CardioVascular and Interventional Radiology

© Springer-Verlag New York, Inc. 1998

Cardiovasc Intervent Radiol (1998) 21:319 –323 DOI: 10.1007/s002709900268

CT-Guided Lumbar Sympathectomy: Results and Analysis of Factors Influencing the Outcome Walter Heindel,1 Stefan Ernst,1 Gudrun Manshausen,1 Michael Gawenda,2 Peter Siemens,1 Thomas Krahe,1 Michael Walter,2 Klaus Lackner1 1

Department of Diagnostic Radiology, University of Cologne Medical School, D-50924 Cologne, Germany Department of Surgery, University of Cologne Medical School, D-50924 Cologne, Germany

2

Abstract Purpose: To prospectively analyze the effectiveness of computed tomography-guided percutaneous lumbar sympathectomy (CTLS) in patients with peripheral arterial occlusive disease in relation to angiographic findings and vascular risk factors. Methods: Eighty-three patients were treated by CTLS. After clinical evaluation of the risk profile and diagnostic intraarterial digital subtraction arteriography, 14 patients underwent unilateral, and 69 bilateral one-level treatment. Follow-up studies took place on the day following the intervention, after 3 weeks, and after 3 months. Results: A total of 152 interventions were performed in 83 patients. After 3 months, clinical examination of 54 patients (5 patients had died, 24 were lost to follow-up) revealed improvement in 46% (25/54), no change in 39% (21/54), and worsening (amputation) in 15% (8/54). There was no significant statistical correlation among any of the analyzed factors (diabetes mellitus, arterial hypertension, smoking, hyperlipidemia, obesity, hyperuricemia, number of risk factors, ankle-arm index, and angiography score) and the outcome after CTLS. Three major complications occurred: one diabetic patient developed a retroperitoneal abscess 2 weeks after CTLS, and in two other patients ureteral strictures were detected 3 months and 2 years after CTLS, respectively. Conclusion: As no predictive criteria for clinical improvement in an individual patient could be identified, CTLS, as a safe procedure, should be employed on a large scale in patients who are unsuitable for treatment by angioplasty or revascularization. Key words: Sympathectomy, lumbar—Computed tomography, guidance—Arteries, extremities—Peripheral arterial Correspondence to: Prof. Dr. med. W. Heindel

occlusive disease—Interventional procedures, complications

Lumbar sympathectomy [1] is an accepted procedure for improving the peripheral blood supply in patients with arterial occlusive disease whose vascular status does not permit practical reconstructive measures [2, 3]. The pathophysiological mechanism of lumbar sympathectomy is based on eliminating the vasoconstriction caused by the sympathetic nerves, resulting in improvement of the collateral perfusion and reduction of the precapillary resistance in the region of reduced perfusion [4]. Contrary to former assumptions, recent studies have shown that a clinically detectable warming of the skin after sympathectomy is not due to a steal effect or a reduction in perfusion of the musculature in the extremities in favor of the skin [5]. At the same time, there is a reduction in the sympathetic pain component in patients with rest pain caused by ischemia. As the procedure does not prevent progress of the underlying disease, the duration of the effect is limited in many cases, and the technique is classified as palliative. Compared with the possibility of surgical lumbar sympathectomy and chemical neurolysis under fluoroscopic control, computed tomography (CT)-guided lumbar sympathectomy (CTLS) can be regarded as the simplest and safest technique [2, 6]. The method is relatively free from complications and side effects. The success rates quoted in the literature range from 21% to 89% [2, 3, 5, 7–10]. One important cause for this wide range might be differences in patient selection. We therefore tried to identify subgroups of patients who profit most from CTLS. Its effectiveness in peripheral arterial occlusive disease was analyzed prospectively in relation to the vascular risk factors and the initial angiographic findings.

320

W. Heindel et al.: CT-Guided Lumbar Sympathectomy

Patients and Methods

correct positioning of the cannula was checked, and the distribution of the injected solution in the region of the sympathetic trunk was verified following injection of a test solution of non-ionic contrast medium and long-acting local anesthetic (bupivacaine hydrochloride; Carbostesin 0.25% 5 ml, Astra Chemicals). Following the test injection, sympathetic nerves were interrupted by injecting 15–20 ml of a neurolytic solution (13 ml 96% alcohol, 5 ml bupivacaine hydrochloride) and 2 ml non-ionic contrast medium). During the injection, the patient was asked to describe the location of the short-duration burning pain. Apparent sensations in the back, abdomen, and groin were regarded as evidence of correct distribution of the neurolytic solution. Pain and paresthesia in the thigh indicate undesirable extension to the motor/ sensory plexus, and were regarded as an indication for stopping the injection and observing the extension by means of a control scan. In each case an interim check was performed after half the volume had been injected to enable the needle to be repositioned or the procedure to be abandoned. This is because different distributions of the alcohol solution were observed, depending on the individual anatomy. In this way the remainder of the solution could be applied in appropriate doses and location. Typically, the neurolytic solution was distributed along the anterior surface of the vertebra, dorsolateral to the abdominal aorta or vena cava, extending to the fascia over the psoas muscle. The total intervention took about 30 min. The patient was told to rest in bed for 3 hr after the intervention. Even where the clinical symptoms were purely unilateral, the intervention was repeated on the contralateral side after 2 or 3 days, as about 40% of the sympathetic nerve fibers cross to the other side [12]. Simultaneous interventions on both sides were avoided, as they are associated with an increased frequency of pain syndromes [2].

Patients A total of 83 patients with peripheral arterial occlusive disease were treated by CTLS in the period from January 1, 1992 to September 9, 1995 (51 men, 32 women; 41–93 years of age, mean 70.8 years). Thirty-four patients (41%) had a vascular surgical procedure and 13 (16%) had a radiological intervention (generally percutaneous angioplasty) in their medical history. Two patients have had an amputation previously. Clinically, 28 patients had stage IIb disease according to the Fontaine classification, 18 had stage III, and 37 stage IV. In 27 patients (33%) the obligatory pre-interventional intraarterial angiography showed isolated changes in the lower femoral artery. All angiograms were numerically scored according to the Rutherford grading scheme [11]. Sixty-eight (82%) of the patients treated were smokers, 41 (49%) had diabetes mellitus and 60 (72%) had arterial hypertension. Hyperlipidemia was present in 33 patients (40%). Only three patients (4%) were found to have no risk factors, 13 (16%) had one risk factor, 32 (38%) had two, and 35 (42%) had three or more risk factors. Bilateral sympathectomy was performed in 69 cases. In the remaining 14 patients only unilateral CTLS was performed as they refused the second intervention.

Methods The basic precondition for the treatment was—as is usual in percutaneous puncture—adequate coagulation status with a Quick value .50%, partial thromboplastin time ,45 sec, and platelets .70,000/ml. The CT-guided lumbar interruption of the lumbar sympathetic nerves was performed according to the following standardized technique, modified according to Haaga et al. [6]. After establishing access to the peripheral venous system, the patient was placed as comfortably as possible in the prone position, supported as necessary by padding under the head, arms, and lower thigh. Following enhancement of the ureters by intravenous administration of 20 ml non-ionic contrast medium (iodine content 300 mg/ml), 8-mm-thick CT slices were obtained for orientation, in a series running cranially from the iliac crest to the upper articular surface of L3. On the basis of these CT slices, a route was selected for the most caudal possible puncture of the region dorsolateral to the abdominal aorta (for left-sided sympathectomy) or dorsolateral to the vena cava (for right-sided sympathectomy). The puncture site was defined and marked on the patient’s skin. The depth and angle of puncture were measured in the CT images. After disinfection of the skin and sterile covering of the puncture site with a cover with access hole, local anesthesia was applied to the epidermis and dermis, down to the psoas muscle, in the selected puncture direction, using 10 ml mepivacaine hypochloride (Scandicain 1%, Astra Chemicals, Wedel/Holstein, Germany). The needle was left in situ in order to verify the puncture angle in a CT control slice. The local anesthetic needle was then replaced by a coaxial fine-needle system (22 G Chiba puncture needle with stylet, length 15 or 20 cm, Angiomed-Bard, Karlsruhe, Germany). The needle was positioned at the same angle, with the tip in the immediate proximity of the L3 or L4 sympathetic ganglion. Next, a further CT control slice was acquired. Following aspiration to exclude the possibility of the needle tip lying in a vessel,

Evaluation Follow-up examinations took place on the day following the intervention, after 3 weeks, and after 3 months. The effect of treatment was assessed according to the following criteria: Improved 1. At least doubling of the walking distance on a moving belt or (where moving belt ergometry was not possible) the subjective walking distance. 2. Elimination of rest pains. 3. Complete healing of ulcers. Patients suffering from both rest pains and ulcers were assessed as improved when only one of the symptoms disappeared. Worsened 1. At least half the walking distance on a moving belt or the subjective walking distance. 2. Appearance of new rest pain. 3. Amputation necessary. All patients who could not be allocated to one or the other of the above categories were assessed as unchanged. In addition to this general evaluation, subgroups were created depending on parameters derived from the patient history, clinical examination, and technical measurements, in order to determine any factors that might have a predictive value. Patients showing improvement were compared with patients in whom the symptoms were unchanged or worsened. The following factors were analyzed in each individual: nicotine abuse, concomitant disease such as

W. Heindel et al.: CT-Guided Lumbar Sympathectomy Table 1. Summarized results of patients with complete follow-up (n 5 54)a

Improved Unchanged Worsened a

321

Table 2. Outcome analysis with respect to initial Fontaine classes

Day 1

3 weeks

3 months

Fontaine class

18 (33.3%) 36 (66.6%) 0

20 (37.0%) 31 (57.4%) 4 (7.4%)

25 (46.3%) 21 (38.9%) 8 (14.8%)

IIb

III

IV

7/17 (41.2%) 10/17 (58.8%)

6/12 (50.0%) 6/12 (50.0%)

12/25 (48.0%) 13/25 (52.0%)

After 3 months, 29 of 84 patients had been lost to follow-up

arterial hypertension or diabetes mellitus, the ankle-arm pressure index, and the angiographic grading scheme. For statistical evaluation the SPSS program (SPSS, Chicago, IL, USA) was used.

Results The results of the follow-up studies are summarized in Table 1. After 3 months, 46.3% of the patients showed clinical improvement, whereas 38.9% had no change in their complaints and 14.8% had worsening of their symptoms. During the follow-up time the number of patients who improved increased. Data analysis of those patients who had elected for only unilateral sympathectomy did not show a statistically significant different outcome as compared with the bilaterally treated cohort (p 5 0.970; chi-squared test). In seven of the eight patients assessed as worsened, an amputation had to be performed during the follow-up period. All these patients were classified as having stage IV disease prior to the intervention. Table 2 compares the results after 3 months for the individual disease stages at the start of treatment. This shows improvement in approximately 50% of cases in Fontaine classes III and IV, but less improvement in Fontaine IIb cases. Table 3 summarizes the statistical analyses for the various factors that could influence the outcome. There was no significant statistical correlation among any of the analyzed factors and the outcome after CTLS. A tendency towards a better response to treatment could be observed in nondiabetics, patients with isolated crural artery changes, and patients with few atherogenic risk factors. Comparison of the angiographic score of improved (n 5 25) and nonimproved (n 5 29) patients showed no statistically significant difference. Analysis of side effects in our series of 152 procedures showed no steal phenomenon or paradoxical reaction. Apart from temporary pain in eight patients, three major complications occurred: an insulin-dependent diabetic patient developed a retroperitoneal abscess requiring surgery 2 weeks after intervention and in two patients a ureteral stricture was observed 3 months and 2 years respectively, after CTLS. In the latter two patients retrospective analysis of the CT scans demonstrated the needle tip to be close to the ureter in one patient and in the other patient, the tip of the needle was 4 mm from the ureter, but the neurolytic solution surrounded the ureter, as demonstrated on the test scan. As no other intervention in this area has been performed in either patient,

Improved Unchanged or worsened

Total

25 29

Table 3. Summary of the statistical analysis for the various factors that could influence the outcome Improved

n

Unchanged or worsened %

n

%

p value

Diabetes mellitus No diabetes mellitus

8 17

36.4 53.1

14 15

63.6 46.9

0.349a

Hypertension No hypertension

16 9

41.0 60.0

23 6

59.0 40.0

0.343a

Smoker Nonsmoker

18 7

46.2 46.7

21 8

53.8 53.3

1.000a

Hyperlipidemia No hyperlipidemia

9 16

37.5 53.3

15 14

62.5 46.7

0.376a

Obesity No obesity

8 17

42.1 48.6

11 17

57.9 51.4

0.865a

Hyperuricemia No hyperuricemia

6 19

46.2 46.3

7 22

53.8 53.7

1.000a

Risk factor 5 0 Risk factors $1

2 23

100 44.2

0 29

0 55.8

0.209b

Risk factor 5 1 Risk factors $2

5 20

62.5 43.5

3 26

37.5 56.5

0.273b

Risk factors #2 Risk factors $3

17 8

54.8 34.8

14 15

45.2 65.2

0.235b

Ankle-arm index Mean

25 0.62

Crural stenoses Combined stenoses (femoral and crural)

10 15

Angiography score Mean

25 11.15

29 0.60 58.8 40.5

7 22 29 12.14

0.869c 41.2 59.5

0.338a

0.391a

a

Pearson chi-squared test Fisher’s exact test c Student’s t-test b

the ureteral injury was definitely related to the lumbar sympathectomy.

Discussion CTLS by injection of neurolytic material is an alternative to surgical procedures, as it allows effective interruption of the sympathetic nerves with significantly less stress for the patient. The frequency of technically inadequate sympathectomies is possibly greater in surgical procedures, because the diffusion of an accurately injected neurolytic has the additional effect of interrupting alternative and anomalous sym-

322

W. Heindel et al.: CT-Guided Lumbar Sympathectomy

pathetic nerve fibers that could be missed in a surgical procedure [13]. The parameters generally used for assessing the success of CTLS are the pain-free walking distance and a rise in skin temperature [14, 15, 16]. A biochemical correlation for assessing the success via the blockade of the sympathetic trunk has been found by the use of phosphorus MR spectroscopy [17]. Our criteria for success were doubling of the preinterventional walking distance, the disappearance of rest pains, or the healing of ulcers. Our therapy results demonstrate that the effect of CTLS may not be evident immediately after the intervention, but improved in some patients during the follow-up interval of 3 months. The main aim of this study, however, was to evaluate possible predictors for the outcome of CTLS. A comparable favorable influence on the clinical symptoms of arterial occlusive disease was observed in each of the various stages. Other authors report success rates of up to 94% in stage II [16, 18]. Success rates of 33%–78% are reported for stage III [3, 16]; those for stage IV lie between 33% and 50% [3, 16]. Our data showed less improvement in Fontaine class IIb, which may be due to the fact that in our patients no special supervised walking training was done. In stage IIb in particular, a training effect and an associated increase in perfusion of the muscle collaterals can be a significant factor affecting the clinical symptoms, independently of CTLS. The effectiveness of CTLS in diabetics is disputed. Because in diabetic disease the resultant neuropathy should already have led to maximum dilation of the vessels [18], McCollum et al. [19] call this effect ’autosympathectomy’. Correspondingly, Ramos et al. [20] found significantly poorer results of lumbar sympathectomy in diabetics. Similar results have been reported by Cross and Cotton [21], Lee et al. [22], Strotzer et al. [10] and Duda et al. [3, 14]. We also observed a tendency to poorer response in our diabetic group, but this was not statistically significant as compared with the nondiabetics. Therefore, taking into account the low complication rate, it is our opinion that sympathectomy is justified even in the case of diabetic gangrene. The predictive value of all analyzed risk factors was shown to be limited from the statistical point of view. This is also true for combinations of multiple risk factors. Less surprising was the fact that the ankle-arm index could not predict the effectiveness of CTLS at all. Analysis of our own data with respect to the location of the occlusion confirms the pathophysiological concept that it is mainly patients with distal occlusions or stenoses who benefit from CTLS. As might be expected, due to the obstructed supply, patients with combined pelvic, upper femoral, and lower femoral changes clearly benefit less. However, similar to the analysis of the angiographic score, no statistically significant difference in outcome could be observed. In addition to analysis of the success rate, the possible complications of a procedure play a decisive role in deter-

mining the indications for treatment. Though the mortality rate for surgical lumbar sympathectomy has been reported as 1%– 6%, and may be even higher in diabetics [23], to our knowledge no death associated with CTLS has been described in the literature. Complications reported for CTLS are ureteral injury with periureteral fibrosis [24 –27], puncture of a lumbar vein [7], and neuralgia or paresthesia [2, 13, 16]. In one of our patients there was the unusual finding of a retroperitoneal abscess in a female patient with a history of many years of insulin-dependent diabetes mellitus, which may be a predisposing factor. The abscess was successfully treated by surgery. In summary, this prospective study is further evidence that CTLS represents a procedure with a low complication rate which, although it does not bring about a cure, can nevertheless lead to an improvement in the quality of life. In spite of a large group of patients being studied, no factor could be detected that could forecast the success of the procedure in individual patients. As any attempt to conserve a leg is certainly justified, even if only on the grounds of cost/benefit considerations [28], our results indicate that CTLS should be employed on a large scale in patients with peripheral occlusive arterial disease in cases where a revascularization procedure or angioplasty is not, or is no longer, possible, as well as to support wound healing following amputation. References ¨ ber paravertebrale Neurektomie am Grenzstrang 1. von Gaza W (1924) U und paravertebrale Injektionstherapie. Wien Klin Wocherschr 3:525– 528 2. Schild H, Gro¨nninger J, Gu¨nther R, Thelen M, Schwab R (1984) Transabdominelle CT-gesteuerte Sympathektomie. Fortschr Rontgenstr 141:504 –508 3. Duda SH, Huppert PE, Heinzelmann B, Schareck W, Seboldt H, Claussen CD (1994) CT-guided percutaneous lumbar sympathicolysis in peripheral arterial occlusive diseases. Fortschr Rontgenstr 160:132–136 4. Lindenauer SM, Cronenwett JL (1982) What is the place of lumbar sympathectomy? Br J Surg 69 [Suppl]: 932–933 5. Fraunhofer S, Henzler M, Strotzer M, von Sommoggy S (1991) Measuring regional blood flow in evaluating CT-controlled lumbar sympathetic neurolysis in arterial occlusive disease. Vasa [Suppl]33:160 –161 6. Haaga J, Reich N, Havrilla T, Alfidi R (1977) Interventional CTscanning. Radiol Clin North Am 15:285–289 7. Redman DR, Robinson PN, Al KM (1986) Computerised tomographyguided lumbar sympathectomy. Anaesthesia 41:39 – 41 8. Coubret P, Lermusiaux P, Becquemin JP, Etienne G, Melliere D (1985) La sympatholyse au phenol a-t-elle des indications? J Mal Vasc 10: 357–361 9. Kurdziel JC, Dondelinger RF (1990) Sympathetic chain. In: Dondelinger RF (ed) Interventional Radiology. Thieme, Stuttgart, pp 773– 780 10. Strotzer M, Lenz M, Fraunhofer S, von Sommoggy S (1992) CT-guided lumbar sympathicolysis: Technique and results. Rontgenpraxis 45:220 –224 11. Rutherford RB, Flanigan DP, Gupta SK, Johnston KW, Karmody A, Whittemore AD, Baker JD, Ernst CB (1986) Suggested standards for reports dealing with lower extremity ischemia. J Vasc Surg 4:80 –94 12. Zenker W, Neuhuber W (1994) Autonomes (viszerales, vegetatives) Nervensystem. In: Drenckhaln D, Zenker W (eds) Makroskopische Anatomie, Embryologie und Histologie des Menschen. Urban and Schwarzenberg, Baltimore 13. Rosen RJ, Miller DL, Imparato AM (1983) Percutaneous phenol sympathectomy in advanced vascular disease. AJR 141:597– 600 14. Duda SH, Huppert PE, Heinzelmann B, Schareck W, Seboldt H, Claus-

W. Heindel et al.: CT-Guided Lumbar Sympathectomy

323

sen CD, Heimig T, Scheffler A, Rieger H, Acevedo A (1993) Regional drug-induced sympathicolysis in patients with arterial occlusive disease. Wien Med Wochenschr 143:184 –185 Duncan JA (1985) Chemical lumbar sympathectomy. (editorial) J Bone Joint Surg Br 67:174 –175 Dondelinger R, Kurdziel JC (1984) Percutaneous phenol neurolysis of the lumbar sympathetic chain with computed tomography control. Ann Radiol Paris 27:376 –379 Melchert UH, Brinkmann G, Fo¨rger K, Gleim M, Wunsch-Binder F, Maier C, Zick G (1992) In-vivo 31-phosphorus MR spectroscopy of the calf musculature in arterial occlusive diseases. Fortschr Rontgenstr 156:346 –352 Walker PM (1982) Is there still a place for lumbar sympathectomy? Can Med Assoc 127:353–354 McCollum PT, Spence VA, Macrae B, Walker WF (1985) Quantitative assessment of the effectiveness of chemical lumbar sympathectomy. Br J Anaesth 57:1146 –1149 Ramos M, Almazan A, Lozano F, Allonso AG (1983) Phenol lumbar sympathectomy in severe arterial disease of the lower limbs: A hemodynamic study. Int Surg 68:127–130 Cross FW, Cotton LT (1985) Chemical lumbar sympathectomy for

ischemic rest pain: A randomized, prospective controlled clinical trial. Am J Surg 150:341–345 Lee BJ, Madden JL, Thoden WR, McCann WJ (1983) Lumbar sympathectomy for toe gangrene: Long-term follow-up. Am J Surg 145: 398 – 401 Imparato AM (1979) Lumbar sympathectomy: Role and treatment of occlusive arterial disease in the lower extremities. Surg Clin North Am 59:719 –735 Daschner H, Allgayer B (1994) Ureteral obstruction following CTguided lumbar sympathetic exclusion. Fortschr Rontgenstr 161:85– 87 Fraser I, Windle R, Smart JG, Barrie WW (1984) Ureteric injury following chemical sympathectomy. Br J Surg 71:349 Trigaux JP, Decoene B, Van Beers B (1992) Focal necrosis of the ureter following CT-guided chemical sympathectomy. Cardiovasc Intervent Radiol 15:180 –182 Ernst S, Heindel W, Fischbach R, Gawenda M, Langen H-J, Neubauer S, Krake T (1998) Complications of CT guided lumbar sympathicolysis: own experience and review of literature. Fortschr Rontgenstr 168: 77– 83 Bruijnen H, Loeprecht H (1989) Cost-benefit analysis of limb salvage. Langenbecks Arch Chir [Suppl II]:617– 619

15. 16. 17.

18. 19. 20. 21.

22. 23. 24. 25. 26. 27.

28.