Spinal Epidural Hematoma after Spinal Anesthesia in a Patient Treated with Clopidogrel and Enoxaparin

A 81-yr-old woman (height, 160 cm; weight, 75 kg) with American Society of Anesthesiologists physical status III had received spinal anesthesia for elective fasciotomy and perforator vein ligation because of a lower limb ulceration. Her medical history included ischemic heart disease grade II (Canadian Cardiovascular Society angina classification) with atrial fibrillation, compensated renal insufficiency (creatinine clearance 34 ml/min), hypertension, and insulin-dependent diabetes. Her oral medication included molsidomine (Corvaton™; Aventis Pharma Deutschland GmbH, Bad Soden, Germany), piretanide/ramipril (Arelix ACE™; Aventis Pharma Deutschland GmbH), spironolactone/furosemide (Furorese™; Hexal AG, Holzkirchen, Germany), metoprolol (Beloc zok™;AstraZeneca GmbH, Wedel, Germany), and xipamid (Aquaphor™;Lilly Deutschland GmbH, Bad Homburg, Germany) for treatment of hypertension, ischemic heart disease, and renal insufficiency. Diabetes was treated with subcutaneous insulin. In addition, she was treated daily with 75 mg clopidogrel (Iscover™;Bristol-Myers Squibb GmbH, München, Germany) for ischemic heart disease which was discontinued 7 days before surgery.

Preoperatively the activated partial thromboplastin time was 39 s (normal range, 30–40 s), the prothrombin time was 85% (normal range, 70–120%) and the international normalized ratio was 1.02 (therapeutic level, 2–4.5), platelet count was 161 × 10⁹/l, at the lower limit of the normal range (150 × 10⁹/l – 400 × 10/l).

Spinal anesthesia was performed by an experienced staff anesthesiologist using a 22-gauge Sprotte needle. However, the first attempt failed at L3–4 because of repeated bone contact as well as a sanguine puncture presumed to be outside the epidural space. At the L4–5 level subarachnoid puncture was performed uneventfully in the first attempt and 3 ml of plain bupivacaine 0.5% was injected. Within 10 min maximum level of sensory block reached the T8 dermatome. Surgery was performed uneventfully and the regression of the block was complete within 4 h. The patient was mobilized on the evening of the day of surgery. For prophylaxis of thromboembolism the patient received two doses of low molecular weight heparin (enoxaparin) 40 mg 8 and 36 h after lumbar puncture. During the second postoperative night, 4 h after the second enoxaparin administration, the patient complained of voiding difficulty, and a urinary catheter was applied. At that time neither sensory nor motor deficiency was observed. On the morning of the second postoperative day the patient complained of numbness and weakness in both lower limbs. No back pain was reported. Emergency magnetic resonance imaging revealed spinal epidural hematoma extending from T12 to L3, with a maximum diameter between T12 and L1 (figs. 1–3). The patient was transferred to the university hospital and presented with lower paraplegia and a Th12 sensory loss. Emergency decompressive laminectomy was performed 20 h after the first signs of bladder dysfunction were observed. Before the second surgical procedure coagulation profile was documented as activated partial thromboplastin time 44, prothrombin time 81, international normalized ratio 1.22, and a platelet count of 216 × 10⁹/l. The operative site showed a partially liquid and partially organized hematoma, which was difficult to resect as organized parts showed strong adhesion to the conus and seemed to obstruct epidural veins. No lesions related to puncture or vessel malformations were seen.

Postoperatively, the patient could partially be mobilized and only slowly recovered from her neurologic deficiencies. Three weeks after laminectomy sensory deficiency up to T12 and motor block Bromage 2 were documented.

Beneficial effects have been demonstrated for new antiplatelet drugs in cardiovascular patients, especially for thienopyridine derivatives in preventing stroke.3–5As an increasing number of patients is treated with these drugs, anesthetists must be familiar with their mechanism of action. In Europe two orally applicable adenosine diphosphate receptor antagonists are available that differ in clinical pharmacology. Both drugs inhibit platelet aggregation by noncompetitive nonreversible interaction with the adenosine diphosphate (P2Y¹²) receptor on the outer platelet surface. Blockade leads to an inhibition of adenosine diphosphate mediated platelet activation via  an increase in intracellular cyclic adenine monophosphate. In addition, release of calcium, fibrinogen, and serotonin that usually boost platelet activation is averted. Concurrently, alteration of GP IIb/IIIa conformation is inhibited.6 

Clopidogrel is six times more potent than ticlopidine7and is therefore effective within 3–7 (ticlopidine, 8–10) days after its application.8Normalization of platelet function is restored accordingly to the half-life time of platelets within 7 days (ticlopidine, 7–14 days).9In an animal model high-dose aprotinin partially reversed the effects of clopidogrel.10Platelet transfusion is the only effective treatment in cases of severe bleeding in humans.6 

For clopidogrel a therapy-free interval of 7 days is recommended before performance of neuraxial blocks.1,11In our patient 7 days, as recommended, were maintained and no sustaining clopidogrel impact was suspected.

Nevertheless, an epidural hematoma developed after spinal anesthesia. No case of epidural hematoma occurrence after central neuraxial block performance has been previously reported in patients shortly after a 7 day clopidogrel therapy-free interval. The first sign of epidural bleeding was bladder dysfunction, which was not considered unusual by the nursing staff at that time after spinal anesthesia in a patient of that age. Onset of sensory and motor deficiencies were delayed by approximately 12 h. Once discovered, magnetic resonance imaging was performed immediately and the patient was transferred to our hospital for laminectomy.

We consider four points to be of possible importance in this case:

• First, sanguine puncture occurred, which has been shown to be a general risk factor in development of spinal hematoma.12,13Even though in the first attempt it was presumed that the epidural space was not reached, puncture of an epidural vein can not be excluded.

• Second, as the platelet count was at the lower limit of normal range, platelet regeneration may have been reduced, hence prolonging the effect of the adenosine diphosphate antagonist.

• Third, the exact response to subcutaneous low molecular weight heparin therapy in our patient remains speculative, as subcutaneous blood flow may vary considerably interindividually in the elderly and may result in therapeutic anticoagulation. A further risk factor might have been the moderate impairment of renal function. A linear correlation between antifactor Xa concentrations and renal function has recently been demonstrated in patients receiving low molecular weight heparin, and the authors concluded that a dose adjustment is necessary in patients with a creatinine clearance of less than 30 ml/min.14In an older retrospective study even clearance rates lower than 40 ml/min were associated with higher antifactor Xa concentrations.15In our patient the clearance was slightly greater than 30 ml/min.

• Finally, a 7–10 day therapy-free interval before operative procedures is recommended according to some manufacturer guidelines.16The additive effect of anticoagulatory drugs acting at different sites of the coagulatory cascade should not be underestimated.

In summary, several factors may have equally contributed to the occurrence of an epidural hematoma in our patient. For thienopyridine derivatives clinical possibilities of testing thrombocyte function are limited.17This case further underscores the necessity of close neurologic monitoring after spinal anesthesia, especially in applying locoregional anesthetic techniques shortly after the recommended therapy-free interval of clopidogrel.