Introduction
Neurolytic blocks of the sympathetic axis were procedures that were widely used in the past for control of chronic upper abdominal pain or pelvic pain in patients with cancer. However, recent studies suggest that these blocks are not effective in treating pain that is not visceral in origin. Consequently, when there is evidence of disease outside the viscera, for example, lymphadenopathy, the success rate decreases significantly. Moreover, a controlled randomized study has shown that even in the best-case scenario, the duration of full pain control is 2 months. Thus we should reconsider the indications for these procedures, and when indicated, they should be performed early in the course of the disease.
Stretching, compressing, invading, or distending visceral structures can result in poorly localized, noxious visceral pain. Patients experiencing visceral pain often describe the pain as vague, deep, squeezing, crampy, or colicky. Other signs and symptoms include referred pain (e.g., shoulder pain that appears when the diaphragm is invaded by tumor) and nausea and vomiting as a result of vagal irritation.
Visceral pain associated with cancer may be relieved by oral pharmacologic therapy, which classically includes combinations of nonsteroidal anti-inflammatory drugs (NSAIDs), opioids, and co-adjuvant therapy. NSAIDs have begun to fall out of favor with some clinicians for the treatment of chronic pain because they have been implicated in thrombotic events. In addition to pharmacologic therapy, neurolytic blocks of the sympathetic axis are also effective in controlling visceral cancer pain and should be considered as important adjuncts to pharmacologic therapy for relief of severe visceral pain. These blocks rarely eliminate cancer pain because patients frequently experience somatic and neuropathic pain as well. Therefore, oral pharmacologic therapy must be continued in the majority of patients with advanced stages of their disease. The goals of performing a neurolytic block of the sympathetic axis are to maximize the analgesic effects of opioid or nonopioid analgesics and reduce the dosage of these agents to alleviate side effects.
Since neurolytic techniques have a narrow risk-to-benefit ratio, undesirable side effects and complications from neurolytic blocks can be minimized by sound clinical judgment and by assessment of the potential therapeutic effect of the technique on each patient. This chapter discusses pertinent information regarding neurolytic block of the celiac plexus, superior hypogastric plexus, and ganglion impar.
Celiac Plexus Block
The celiac plexus is situated retroperitoneally in the upper part of the abdomen. It is at the level of the T12 and L1 vertebrae, anterior to the crura of the diaphragm. The celiac plexus surrounds the abdominal aorta and the celiac and superior mesenteric arteries. The plexus is composed of a network of nerve fibers from both the sympathetic and parasympathetic systems. It contains two large ganglia that receive sympathetic fibers from the three splanchnic nerves (greater, lesser, and least). The plexus also receives parasympathetic fibers from the vagus nerve. Autonomic nerves supplying the liver, pancreas, gallbladder, stomach, spleen, kidneys, intestines, and adrenal glands, as well as blood vessels, arise from the celiac plexus.
Neurolytic block of the celiac plexus has been used for malignant and chronic nonmalignant pain. In patients with acute or chronic pancreatitis, celiac plexus block has been used with variable success. Likewise, patients with upper abdominal cancer who have a significant visceral pain component have responded well to this block.
The Cochrane collaboration performed a systematic review of all randomized controlled studies investigating the effectiveness of celiac plexus block in the management of pancreatic cancer pain. They concluded that despite only a small amount of evidence for an increase in pain relief, the improved side effect profile as a result of the reduced need for opiates is desirable. It must be considered that many of the studies comparing the various interventional procedures with oral pharmacotherapy included NSAIDs. If more protocols in the future exclude NSAIDs, there may become a bigger statistical difference between the interventional and pharmacotherapy groups.
Three approaches to block nociceptive impulses from viscera in the upper abdominal region include the retrocrural (or classic) approach, the anterocrural approach, and neurolysis of the splanchnic nerves. Regardless of the approach, the needle or needles are inserted at the level of the first lumbar vertebra, 5 to 7 cm from the midline. Then the tip of the needle is directed toward the upper third of the body of L1 for the retrocrural and toward the lower third of the body of L1 for the anterocrural technique ( Fig. 59.1 ). In the case of the retrocrural approach, the tip of the needle is advanced no more than 0.5 cm anterior to the anterior border of L1, and with the anterocrural approach (see Fig. 59.1 ), the tip of the needle is advanced through the aorta on the left side until no more blood flow is noted through the needle. This is why the anterocrural approach is also known as transaortic approach. In the case of a splanchnic nerve block, the needle is directed toward the body of T12 ( Fig. 59.2 ). Perfect needle positioning in this case is achieved when the tip of the needle is at the anterior portion of the T12 vertebral body on the lateral view ( Fig. 59.3 ).
More recently, computed tomography (CT) and ultrasound techniques have allowed pain specialists to perform neurolysis of the celiac plexus via a transabdominal approach. This approach is frequently used when patients are unable to tolerate either the prone or lateral decubitus position or when their liver is so enlarged that a posterior approach is not feasible. Moreover, CT guidance allows an anterocrural technique to be performed without piercing the aorta, thus adding an element of safety in this case (see the section “Complications”).
Drugs and Dosing
For neurolytic blocks performed via the retrocrural or anterocrural approaches, 50% to 100% alcohol is used. Injected by itself, alcohol can produce severe pain. Thus, it is recommended that 5 to 10 mL of 0.25% bupivacaine first be injected 3 to 5 minutes before injection of the alcohol or at the time of the injection by diluting 100% alcohol to a 50% concentration with the same amount of local anesthetic (0.25% bupivacaine). Phenol in a 6% to 10% final concentration may also be used; it has the advantage of being painless on injection, and both agents seem to have the same clinical efficacy. The dose of alcohol or phenol administered varies with the approach to be used. For the retrocrural approach, 20 to 25 mL of alcohol is injected on each side. Consequently, the need to inject this high volume precludes the use of phenol in the retrocrural approach. For the anterocrural approach, 8 to 10 mL of either neurolytic agent is used per side. For splanchnic nerve blocks, 6 to 8 mL of phenol per side is recommended.
Complications
The incidence of complications from neurolytic celiac plexus blocks (NCPBs) was evaluated by Davis in 2730 patients who had blocks performed from 1986 to 1990. The overall incidence of major complications (e.g., paraplegia, bladder and bowel dysfunction) was 1 in 683 procedures. However, the report does not describe which approach or approaches were used to perform the blocks.
Complications associated with celiac plexus block appear to be related to the technique used: retrocrural, transcrural, or transaortic. In a prospective, randomized study of 61 patients with cancer of the pancreas, Ischia and colleagues compared the efficacy and incidence of complications associated with these three approaches with those of celiac plexus neurolysis. Orthostatic hypotension occurred more often when the retrocrural (50%) or splanchnic (52%) technique was used, thus suggesting associated neurolysis of the sympathetic chain. In contrast, the anterocrural approach produced a 10% incidence of hypotension. Conversely, transient diarrhea was more frequent with the anterocrural approach (65%) than with the splanchnic nerve block technique (5%) or the retrocrural approach (25%). The incidence of dysesthesia, interscapular back pain, reactive pleurisy, hiccups, or hematuria was not statistically different among the three groups.
The following paragraphs discuss several aspects involved in the diagnosis and management of specific complications.
Malposition of the needle is avoided with radiologic imaging before injection of a neurolytic agent inasmuch as the tip of the needle may be intravascular, in the peritoneal cavity, or in a viscus. Imaging guidance techniques currently used include biplanar fluoroscopy, CT, and ultrasound. However, no study has evaluated the superiority of one technique over the others. Wong and Brown suggested that the use of radiologic imaging does not alter the quality of the block or the incidence of complications based on a retrospective study of 136 patients with pancreatic cancer pain treated with a celiac plexus block with or without radiologic control of the position of the tip of the needle. However, it is not clear how many of these patients underwent radiologic imaging. Assuming that half the patients did not, the upper 95% confidence limit for complications would be 5%.
Orthostatic hypotension may occur up to 5 days after the block. Treatment includes bed rest, avoidance of sudden changes in position, and replacement of fluids. Once the compensatory vascular reflexes are fully activated, this side effect disappears. Wrapping the lower extremities from the toes to the upper part of the thighs with elastic bandages has been successful in patients in whom orthostatic hypotension developed and thus enabled them to walk during the first week after the block.
Backache may result from local trauma during needle placement and subsequent retroperitoneal hematoma or from irritation of retroperitoneal structures by alcohol. Patients with a backache should have at least two hematocrit measurements at a 1-hour interval. If there is a decrease in the hematocrit, radiologic imaging is indicated to rule out a retroperitoneal hematoma. Urinalysis positive for red blood cells suggests renal injury.
Retroperitoneal hemorrhage is rare; however, in patients with orthostatic hypotension, the possibility of hemorrhage must be ruled out before assuming that it is a physiologic response to the block. Patients in whom backache and orthostatic hypotension develop after a celiac plexus block should be admitted to the hospital for serial hematocrit monitoring. If the hematocrit level is low or decreasing, patients should undergo radiologic evaluation to rule out injury to the kidneys, the aorta, or other vascular structures. Surgical consultation should be obtained as soon as feasible.
Diarrhea may occur as a result of sympathetic block of the bowel. Treatment includes hydration and antidiarrheal agents. Oral loperamide is a good choice, although any anticholinergic agent may be used. Matson and colleagues reported nearly fatal dehydration from diarrhea following this block. In debilitated patients, diarrhea must be treated aggressively.
Abdominal aortic dissection has also been reported. The mechanism of aortic injury is direct damage with the needle during performance of the block. As expected, the anterocrural approach is more frequently associated with this complication. Thus, this approach should be avoided in patients with atherosclerotic disease of the abdominal aorta.
Paraplegia and transient motor paralysis have occurred after celiac plexus block. These neurologic complications may be due to spasm of the lumbar segmental arteries that perfuse the spinal cord. In fact, canine lumbar arteries undergo contraction when exposed to both low and high concentrations of alcohol. Thus, these data suggest that alcohol should not be used if there is evidence of significant atherosclerotic disease of the aorta because the circulation to the spinal cord may also be impaired and be dependent only on lumbar artery flow. However, there is also a report of paraplegia after the use of phenol, which suggests that other factors (e.g., direct vascular or neurologic injury or retrograde spread to the spinal cord) may come into play. These complications further support the need for radiologic imaging when performing these blocks.
Efficacy
To date, only three randomized, controlled trials and one prospective study have evaluated the efficacy of celiac plexus neurolysis in relieving pain caused by upper abdominal cancer. In a prospective, randomized study, Ischia and coworkers evaluated the efficacy of three different approaches to celiac plexus neurolysis for pancreatic cancer. Of 61 patients with pancreatic cancer pain, 29 (48%) experienced complete relief of pain after the neurolytic block. The remaining 32 patients (52%) required further therapy for residual visceral pain secondary to technical failure in 15 patients and neuropathic or somatic pain in 17 patients. The second trial, which compared the procedure with oral pharmacologic therapy in 20 patients, concluded that celiac plexus neurolysis results in an equal reduction in visual analog scale (VAS) pain score as does therapy with a combination of NSAIDs and opioids. However, opioid consumption was significantly lower in the group of patients who underwent neurolysis than in the group receiving oral pharmacologic therapy during the 7 weeks of the study. Moreover, the incidence of side effects was greater in patients who received oral pharmacologic therapy than in those who underwent neurolytic block. Regarding the third randomized controlled study, Wong and collaborators are to be congratulated for designing and completing this study. Their results are welcome in light of a lack of properly designed comparative studies between neurolytic techniques and comprehensive medical management (CMM). However, several issues in the design and results of this study need to be highlighted:
- 1.
Patients enrolled in the study did not have severe pain at study entry. Mean pain scores at baseline were 4.4 ± 1.7 in the NCPB group and 4.1 ± 1.8 in the CMM group. This is a surprising finding in patients with this type of malignancy and may reflect ethnic and racial differences in pain perception and reporting by the population enrolled in the study.
- 2.
Although the authors reported a significant statistical reduction in pain scores 1 week after therapy when comparing the NCPB group with the CMM group, the difference between the two groups may not be clinically important. Patients assigned to the NCPB group reported mean pain scores of 2.1 ± 1.4, whereas those randomized to the CMM group reported pain scores of 2.7 ± 2.1 at that time interval. Moreover, statistical difference was found only when the percent reduction from baseline in the NCPB and the CMM group was analyzed separately (53% reduction from baseline in the NCPG group, P = 0.05, versus a 27% reduction observed in the CMM group, P = 0.01).
- 3.
In analyzing these results it is critically important to note that most patients (93%) took opiates during the first week of therapy, with similar amounts of opiates being administered to the two treatment groups. In fact, opiate consumption increased with time with no differences between groups at the different time intervals during the study. Furthermore, the incidence of side effects was not different between the two treatment groups at any point in time.
- 4.
Likewise, quality-of-life measurements and the physical and functional well-being subscales of the Functional Assessment of Cancer Therapy for Prostate Cancer did not differ between the two groups at any evaluation point.
Two important questions stem from these results:
- 1.
Can the authors truly conclude that the major finding of the study was that NCPB significantly improves relief of pain in patients with advanced pancreatic cancer when compared with those who received optimized CMM?
- 2.
Based on these results, are we justified to submit a patient with advanced pancreatic cancer to NCPB in view of the potential side effects and complications associated with this procedure?
We do not believe that the authors can conclude that NCPB significantly improves pain relief in patients with advanced pancreatic cancer . This is partly because the levels of analgesia achieved by the patients assigned to either group after 1 week of therapy can be considered clinically acceptable. Additionally, statistical difference was found only when the authors analyzed the percent reduction in pain from baseline in each of the treatment groups.
Likewise, based on these results we do not believe that we would recommend NCPB to a patient with advanced pancreatic cancer because in this not so perfect world, complications do occur, as has been addressed earlier.
Given these reservations, does this mean that we should not perform NCPB in patients with pancreatic malignancy? As with every clinical study, the results of Wong and associates apply only to the population studied and under the conditions of the study protocol design. The critical issue is that all patients had nonresectable disease, which suggests that the patients were likely to have other pain components such as somatic or neuropathic, which are not responsive to NCPB. This is because neurolytic blocks of the sympathetic axis are effective in treating visceral pain only. Moreover, previous studies have suggested that in patients with evidence of disease outside the pancreas, such as celiac or portal adenopathy, the success rate of this block decreases significantly. In the study by De Cicco and collaborators, long-lasting pain relief was described in 9 of 9 patients (95% confidence interval of 60 to 100) when contrast medium spread into the four quadrants and in 10 of 21 patients (95% confidence interval of 26 to 70) when contrast medium spread into three quadrants. None of the 75 patients with spread of contrast agent into two or one quadrant experienced long-lasting pain relief. Thus, the presence of adenopathy secondary to metastasis is a poor prognostic factor for success of the block. This decrease in effectiveness is not due to a mechanical factor such as big lymph nodes blocking spread of the neurolytic agent; rather, it is simply a marker of more extensive disease that often includes nonvisceral pain components and is therefore less amenable to NCPB. The results of the study by Wong and collaborators further support the notion that NCPB should not be performed in patients with advanced unresectable carcinoma of the pancreas. This block should be reserved for patients without evidence of disease outside the viscera so that one is guaranteed that the patient has a visceral pain component only.
A prospective, nonrandomized study compared 41 patients treated according to the World Health Organization guidelines for relief of cancer pain with 21 patients treated by NCPB. The authors concluded that this technique can play an important role in managing pancreatic cancer pain.
Since one of the three randomized, controlled studies compared different approaches to the celiac plexus and had no control group and the other study compared the procedure with an analgesic drug, it is not possible to estimate the success rate of this technique. In contrast, the results of a meta-analysis that evaluated the findings of 21 retrospective studies with a total of 1145 patients concluded that adequate to excellent pain relief can be achieved in 89% of patients during the first 2 weeks following the block. Partial to complete pain relief continued in approximately 90% of the patients who were alive at the 3-month interval and in 70% to 90% of patients during the 3-month interval before death. Moreover, its efficacy was similar in patients with pancreatic cancer as in those with other upper intra-abdominal malignancies. However, these results are based on retrospective evaluation, which may not yield reliable information or may be subject to publication bias. In addition, the statistical techniques used for the analysis must account for the heterogeneity produced by patient selection criteria, technical differences in performance of the blocks, choice of neurolytic agents and doses used, diversity in the tools for evaluation of pain, goals of therapy, and other factors. Thus, the meta-analysis must be interpreted with caution because the report may be overly enthusiastic.
As discussed previously, oral pharmacologic therapy with opioids, NSAIDs, and co-adjuvants is used frequently for the treatment of cancer pain. However, the evidence suggests that chronic use of high doses of opioids may have a negative effect on immunity. Thus, analgesic techniques that lower opioid consumption may have a positive effect on patient outcomes. Lillemoe and colleagues showed in a prospective, randomized trial that patients with nonresectable cancer of the pancreas who underwent splanchnic neurolysis lived longer than did those who did not undergo neurolysis. These findings may be the result of lower opioid use in the neurolysis patients, who not only had better-preserved immune function but also experienced fewer side effects (e.g., nausea and vomiting), thus allowing them to eat better. Although the study by Wong and colleagues did not show that patients randomized to the neurolytic arm of the study lived longer, this may be explained by their high intake of opioids during the study period, which negated the effect of the blocks. Consequently, the effects of this procedure on long-term survival will need further experimental investigation.
Celiac Plexus Block
The celiac plexus is situated retroperitoneally in the upper part of the abdomen. It is at the level of the T12 and L1 vertebrae, anterior to the crura of the diaphragm. The celiac plexus surrounds the abdominal aorta and the celiac and superior mesenteric arteries. The plexus is composed of a network of nerve fibers from both the sympathetic and parasympathetic systems. It contains two large ganglia that receive sympathetic fibers from the three splanchnic nerves (greater, lesser, and least). The plexus also receives parasympathetic fibers from the vagus nerve. Autonomic nerves supplying the liver, pancreas, gallbladder, stomach, spleen, kidneys, intestines, and adrenal glands, as well as blood vessels, arise from the celiac plexus.
Neurolytic block of the celiac plexus has been used for malignant and chronic nonmalignant pain. In patients with acute or chronic pancreatitis, celiac plexus block has been used with variable success. Likewise, patients with upper abdominal cancer who have a significant visceral pain component have responded well to this block.
The Cochrane collaboration performed a systematic review of all randomized controlled studies investigating the effectiveness of celiac plexus block in the management of pancreatic cancer pain. They concluded that despite only a small amount of evidence for an increase in pain relief, the improved side effect profile as a result of the reduced need for opiates is desirable. It must be considered that many of the studies comparing the various interventional procedures with oral pharmacotherapy included NSAIDs. If more protocols in the future exclude NSAIDs, there may become a bigger statistical difference between the interventional and pharmacotherapy groups.
Three approaches to block nociceptive impulses from viscera in the upper abdominal region include the retrocrural (or classic) approach, the anterocrural approach, and neurolysis of the splanchnic nerves. Regardless of the approach, the needle or needles are inserted at the level of the first lumbar vertebra, 5 to 7 cm from the midline. Then the tip of the needle is directed toward the upper third of the body of L1 for the retrocrural and toward the lower third of the body of L1 for the anterocrural technique ( Fig. 59.1 ). In the case of the retrocrural approach, the tip of the needle is advanced no more than 0.5 cm anterior to the anterior border of L1, and with the anterocrural approach (see Fig. 59.1 ), the tip of the needle is advanced through the aorta on the left side until no more blood flow is noted through the needle. This is why the anterocrural approach is also known as transaortic approach. In the case of a splanchnic nerve block, the needle is directed toward the body of T12 ( Fig. 59.2 ). Perfect needle positioning in this case is achieved when the tip of the needle is at the anterior portion of the T12 vertebral body on the lateral view ( Fig. 59.3 ).
