European Journal of Anaesthesiology 2005; 22: 907912

2005 European Society of Anaesthesiology ISSN 0265-0215

Original Article Neurotoxicity with single dose intrathecal midazolam administration

B. Ugur*, K. Basaloglu, T. Yurtseven, U. Ates, O. N. Aydin*, D. zen, M. Yurtseven, A. Grel* Adnan Menderes University Medical Faculty, Departments of *Anaesthesiology and Reanimation, Histology and Embryology, Aydin; Ege University Medical Faculty, Departments of Neurosurgery, Histology and Embryology, Izmir, Turkey

Summary Background and objective: The aim of study was to investigate the electron microscopic changes in the medulla of the spinal cord that occur with intrathecal midazolam administration. Methods: Twenty-eight albino rabbits of New Zealand type were randomized into two groups. Following anaesthesia, 16 rabbits were given 300 g of midazolam (Group M) and 12 rabbits were given 0.3 mL of normal saline solution (Group C) intrathecally. Eight rabbits from Group M (Group M1) and 6 rabbits from Group C (Group C1) were sacrificed 24 h after the anaesthesia and 8 rabbits from Group M (Group M2) and 6 rabbits from Group C (Group C2) were sacrificed 6 days after the anaesthesia. The lumbosacral portion was removed by laminectomy and thin sections were examined microscopically. Results: Severe separation in myelin lamella of the large axons, honeycomb appearance, slight separation in myelin lamella of small to moderately large axons, degenerate vacuoles in the cytoplasm and nuclear membrane irregularity were observed in neurons of Groups M1 and M2. Myelin lamella and nuclear membranes were found to be regular, vacuoles and oedema were observed in the neurons in the Groups C1 and C2. Conclusion: Midazolam administered at single dose by the intrathecal route may have neurotoxic effects on the neurons and myelinated axons at 24 h and 6 days following administration.
Keywords: MICROSCOPY ELECTRON; MIDAZOLAM; INJECTIONS SPINAL, intrathecal; NEUROTOXICITY; RABBITS.

Introduction Intrathecal and epidural drug administration for postoperative analgesia and relief of cancer pain is widely used. Various drugs have been administered and evaluated for intrathecal or epidural routes in animals and human beings [13]. The first drugs were opioids which were followed by -2 agonists such as clonidine, local anaesthetics and several peptides [46]. The development of spinal drugs such as midazolam
Correspondence to: Bakiye Ugur, Adnan Menderes Universitesi, Tip Fakltesi, Anesteziyoloji ve Reanimasyon Departmani 09100, Aydin, Turkey. E-mail: bakiyeugur@hotmail.com; Tel: 90 256 444 1256; Fax: 90 256 214 6495 Accepted for publication September 2005 EJA 3058

that optimize antinociceptive effects and minimize adverse effects therefore seems desirable as an alternative to opioids or local anaesthetics in morphinetolerant patients [7]. Midazolam is a widely used systemic adjuvant delivered for its sedative, anxiolytic and amnesic effects [8,9]. Benzodiazepines may modulate the affinity of -aminobutyric acid (GABA) at its receptors while enhancing its control of chloride channel activity [10]. When intrathecally administered, they have relieved pain of somatic origin, produced selective sensory blockade and blocked somatosympathetic reflexes [7]. It is important to perform neurotoxicity screening in animals of compounds intended for spinal administration before performing clinical experiments in

908

B. Ugur et al

human beings. However, there is still controversy regarding its possible neurotoxicity to the spinal cord. Some animal studies have shown neurotoxicity of spinally [2,11] and epidurally [12] administered midazolam in rabbits; but some others not [13,14]. Possible toxic reactions such as demyelination, axonal degeneration, oedema, necrosis, arachnoiditis, vascular changes [15], vacuolization and fibrosis [12] may be related to the drug itself or to chronic administration of drugs via long-term cannulation during the study. Thus, it is difficult to distinguish the damage induced by midazolam from that induced by the catheter. Furthermore, it is not clear whether this neurotoxicity is due to the drug itself, preservative hypoxia or ischaemia [4]. In addition, the efficacy of light microscopy in determining neurotoxicity of a drug is a matter of debate [16]. There are few studies investigating the histological changes by electron microscopy. The purpose of the present study was to investigate the histological reactions of the spinal cord at 24 h and 6 days to directly administered single dose midazolam in rabbits by electron microscopy.

Methods Following approval of the animal Ethics Committee, 28 New Zealand albino rabbits of either sex, weighing 2.32.5 kg were randomly assigned into two groups: Group midazolam (M) received 300 g (0.3 mL) midazolam (Dormicum; Roche, pH 3.3, 1 mg mL1, n 18), and group control (C) received 0.9% saline (pH 5.5, 0.3 mL, n 12) intrathecally following laminectomy under anaesthesia. The rabbits were anaesthetized with 5 mg kg1 xylazine (Alfazyne 20 mg mL1; Alfasan International B.V.) and 10 mg kg1 ketamine (Ketalar; ParkeDavis, 50 mg mL1) by ear veins. All rabbits were kept in identical conditions with respect to food regimen, temperature, humidity and periods of daylight. A pulse oximeter probe was placed on the rabbits ear after shaving the area and heart rate (HR) and peripheral oxygen saturation (SPO2) were monitored. The exclusion criteria were an HR of lower than 130 min1 and an SPO2 below 90. Eight randomly selected rabbits from the midazolam group (Group M1) and six rabbits from the control group (Group C1) were sacrificed after 24 h. The remaining eight rabbits from the midazolam group (Group M2) and six rabbits from the control group (Group C2) were sacrificed after 6 days by administration of transcardiac 1% formaldehyde solution and 2% gluteraldehyde with 0.1 mol phosphate buffer. The spinal medulla of the rabbits were removed by laminectomy.

Surgical procedure The lumbosacral areas and surgical sites of the rabbits were incised on the midline following sterilization with 10% iodine solution. The laminectomy procedure was carried out with the aid of an operating microscope through L7S1 intervertebral space following exploration of the lumbosacral junction. The dura was fixed and the injection was performed intrathecally through a 22-G needle with an insulin syringe. Muscle, fascia and skin layers were closed and sterilized again with iodine solution. Lumbosacral specimens removed from the site of intrathecal administration and exposed to post-fixation with 1% osmium tetroxide following fixation with 2.5% gluteraldehyde. They were embedded into epon (812) following routine electron microscopic procedures. Thin sections of epon blocs obtained from a Reicter ultra microtome were stained by uranyl acetate and lead citrate. Specimens were examined by a histologist who was blinded to the injected solution and the time of sacrifice. All rabbits were observed for sleeping habits, food intake and motor activity. The motor activity was assessed by observing the rabbits moving freely in their cages and the reaction provoked by sudden movement or the clapping of hands. Results No behavioural changes were observed in the animals following midazolam or normal saline administrations. In none of the rabbits SPO2 values reduced below 90 and HR reduced below 130 min1 during the administration. SPO2 varied from 95 to 100. The electron microscopic evaluations of Group M1 (n 8) revealed severe separation and degeneration in myelin rings of large myelinated axons, vacuole formation and oedema in the cytoplasm (Fig. 1), a honeycomb appearance in the myelin lamella, disorganization, slight separation in myelin lamella of moderately large and small axons, vacuole formation and oedema in the cytoplasm (Fig. 2), irregularities and invaginations of nuclear membrane (Fig. 3) at 24 h in all preparations. In Group C1 (n 6) specimens, myelin lamella and nuclear membranes were regular and a few small vacuoles were observed in the neurons. A specimen from a rabbit sacrificed at 24 h after saline injection is shown in Figure 4. In Group M2 (n 8) specimens, severe separation and degeneration in myelin lamella of large axons, vacuoles and oedema in the cytoplasm (Fig. 5), honeycomb view in myelin lamella of large axons, slight separations and disorganizations in the moderately large axons, very slight degeneration in the

2005 European Society of Anaesthesiology, European Journal of Anaesthesiology 22: 907912

Neurotoxicity with midazolam

909

Figure 1. In the Group M1 separation of myelin rings ( oedema ().

) vacuoles and

Figure 3. In the Group M1 nuclear membrane irregularities, invaginations ( ).

Figure 2. In the Group M1 honeycomb view of myelin lamella, slight separation in myelin lamella of small axon, vacuole and oedema in the cytoplasm.

myelin lamella of small axons (Fig. 6), irregularity and invaginations of nuclear membrane and a few, moderately large degenerated vacuoles and oedema (Fig. 7) were observed in the grids taken on the 6th day. In Group C2 (n 6) specimens, myelin lamella of axons and nuclear membranes were regular (Fig. 8), and vacuoles and oedema in the cytoplasm (Fig. 9) were observed in the neurons. Vacuole formation and oedema in the cytoplasm was the common histological changes observed in all groups.

Figure 4. In the Group C1 regular nuclear membrane ( ), regular myelin lamella ( ), vacuoles and oedema in the cytoplasm () (7200).

Discussion The lumbar spinal canal is a region which allows injection of various solutions including the anaesthetic agents and of which subarachnoid and epidural spaces allow catheter insertion. Various studies have

2005 European Society of Anaesthesiology, European Journal of Anaesthesiology 22: 907912

910

B. Ugur et al

Figure 7. In the Group M2 nuclear membrane irregularity and invaginations ( ), vacuoles and oedema () (5400).

Figure 5. In the Group M2 severe separation in myelin lamella ( uoles and oedema in the cytoplasm () (7200).

), vac-

Figure 8. In the Group C2 regular nuclear membrane and myelin lamella.

Figure 6. In the Group M2 honeycomb view in myelin lamella ( ), slight separations in the small to moderately large axons ( ) (7200).

shown that analgesia might be provided by intrathecal or epidural administration of midazolam in human beings in the postoperative period and chronic back pain [3,1720]. Studies on neurotoxic effects of intraspinally administered drugs are of utmost importance. Interestingly,

these drugs are generally tested for clinical side-effects prior to their administration, but not for neurotoxic effects. There are few studies on the neurotoxicity of intrathecally administrated midazolam in animals. Although animal toxicity studies have yielded conflicting results, the present study is consistent with those studies that found that intrathecal midazolam did cause neurotoxicity [2,11,12] and were different from those that suggested it did not [13,14,21]. Our results have shown that intrathecal single dose midazolam without chronic catheter administration caused clear histological changes in specimens obtained from spinal medulla by laminectomy after 24 h and 6 days following administration. We found separation in myelin lamella of small, moderate and large axons and honeycomb appearance, disorganization, nuclear membrane irregularity and vacuole formation in all of the sections after administrating midazolam. Myelin and nuclear membranes were regular but there were vacuoles in

2005 European Society of Anaesthesiology, European Journal of Anaesthesiology 22: 907912

Neurotoxicity with midazolam

911

Figure 9. In the Group C2 regular nuclear membrane, vacuoles in the cytoplasm, regular myelin lamella.

the cytoplasm in the sections from the saline groups. Vacuole formation and oedema in the cytoplasm were the common findings in both saline and midazolam groups. We believe that these changes might be attributable to the midazolam administration. Malinovsky and colleagues [2] reported necrosis, haemorrhage and other histopathological changes in two of nine spinal cords of rabbits that had received a single intrathecal injection of midazolam of 300 g (0.3 mL) by the intracisternal route by light and fluorescence microscopy. Although it is impossible to perform blind percutaneous puncture in the rabbit lumbar region without causing trauma, this is easier in the larger cervical subarachnoid space. Therefore, the authors preferred the intracisternal route. In contrast, Nishiyama and colleagues [14] reported no inflammatory reaction of the spinal cord in cats by light microscopy, in the acute phase up to 6 h after the direct administration of high doses of midazolam of 10 mg. They suggested that 6 h might not be long enough to see acute histological changes that would result in chronic changes. Also by light microscopy, Serrao and colleagues [13] did not find damage to the spinal cord, nerve roots or meninges and evidence of toxicity by administering to the rats midazolam for 48 weeks via an indwelling intrathecal catheter. Although there was no statistically significant difference when midazolam was compared with saline, they observed pathological changes to some extent in both groups and they attributed this finding to the mechanical presence of the catheter as in studies by Johansen and colleagues [21] in their sheep model by light microscopy. Indeed, chronic spinal cannulation has been shown to lead to changes, including fibrosis and lymphocytic infiltrations around the catheter in the subarachnoid space [4]. We administered drugs

intrathecally by a laminectomy method under microscopic control in order to minimize the trauma. Bozkurt and colleagues [12] administered a single dose of acidic midazolam (commercially available preparation) epidurally in neonatal rabbits. In this electron microscopic study, an acidic saline vehicle control (matching the pH of the midazolam) was used, and pathology was observed with the acidic vehicle formulations that exceeded the pathology observed with a non-acidified vehicle. They pointed out that one of the concerns when applying midazolam to the spinal cord in animal trials was the species and age differences of animals because the developing spinal cord where myelination is not complete might be more sensitive to midazolam. Johansen and colleagues [21] demonstrated no spinal cord toxicity after 43 days of daily 515 mg doses of intrathecally administered midazolam in a sheep model by light microscopy. The single histological finding in this study was characterized as a mild to moderate infiltration of inflammatory cells surrounding the catheter tract; this was present in all animals. They suggested that this inflammation was likely attributable to the mechanical presence of the catheter. In contrast to Johansen and colleagues [21] work, Erdine and colleagues [11] concluded that chronic intrathecal administration of midazolam should be avoided in human beings. They also suggested that the histological and vascular lesions in the spinal cord were well correlated with the use of midazolam and preservative free midazolam. In conclusion, we should consider that while intrathecal midazolam in human beings might provide positive effects on the analgesia, it may produce histological changes in the spinal cord. In our opinion, midazolam may cause neurotoxic effects on myelinated axons and nuclear membranes, but we also believe that further studies are required for determining whether these changes could cause irreversible neurological damage.

Acknowledgements This study was funded by Research Fund Accountancy of Adnan Menderes University. This study was presented orally at XXXVII. Turkish Anaesthesiology and Reanimation Society National and II. International Congress. References
1. Yanez A, Sabbe MB, Stevens CW, Yaksh TL. Interaction of midazolam and morphine in the spinal cord of the rat. Neuropharmacology 1990; 29: 359364.

2005 European Society of Anaesthesiology, European Journal of Anaesthesiology 22: 907912

912

B. Ugur et al
14. Nishiyama T, Matsukawa T, Hanaoka K. Acute phase histopathological study of spinally administered midazolam in cats. Anesth Analg 1999; 89: 717720. 15. Abolueish E, Barmada BA, Nemato EM et al. Acute and choronic effects of intrathecal morphine in monkeys. Br J Anaesth 1981; 53: 10271032. 16. Madsen JB, Jensen FM, Faber T et al. Choronic catheterization of the spinal epidural space in rabbits: a model for behavioural and histopathologic studies. Examination of meptazinol neurotoxicity. Acta Anaesthesiol Scand 1993; 37: 307313. 17. Gle S, BykkMdan B, Oral N, zcan N, TanrMverdi B. Comparison of caudal bupivacaine, bupivacainemorphine and bupivacainemidazolam mixures for postoperative analgesia in children. Eur J Anaesthesiol 1998; 15: 161165. 18. Kim MH, Lee YM. Intrathecal midazolam increases the analgesic effects of spinal blockade with bupivacaine in patients undergoing haemorrhoidectomy. Br J Anaesth 2001; 86: 7779. 19. Nishiyama T, Hanaoka K. Effect of diluent volume on post-operative analgesia and sedation produced by epidurally administered midazolam. Eur J Anaesthesiol 1998; 15: 275279. 20. Serrao M, Marks RL, Morley SJ, Goodchild CS. Intrathecal midazolam for the treatment of choronic mechanical low back pain: a controlled comparison with epidural steroid in a pilot study. Pain 1992; 48: 512. 21. Johansen MJ, Gradert TL, Satterfield WC et al. Safety of continuous intrathecal midazolam infusion in the sheep model. Anesth Analg 2004; 98: 15281535.

2. Malinovsky JM, Cozian A, Lepage JY, Mussini JM, Pinaud M, Souron R. Ketamine and midazolam neurotoxicity in the rabbit. Anesthesiology 1991; 75: 9197. 3. Valentine JMJ, Lyons G, Bellamy MC. The effect of intrathecal midazolam on postoperative pain. Eur J Anaesthesiol 1996; 13: 589593. 4. Cooms DW, Fratkin JD. Neurotoxicology of spinal agents. Anaesthesiology 1987; 66: 724726. 5. Abram SE. Spinal cord toxicity of epidural and subarachnoid analgesics. Reg Anaesth 1996; 21: 8488. 6. Eisenach JC. Three novel spinal analgesics; clonidine, neostigmin, amitriptyline. Reg Anaesth 1996; 21: 81133. 7. Goodchild CS, Noble J. The effects of intrathecal midazolam on sympathetic nervous system reflexes in man: a pilot study. Br J Clin Pharmacol 1987; 23: 279285. 8. Young CC, Prielipp RC. Benzodiazepines in the intensive care unit. Crit Care Clin 2001; 17: 843862. 9. Liu LL, Gropper MA. Postoperative analgesia and sedation in the adult intensive care unit: a guide to drug selection. Drugs 2003; 63: 755767. 10. Niv D, Whitman JG, Loh L. Depression of nociceptive sympathetic reflexes by the intrathecal administration of midazolam. Br J Anaesth 1983; 55: 541547. 11. Erdine S, Ycel A, zyalin S et al. Neurotoxicity of midazolam in the rabbit. Pain 1999; 80: 419423. 12. Bozkurt P, Tunali Y, Kaya G, Okar I. Histological changes following epidural injection of midazolam in the neonatal rabbit. Pediatr Anaesth 1997; 7: 385389. 13. Serrao JM, Mackenzie JM, Goodchild CS, Gent JP. Intrathecal midazolam in the rat: an investigation of possible neurotoxic effects. Eur J Anaesthesiol 1990; 7: 115122.

2005 European Society of Anaesthesiology, European Journal of Anaesthesiology 22: 907912