Shown isthe relative frequency of post–lumbar puncture outcomes along the y-axis plotted against cerebrospinal fluid (CSF) volume collected (A), needle size (B), and participant age (C) along the x-axis. From bottom to top, colored segments within each bar show the proportion of participants with the listed outcomes. The area of each graph segment is directly proportional to the number of individuals within that category. PDPH indicates post–dural puncture headache. Among all participants, the mean (SD) CSF volume collected was 22.32 (4.93) mL, and the mean (SD) participant age was 39.08 (10.31) years.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Monserrate AE, Ryman DC, Ma S, et al. Factors Associated With the Onset and Persistence of Post–Lumbar Puncture Headache. JAMA Neurol. 2015;72(3):325–332. doi:10.1001/jamaneurol.2014.3974
This study assesses factors associated with the most common adverse event following lumbar puncture.
To identify factors associated with the risk, onset, and persistence of post–dural puncture headache (PDPH).
Design, Setting, and Participants
We performed univariate and multivariable analyses of 338 lumbar punctures in the Dominantly Inherited Alzheimer Network observational study using linear mixed models, adjusting for participant-level and family-level random effects.
Main Outcomes and Measures
We directly evaluated associations of 3 post–lumbar puncture outcomes (immediate postprocedural headache, PDPH at 24-hour follow-up, and PDPH receiving a therapeutic blood patch) with participant age and sex, positioning, collection method, needle size, needle insertion site, and cerebrospinal fluid (CSF) volume collected.
The incidence of adverse events included 73 immediate postprocedural headaches (21.6%), 59 PDPHs at 24-hour follow-up (17.5%), and 15 PDPHs receiving a therapeutic blood patch (4.4%). Greater volume of CSF collected was associated with increased risk of immediate postprocedural headache, largely owing to a nonlinear increase in risk on collection of volumes above 30 mL (odds ratio, 3.73 for >30 mL and 0.98 for <17 mL). In contrast, collection of higher volumes showed a protective effect in decreasing rates of PDPH at 24-hour follow-up and rates of PDPH receiving a therapeutic blood patch (odds ratio, 0.35 per 10 mL). Although differences in needle size did not reach statistical significance, no participant in the 24G needle group received a therapeutic blood patch compared to 8 of 253 for the larger 22G needles.
Conclusions and Relevance
Factors that acutely lower CSF pressure (eg, seated positioning or extracting very high volumes of CSF) may be associated with transient post−lumbar puncture headache, without increasing rates of persistent PDPH or therapeutic blood patch. Collection of up to 30 mL of CSF appears to be well tolerated and safe.
Lumbar punctures (LPs) are widely used for diagnostic purposes in clinical neurology and are becoming increasingly important for measuring various cerebrospinal fluid (CSF) biomarkers in clinical research. The most common adverse event associated with LP is post–dural puncture headache (PDPH), which can be incapacitating and cause significant morbidity.1 Post–dural puncture headache is defined as a positional headache arising within 7 days of a dural puncture that becomes worse when standing and is relieved on lying down.2 Hypotheses suggest that PDPH results from low CSF pressure causing meningeal vasodilation, in addition to mechanical traction on cranial nerves and other pain-sensitive structures when in the upright position.3 Because the volume withdrawn during most LPs will be regenerated within several hours based on estimated CSF production of 300 to 1000 mL/d,4 PDPH manifesting outside of this window is believed to be consistent with continued CSF leakage through a dural tear, although such leakage does not always result in clinical symptoms.5 For persistent PDPH, a therapeutic blood patch can be performed, which consists of injecting 15 to 60 mL of autologous blood at the site of LP and is believed to occlude the dural tear and prevent further CSF leakage.6
Previous studies6-11 have evaluated factors associated with overall rates of PDPH in several case series, but the association of procedural variables with the onset and persistence of PDPH has not yet been evaluated directly. In this study, we performed a retrospective analysis of 338 LPs performed in the Dominantly Inherited Alzheimer Network (DIAN) observational study to identify factors associated with immediate postprocedural headache, PDPH at 24-hour follow-up, and PDPH receiving a therapeutic blood patch.
Institutional review board approval was obtained from Washington University School of Medicine. Written informed consent was obtained from all participants. The DIAN observational study is an international multicenter research trial enrolling individuals at risk of autosomal dominant Alzheimer disease.12 Cerebrospinal fluid is collected from DIAN participants to measure concentrations of β-amyloid and tau proteins, in addition to storage for the development of novel biomarkers.13 More than 80% of participants complete the LP procedure. We performed a retrospective study of 338 LPs performed in DIAN participants.
The DIAN study protocols include specific guidelines for LP performance but allow for some flexibility in procedural variables. All participants are required to fast overnight at least 8 hours before the procedure but are allowed to drink water. Participants can be seated (flexed forward) or lying (lateral decubitus with knees flexed) for insertion of the needle. The protocols specify a 22G atraumatic Sprotte needle with gravity collection, or a 24G atraumatic Sprotte needle with aspiration can be used. However, the use of a nonstandard needle (20G, 23G, 25G, or 26G) was reported in 20 cases herein. The needle can be inserted in the L2-L3, L3-L4, or L4-L5 interspace. The DIAN protocol requests that a minimum threshold of 17 mL of CSF should be collected, but actual volumes collected varied between 3 and 33 mL.
Participants are assessed for the presence of immediate postprocedural headache following LP and are advised to refrain from strenuous activity for 24 hours. After 24 hours, the participant is reassessed by telephone interview for the presence of PDPH and any other adverse events. Participants with PDPH can be offered a therapeutic blood patch (at the option of the participant), and this is recorded if performed.
In this retrospective study, we investigated factors associated with immediate postprocedural headache, PDPH at 24-hour follow-up, and PDPH receiving a therapeutic blood patch. Data on participant age and sex, positioning (seated or lateral decubitus), CSF collection method (gravity or aspiration), needle size, needle insertion site, and CSF volume collected were obtained from the DIAN study database and a manual audit of study records. Sixty-seven LPs were excluded from the study owing to missing information regarding participant age and sex and Clinical Dementia Rating14 score.
Generalized linear mixed models were used to estimate significance and effect size for each variable, including participant-level and family-level random effects. Statistical analysis was performed separately for all 3 outcomes (immediate postprocedural headache, PDPH at 24-hour follow-up, and PDPH receiving a therapeutic blood patch), and the odds ratio (OR) for each statistically significant predictive variable was determined. Each model included participant age and sex, positioning, collection method, needle size, and CSF volume collected. To assess the effect of some LPs that collected less than the study guideline threshold of 17 mL and to investigate a potential nonlinear increase in PDPH with CSF threshold above 30 mL that became apparent on inspection of the data, additional analyses were performed that analyzed CSF volume collected as a categorical variable (with thresholds at <17, 17-30, and >30 mL). Because the data set includes some individuals who have had multiple LPs, a longitudinal analysis was performed, adjusting for the interval between each visit date and the baseline visit. The PROC GLIMMIX procedure (SAS, version 9.3; SAS Institute Inc) was used to conduct the analyses, and statistical significance was defined as P < .05.
Characteristics of the study population are summarized in Table 1. We performed univariate analyses investigating potential effects of global Clinical Dementia Rating14 scores on longitudinal data and initial visit data for each of the 3 post-LP outcomes. We found no significant effects of this well-validated measure of clinical status on rates of postprocedural or follow-up headache or a therapeutic blood patch (P > .5). The frequency of post-LP outcomes stratified by the categorical variables of participant sex, positioning, collection method, and needle insertion site is summarized in Table 2, and the frequency of outcomes by CSF volume collected, needle size, and participant age is shown in the Figure.
Table 3 summarizes results of statistical analyses on the incidence of immediate postprocedural headache present during participant assessment following LP. When analyzed as a continuous variable, greater volume of CSF collected was associated with increased risk of immediate postprocedural headache (OR, 2.67 per 10 mL; P = .007). This increased risk appeared to be largely owing to a nonlinear increase in risk occurring above approximately 30 mL of volume collected (Figure). When analyzed as a categorical variable, rates of immediate postprocedural headache for the thresholds of below 17 mL and 17 to 30 mL were similar (OR, 0.98; 95% CI, 0.30-3.22; P = .98), while only the threshold above 30 mL showed significantly greater risk (OR, 3.76; 95% CI, 1.46-9.67; P = .007) compared with the threshold of 17 to 30 mL (Table 3). Participant sex (OR, 0.95; 95% CI, 0.53-1.70; P = .85) and age (OR, 0.98; 95% CI, 0.96-1.02; P = .42) showed no significant effects for the immediate postprocedural headache outcome. Seated positioning showed a trend toward more frequent immediate postprocedural headaches relative to lateral decubitus positioning (OR, 2.51; 95% CI, 0.98-6.41; P = .05). Needle size reached overall significance in the multivariable model (P = .02), but this was principally mediated by higher rates of headache in LPs performed with nonstandard needles, for which no further conclusions could be reached owing to small numbers. The 24G needles showed a protective effect in reducing risk relative to 22G needles, although this did not reach significance for 22G vs 24G (OR, 3.21; 95% CI, 0.74-13.97; P = .12). We were unable to calculate estimates of an effect of needle insertion site owing to the small sample size of some subgroups.
Table 4 summarizes results of statistical analyses for PDPH at 24-hour follow-up. In contrast to the association observed for immediate postprocedural headache, collection of higher volumes of CSF showed a protective effect toward reduced risk of headache at follow-up (OR, 0.58; 95% CI, 0.31-1.06 per 10 mL; P = .08). The threshold below 17 mL showed significantly higher rates of follow-up headache relative to the threshold of 17 to 30 mL (OR, 3.07; 95% CI, 1.11-8.49; P = .04). Older participant age also showed a protective effect that approached significance (OR, 0.74; 95% CI, 0.53-1.01 per 10 years; P = .06), while participant sex again showed no significant effect (P = .35). Consistent with what was observed for immediate postprocedural headache, seated positioning again showed a trend toward increased risk, although this did not reach significance for the 24-hour follow-up outcome (OR, 1.90; 95% CI, 0.75-4.78; P = .17). The use of nonstandard needles again was associated with greater risk (P = .01), and the use of 22G needles showed a trend toward greater risk of headache at follow-up relative to the use of24G needles, which did not reach significance (OR, 2.24; 95% CI, 0.58-8.66; P = .24).
Table 5 summarizes results of statistical analyses for variables associated with PDPH receiving a therapeutic blood patch. Greater CSF volumes collected again showed a protective effect toward reduced risk of receiving a therapeutic blood patch (OR, 0.35; 95% CI, 0.12-1.05 per 10 mL; P = .06), and this effect was most apparent in substantially increased risk for the lowest threshold of below 17 mL relative to the threshold of 17 to 30 mL (OR, 5.59; 95% CI, 1.23-25.30; P = .03). Older participant age again showed a protective effect (OR, 0.54; 95% CI, 0.29-1.03; P = .06), while participant sex was not significantly associated with rates of receiving a therapeutic blood patch (OR, 1.08; 95% CI, 0.34-3.45; P = .89). The use of nonstandard needles was again associated with greater risk (P < .001). None of the 65 LPs performed using a 24G needle received a therapeutic blood patch compared with 8 of 253 (3.2%) using a 22G needle.
Lumbar puncture is becoming a more widely used procedure in clinical research and clinical practice in neurology. Because persistent PDPH can cause significantly greater morbidity than transient post-LP headache and because a therapeutic blood patch can result in substantial cost and inconvenience to those affected, our analysis took the approach of directly investigating potential factors associated with each of these LP-associated adverse events.
We found that collection of greater volumes of CSF was associated with higher rates of immediate postprocedural headache (particularly at volumes >30 mL) but trended toward lower rates of PDPH at 24-hour follow-up and lower rates of PDPH receiving a therapeutic blood patch, suggesting that these outcomes may result from different mechanisms. Physiologically, immediate postprocedural headache can result from an acute decrease in intracranial CSF pressure owing to extraction of larger volumes of CSF, triggering meningeal vasodilation and positional traction on intracranial structures. If no further leakage of CSF occurs following the procedure, the amount of CSF removed should be replenished within several hours at physiological rates of CSF production.4 This implies that PDPH at 24-hour follow-up is consistent with continued leakage of CSF through the dural puncture. Cerebrospinal fluid leakage has been visualized directly by magnetic resonance myelography in patients with PDPH, although the presence of leakage is not always associated with clinical symptoms.5 Persistent failure of the dural puncture to close may cause a prolonged and severe positional headache requiring treatment with a therapeutic blood patch. One potential mechanism for the lower rates of PDPH and therapeutic blood patch seen with extraction of higher volumes of CSF may be that transiently lower CSF pressures immediately following the procedure decrease continued leakage through the dural puncture, facilitating dural closure.
Because the DIAN study guidelines request that at least 17 mL of CSF be collected, worse overall outcomes observed below this threshold could be an artifact of premature discontinuation of the procedure (for reasons such as loss of access to the dural space or discontinuation owing to a participant’s report of symptoms), which is a limitation of this study. In addition, we analyzed CSF volume collected as a categorical variable, confirming that much of the increased risk associated with low extraction volumes was attributable to significantly higher rates of follow-up headache and therapeutic blood patch for the threshold below 17 mL. Rates of immediate postprocedural headache were not significantly greater in this group (P = .98) compared with the threshold of 17 to 30 mL, suggesting that higher rates of persistent headache after extraction of low volumes were not solely attributable to early discontinuation from participant reports of symptoms during the procedure. Conversely, higher rates of immediate postprocedural headache with extraction of high volumes appeared to be associated with a potential nonlinear increase in risk on extraction of very high volumes above approximately 30 mL. The other groups (<17 mL and 17-30 mL) did not differ significantly (P = .98) with respect to immediate postprocedural headache. Although we can reach no strong conclusions regarding the lowest threshold, collection of very low volumes did not appear to benefit participants, and CSF collection within the range of 17 to 30 mL seems to be the safest and best tolerated.
Although the generalizability of these findings to other populations remains to be established, the effects of procedural variables such as CSF volume collected appeared to be substantial relative to potential demographic effects of participant age or sex. Previous investigators have reported a higher incidence of PDPH in women than in men.7 We observed no significant independent effect of participant sex on rates of any of the 3 outcomes in this study. Older age has also been previously reported to be inversely associated with overall rates of PDPH.8 We observed a trend toward a protective effect of older age against PDPH at 24-hour follow-up and PDPH receiving a therapeutic blood patch, but we found no indication of a significant effect of participant age on immediate postprocedural headache. It has previously been hypothesized that the protective effect of age may result from greater baseline CSF volume owing to age-related cerebral atrophy; however, such an effect might be expected to be more pronounced in the acute postprocedural period, contrary to what was observed in this study.
Studies8,13 have shown that PDPH risk is greater with cutting needles (eg, a Quincke needle) than with noncutting needles (eg, Sprotte or pencil-point needles). As a result, the DIAN study protocol12,17 uses noncutting Sprotte needles exclusively. Although study protocols call for 22G or 24G needles, 20 procedures were performed with other needle types, and these were associated with significantly higher rates of immediate postprocedural headache, PDPH at 24-hour follow-up, and PDPH receiving a therapeutic blood patch. This was the case for smaller (25G) and larger (20G) needles. When comparing 22G and 24G needles directly, we observed trends toward decreased risk of immediate postprocedural headache and PDPH at 24-hour follow-up with the smaller 24G needles. Although these trends did not reach statistical significance, none of the 65 LPs performed using a 24G needle received a therapeutic blood patch compared with 8 of 253 (3.2%) using a 22G needle.
Recent evidence suggests that LPs performed in a seated position may result in higher overall risk of PDPH than in the lateral decubitus position.8 We observed a trend (OR, 2.51; P = .05) toward higher risk of immediate postprocedural headache when LP was performed in a seated position. Although rates of PDPH and therapeutic blood patch were slightly higher in LPs performed when seated, these associations were not significant. When LP is performed in a seated position, the acute decrease in CSF pressure may cause position-dependent symptoms from vasodilation or from mechanical traction on intracranial structures. Positioning may have less effect on rates of persistent PDPH that result from continued CSF leakage through a dural tear. In this study, we observed higher rates of immediate postprocedural headache after seated LPs, despite recommendation for participants to avoid strenuous activity, and a recent meta-analysis16 found no significant effect of post-LP bed rest on the risk of PDPH.
Additional limitations of this study include that it is retrospective in nature and could not control for all potential confounding factors. Because the DIAN observational study comprises 12 sites with more than 35 physicians performing LPs and because many physicians performed few procedures, we could not directly control for potential effects unique to individual physicians or study sites. However, adjustment for family as a random effect in the statistical models may somewhat mitigate potential confounding owing to site-based effects. Because randomized clinical trials focused on prospectively investigating the effects of multiple procedural variables on post-LP outcomes are unlikely to be undertaken, analysis of well-standardized clinical research data provides a reasonable approach to assessing the safety and tolerability of study procedures, as well as identifying factors that may be associated with postprocedural morbidity.
Our study has several key findings. First, factors associated with immediate postprocedural headache and PDPH at 24-hour follow-up may not be identical. Factors that acutely lower CSF pressure (eg, seated positioning or extracting very high volumes of CSF) were associated with transient immediate postprocedural headache, without increasing the risk of PDPH at 24-hour follow-up or PDPH receiving a therapeutic blood patch.
Second, collection of 17 to 30 mL of CSF was safe and well tolerated, with extraction of greater volumes within this range not associated with higher morbidity. Collecting very low volumes of CSF (<17 mL) did not benefit participants. We observed a nonlinear increase in rates of immediate postprocedural headache at very high volumes of CSF collection (>30 mL), which did not result in higher rates of PDPH at 24-hour follow-up or PDPH receiving a therapeutic blood patch.
Third, rates of adverse events are comparable to the incidence of adverse events in previous studies.1,6-8,13,15 The incidence of adverse events herein included 73 immediate postprocedural headaches (21.6%), 59 PDPHs at 24-hour follow-up (17.5%), and 15 PDPHs receiving a therapeutic blood patch (4.4%).
Accepted for Publication: October 30, 2014.
Corresponding Author: Randall J. Bateman, MD, Department of Neurology, Washington University School of Medicine, 660 S Euclid Ave, Campus Box 8111, St Louis, MO 63110 (firstname.lastname@example.org).
Published Online: January 26, 2015. doi:10.1001/jamaneurol.2014.3974.
Author Contributions: Mr Monserrate had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Monserrate, Ryman, Masters, Salloway, Bateman.
Acquisition, analysis, or interpretation of data: Monserrate, Ryman, Ma, Xiong, Noble, Ringman, Morris, Danek, Müller-Sarnowski, Clifford, McDade, Brooks, Darby, Weston, Farlow, Graff-Radford, Salloway, Fagan, Oliver, Bateman.
Drafting of the manuscript: Monserrate, Ryman, Bateman.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Monserrate, Ryman, Ma, Xiong, Müller-Sarnowski, Bateman.
Obtained funding: Morris, Masters, Bateman.
Administrative, technical, or material support: Noble, Morris, Clifford, Farlow, Fagan, Oliver, Bateman.
Study supervision: Ryman, Bateman.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported by Washington University Institute of Clinical and Translational Sciences grants UL1 TR000448 and TL1 TR000449 from the National Center for Advancing Translational Sciences. The Dominantly Inherited Alzheimer Network (DIAN) is supported by grant U19 AG032438 from the National Institute on Aging (Dr Morris) and by the generous support of F. Simmons and O. Mohan, in addition to DIAN site support from the German Center for Neurodegenerative Diseases, the National Institute for Health Research Queen Square Dementia Biomedical Research Unit, and J. O. and J. R. Wicking Trust grants 13026 and 20821.
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Group Information: The Dominantly Inherited Alzheimer Network investigators are listed at http://www.dian-info.org.17
Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Additional Contributions: We gratefully acknowledge the contributions of all participants and family members involved in this research.