Severe Bleeding following Lumbar Sympathetic Blockade in Two Patients under Medication with Irreversible Platelet Aggregation Inhibitors

A 71-yr-old male patient was admitted to our pain clinic for LSB. He presented with intermittent claudication (left leg) and a trophic leg ulcer on the right foot due to a progressive PAD as shown by angiography. He reported a pain-free walking distance of 70 m. His medication included 500 mg/day ticlopidine (Tiklyd^®; Sanofi-Synthelabo, Berlin, Germany) for stroke prevention in the presence of carotid artery stenosis. History and physical examination revealed no signs of an increased bleeding disposition, and, thus, no coagulation parameters were obtained. Left-sided LSB was performed (5 ml bupivacaine, 0.5%) using radiographic control. No vascular puncture was noted. Within the next 12 h, his pain-free walking distance increased to 113 m. However, he complained of numbness at the medial side of his left thigh as well as groin pain. Two days later, a widespread skin hematoma was recognized (fig. 1A). Hemoglobin decreased from 13.5 g/dl to 10.3 g/dl. Ticlopidine was stopped; however, no bleeding time or other anticoagulation test was performed. Four days later (6 days after the first block), a second block was performed. Direct radiographic control using contrast medium confirmed an intravascular needle position. Redirecting the cannula resulted in the correct spread of solution, and 1.5 ml ethanol, 95%, was then administered. The following night, the patient complained of severe pain in the left groin with a simultaneous decrease in blood pressure and hemoglobin (8.9 days/dl) due to a large retroperitoneal hematoma (fig. 1B). Transfusion stabilized the patient's hemodynamic condition. He was discharged 5 days later without major complaints and with an improved walking distance.

A 79-yr-old female patient with generalized PAD, severe pain in her lower extremities, polyneuropathy, coronary artery sclerosis, and a history of stroke some years ago presented with symptoms of an acute obstruction of the left femoral arteries. Angiography demonstrated multiple occlusions predominately on the left side but did not reveal an indication for angioplasty or surgical reconstruction. After failure of prostaglandin E1 treatment, the patient was referred for diagnostic LSB 2 weeks later, since her foot pain was unbearable despite morphine medication. In the following 24 h, she complained of an increasing numbness of the entire left leg. Even in the absence of a history of any bleeding disorder, 75 mg/day clopidogrel (Plavix^®; Sanofi-Synthelabo, Berlin, Germany) was discontinued 3 days prior to LSB. Seventy-two hours later, the coagulation parameters, including the bleeding time (160 s), were within normal limits. LSB was performed at the L3 level using a 26-gauge needle and checking for any blood aspiration. The first injection of contrast medium revealed a paravertebral displacement. After redirecting the needle position, the typical spread of the solution was demonstrated. Intravascular injection had not been recognized. The application of 5 ml bupivacaine, 0.5%, led to a minor increase in skin temperature (1.5°C) without significant pain relief.

The patient spent her first 9 h on the ward uneventfully and then suddenly complained of burning groin and medial thigh pain. There were no other pathologic hints (soft abdominal wall; normal peristaltic sounds; blood pressure, 120–130/75 mmHg through 1 h; heart rate, 70–80 beats/min). Administration of a low-dose opiate decreased her pain remarkably, and 2 h later she was walking on the ward without complaints. However, another hour later (i.e. , 12 h after LSB), she was found pulseless. Resuscitation attempts were unsuccessful.

The autopsy revealed a massive coagulated hematoma (about 2.3–3 l) beneath the left psoas muscle with enlarged retroperitoneal hemorrhage (fig. 2). There were no visible puncture lesions in larger vessels. Severe atherosclerosis was detected in the aorta and in most arteries of the left leg. Due to a recent thrombosis of the left iliac artery, collateral vessels were enlarged.

The blockade of the lumbar sympathetic chain versus  a surgical lumbar sympathectomy has proven to be a safe technique. It has reduced mortality when using radiographic control or computed tomography (CT) guidance combined with application of contrast medium for evaluation of the spread of the solution. However, there are a few reports of serious complications of LSB due to failed puncture or application of the neurolytic agent too close to the spinal cord or the ureter. 8Out of more than 2,100 patients with LSB, 5 deaths have been reported, 2 of which were due to severe bleeding under anticoagulation. 8–11Although puncture of arterial or venous vessels may be more frequent than reported, large hematomas are obviously uncommon and of minor consequence in the absence of other risk factors. 8 

We have performed more than 2,000 LSBs since 1985, performing the conventional dorsal–lateral approach on one or two levels under plain radiographic control or CT guidance. All injections but the two reported cases were without any clinically relevant bleedings. Although we did not specifically look for these complications, all patients were monitored on a ward for 24–48 h. In addition, an internal quality assurance was performed, including the routine measurement of blood hemoglobulin and urine screening for erythrocytes. No cases of injury to the ureter were revealed, as determined by urine screening for erythrocytes. The incidence of other side effects was also minimal. Genitofemoral neuralgia after neurolytic LSB was less than 0.8%versus  1–30% reported elsewhere. 9,12Therefore, the dramatic course of these two patients taking antiplatelet agents suffering from severe bleeding after LSB was unexpected and alarming, more so since a clear cause-and-effect relationship could not be established.

Ticlopidine and clopidogrel are irreversible inhibitors of adenosine diphosphate–induced platelet aggregation. As a consequence, coagulation will gradually return to its normal function only after the affected platelets are being replaced within 5–7 days of discontinuation of treatment. These drugs have been reported to be more effective than aspirin in the prevention of atherosclerotic events in patients with a history of a recent stroke, myocardial infarction, severe PAD, or vascular stent implantation. 1,2Although purpura or epistaxis was observed in 2.9–5.3% of patients receiving clopidogrel, there was a significant reduction in the total number of bleeding events compared to aspirin. 3The new irreversible platelet aggregation inhibitors are therefore considered to be the gold standard for patients with manifested symptoms of atherosclerotic disease.

It remains unclear, however, in which way platelet function is affected in patients taking these new drugs undergoing surgery or other invasive therapeutic approaches. There are still no reliable tests available, including testing for bleeding time, to address this problem, to guide antiplatelet therapy, to calculate the individual risk of bleeding, and, especially, to determine if the coagulation has returned to normal after cessation of the platelet inhibitors. 3, **Drug companies therefore recommend stopping ticlopidine and clopidogrel 7 days prior to elective surgery. There are no studies regarding the safety of epidural or spinal anesthesia in patients during or immediately after discontinuation of antiplatelet drug treatment. **Only one case report describes a subarachnoidal hematoma following a very difficult lumbar puncture in a patient receiving ticlopidine. 5 

In the presented cases, it remains uncertain whether the platelet inhibitors were the direct cause of the retroperitoneal bleeding. An intravascular positioning of the needle was noted in the first patient, although puncturing a vessel with a 25-gauge needle is usually of little clinical consequence. Lumbar arteriography used to be performed with 16- to 20-gauge cannulas but with low rates of bleeding complications.

In the second patient, autopsy revealed no vascular lesions. The prolonged interval of 9 h between LSB and the onset of early signs (e.g. , groin and femoral pain) and the patient's general good health until then are strong arguments against the assumption of a needle-induced vessel lesion with consecutive bleeding of more than 2.5 l. The hemorrhagia could have been a consequence of increased pressure in the paraaortal collateral vessels due to the progressive as well as recent thrombotic occlusion of the left femoral artery revealed on autopsy. Although clopidogrel had been discontinued 4 days pre- viously, it still may have exerted some effects and, as such, may have been a synergistic factor in this scenario.

Since LSB is a common pain clinic procedure, it is important for other clinicians to be aware of the potential for severe hematomas following nerve blocks in patients taking platelet aggregation inhibitors. Still, LSB is considered to be a very safe technique with great benefits for high-risk patients with otherwise untreatable pain due to PAD. Therefore, the presented cases have led us to make the following recommendations:

1. Since there is a lack of reliable and sensitive tests to monitor coagulation, 4, **irreversible platelet inhibitors should be discontinued for at least 7 days prior to any invasive technique, in case any hemorrhage is not controllable by external compression. Following the recent addendum of the American Society of Regional Anesthesia for epidural and spinal injection, this precaution should be extended to LSB and celiac block. **In case of a high thrombotic risk, patients should be switched to a different therapeutic regimen with heparin up to 6 h prior to the blockade.

2. Patients with an increased risk of bleeding, i.e. , taking any platelet inhibitor or with a history of recent progression of PAD, should be closely monitored as inpatients for 24 h after LSB. 4 

3. Groin pain and, particularly, pain at the medial side of the thigh seem to be early signs of a bleeding complication after LSB. The side of the thigh pain indicates which side is affected in cases of bleeding into the iliopsoas muscle, as explained by the distribution area of the genitofemoral nerve. If the patient experiences a painful sensation such as that described in this report, an ultrasound examination or a more sensitive CT scan of the retroperitoneum should be performed immediately.