| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The epidural blood patch is considered effective in treating postdural puncture headache. We have developed a postdural puncture model in rats for quantitative evaluation of the magnitude and duration of changes in cerebrospinal fluid (CSF) pressure in the cisterna magna in response to the administration of epidural blood or other moieties. This model was used to compare the efficacy of various methods of epidural injection for restoring and maintaining CSF pressure for up to 240 min. After lumbar dural puncture, CSF pressure declined 3.6 ± 0.2 mm Hg. Epidural saline (100 µL) injected at the puncture site initially increased pressure by 7.2 ± 0.7 mm Hg, but it rapidly (7.8 ± 0.6 min) returned to postdural puncture baseline. A similar initial increase of CSF pressure was observed with equal volumes of all other epidural injectates, but the duration of pressure increase varied greatly. Hetastarch and dextran 40 produced results similar to saline. Only whole blood or fibrin glue consistently increased CSF pressure for the entire 240-min observation period. Whole blood mixed with anticoagulant or injected 20-mm cephalad to the puncture site did not sustain pressure. After laminectomy, direct application of blood or adhesive to the dural defect caused no pressure increase. Continuous infusion of saline after bolus could maintain pressure increase for 180 min, but within 60 min of stopping infusion, pressure returned to baseline. These results confirm the efficacy of the epidural administration of blood or fibrin glue to correct CSF hypotension after dural puncture and also provide insight into the mechanisms of intracranial pressure modulation. Sealing the dural defect does not effectively correct CSF pressure unless an epidural tamponade effect is also maintained. IMPLICATIONS: A rat model was developed to evaluate different drugs that may be injected epidurally to treat postdural puncture headache. Epidural injection of blood or fibrin glue was the most effective method of maintaining increased cerebrospinal fluid pressure after dural puncture. Sealing the dural defect does not effectively correct cerebrospinal fluid pressure unless an epidural tamponade effect is maintained.
Postdural puncture headache (PDPH) is a common clinical problem associated with intentional or inadvertent creation of a defect in the integrity of the dura-arachnoid. Regardless of the etiology, PDPH is annoying to patients and in some cases can be debilitating and require specific therapeutic intervention. The mechanism(s) by which a dural puncture leads to headache symptoms is controversial (14). It is generally believed that headache results from the loss of cerebrospinal fluid (CSF), causing a downward shift of the brain and stimulating pain-sensitive nerve fibers. In addition, the loss of CSF volume may produce a compensatory cerebral vasodilation that has been classically associated with other types of headaches. Biochemical mechanisms, such as stimulation of adenosine receptors, may also be involved (5). The epidural blood patch (EBP) is often considered the "gold standard" for treatment of PDPH, and is used when there is not prompt spontaneous resolution or when there is failure of pharmacologic management, such as caffeine. Two mechanisms have been proposed for the effectiveness of the EBP (69). One hypothesis is that the transmitted pressure from the epidural injection (a tamponade effect) causes an increase in intracranial pressure (ICP) that restores the brain to its normal position. The other possibility is that epidurally injected blood seals the dural defect and normal CSF production eventually replenishes the lost CSF and restores brain homeostasis. However, the two mechanisms may be complementary with the almost immediate relief of headache symptoms after EBP readily explained by the first mechanism (because human CSF production is only 0.35 mL/min), whereas the continued maintenance of headache relief, even after the epidural blood mass has resolved within a few hours (6), is consistent with the second mechanism. Although an EBP using autologous blood is relatively easy to perform, there may be circumstances (such as systemic or localized sepsis) when this should be avoided. For this reason, the injection or infusion of nonblood substitutes (1012) or commercially available fibrin glue (13,14) have been evaluated as replacement for the traditional autologous EBP. Although there are reports of successful outcomes with these alternatives, there are limited data from controlled trials detailing relative efficacy of alternative epidural injectates after dural puncture, and comparisons of case studies are hindered by inhomogeneity of patient characteristics and epidural injection methods for PDPH. Considering the likely mechanism(s) by which EBP is effective after dural puncture, we designed an animal model to systemically evaluate the change in CSF pressure in response to different types of epidural administration of blood, blood products, colloids, and crystalloids via bolus injections or continuous infusions. The primary goal of this study was to compare the relative efficacy of these methods for achieving sustained correction of intracranial hypotension after dural puncture in a controlled model. A secondary goal was to gain insight into the mechanism(s) by which EBP modulates CSF pressure.
After approval from the Institutional Animal Care and Use Committee, experiments were performed on male Sprague-Dawley rats (Sasco, Wilmington, MA) weighing 300400 g. Animals were anesthetized with 1.5% halothane in oxygen through a nose cone and mounted in a rat stereotactic head holder (David Kopf, Tujunga, CA). The L4-6 vertebrae were exposed and the L5 and L6 spinal processes removed (Fig. 1). A small hole was made in the L5-6 ligament and a saline-filled polyethylene catheter (0.28-mm inner diameter, 0.61-mm outer diameter) was inserted through the ligament and advanced 5 mm cranially into the epidural space (15). The catheter was secured with cyanoacrylate tissue adhesive (VetbondTM; 3M, St. Paul, MN). In one experimental group, the catheter was advanced 25 mm. A stylet was used to seal the external end of the catheter. The muscle and fascia covering the cisterna magna (CM) were retracted for later access to monitor CSF pressure.
The animal was then placed head up at 40° to elevate the brain above the lumbar spine. This was done to simulate postural conditions in which PDPH symptoms are worsened in patients (16). A 22-gauge beveled cannula was inserted 1 mm into the CM and secured with Vetbond adhesive. The cannula was previously attached to a pressure transducer (model P23XL-1; Harvard Instruments, Holliston, MA) via 15 cm of stiff polyvinyl tubing. Saline was used to fill the entire measurement system. The level of the transducer was adjusted so that the pressure was 0.0 mm Hg when the cannula tip was just above the CM. At 20-min intervals throughout the experiment, a transducer port was opened to the atmosphere and rezeroed if necessary to compensate for any transducer drift over the course of the experiment. The pressure transducer was connected to a preamplifier and chart recorder (Gould, Cleveland, OH), with the preamplifier output transferred to a data acquisition system (Digidata 1200 analog/digital board and Axoscope software, version 7.0; Axon Instruments, Foster City, CA). In addition, electrocardiogram leads were also connected to monitor the cardiac pulse in synchrony with the CSF pulse pressure. Once a stable CM pressure reading was attained, a dural puncture was performed in the L4-5 interspace using a 22-gauge Quincke type spinal needle (B-D, Rutherford, NJ). Although this may not be considered a large-gauge needle for human dural puncture, for the much smaller rat spinal column, this creates a relatively large hole and CSF leak. In preliminary experiments, we tried using smaller-gauge needles (e.g., 25-gauge) but they did not consistently reduce CM pressure by at least 2 mm Hg. The location of the puncture corresponds approximately to the position of the tip of the epidural catheter (or 20 mm below the catheter tip in one experiment). The subsequent decline in CM pressure was then monitored for 30 min, by which time it had stabilized at a significantly lower level. If the CM pressure in any animal declined by <2 mm Hg, that animal was excluded from further study because the dural puncture was considered insufficient to produce intracranial hypotension. At 5 min after the dural puncture, Vetbond adhesive was placed over the L4-5 interspace, to simulate the pressure sealing effect that vertebral muscle and skin have over lumbar punctures performed percutaneously in humans.
Bolus Injection Studies To evaluate the time course of CM pressure changes after lumbar dural puncture without intervention, a group (n = 5) was included in which no injections were performed after dural puncture. In two other groups (n = 5/group), a laminectomy was performed at the L4-5 vertebrae, a lumbar dural puncture performed, and 30 min later, the dura-arachnoid defect sealed by direct application of either 100-µL Vetbond adhesive or whole donor blood. This was to investigate the effect of sealing the CSF leak on CM pressure, without the tamponade effect of the fluid bolus in the closed epidural space. An additional group (n = 5) received a subcutaneous injection of 3 mg of caffeine sodium benzoate after dural puncture. The criterion for evaluating efficacy was the duration of time (up to 240-min measurement interval) that CM pressure was >1 mm Hg above preinjection postdural puncture pressure. This definition was chosen because small physiologic drifts in CSF pressure >4 h could cause a slight increase in CM pressure that could be misleadingly interpreted as a maintenance of increased pressure unless some threshold value is chosen. In all of the above experiments, there were five animals in each group, with the results recorded as the mean ± SE. CM pressures before and after lumbar dural puncture, and peak CM pressures after bolus injections were compared using analysis of variance for repeated measures and Bonferroni corrected paired t-tests. The durations of CM pressure increase were compared among groups using a one-way analysis of variance and the Tukey-B multiple-comparison method.
Continuous Infusion Studies
Postdural Puncture Model The initial CM pressure, in the halothane-anesthetized rat on a 40° incline, was 6.2 ± 0.2 mm Hg (n = 55), and decreased by 3.6 ± 0.2 mm Hg after lumbar dural puncture (Fig. 2). Without further intervention, the pressure declined an additional 0.7 ± 0.4 mm Hg over the next 240 min. Recovery of CM pressure toward prepuncture baseline was not observed in any animal.
Bolus Injection Studies There was no difference in the postdural puncture pressure among the different epidural injectant groups. The initial effect of the 100-µL epidural saline injection was a rapid increase in CM pressure, with a peak increase of 7.2 ± 0.7 mm Hg (n = 35) (Fig. 3, left). However, this effect was nonsustained, and the pressure rapidly returned to baseline (duration, 7.8 ± 0.6 min). The initial effect of a 100-µL epidural whole blood injection was a similar rapid increase in CM pressure (Fig. 3, right). There was no statistical difference in the magnitude of the initial increase in CM pressure after the saline injection compared with any of the other epidural injectates. However, the duration of pressure increase was considerably different among injectates, with epidural whole blood showing a sustained pressure increase after the initial peak (Fig. 3, right).
The effect of the tested injectates on the duration of CM pressure elevation is shown in Figure 4. There was no difference between saline and two of the colloids, hetastarch, and 10% dextran 40. The more viscous colloidal solution, 32% dextran 70, sustained the increase in CM pressure for over an hour, which was longer than all of the other nonblood products. Whole blood injected 20 mm above the lumbar dural puncture site or whole blood premixed with EDTA had durations of pressure increase near one-half hour. Only whole blood itself or fibrin glue injected at the level of the dural puncture maintained CM pressure above the postdural puncture baseline for the entire 240-min period. The magnitudes of the pressure increase were similar at 30 and 240 min postinjection with whole blood (2.9 ± 0.5 and 2.5 ± 0.5 mm Hg) or fibrin glue (3.3 ± 0.3 and 2.9 ± 0.5 mm Hg), respectively, and were greater than the untreated dural puncture animals at the same times.
In the groups of animals with the epidural space opened by laminectomy, the application of either 100 µL of Vetbond adhesive or whole blood applied directly over the dural defect did not produce an initial increase in CM pressure. Also, in these groups, after 240 min CM pressure had declined an additional 0.7 ± 0.2 and 0.8 ± 0.2 mm Hg, respectively, from the postdural puncture baseline. These values were not different from that of the untreated dural puncture group. In the group that received subcutaneous caffeine 30 min after lumbar dural puncture, the CM pressure declined an additional 1.6 ± 0.1 mm Hg after 10 min (which was different from the untreated dural puncture group over an equivalent 10-min period), and then slowly recovered by 240 min to the postdural puncture baseline.
Continuous Infusion Studies
Normal CSF pressure in humans ranges between 5.1 and 13.2 mm Hg in the lateral recumbent position (17). The mean pressure in our rat study (40° angle) was 6.2 mm Hg. In patients with typical PDPH, the lumbar subarachnoid pressure was 1.7 mm Hg in the lateral decubitus position (18). The postdural puncture pressure in our rat study was 2.6 mm Hg (6.2 - 3.6). Although the values of CSF pressure between human studies and the rat model seem similar, no exact comparison can be made because the human data are derived from two separate populations (normal subjects versus PDPH patients). In 1967, Usubiaga et al. (19) injected 20 mL of saline epidurally in patients with PDPH and measured the pressure in the lumbar subarachnoid space. There was a large increase in pressure initially, but it returned to baseline within 310 minutes. After the initial report of the successful treatment of PDPH by Gormley (20) with EBP, Coombs and Hooper (18) injected 15 mL of autologous blood epidurally and measured lumbar subarachnoid pressure. Initially, the pressure increased to 6.8 mm Hg, from a postdural puncture baseline of 1.7 mm Hg, but instead of declining rapidly, it was still at 4.8 mm Hg at 15 minutes postinjection. We intentionally modeled our experiments after that study. The 100-µL epidural injection volume we chose in the rat corresponds approximately to a 15-mL volume in a human, because based on body weight (e.g., 60 kg human/0.4 kg rat) or CSF turnover rate, the scaling factor from rat to human is about 150. Our experiments in the rat showed an initial increase in CM pressure of 7.2 mm Hg after epidural bolus injection, which is slightly higher than the 5.1 mm Hg increase in lumbar subarachnoid CSF pressure in patients. Our experience with bolus epidural saline injections in the rat, showing a return to baseline pressure in 7.8 minutes, is similar to the duration measured in patients (19). Single epidural injections of 2 commonly used blood substitutes, 6% hetastarch and 10% dextran 40 proved to be no better than equal volumes of saline in restoring CM pressure. It should be cautioned, however, that with n = 5/group, there may actually be a difference between these blood substitutes and saline, but our statistical tests do not have sufficient power to detect it. One clinical report indicates that epidural dextran 40 injection provides relief from PDPH (10). Because additional clinical reports used a protocol of a bolus injection of dextran 40 followed by a maintenance infusion at 15% of the initial volume/hour (1112), we evaluated that strategy in our rat model. As Figure 5 shows, there was little success in increasing CM pressure even with a maintenance dextran 40 infusion of 120% initial volume/hour. It is noteworthy that at very fast rates of infusion (300% initial volume/hour), a sizeable proportion of animals had increased CM pressure at 180 minutes (only significant for the saline group). However, when epidural infusion was terminated, the CM pressure returned to postdural puncture baseline within an hour, indicating that in this model, no permanent recovery was achieved by maintenance infusion of 10% dextran 40 or saline. Our rat model provides insight into the mechanisms involved in the correction of intracranial hypotension by the EBP technique. When the epidural injection site was between the lumbar dural puncture (20 mm above the puncture) and the CM monitoring site, rather than contiguous with the puncture site, the same initial increase in CM pressure followed the 100-µL injection of blood or saline. However, that method of epidural blood injection maintained increased CM pressure for a significantly shorter interval (27.0 minutes) than blood injected in continuity with the dural puncture. Therefore, it seems to be essential that the EBP be delivered close to the dura-arachnoid defect. Presumably, the increased "mass effect" of blood compressing the CSF space cannot solely compensate for the continued leakage of CSF at the lumbar site. The addition of EDTA to whole blood also resulted in a significant reduction of the duration of CM pressure increase. Although the mean time to return to postdural puncture baseline (33.4 minutes) was longer than with epidural saline, the overall results were inferior compared with unmodified whole blood. Therefore, the coagulability of blood also seems to be essential for producing the long-term effect of maintaining CM pressure correction. Finally, opening the epidural space with a laminectomy, and applying whole blood or tissue adhesive caused no increase in CM pressure, confirming the importance of the tamponade effect (6,7). Our results in this rat model comport with the widely held opinion that EBP is the "gold standard" for treating PDPH. The injection of 100 µL of fresh whole blood consistently maintained an increased CM pressure over the 240-minute measurement period. This is consistent with magnetic resonance imaging studies showing that the mass effect from EBP in humans lasts 37 hours (6). It would have been desirable to follow the CM pressure for a longer time interval to demonstrate a permanent sealing effect in the dura-arachnoid of the whole blood injection, and maintenance of CM pressure related to CSF production. However, attempts to extend the total experiment time were not fruitful because after 6 hours of halothane anesthesia, we began to observe a decreased heart rate and a change in the pattern of the CM pressure from a typical pulse pattern to an atypical sinusoidal waveform. The use of fibrin glue was a successful method of producing a sustained increase of CM pressure, and may be a viable alternative to an autologous blood patch if there is systemic sepsis. Fibrin glue seals lumbar CSF leakage in patients with intrathecal catheters (13), and in pigs after dural puncture (14). Our 100-µL injection in a rat corresponds to a 15-mL epidural injection in a human, so the relative amount of fibrin glue evaluated in the study is larger than has been used in humans for a comparable purpose (13). Nonetheless, the fibrin glue injection in our rat model demonstrates that a substance with both mass effect and sealing ability can maintain increased CM pressure. Methods involving mechanisms other than restoring ICP to normal can be effective for treating PDPH. An example of this is the systemic injection of caffeine, which produced an immediate negative effect in our model by even further reducing CM pressure after dural puncture. The latter observation suggests that this common pharmacologic treatment is effective for PDPH because of its cerebral vasoconstrictive properties (5,21), without increasing CSF pressure. Another important consideration when evaluating the effects of interventions used for PDPH is the anatomical differences between species. The much thicker human dura-arachnoid membrane compared with the rat may have better self-sealing properties in response to external pressure or it may be more easily plugged by colloids such as dextran 40. This may explain why saline (19) or dextran 40 (1012) epidural application has been used successfully in patients, but seems to be ineffective in the rat model. One aspect of PDPH not addressed in our acute experiments is the time delay between a dural puncture and the appearance of headache symptoms that can occur days after the precipitating event (1,4). In the rat model, the decrease in CM pressure is almost immediate and the epidural injections were performed soon afterward. These differences mandate caution in translating the results of this controlled animal model to the clinical spectrum of PDPH. Despite certain limitations, our rat model seems to be a useful method for evaluating mechanisms and quantifying relative efficacy of approaches designed to restore ICP after dural puncture.
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|