Chapter 73. Postdural Puncture Headache

Postdural puncture headache (PDPH) was perhaps the first recognized complication of regional anesthesia. Dr. August Bier noted this adverse effect in 1898 in the first patient to undergo successful spinal anesthesia, a 34-year-old laborer undergoing resection of an ulceration of the foot. Bier observed: “Two hours after the operation his back and left leg became painful and the patient vomited and complained of severe headache. The pain and vomiting soon ceased, but headache was still present the next day.”1

The following week, Bier and his assistant, Dr. August Hildebrandt performed experiments on each other. Following the “cocainization of the spinal cord", Bier described firsthand his experience in the days to follow: “I had a feeling of very strong pressure on my skull and became rather dizzy when I stood up rapidly from my chair. All these symptoms vanished at once when I lay down flat, but returned when I stood up … I was forced to take to bed and remained there for nine days, because all the manifestations recurred as soon as I got up … The symptoms finally resolved nine days after the lumbar puncture.”1 Interestingly, since Bier's time, the clinical picture of PDPH has remained unchanged and it remains the most common complication of spinal anesthesia.

Portions of this text originally appeared in the following publication: Neal, Rathmell (eds): Complications in Regional Anesthesia and Pain Management, copyright 2006, published by Elsevier, Philadelphia, PA, USA, with permission.

PDPH occurs following procedures that disrupt the integrity of the meninges. Most cases of PDPH are characterized by their typical onset, quality, and associated symptoms.

Onset

Onset of symptoms is generally delayed, with headache usually beginning 12–48 h (rarely > 5 days) following dural puncture. Following dural puncture, 65% experience symptoms within 24 h and 92% within 48 h.2 An onset of symptoms within 1 h is suggestive of pneumocephalus, especially in the context of an epidural loss-of-resistance technique using air.3

Quality

The hallmark of PDPH is its postural nature, with symptoms worsening in the upright position and improving with recumbency. The International Classification of Headache Disorders further describes this positional quality as worsening within 15 min of sitting or standing and improving within 15 min after lying down.4 Headache is nearly always bilateral, with a distribution that is frontal (25%), occipital (27%), or both (45%).2 Headaches are typically described as “dull/aching,” “throbbing,” or “pressure-type” in nature.

The severity of PDPH symptoms is directly related to needle size. Larger needles are associated with more severe headaches5 and a greater need for definitive treatment measures.6 The severity of PDPH after spinal anesthesia varies from mild (11%) to moderate (23%), to severe (67%).2

Associated Symptoms

If headaches are severe, they are more likely to be accompanied by a variety of other symptoms. Pain and stiffness in the neck and shoulders is common and seen in nearly half of all patients experiencing PDPH.7 With careful questioning, nausea may be reported by a majority of patients and can lead to vomiting.2

Uncommonly, patients may experience auditory or visual symptoms,8 and the risk for either appears to be directly related to needle size.9,10 In a large study of PDPH, these symptoms were present to a clinically significant degree in 0.4% of patients.11 Auditory symptoms are frequently unilateral, and include hearing loss, tinnitus, and even hyperacousis. Subclinical hearing loss, especially in the lower frequencies, appears to be common following spinal anesthesia, even in the absence of PDPH.10 Closely associated with auditory function, vestibular disturbances (dizziness or vertigo) may also occur. Visual problems include blurred vision, difficulties with accommodation, diplopia, and mild photophobia.9 In contrast to headache complaints, nearly 80% of episodes of diplopia involve unilateral cranial nerve palsies.

Incidence

The reported incidence of PDPH varies considerably depending on the technique employed, patient population, definition of PDPH, and duration of follow-up. PDPH is primarily discussed in association with spinal anesthesia; however, it is also frequently observed following diagnostic lumbar puncture, myelography, and accidental dural puncture (ADP) during epidural techniques.

Rates of PDPH are particularly high (up to 10%) after myelography and diagnostic lumbar puncture, in which large-gauge needles are considered necessary due to the viscosity of contrast material and to facilitate the timely collection of cerebrospinal fluid (CSF).

Rates of PDPH following spinal anesthesia have steadily declined; from an incidence exceeding 50% in Bier's time, to around 10% in the 1950s,11 until currently a rate of 1% or less can be reasonably expected. This reduction is clearly due to modifications in practice that have followed the identification of risk factors (discussed in Risk Factors for Developing PDPH). In obstetric patients, a population known to be at high risk for PDPH, meta-analysis demonstrated an incidence of 1.7% using 27-gauge Whitacre needles.12 Combined spinal-epidural techniques have also been associated with a low incidence of PDPH (1.6% in one study of over 2000 obstetric patients).13 This observation may be due to several factors, including the ability to successfully use needles of very small diameters (ie, 27 gauge) with a noncutting tip design as well as possible tamponade provided by epidural infusions. Continuous spinal anesthesia using standard 20-gauge catheters (macrocatheters) has been reported by some to be associated with surprisingly low incidences of PDPH (3.4%) compared with single-dose spinal techniques using similar gauge needles.14 This observation has been attributed to reaction to the catheter, which may promote better sealing of a breach in the meninges. However, deliberate continuous spinal anesthesia has usually been investigated in low-risk populations.

Epidural techniques are an attractive alternative to spinal anesthesia due to the perceived ability to avoid puncture of the dura. The problem is, however, that ADP cannot be reliably prevented and may be unrecognized at the time in over 25% of patients who eventually develop PDPH.15 ADP is of greatest concern in the obstetric anesthesia setting, where the incidence may be up to 1.5%.12 Of these patients, 52–80% will develop PDPH.

Medicolegal Considerations

PDPH is a significant source of concern and dissatisfaction for patients. In parturients who had experienced an ADP, 70% would not choose to have epidural analgesia again, and only 28% would recommend it to a friend or relative.16 Iatrogenic headaches continue to represent a significant medicolegal liability, as noted by the high frequency of malpractice claims for headache in the American Society of Anesthesiologists Closed Claims Project database. For instance, for obstetric anesthesia claims, headache was the second most common maternal injury (after death) and resulted in payment in 56% of cases.17,18 Similarly, headache following epidural steroid injections was the most common malpractice claim in pain management practices.19 For these reasons, the potential for this complication necessitates a proper discussion and informed consent for any procedure that may result in PDPH.

Pathophysiologic Mechanisms behind PDPH

Despite a great deal of research and observational data, the pathophysiology of PDPH remains incompletely understood.20 It is generally accepted that PDPH results from a disruption of normal CSF homeostasis. CSF is produced primarily in the choroid plexus at a rate of approximately 0.35 mL/min and reabsorbed through the arachnoid villa. The total CSF volume in adults is maintained around 150 mL, of which approximately half is extracranial, and gives rise to normal lumbar opening pressures of 5 to 15 cm H2O in the horizontal position (40–50 cm H2O in the upright position). A sudden loss of CSF results in the development of typical PDPH symptoms, which resolve promptly with reconstitution of lost CSF volume.21 It is thought that PDPH is due to the loss of CSF through a persistent leak in the meninges. In this regard, it has been postulated that the arachnoid mater may be at least as and perhaps more important than the dura mater in the genesis of PDPH.22 However, the actual mechanism by which CSF hypotension generates headache is controversial and currently ascribed to a bimodal mechanism involving both loss of intracranial support and cerebral vasodilation (predominantly venous). Diminished hydrostatic support is thought to cause the brain to sag in the upright position, resulting in traction and pressure on pain-sensitive structures within the cranium (dura, cranial nerves, bridging veins, and venous sinuses). In addition. vasodilation may occur secondary to diminished intracranial CSF volume and reflexively secondary to traction on intracranial vessels.

The neural pathways primarily involved in PDPH include the ophthalmic branch of the facial nerve (CN V1) in frontal pain, cranial nerves IX and X in occipital pain, and cervical nerves C1 through C3 in neck and shoulder pain.23 Nausea is attributed to vagal stimulation (CN X). Auditory and vestibular symptoms are secondary to the direct communication between the CSF and the perilymph via the cochlear aqueduct, which results in decreased perilymphatic pressures in the inner ear and an imbalance between the endolymph and perilymph.8 Visual disturbances usually represent a transient palsy of the nerves that supply the extraocular muscles of the eye (CN III, IV, and, VI). The lateral rectus muscle is most often involved, attributed to the long, vulnerable intracranial course of the abducens nerve (CN VI).9

Risk factors for PDPH can be broadly categorized into patient characteristics and procedural details.

Patient Characteristics

Although uncommon in children younger than 10 years of age, PDPH has a peak incidence in the teens and early 20s.24 The incidence then declines over time, becoming infrequent in patients older than 50 years of age. Female gender has long been thought to impart an increased risk for PDPH.11 However, gender has not always constituted an independent risk factor when age-matched nonpregnant populations are studied.24 Pregnancy is generally regarded as a risk factor for PDPH,11 but this consideration partially reflects the young age as well as the high incidence of ADP in the gravid population. Pushing during the second stage of labor, thought to promote the loss of CSF through a hole in the meninges, has been shown to be a significant risk factor for PDPH following ADP. Angle and colleagues noted that the cumulative duration of bearing down correlated with the risk of developing PDPH in patients who had experienced an ADP and that patients who avoided pushing altogether (ie, proceeded to cesarean delivery prior to reaching second stage labor) had a much lower incidence of PDPH (10%) than those who pushed (74%).25 Observed variations in the thickness of the dura (from 0.5 to nearly 2.0 mm) have also been proposed to influence the risk of PDPH, as punctures made where the dura is thicker appear to be less prone to allowing the loss of CSF.26

PDPH appears to have an interesting association with other headaches. Patients who report having had a headache within the week prior to lumbar puncture have been observed to have a higher incidence of PDPH.27 On further analysis, only those with chronic bilateral tension-type headaches were found to be at increased risk.28 A history of unilateral headache28 or migraine29 has not been linked with an increased risk of PDPH. Menstrual cycle, a factor in migraine headaches, did not influence the rate of PDPH in one underpowered pilot study.30 A small but statistically significant increased incidence of PDPH following spinal anesthesia has been reported if patients have a history of previous PDPH.24 Patients with a history of ADP have been shown to have a slightly increased risk of another ADP (and subsequent PDPH).31

A number of other patient characteristics have been demonstrated to be minor risk factors for PDPH, including lower opening pressures,32 lower patient body mass index,27,33 low CSF substance P concentration,34 and low baseline pain sensitivity.35 Although such findings would appear to be of academic interest only, they may hold important keys to the future prevention and treatment of PDPH.

Risk Factors Related to the Procedure and Equipment

Needle size and tip design are the most important procedural factors.5 Needle size is directly related to the risk of PDPH, with larger needles resulting in a higher incidence of symptoms.11 “Noncutting” needles are associated with a reduced incidence of PDPH than occurs with “cutting” (eg, Quincke) needles of the same size. Noncutting needles, also referred to as pencil-point, blunt tip, atraumatic, or conical tip, include the Whitacre, Sprotte, European, Pencan, and Gertie Marx needles. Although originally thought to be less traumatic than cutting needles, electron microscopy has shown noncutting needles to produce a more traumatic hole in the dura, perhaps resulting in a better inflammatory healing response.36 The influence of needle size on risk of PDPH appears to be less for noncutting needles than cutting needles (eg, the reduction in the incidence of PDPH between 22- and 26-gauge sizes is greater for cutting than noncutting needles). Some needles, such as the Greene and Atraucan, appear to be acceptable combinations of cutting and noncutting features.37

Inserting cutting needles with the bevel parallel to the long axis of the spine reduces the risk of PDPH.24,38,39 This observation was for many years attributed to a “spreading” rather than cutting of longitudinally oriented dural fibers. However, scanning electron microscopy reveals the dura to be made of many layers of concentrically directed fibers,40 and the importance of needle bevel insertion is most likely due to the longitudinal tension on the dura and its influence on CSF leakage through holes with differing orientations.

A greater number of dural punctures increases the rate of PDPH.41 The angle of approach to the dura may be another important consideration, due to the dura–arachnoid relationship. Experimental and clinical evidence suggest that paramedian42 (vs midline) and oblique43 (vs 90-degree) approaches may allow for better sealing of a dural puncture, reducing CSF loss and resulting in a lower incidence of PDPH. The operator's experience, comfort, and skill are clearly associated with the incidence of ADP during epidural procedures, but have not consistently been a significant risk factor for other central neuraxial techniques. The possible contribution of operator fatigue, especially for ADP, remains to be determined.

A number of procedural details do not appear to influence the rate of development of PDPH, including patient position at the time of dural puncture,27 bloody tap during spinal anesthesia,44 addition of narcotics to spinal block,45 and volume of spinal fluid removed (for diagnostic purposes).27 The specific local anesthetic used may affect the rate of nonspecific headache but does not appear to influence the need for epidural blood patch (EBP).46

Practical Approach to Evaluating Patients Presenting with Headache after a Neuraxial Anesthetic

A careful history with a brief consideration of other possible diagnoses is usually sufficient to differentiate PDPH from other causes of headache (Table 73–1). Most non-postward puncture headaches will not have a strong positional nature, as seen in PDPH. Vital signs (normal blood pressure and absence of fever) and a basic neurologic exam should be documented. Bilateral jugular venous pressure, briefly applied, tends to worsen headaches secondary to intracranial hypotension. Laboratory studies are usually not necessary for the diagnosis of PDPH and, if obtained, are generally unremarkable (most commonly, magnetic resonance imaging may show meningeal enhancement, and lumbar puncture reveal increased CSF protein and low opening pressures).

It should be noted that benign headaches are common in the perioperative setting, even in the absence of dural puncture, and have generally been noted to be less severe than PDPH7 (causes may include anxiety, dehydration, hypoglycemia, anxiety, and caffeine withdrawal). It has even been proposed that some nonpostdural puncture headaches may be due to contamination of the central neuraxis with skin preparation solutions.47 The majority of headaches following dural puncture will be benign nonspecific headaches. In a careful analysis of headache following spinal anesthesia using strict criteria for PDPH, Santanen and colleagues found an incidence of nonspecific headache of 18.5%, with an incidence of true PDPH of only 1.5%.48 In an obstetric population, Grove noted postpartum headache in 23% of patients who had not had any regional anesthesia.49

Exacerbation of chronic headache (eg, tension, cluster, or migraine) is usually notable for a history of similar headaches. Hypertensive headache has a clear association with significant hypertension. Pneumocephalus can produce a positional headache that does not respond to EBP and can be difficult to distinguish from PDPH, but is readily diagnosed with computed tomography.50 Sinusitis may be associated with purulent nasal discharge and tenderness over the affected sinus and is often improved with assuming an upright position. A number of other benign causes are possible.

Serious causes of headache are rare but must be excluded. It is important to remember that lateralizing neurologic signs, fever/chills, seizures, or changes in mental status are not consistent with a diagnosis of PDPH. Meningitis tends to be associated with fever, leukocytosis, changes in mental status, and meningeal signs (eg, nuchal rigidity).51 Subdural hematoma (SDH), long recognized as a potential complication of dural puncture, is believed due to intracranial hypotension resulting in excessive traction on cerebral vessels, leading to their disruption. Practioners should maintain a high index of suspicion for SDH, which is often preceded by classic PDPH symptoms but progresses to lose its postural component and may include disturbances in mentation and focal neurologic signs.52 It has been proposed that early definitive treatment of severe PDPH may serve to prevent SDH.53 Subarachnoid hemorrhage, most commonly due to rupture of a cerebral aneurysm or arteriovenous malformation, is usually associated with the sudden onset of excruciating headache followed by a decreased level of consciousness or coma.54 Preeclampsia/eclampsia often presents with headache and may only become evident in the postpartum period.55,56 Dural venous sinus thrombosis (DVST) is usually seen in the postpartum obstetric population, in whom headache symptoms may progress to seizures, focal neurologic signs, and coma.57 Predisposing factors for DSVT include hypercoagulability, toxemia, and dehydration. Reports of other intracranial pathology (intracranial tumor, intracerebral hemorrhage, etc) misdiagnosed as PDPH are extremely uncommon and will be detected with a thorough neurologic evaluation.54

Diagnosis of PDPH can be particularly challenging in patients who have undergone lumbar puncture as part of a diagnostic work-up for headache. In these situations, headache should have changed in quality, the most common difference being a new postural nature. Occasionally, if the benign diagnostic possibilities cannot be narrowed down with certainty, a favorable response to EBP can provide definitive evidence for a diagnosis of PDPH.

A number of recommendations can be made to decrease the risk of PDPH, but several long-standing practices, namely prolonged recumbency and aggressive hydration, do not appear to be useful.58,59

Subarachnoidal Needle Placement

Prevention of PDPH in the setting of spinal anesthesia involves appropriate patient selection and careful attention to technique. Patients younger than 40 years of age are at particularly high risk for PDPH, and alternatives to spinal anesthesia should perhaps be sought in these individuals unless the benefits are sufficiently compelling (as they appear to be in the obstetric population). Using needles with a lower risk of PDPH is the most important technical means of reducing the risk of PDPH with spinal anesthesia. As a rule, the smallest noncutting needles feasible should be employed. However, extremely small needles are more difficult to place, have a slow return of CSF, may be associated with multiple punctures of the dura, and may result in a higher rate of unsuccessful block. Therefore, it appears that the best compromise between the risk of PDPH and ease of placement is a 24- to 27-gauge noncutting needle. If cutting tip needles are used, the bevel should be directed parallel to the long axis of the spine.

It has been suggested that intravenous caffeine (500 mg caffeine sodium benzoate within 90 min after spinal anesthesia) reduces the incidence of moderate to severe headache.60 However, the applicability of these data is questionable as this study involved the use of 22-gauge Quincke needles in a relatively young patient population.

Epidural Needle Placement

Unlike spinal anesthesia, patient selection appears to play a minimal role in the prevention of ADP during attempted epidural techniques. The use of a noncompressible medium (saline or local anesthetic) for loss-of-resistance generally results in a lower incidence of ADP than using air.61 The risk of PDPH following ADP can probably also be reduced by always using the smallest gauge epidural needle possible (yet equipment options in this regard are limited with catheter techniques.

Optimal bevel orientation for epidural needle insertion remains a matter of debate. Although incidence of PDPH following ADP may be reduced by directing the needle bevel parallel to the long axis of the spine,38 this technique necessitates a 90-degree rotation of the needle for catheter placement. Consequently, the possibility of dural trauma with needle rotation has prompted some to suggest always inserting epidural needles with the bevel facing the direction of desired catheter placement.63

Strategies to Reduce the Risk of PDPH after ADP

Since not all patients who experience an ADP will develop PDPH, and only a portion of those who do will require definitive treatment, a cautious approach in this regard is suggested. Several immediate measures following ADP have been proposed to prevent the development of PDPH. Replacing the stylet prior to needle withdrawal is a simple and possibly effective means of lowering the incidence of PDPH. This recommendation is based on observations of lumbar punctures performed using 21-gauge Sprotte needles and is theorized to decrease the possibility of an arachnoid mater “wick” through the dura.64 Charsley and Abrams performed a subarachnoid injection of 10 mL of saline following ADP and noted a significantly reduced need for EBP in the treated group.65 Their analysis is limited by the small number of patients studied, and further investigation is clearly indicated. Ayad and colleagues recommend placement of a subarachnoid catheter for 24 h following ADP.66 In their obstetric population, catheter placement resulted in a rate of PDPH of only 6.2%, with an expected incidence under these conditions of over 50%. This impressive reduction in the incidence of PDPH, however, has not been noted in studies in which catheters have been left in place for less than 24 h.

Provided an epidural catheter can be successfully placed following ADP, prophylactic epidural saline has been used in hopes of reducing the incidence and severity of PDPH. Efforts have included both bolus (eg, 50 mL as a single injection or every 6 h for four doses) and continuous infusion techniques (commonly 600–1,000 mL over 24 h). In the largest analysis of this approach (n = 241), Stride and Cooper reported a reduction in the incidence of PDPH from 86% in a conservatively treated control group to 70% with epidural saline infusion.67 Trivedi and colleagues noted a similar reduction in PDPH (from 87% to 67%) in 30 patients who received a single prophylactic “saline patch” (40–60 mL) following completion of the obstetric procedure.68 Other studies of epidural saline have noted this modest decrease in the incidence of PDPH. Stride and Cooper also reported a lower incidence of severe headache (from 64% to 47%), but this effect has been inconsistently seen by other investigators, and there is no definitive evidence that epidural saline reduces the eventual need for EBP. Despite the paucity of data, a survey in North America published in 1998 reported 25% of tertiary care obstetrical centers using prophylactic epidural saline infusions.69 Such observations appear to be due more to a desire to “do something” and the perceived safety of the practice rather than any proven efficacy. The 1999 National Obstetric Anesthetic Database in the United Kingdom reported epidural saline infusion/bolus used as immediate management in only 7% of cases of ADP.7 Importantly, epidural saline does not appear to reduce the success rate of EBP. Although the epidural administration of saline is generally benign, transmitted pressure not uncommonly results in back and eye pain and has caused retinal hemorrhage.

Prophylactic Epidural Blood Patch: When, How, in Whom

The risk–benefit of prophylactic measures should be most favorable in those having the greatest likelihood of developing PDPH, such as obstetric patients experiencing an ADP with an epidural needle. However, investigations into the efficacy of the prophylactic EBP in this setting have yielded mixed results.70 Scavone and colleagues, for example, in a prospective, randomized, double-blind study in an obstetric population found that the prophylactic EBP shortened the total duration of symptoms but failed to reduce the incidence of PDPH or subsequent need for EBP.71 Currently, due to concerns of exposing patients to a potentially unnecessary and marginally beneficial procedure, most centers do not utilize the prophylactic EBP as a routine measure.69

If employed prophylactically, the EBP should be performed only after any spinal or epidural local anesthetic has worn off, as premature administration has been associated with excessive cephalad displacement of local anesthetic.72 Residual epidural local anesthetic may also inhibit coagulation of blood, further decreasing the efficacy of EBP.73

Once the diagnosis has been made, patients should be provided a straightforward explanation of the cause, natural history, and treatment options for PDPH. A treatment algorithm, based primarily on the severity of symptoms, can serve as a useful guide for management (Figure 73–1).

Time

PDPH is a complication that tends to resolve spontaneously. Prior to the introduction of definitive therapy for PDPH (ie, the EBP), the natural history of the disorder was documented by Vandam and Dripps as they followed 1011 episodes of PDPH after spinal anesthesia using cutting needles of various sizes.11 Although their analysis is flawed by a lack information on duration in 9% of patients, if one considers their observed data, spontaneous resolution of PDPH was seen in 59% of cases within 4 days and 80% within 1 week. More recently, Lybecker and coworkers followed 75 episodes of PDPH and, although providing an EBP to 40% of their patients, noted in the untreated patients a median duration of symptoms of 5 days with a range of 1 to 12 days.2 These data, although generally reassuring, also illustrate the sometimes prolonged duration of untreated PDPH. Indeed, Vandam and Dripps noted 4% of patients still experiencing symptoms 7–12 months after spinal anesthesia.11 Similar observations of prolonged symptoms have been reported following diagnostic lumbar puncture74 and ADP.75 A number of case reports exist of successful treatment of PDPH months or years after known76 or even occult dural puncture.77

Primarily due to the self-limited nature of PDPH, the optimal time course of treatment has not been well-defined. Although many practitioners have advocated 24–48 h of conservative therapy, the practicality of this approach is questionable, given the often severely disabling nature and prolonged duration of symptoms. Notably, the duration of PDPH appears to be directly related to needle size,78 which would support early definitive therapy for severe headaches following dural punctures made with large (eg, epidural) needles.

Supportive Measures

Reassurance and supportive measures, although not expected to alter the duration of symptoms, are advised for all patients. Most patients with severe PDPH will naturally seek a recumbent position for symptomatic relief. Although excessive hydration does not appear to influence the duration of symptoms,59 patients should be encouraged to avoid dehydration. Oral analgesics (acetaminophen, NSAIDs, or narcotics) may be utilized, yet the relief obtained is often unimpressive, especially with severe headaches. Antiemetics and stool softeners should be prescribed when indicated. Abdominal binders can provide some degree of relief, but are uncomfortable and seldom used in modern practice.

Pharmacologic Treatments

A number of pharmacologic agents have been advocated as treatments for PDPH.79 However, these options have generally been poorly studied and are of questionable efficacy due to the small numbers of patients treated, methodologic flaws in published reports, and the self-limited nature of PDPH.

Methylxanthines are used for their cerebral vasoconstrictive effects and include aminophylline,80 theophylline, and the best studied, caffeine.79 Caffeine has been used intravenously (500 mg caffeine sodium benzoate, which contains 250 mg caffeine) and orally (300 mg). Published studies of caffeine used for PDPH consistently demonstrate improvement at 1 to 4 h in over 70% of patients treated. However, a single oral dose of caffeine for treatment of PDPH was statistically no better than placebo at 24 h.81 With a terminal half life of 6 h, repeated doses of caffeine would seem necessary for treatment of PDPH; however, no studies have evaluated more than two doses for efficacy or safety (of particular concern in the breastfeeding patient). Furthermore, there is no convincing evidence that caffeine reduces the eventual need for EBP. The temporary benefit observed with caffeine would indicate that it is perhaps most useful for relief of moderate symptoms while awaiting the spontaneous resolution of PDPH. The familiarity of caffeine for nonmedical purposes would argue for its general safety, but its use is contraindicated in patients with seizure disorders, pregnancy-induced hypertension, or a history of supraventricular tachyarrhythmias.

Sumatriptan, a serotonin type-1d receptor agonist that causes cerebral vasoconstriction, is commonly used for migraine headache and has also been used to treat PDPH. However, sumatriptan was not effective in a small randomized, prospective study for treatment of severe PDPH.82 Corticosteroidogenics (ACTH and its synthetic form, cosyntropin/tetracosactin83) have also been proposed as treatments for PDPH. Reports of the successful use of these agents are intriguing, but their role in the management of PDPH awaits further study.

Epidural Interventions

Epidural Saline

Bolus injections of epidural saline (usually 20–30 mL, repeated as necessary if a catheter is present) have been reported to produce prompt and virtually universal relief of PDPH, yet the practice is plagued by an extremely high rate of headache recurrence. This transient effect is not surprising as increases in epidural pressure following bolus administration of saline have been shown to return to baseline within 10 min.84 Favorable results achieved with this approach have been speculated to represent the mechanical reapproximation of a dural flap. A combination of saline bolus followed by infusion has occasionally been successful in the treatment of PDPH under exceptional circumstances.85

However, bolus administration of saline for treatment of PDPH is of questionable merit when compared with the EBP, especially when headaches are secondary to large-bore punctures of the dura.86 Overall, epidural saline appears to be of limited value for established PDPH.

Epidural Blood Patch: When, Why, and How

During the past several decades, the EBP has emerged as the “gold standard” of treatment for PDPH.44 This procedure has been well-described and consists of a sterile injection of autologous blood at or below the level of previous dural puncture (due to the preferential cephalad spread of blood in the lumbar epidural space87). The mechanism of the EBP, although not entirely understood, appears to be related to the ability to stop further CSF loss by the formation of clot over the defect in the dura as well as a tamponade effect with cephalad displacement of CSF (the “epidural pressure patch”).88

A Cochrane Review (a systematic assessment of the evidence) of the EBP recently concluded that the role of this procedure in the prevention and treatment of PDPH was uncertain, primarily due to a lack of reliable data.89 Reflecting this absence of adequately powered, randomized trials, a number of controversies surround the EBP.90

The optimal timing of the EBP is undetermined. A number of studies suggest that the procedure becomes more effective with the passage of time,91–93 but this observation may simply be due to larger, harder to treat CSF leaks, which demand earlier attention. Although prophylactic use of the EBP does not appear to be warranted, there is little rationale for delaying EBP if symptoms are severe.

The ideal volume of blood for EBP is also unclear, as extensive epidural spread is visualized with volumes greater than 10 mL.87 Currently, volumes of up to 20 mL are commonly employed, with the injection stopped when the patient complains of discomfort or fullness in the back, buttocks or head. There is no clear association between success rates for EBP and volume of blood used (between 10 and 20 mL), and there are few data to encourage the use of volumes greater than 20 mL.

Early observations of the EBP reported success rates well over 90%, but the true efficacy of the procedure appears to be significantly lower. Although the EBP is associated with nearly immediate symptomatic relief in approximately 90% of cases, follow-up reveals a significant number of patients experiencing incomplete relief, failure, and recurrence. In an analysis of 504 patients receiving EBP following dural puncture with needles of various sizes, Safa-Tisseront and colleagues noted complete relief in 75%, incomplete relief in 18%, and failure in 7%.92 PDPH secondary to needles larger than 20 gauge were an independent risk factor for failure of EBP, an observation supported by studies in obstetric patients following ADP with epidural needles. Under these circumstances, Williams and colleagues noted complete relief of symptoms with EBP in only 33% of patients, partial relief in 50%, and no relief in 12%.94 If performed, a second EBP resulted in complete relief in 50%, partial relief in 36%, and no relief in 14%. In a similar patient population, Banks and colleagues, despite initially observing complete or partial relief with EBP in 95% of patients, reported the return of moderate-to-severe symptoms in 31%, with a mean time to development of recurrent headache of 31.8 h (range 12–96 h).93 The rates of repeat EBP for the Williams and Banks studies were 27% and 19%, respectively. These studies clearly demonstrate the reduced efficacy of the EBP following dural punctures made with large needles, which will not uncommonly make it necessary for clinicians to consider repeating the procedure. Success rates of a second EBP appear to be equal to that of a first.

Following the successful performance of an EBP, maintenance of the decubitus position for at least 1 and preferably 2 h may result in a more complete resolution of symptoms.95 Avoidance of lifting, Valsalva maneuvers, and air travel for 24 to 48 h are commonly advised to minimize the risk of patch disruption. Finally, patients should be provided clear instructions for the provision of timely medical attention should they experience a recurrence of symptoms.

Contraindications and Risks

Contraindications to the EBP are similar to those of any epidural needle placement: patient refusal, coagulopathy, systemic sepsis, fever, and infection at the site. Theoretical concerns have been expressed regarding the possibility of neoplastic seeding of the central nervous system in patients with cancer.96 Although not free of controversy,97 the EBP has been safely provided to patients with HIV infection98 and acute varicella.99 The EBP may also be indicated and performed, with decreased volumes of blood, in the pediatric population100 and at extralumbar (eg, cervical101) sites. Modifications of usual EBP technique have been suggested to accommodate the special needs of Jehovah's Witness patients.102

Minor side effects are common following the EBP. Patients should be warned to expect aching in the back, buttocks, or legs (seen in approximately 25% of patients).93 Although usually short-lived, backache was noted to be persistent in 16% of patients following EBP, lasting 3–100 days (with a mean duration of 27.7 days).103 Despite these lingering symptoms, patient satisfaction with the EBP is high. Other frequent but benign after effects of the EBP include transient neckache,103 bradycardia,104 and modest temperature elevation.103 Significant risks of EBP are essentially the same as with other epidural procedures (infection, bleeding, nerve damage, and ADP). Occasionally, the temporary back and lower extremity radicular pain mentioned earlier has been reported to be severe. With proper technique, infectious complications are vanishingly rare. Although controversial, it appears that a previous EBP does not significantly influence the success of future epidural interventions.105 Serious complications secondary to the EBP do occur but have usually consisted of isolated case reports and have often been associated with significant deviations from standard practice.44

Finally, a history of significant technical difficulties with attempted central neuraxial techniques, although not a contraindication to epidural treatments, should naturally encourage a trial of less invasive measures.

Other Epidural Treatments

A variety of alternatives to blood have been promoted as patch materials, including dextran-40,106 hydroxyethyl starch,107 gelatin,107,108 and fibrin glue.109 Although not without merit, these options remain poorly defined, and reports of their use should be considered preliminary.

Neurologic consultation is indicated when serious non-PDPH is suspected or cannot reasonably be ruled out. Consultation is also appropriate for any headache with atypical features. As noted earlier, lateralizing neurologic signs, fever/chills, seizures, or change in mental status are not consistent with a diagnosis of PDPH. Proceeding with treatment measures directed toward PDPH under uncertain circumstances may hinder correct diagnosis, cause critical delays in proper treatment, and can prove harmful (the EBP, for example, by producing increases in intracranial pressure).

Persistent or recurrent headaches, although not necessarily requiring consultation, warrant follow-up and thoughtful reevaluation. As PDPH tends to resolve over time, headaches that worsen over time into headaches which no longer have a positional nature should be suspected to possibly be secondary to SDH (especially if there are focal neurologic signs or decreases in mental status).53 Under these circumstances, a neurologic consultation should be obtained and diagnostic radiologic studies performed.

Although headache and most associated symptoms, including auditory symptoms,10 resolve quickly following EBP, it should be noted that cranial nerve palsies generally resolve slowly (within 6 months).9 Therefore, residual cranial nerve palsy, although usually not requiring any specific therapy, may appropriately prompt a neurology consult for reassurance and management.

Since the EBP has a reasonable success rate and PDPH tends to improve even without treatment, many practitioners seek neurologic consultation when symptoms have failed to resolve after a logical but arbitrary number of EBPs (commonly three).

In summary, PDPH after neuraxial therapeutic, anesthetic, or diagnostic procedures continues to be a significant clinical concern. Prevention is obviously much preferable to the available treatment. PDPH of mild-to-moderate severity will resolve in a timely manner without specific treatment. Pharmacologic agents may be useful in the management of PDPH of mild-to-moderate severity. Despite proposed alternatives, the EBP remains the sole definitive treatment for PDPH, and its proper role in clinical practice continues to become more clearly defined. The prophylactic use of the EBP, for example, is of questionable value. PDPHs secondary to the use of large-gauge (ie, epidural) needles present a particular challenge, as they tend to be more severe, of longer duration, and more difficult to treat.