [11]. The first suggests a reduction in CSF pressure leading to a downward shift of the brain and subsequent tension on the meninges and other pain-sensitive structures when the individual sits upright. The second hypothesis proposes that the CSF leak leads to compensatory cerebral vasodilation, resulting in PLPHA. It is possible that both mechanisms play a role. The link between damage to the dura, CSF leak and PLPHA was recognized many decades ago, leading to the invention of the ATN [5]. ATNs were designed to separate, rather than cut through, the longitudinal fibres of the dura. By 1951 the needle design had been refined, and a first successful reduction in PLPHA incidence using ATNLP was reported [12]. The needle type is not the only recognized risk factor for PLPHA. Previous studies have shown an increased incidence of PLPHA associated with young age, female gender and low body mass index [13]. Procedure technique may also play a role, with lower rates of PLPHA observed when (i) the needle bevel is inserted parallel to longitudinal dural fibres [14] and (ii) the needle stylet is replaced prior to withdrawing the needle [15]. Bed rest has traditionally been thought to have a preventive role in the development of PLPHA; however, evidence supporting this hypothesis is lacking [16,17]. Given that dural leak is one of the proposed mechanisms underlying PLPHA it is not surprising that a larger needle diameter has been linked to increased risk of PLPHA [18]. In the majority of cases, PLPHA resolves spontaneously with an estimated remission rate of 72% at 5 days [18,19]. However, such a seemingly high resolution rate of an iatrogenic condition should not lead to complacency. Given the importance of LP, and therefore the frequency at which it is being performed on a daily basis in any hospital across the UK, the socioeconomic impact of more than one in four people still suffering from PLPHA 5 days after their LP is massive (Table 3). Whilst simple analgesia and intravenous caffeine have been shown sometimes to reduce the need for more aggressive intervention [18,20], evidence-free measures such as increased fluid intake and bed rest continue to be encouraged [10,16,17]. Only recently a study suggested that clinicians would prefer such ‘established practices’ rather than switching from TNLP to ATNLP [10]. By far the most effective treatment once PLPHA has developed is epidural blood patching with a success rate of up to 98% [18,21]. Epidural patches are not routinely performed by neurologists (who often
310
A. Davis et al.
do diagnostic LP), however, at least in the UK, but rather by anaesthetists, which often results in delays due to the required coordination between medical disciplines. How can a culture of avoidance be turned into a culture of adoption of ATNLP in line with the evidence? The results of our survey suggest that neurologists in the UK are generally aware of ATNs; however, they either avoid using them or are not in the position to use them as their Trust does not provide them. Of the neurologists and trainees completing our survey 35% cited lack of availability and cost as main reasons for not using ATNLP. However, in one study a decision analytical model predicted $10.4 million savings to the US healthcare system should ATNLP replace TNLP in clinical practice [22]. In our study of nearly 100 patients, savings of £2903 were calculated for using ATNLP instead of TNLP provided not only the price of the needle is taken into account but also the cost for managing PLPHA. The reluctance to use ATNLP is not unique to the UK. A recent study from Denmark detected a similarly low rate of ATNLP amongst neurologists [10]. However, the reasons for this reluctance may vary, for example due to different national training and funding streams. Further research exploring this area is warranted. A further reason not to adopt ATNLP was the lack of training. Whilst this could easily be taken care of, the lack of a significant learning curve in our study suggests that ATNLP can be learned very quickly and does not pose any significant difficulty. Of note, a meta-analysis by Halpern and Preston [23] demonstrated a higher incidence of operator difficulty when using TNs compared with ATNs, suggesting that ATNLP should be easy to adopt. Nevertheless, this problem should be addressed and training provided. At our institution an online demonstration video is being used to help train junior doctors in ATNLP (http://www.youtube.com/watch?v=cfGKRb WvloA). Whilst LP is an essential diagnostic procedure in clinical practice, it is also emerging as a means to collect CSF biomarkers to predict outcome in clinical trials of neuro-inflammatory and neurodegenerative diseases [24]. However, studies requiring patients to undergo serial LP may be threatened by high dropout rates or refused institutional review board approval should modifiable risk factors such as PLPHA not be minimized. The debate about causes and modifiable risk factors of PLPHA has lasted for over a century. Surely it is now time to finally make ATNLP standard clinical practice.
Ethical approval This service development study was approved by the Clinical Effectiveness Department at Barts Health NHS Trust (project registration number 827).
Disclosure of conflicts of interest AD, RD, SK, ME and AB have no conflicts of interest to declare. GG has received research grant support from Bayer-Schering Healthcare, Biogen-Idec, GW Pharma, Merck Serono, Merz, Novartis, Teva and Sanofi-Aventis. GG has received personal compensation for participating on Advisory Boards in relation to clinical trial design, trial steering committees and data and safety monitoring committees from Bayer-Schering Healthcare, Biogen-Idec, Eisai, Elan, Fiveprime, Genzyme, Genentech, GSK, Ironwood, Merck-Serono, Novartis, Pfizer, Roche, SanofiAventis, Synthon BV, Teva, UCB Pharma and Vertex Pharmaceuticals. KS is a PI on trials sponsored by Novartis and Roche. He has received speaking honoraria from, and served on advisory boards for, Novartis, Merck-Serono and Merck Inc.
Acknowledgements RD is funded by an Association of British Neurologists/MS Society of Great Britain Clinical Research Fellowship. GG receives grant support from the MRC, National MS Society, MS Society of Great Britain and Northern Ireland, AIMS2CURE and the Roan Charitable Trust. KS is in receipt of an HEFCE Clinical Senior Lectureship and grant support from Barts Charity, the US National MS Society, Novartis and Biogen.
Supporting Information Additional Supporting Information may be found in the online version of this article: Data S1. Association of British Neurologists annual conference, Brighton 2012 Delegates Survey.
References 1. Lavi R, Yarnitsky D, Rowe JM, et al. Standard vs atraumatic Whitacre needle for diagnostic lumbar puncture: a randomized trial. Neurology 2006; 67: 1492–1494. 2. Barker P. Headache after dural puncture. Anaesthesia 1989; 44: 696–697. 3. Flaatten H, Krakene J, Vedeler C. Post-dural puncture related complications after diagnostic lumbar puncture, myelography and spinal anaesthesia. Acta Neurol Scand 1998; 98: 445–451.
© 2013 The Author(s) European Journal of Neurology © 2013 EFNS
Preventing post lumbar puncture headache
4. Lavi R, Rowe JM, Avivi I. Traumatic vs. atraumatic 22 G needle for therapeutic and diagnostic lumbar puncture in the hematologic patient: a prospective clinical trial. Haematologica 2007; 92: 1007–1008. 5. Greene HM. A technique to reduce the incidence of headache following lumbar puncture in ambulatory patients with a plea for more frequent examination of cerebrospinal fluids. Northwest Med J 1923; 22: 240–245. 6. Adrendt K, Demaerschalk BM, Wingerchuk DM, et al. Atraumatic lumbar puncture needles: after all these years, are we still missing the point? Neurologist 2009; 15: 17–20. 7. De Almeida S, Shumaker S, LeBlanc S. Incidence of post-dural headache in research volunteers. Headache 2011; 51: 1503–1510. 8. Birnbach DJ, Kuroda MM, Sternman D, et al. Use of atraumatic needles among neurologists in the United States. Headache 2001; 41: 385–390. 9. Armon C, Evans RW. Addendum to assessment: prevention of post-lumbar puncture headaches. Neurology 2005; 65: 510–512. 10. Stendell L, Fomsgaard JS, Olsen KS. There is room for improvement in the prevention and treatment of headache after lumbar puncture. Dan Med J 2012; 59: A4483. 11. Bezov D, Lipton RB, Ashina S. Post-dural puncture headache: Part I. Diagnosis, epidemiology, etiology and pathophysiology. Headache 2010; 50: 1144–1152. 12. Hart JR, Whitacre RJ. Pencil point needle in prevention of postspinal headache. J Am Med Assoc 1951; 147: 657–658. 13. Lybecker H, Møller JT, May O, et al. Incidence and prediction of postdural puncture headache. A prospective study of 1021 spinal anesthesias. Anesth Analg 1990; 70: 389–394.
© 2013 The Author(s) European Journal of Neurology © 2013 EFNS
311
14. Richman JM, Joe EM, Cohen SR, et al. Bevel direction and postdural puncture headache: a meta-analysis. Neurologist 2006; 12: 224–228. 15. Strupp M, Brandt T. Should one reinsert the stylet during lumbar puncture? N Engl J Med 1997; 336: 1190–1198. 16. Sudlow C, Warlow CP. Posture and fluids for preventing post-dural puncture headache. Cochrane Database Syst Rev 2002; 2: CD001790. 17. Thoennissen J, Herkner H, Lang W, et al. Does residual bed after cervical or lumbar puncture prevent headache? CMAJ 2001; 165: 1311–1316. 18. Turnbull DK, Shepherd DB. Post-dural puncture headache: pathogenesis, prevention and treatment. Br J Anaesth 2003; 91: 718–729. 19. Lybecker H, Djernes M, Schmidt JE. Post dural puncture headache onset, duration, severity and associated symptoms: an analysis of 75 consecutive patients with PDPH. Acta Anaesthesiol Scand 1995; 39: 605–612. 20. Basurto OX, Martinez GL, Sola I, et al. Drug therapy for treating post-dural puncture headache. Cochrane Database Syst Rev 2011; 8: CD007887. 21. Boonmak P, Boonmak S. Epidural blood patching for preventing and treating post-dural puncture headache. Cochrane Database Syst Rev 2010; 1: CD001791. 22. Tung CE, So YT, Lansberg MG. Cost comparison between the atraumatic and cutting lumbar puncture needles. Neurology 2012; 78: 109–113. 23. Halpern S, Preston R. Postdural puncture headache and spinal needle design. Metaanalyses. Anaesthesiology 1994; 81: 1376–1383. 24. Gunnarsson M, Malmestr€ om C, Axelsson M, et al. Axonal damage in relapsing multiple sclerosis is markedly reduced by natalizumab. Ann Neurol 2011; 69: 83–89.
