Chronic visceral pain is very common. Abdominal pain is among the main reasons for physician visits, with more than 12 million consultations occurring each year in the United States. Patients with visceral pain present unique challenges because the pain is often poorly localized and is associated with strong autonomic reactions and changes in visceral function. Pain management, in turn, may further alter visceral function, with opioid effects on the gastrointestinal tract providing a good example. These unintended treatment effects on visceral function can exacerbate the pain or lead to additional discomfort, thus showing that rational and effective pain management needs to be based on an understanding of the anatomic and physiologic basis of visceral function and pain.
Physiologic Basis of Visceral Sensation and Pain
Anatomy and Physiology of Visceral Afferent Pathways
Most viscera arise from midline structures and thus receive bilateral innervation. As a result, visceral stimuli activate both hemispheres of the brain, with a predominance of the left side in most right-handed individuals. In addition, organs in the chest cavity and most viscera within the abdomen receive dual afferent innervation, with vagal and spinal nerves conveying sensory input to the central nervous system ( Fig. 32.1 ). Although vagal fibers do not reach the pelvic organs, the distal part of the colon, bladder, prostate, and uterus also have a complex sensory innervation, with afferents projecting through the thoracolumbar (hypogastric) and lumbosacral (pelvic) nerves to the spinal cord.
The vagus nerve is predominantly composed of afferent fibers (≥80%) that project via the nodose and jugular ganglia to the nucleus of the solitary tract in the brainstem. Spinal afferents pass through prevertebral (sympathetic) ganglia to the dorsal root ganglia and send their endings to the dorsal horn and central gray matter of the spinal cord. Information about noxious stimuli is relayed rostrally through the spinothalamic tract. In addition, postsynaptic dorsal horn neurons within the central gray send their central processes through the medial aspect of the dorsal columns, a pathway that has only recently been recognized as being uniquely important in visceral pain.
Most physiologic studies characterizing the properties of visceral sensory pathways rely on responses to defined mechanical stimuli. Vagal afferents form a relatively homogeneous group that is activated by low-intensity mechanical stimuli and encodes stimulus intensity over a wide range. The low threshold of activation is consistent with the presumed role of these sensory pathways in the regulation of physiologic processes. In contrast, two classes of mechanosensitive spinal afferents can be distinguished by their response characteristics: low-threshold fibers are similarly activated by low-intensity stimuli and continue to encode stimulus intensity over a wide range, whereas high-threshold fibers are activated by intense, potentially noxious intensities of mechanical stimuli. Thus, spinal high-threshold fibers resemble specialized nociceptors that have been described and characterized best in the skin, thus suggesting that they play a primary role in visceral nociception. However, because mechanosensitive visceral afferents encode stimulus intensity over a wide range and can become sensitized after a visceral insult, it is likely that both low- and high-threshold mechanosensitive afferents can contribute to visceral pain conditions. Finally, most visceral sensory neurons are polymodal, which means that they respond to multiple stimulus modalities, including endogenous and exogenous chemicals contained in luminal contents, temperature (heat or cold), and stretch.
Mucosal Signaling and Visceral Sensation
In the airways, gastrointestinal tract, and urinary bladder, nerve fibers are found in close proximity to epithelial cells, which often exhibit specializations consisting of secretory vesicles on the basal surface. This structural organization and functional studies suggest that visceral epithelial cells function as an interface between chemical or mechanical stimuli and the nervous system. The best example is serotonin released from enteroendocrine cells in the gastrointestinal tract. The gut is the major source of this signaling molecule and contains 95% of the body’s serotonin. Much of this serotonin is stored in specialized enteroendocrine (enterochromaffin) cells and can be released by chemical or mechanical stimuli to activate intrinsic and extrinsic neurons. Within the urinary tract, epithelial cells release adenosine triphosphate, which acts on purinergic receptors (P2X receptors) and is involved in normal micturition, as well as in bladder pain. Finally, bladder epithelial cells also express the capsaicin receptor transient receptor potential vanilloid-1 [TRPV1]), an ion channel that is activated by acid, temperature, endogenous lipid mediators, and the pungent substance (capsaicin) contained in hot peppers. Animals with a targeted deletion of this channel display altered micturition behavior, again pointing to the importance of epithelial cells in visceral sensation.
Sensitization of Visceral Afferent Pathways
Visceral afferents can become sensitized in response to inflammation or injury. The release of mediators, such as prostaglandins or bradykinin, rapidly alters the properties of ion channels and thereby leads to an increase in neuron excitability. In addition, cytokines and growth factors may trigger transcriptional changes that affect the properties of sensory neurons through changes in gene expression. The increased peripheral input ( peripheral sensitization ) may secondarily alter sensory processing in the central nervous system ( central sensitization ), with both contributing to visceral pain syndromes. One easily recognized consequence of sensitization of visceral afferent pathways is increased tenderness to palpation over a larger area than normal. Both peripheral and central mechanisms contribute to the increased sensitivity and expanded areas of referral, both of which are common in patients with irritable bowel syndrome, dyspepsia, and interstitial cystitis.
Central Processing of Visceral Sensation and Pain
Vagal afferents project to the nucleus of the solitary tract and from there via the parabrachial nucleus and ventromedial thalamus to the insular cortex. They form many connections with the hypothalamus, supraoptic nucleus, anterior cingulate cortex, and amygdala, which is essential for autonomic and emotional responses to visceral stimulation. Spinal afferents also project to the thalamus but are preferentially found in the ventral posterolateral nucleus, which is connected to cortical areas, including the insula. The lateral components primarily serve discriminative functions associated with pain perception (e.g., location and intensity), whereas the medial thalamic nuclei, the main target of vagal input, are more closely linked to the emotional and autonomic responses triggered by pain. Consistent with the bilateral innervation of organs originating from midline structures, most visceral stimuli activate both cerebral hemispheres, albeit with a preferential activation of one—mostly the left—side. Studies using functional brain imaging have not demonstrated striking differences in the processing of visceral and nonvisceral pain. However, visceral pain preferentially activates the perigenual portion of the anterior cingulate cortex, whereas nonvisceral pain is primarily represented in the midcingulate cortex. The physical proximity or even overlap between processing of visceral pain and emotion in the perigenual portion of the anterior cingulate cortex provides a potential explanation for the stronger emotional response to painful visceral than to nonvisceral stimuli.
Visceral Pain Stimuli
Visceral events that can produce conscious sensation or acute pain in humans include traction on the mesentery, distention of hollow organs, strong contractions of muscle layers surrounding such hollow organs, ischemia, and chemical irritants. Routine endoscopic interventions have established that cutting or burning—two clearly noxious stimuli when applied to the skin—are not perceived when applied to viscera, thus setting visceral apart from nonvisceral sensation. Because of its sensitizing influences on sensory pathways, visceral inflammation can trigger pain or increase the excitability of visceral afferent pathways (e.g., enhancing low-threshold input into a noxious range), which can become chronic (e.g., chronic pancreatitis). Finally, though not studied in as much detail, malignancies can trigger chronic pain as a result of direct effects of the tumor on afferent nerves (e.g., nerve compression, release of chemicals) or as a result of indirect effects such as distention of a hollow organ.
Current evidence suggests that spinal afferents primarily serve the discriminatory function of nociception, encoding location and intensity of visceral pain. Thus, treatment strategies, such as regional block or surgical dissection, generally target spinal afferent pathways.
Unilateral nerve blocks are often ineffective, which is partly due to the fact that most viscera receive bilateral innervation and are innervated by two sets of nerves.
Visceral pain is a complex experience associated with strong emotional and autonomic reactions. The emotive component of visceral pain is at least in part due to the central projections of spinal and vagal sensory pathways, both of which can activate the perigenual area of the anterior cingulate cortex, an area closely associated with emotional processing. The complex innervation of most viscera by spinal and vagal afferent pathways provides an explanation for the often only partial or temporary effects of regional blocks.
Impaired function of viscera, such as decreased transit of material in the gut, may significantly contribute to pain due to distension or strong contractions of visceral muscles or the composition of luminal contents. Thus, effective pain management needs to combine analgesic therapies with treatment strategies targeting specific visceral function.
Regional Block and Neural Ablation
Surgical and nonoperative approaches have been developed to perform transient nerve blocks or more permanently destroy the sensory pathways involved in visceral pain. In view of concerns about the irreversible nature of nerve ablation or neurolysis, most studies targeting visceral sensory pathways have involved patients with inoperable cancer and a relatively short life expectancy. Because these patients are typically seen in advanced stages of their disease with poor performance status, surgical denervation plays only a minor role in treatment. Chemical neurolysis, generally performed with high concentrations of ethanol, can achieve comparable results without subjecting patients to the risk associated with an operation. Since the first description by Kappis nearly 100 years ago, several techniques have been developed to optimize the targeted delivery of neurolytic agents or minimize the likelihood of adverse effects, or both. Most of these approaches use imaging methods, such as fluoroscopy, ultrasound, or computed tomography (CT). However, despite this wealth of literature, few systematic studies have addressed the efficacy, outcomes, or adverse effects of these interventions. Problems assessing the effects of visceral pain management with peripheral blocks are further confounded by differences in type, concentration, and amount of the neurolytic agent, primarily ethanol; differences in both definition and measurement end points; and differences in follow-up time. Only five randomized controlled trials comparing ablative therapies with conventional pharmacologic pain management have been published. All but one of them showed at least transiently improved pain levels and a concomitant decrease in opioid use. A recently conducted meta-analysis concluded that at least for pancreatic cancer, the pain relief achieved with neurolytic blocks is comparable to that achieved with conventional analgesic therapy but may be associated with fewer side effects.
Celiac Plexus and Splanchnic Nerve Block
Spinal afferents innervating organs in the upper part of the abdomen traverse the celiac plexus, with two distinct ganglia located caudal to the origin of the celiac artery. The afferents travel centrally behind the crura of the diaphragm in the splanchnic nerves. The traditional dorsal approach uses the 12th rib and spinal process of the first lumbar vertebra as landmarks. With the patient in the prone position, a needle is advanced about 7 cm lateral to the midline at a 30- to 45-degree angle and tilted slightly cranially to reach the lateral wall of the body of the first lumbar vertebra. The needle is then moved anteriorly by about 2 cm. If aspiration does not yield return of blood, water-soluble contrast material (3 to 5 mL), often mixed with a local anesthetic, is injected under fluoroscopic control. To better directly target the celiac plexus, the needle can be placed about 2 cm more anterior, which requires piercing the diaphragmatic crura and positions the needle close to the anterolateral aspect of the aorta. Severely ill patients with respiratory compromise and patients with significant ascites or recent abdominal surgery often poorly tolerate being in the prone position for the time required to complete this procedure. Therefore, an anterior approach has been developed in which the needle is advanced from the epigastric area toward the body of the first lumbar vertebra. If the appropriate position is confirmed, the neurolytic agent, generally phenol or ethanol, is administered. Current approaches rely mostly on the tissue destructive properties of ethanol, which is used in concentrations between 50% and 99% and volumes between 20 and 50 mL per injection. Injections are generally performed bilaterally to effectively destroy the afferent pathways. A recent report suggested that radiofrequency ablation may provide an alternative to chemical neurolysis. Although this approach has since been adopted by others, well-designed studies on visceral pain syndromes are still missing.
Various imaging techniques have been used with the intent of improving the efficacy of neurolysis and decreasing the potential for adverse effects. The celiac ganglia are too small to allow direct visualization with CT or transabdominal ultrasound. Using endoscopic ultrasonography, the scanner can be brought into close proximity to the plexus. However, the ganglia do not differ in echogenicity from surrounding structures, thus again not allowing direct imaging of the target structure. Therefore, all these approaches rely on identifying the celiac artery as the main landmark. CT allows guidance of the needle to the target area and three-dimensional reconstruction of the area affected by the neurolytic agent based on the spread of radiopaque contrast material. Although ultrasound allows real-time guidance without exposure to radiation, air in the overlying structures often interferes with sound penetration and imaging, thus limiting its utility in patients undergoing celiac plexus block. Endoscopic ultrasound with an endoscopically advanced needle has been used successfully. Close inspection may identify tumor infiltration as a negative prognostic criterion and hence help in appropriate patient selection. In addition, ethanol injection alters the echogenicity of the affected tissue, thus allowing direct visualization of spread of the neurolytic agent. Only one trial directly compared the endoscopic and conventional transcutaneous approaches for celiac plexus block, and it suggested superiority of the endoscopic approach.
Surgical Nerve Ablation
One randomized controlled trial and several small case series have used surgical approaches to ablate the celiac ganglia or splanchnic nerves. The direct visualization during surgical exploration allows targeted injection of neurolytic agents into the celiac plexus if curative resection cannot be performed. Because of advances in preoperative imaging, fewer patients currently undergo exploratory laparotomy, thus limiting the number of patients in whom an intraoperative celiac plexus block might be performed. With the advent of minimally invasive surgery, thoracoscopic resection of splanchnic nerves has been reported. As is true for the less invasive procedures, approaches have not been standardized, with unilateral and bilateral resection with or without vagotomy being performed to achieve pain control in these patients.
Efficacy of Neural Block and Ablation
Differences in patient selection, techniques, and outcome measures complicate comparison of published results. A meta-analysis published in 1995 concluded that nearly 90% of patients with various malignancies experienced good pain relief for about 3 months after the procedure. Even though most case series report similar results, it is important to determine whether ablative therapy actually improves pain control or quality of life (or both) over that of conventional treatment involving the systemic use of analgesic substances. Two small randomized controlled trials demonstrated similar pain control but fewer adverse effects as a result of opioid consumption in patients treated with neurolytic blocks. Subsequent studies have pointed to a slight advantage of ablative treatment over pharmacotherapy. Intraoperative celiac plexus block with 50% ethanol led to stable pain levels in patients with inoperable pancreatic cancer, whereas patients receiving saline injection as control experienced a significant increase in pain scores during follow-up. This improvement in pain control was associated with significantly lower use of opioids. In a post hoc analysis, patients with significant pain lived longer than controls did if they underwent celiac plexus block. However, there was no overall survival benefit when the comparison involved use of the original study design as a template for the analysis. Three subsequent trials comparing neurolytic celiac plexus block with analgesic therapy confirmed better pain relief after ablative therapy. In two studies, the improvement was transient, with progressive recurrence of pain after about 1 to 2 months. Most patients still required opioids, albeit in lower dosages than needed by controls in the two smaller studies. A recent study by Wong and coworkers did not confirm this decrease in opioid consumption. Consistent with the results shown by Kawamata and colleagues, their results also demonstrated that appropriate dosing of opioids achieves good pain control and a comparable quality of life that does not differ between treatment groups. None of these investigations demonstrated a significant effect of neurolytic plexus block on patient survival.
Because cancer progression may affect the efficacy of ablative procedures, one small study randomly enrolled patients to receive early or late neurolytic blocks, with stages being defined by the use of low or high dosages of opioids, respectively. Although both groups reported better pain control than medically treated controls did, there were no significant differences between the groups.
A randomized controlled trial reported better pain control after bilateral splanchnic nerve blockade than after celiac plexus blockade. This assessment was based on a more significant decrease in pain intensity from baseline scores rather than a difference in the primary end point—pain measured by visual analog scale. Thus, confirmatory studies are needed to establish whether splanchnic nerve destruction is indeed superior to neurolysis of the celiac plexus. De Cicco and associates used CT scanning to examine the spread of contrast material injected before the neurolytic agent. In their retrospective study, the pattern was a good predictor of pain control, with optimal results being achieved in patients in whom the contrast agent spread bilaterally above and below the origin of the celiac artery. However, case series using CT guidance do not report better response rates than when the conventional approach is used. The more recently introduced endoscopic ultrasonography allows visualization of the area of interest during injection of the neurolytic agent. Only one small and underpowered study compared this approach with CT guidance and reported better results with endoscopic ultrasonography. Published case series reveal results that fall within the range achieved with other techniques.
Similarly, the results of surgical interventions with neurolytic or neuroablative therapy remain inconclusive. Although the only randomized controlled trial demonstrated good pain relief and a decrease in opioid consumption, a case series evaluating different palliative operations for pancreatic cancer did not confirm the decrease in opioid use after celiac plexus block. Smaller case series report improved pain control in about 60% to 80% of patients after thoracoscopic splanchnicectomy performed unilaterally or bilaterally with or without vagotomy. Poor definition of end points, limited assessment of analgesic effects, lack of appropriate control groups, and the fact that comparable results were obtained with a variety of often quite different approaches clearly demonstrate the need for appropriately designed studies. Splanchnicectomy reduces opioid requirements in comparison to systemic pharmacotherapy, but not in comparison to celiac plexus block. Overall, the current data suggest acceptable pain relief with at least a transient decrease in opioid requirements but do not support the superiority of one ablative approach over another. Thus, the choice that patients and physicians face should primarily focus on the available expertise when selecting one technique over another.
Nerve Block or Ablative Therapy for Benign Disorders
As described earlier, the pain relief after a neurolytic block is often transient, which decreases the enthusiasm to use such approaches in patients with benign disorders and long life expectancy. Therefore, less information is available about the efficacy of nerve blocks in such patients. Several case series combining about 400 patients with chronic pancreatitis have been published, thus making this the largest patient group with benign diseases. Initial response rates varied between 30% and 90%, with limited follow-up in the majority of cases. Because of concerns about the use of neurolytic agents, the largest series combined bupivacaine with triamcinolone. Although about half the patients reported an initial benefit, sustained responses after 24 weeks were seen in only 10%. Though commonly practiced, a recently completed randomized controlled trial showed no additional benefit of triamcinolone, with response rates of about 15% 1 month after celiac plexus block with bupivacaine. Moore and coauthors reported the results of stellate ganglion and paravertebral blocks with bupivacaine in 59 patients with refractory angina, a thoracic visceral pain, secondary to coronary artery disease. About 60% of the patients experienced pain relief for more than 2 weeks. However, the benefit was transient and most patients required multiple interventions to maintain some benefit. Interestingly, the limited experience with differential neuraxial blocks suggests that primary visceral pain was present only in about a fifth of patients with chronic pancreatitis, with the majority experiencing “central pain,” defined as persistent pain despite surgical anesthesia through the epidural administration of lidocaine. Overall, the data highlight the complex etiology of chronic pain syndromes in patients with benign disorders and do not support the routine use of ablative therapies in these patients.
Adverse Effects of Neurolytic Blocks
Transient pain is the most common side effect reported when ethanol is used as the neurolytic agent. Although the use of local anesthetic before the ethanol injection decreases the incidence of this adverse effect, at least 10% to 30% of patients experience significant pain within the first hours after the procedure. The destruction of sympathetic efferent pathways causes vasodilation of the splanchnic vessels, which results in hypotension in up to 20% of patients. Similarly, the unopposed parasympathetic drive can lead to diarrhea, which again is reported by about a fifth of patients. All these problems can generally be managed medically with appropriate premedication, hydration, postprocedural observation, and appropriate symptomatic therapy. More significant adverse effects are rare but have been reported for all the different approaches.
The posterior approach may traverse the kidney, as well as the pleural space, thereby potentially leading to hematuria or pneumothorax. The anterior approach requires advancing the needle through the liver, stomach, and colon. Although this is largely inconsequential, clinically relevant perforation can occur and may be difficult to diagnose because of the neurolytic block and often concomitantly administered analgesia. Therefore, all patients should be well hydrated, receive an intravenous fluid bolus before performance of the nerve block, and remain under observation for at least 2 to 4 hours after the procedure. Delivery of neurolytic agents in close proximity to major vessels is another potential source of complications. In all cases, aspiration should ensure that the needle is not inside a vessel before the neurolytic agent is injected. However, indirect effects, such as mesenteric venous thrombosis, may occur rarely despite these precautions. Spread of the neurolytic agent within the retroperitoneal space with injury to the lumbar nerves has been reported. Paraparesis is the most feared and mostly irreversible complication that is thought to be caused by injury to the nutrient vessels supplying the spinal cord. Even though sufficiently powered studies addressing the likelihood of these adverse effects are not available, current evidence suggests that the incidence does not differ between the different approaches.
Neurolytic celiac or splanchnic nerve block is a moderately effective method to decrease pain in patients with intra-abdominal malignancies, and has been best studied in patients with pancreatic cancer. The benefit is often transient, and may decrease the need for opioids, but does not affect life expectancy or quality of life compared to conventional pain management. Adverse effects occur in about 20% of patients, are largely minor and transient and can be managed by appropriate medical therapy. Approaches using different imaging techniques or even direct operative visualization do not seem to affect outcome or the incidence of side effects.
Regional Block and Neural Ablation
Surgical and nonoperative approaches have been developed to perform transient nerve blocks or more permanently destroy the sensory pathways involved in visceral pain. In view of concerns about the irreversible nature of nerve ablation or neurolysis, most studies targeting visceral sensory pathways have involved patients with inoperable cancer and a relatively short life expectancy. Because these patients are typically seen in advanced stages of their disease with poor performance status, surgical denervation plays only a minor role in treatment. Chemical neurolysis, generally performed with high concentrations of ethanol, can achieve comparable results without subjecting patients to the risk associated with an operation. Since the first description by Kappis nearly 100 years ago, several techniques have been developed to optimize the targeted delivery of neurolytic agents or minimize the likelihood of adverse effects, or both. Most of these approaches use imaging methods, such as fluoroscopy, ultrasound, or computed tomography (CT). However, despite this wealth of literature, few systematic studies have addressed the efficacy, outcomes, or adverse effects of these interventions. Problems assessing the effects of visceral pain management with peripheral blocks are further confounded by differences in type, concentration, and amount of the neurolytic agent, primarily ethanol; differences in both definition and measurement end points; and differences in follow-up time. Only five randomized controlled trials comparing ablative therapies with conventional pharmacologic pain management have been published. All but one of them showed at least transiently improved pain levels and a concomitant decrease in opioid use. A recently conducted meta-analysis concluded that at least for pancreatic cancer, the pain relief achieved with neurolytic blocks is comparable to that achieved with conventional analgesic therapy but may be associated with fewer side effects.
Celiac Plexus and Splanchnic Nerve Block
Spinal afferents innervating organs in the upper part of the abdomen traverse the celiac plexus, with two distinct ganglia located caudal to the origin of the celiac artery. The afferents travel centrally behind the crura of the diaphragm in the splanchnic nerves. The traditional dorsal approach uses the 12th rib and spinal process of the first lumbar vertebra as landmarks. With the patient in the prone position, a needle is advanced about 7 cm lateral to the midline at a 30- to 45-degree angle and tilted slightly cranially to reach the lateral wall of the body of the first lumbar vertebra. The needle is then moved anteriorly by about 2 cm. If aspiration does not yield return of blood, water-soluble contrast material (3 to 5 mL), often mixed with a local anesthetic, is injected under fluoroscopic control. To better directly target the celiac plexus, the needle can be placed about 2 cm more anterior, which requires piercing the diaphragmatic crura and positions the needle close to the anterolateral aspect of the aorta. Severely ill patients with respiratory compromise and patients with significant ascites or recent abdominal surgery often poorly tolerate being in the prone position for the time required to complete this procedure. Therefore, an anterior approach has been developed in which the needle is advanced from the epigastric area toward the body of the first lumbar vertebra. If the appropriate position is confirmed, the neurolytic agent, generally phenol or ethanol, is administered. Current approaches rely mostly on the tissue destructive properties of ethanol, which is used in concentrations between 50% and 99% and volumes between 20 and 50 mL per injection. Injections are generally performed bilaterally to effectively destroy the afferent pathways. A recent report suggested that radiofrequency ablation may provide an alternative to chemical neurolysis. Although this approach has since been adopted by others, well-designed studies on visceral pain syndromes are still missing.
Various imaging techniques have been used with the intent of improving the efficacy of neurolysis and decreasing the potential for adverse effects. The celiac ganglia are too small to allow direct visualization with CT or transabdominal ultrasound. Using endoscopic ultrasonography, the scanner can be brought into close proximity to the plexus. However, the ganglia do not differ in echogenicity from surrounding structures, thus again not allowing direct imaging of the target structure. Therefore, all these approaches rely on identifying the celiac artery as the main landmark. CT allows guidance of the needle to the target area and three-dimensional reconstruction of the area affected by the neurolytic agent based on the spread of radiopaque contrast material. Although ultrasound allows real-time guidance without exposure to radiation, air in the overlying structures often interferes with sound penetration and imaging, thus limiting its utility in patients undergoing celiac plexus block. Endoscopic ultrasound with an endoscopically advanced needle has been used successfully. Close inspection may identify tumor infiltration as a negative prognostic criterion and hence help in appropriate patient selection. In addition, ethanol injection alters the echogenicity of the affected tissue, thus allowing direct visualization of spread of the neurolytic agent. Only one trial directly compared the endoscopic and conventional transcutaneous approaches for celiac plexus block, and it suggested superiority of the endoscopic approach.
Surgical Nerve Ablation
One randomized controlled trial and several small case series have used surgical approaches to ablate the celiac ganglia or splanchnic nerves. The direct visualization during surgical exploration allows targeted injection of neurolytic agents into the celiac plexus if curative resection cannot be performed. Because of advances in preoperative imaging, fewer patients currently undergo exploratory laparotomy, thus limiting the number of patients in whom an intraoperative celiac plexus block might be performed. With the advent of minimally invasive surgery, thoracoscopic resection of splanchnic nerves has been reported. As is true for the less invasive procedures, approaches have not been standardized, with unilateral and bilateral resection with or without vagotomy being performed to achieve pain control in these patients.
Efficacy of Neural Block and Ablation
Differences in patient selection, techniques, and outcome measures complicate comparison of published results. A meta-analysis published in 1995 concluded that nearly 90% of patients with various malignancies experienced good pain relief for about 3 months after the procedure. Even though most case series report similar results, it is important to determine whether ablative therapy actually improves pain control or quality of life (or both) over that of conventional treatment involving the systemic use of analgesic substances. Two small randomized controlled trials demonstrated similar pain control but fewer adverse effects as a result of opioid consumption in patients treated with neurolytic blocks. Subsequent studies have pointed to a slight advantage of ablative treatment over pharmacotherapy. Intraoperative celiac plexus block with 50% ethanol led to stable pain levels in patients with inoperable pancreatic cancer, whereas patients receiving saline injection as control experienced a significant increase in pain scores during follow-up. This improvement in pain control was associated with significantly lower use of opioids. In a post hoc analysis, patients with significant pain lived longer than controls did if they underwent celiac plexus block. However, there was no overall survival benefit when the comparison involved use of the original study design as a template for the analysis. Three subsequent trials comparing neurolytic celiac plexus block with analgesic therapy confirmed better pain relief after ablative therapy. In two studies, the improvement was transient, with progressive recurrence of pain after about 1 to 2 months. Most patients still required opioids, albeit in lower dosages than needed by controls in the two smaller studies. A recent study by Wong and coworkers did not confirm this decrease in opioid consumption. Consistent with the results shown by Kawamata and colleagues, their results also demonstrated that appropriate dosing of opioids achieves good pain control and a comparable quality of life that does not differ between treatment groups. None of these investigations demonstrated a significant effect of neurolytic plexus block on patient survival.
Because cancer progression may affect the efficacy of ablative procedures, one small study randomly enrolled patients to receive early or late neurolytic blocks, with stages being defined by the use of low or high dosages of opioids, respectively. Although both groups reported better pain control than medically treated controls did, there were no significant differences between the groups.
A randomized controlled trial reported better pain control after bilateral splanchnic nerve blockade than after celiac plexus blockade. This assessment was based on a more significant decrease in pain intensity from baseline scores rather than a difference in the primary end point—pain measured by visual analog scale. Thus, confirmatory studies are needed to establish whether splanchnic nerve destruction is indeed superior to neurolysis of the celiac plexus. De Cicco and associates used CT scanning to examine the spread of contrast material injected before the neurolytic agent. In their retrospective study, the pattern was a good predictor of pain control, with optimal results being achieved in patients in whom the contrast agent spread bilaterally above and below the origin of the celiac artery. However, case series using CT guidance do not report better response rates than when the conventional approach is used. The more recently introduced endoscopic ultrasonography allows visualization of the area of interest during injection of the neurolytic agent. Only one small and underpowered study compared this approach with CT guidance and reported better results with endoscopic ultrasonography. Published case series reveal results that fall within the range achieved with other techniques.
Similarly, the results of surgical interventions with neurolytic or neuroablative therapy remain inconclusive. Although the only randomized controlled trial demonstrated good pain relief and a decrease in opioid consumption, a case series evaluating different palliative operations for pancreatic cancer did not confirm the decrease in opioid use after celiac plexus block. Smaller case series report improved pain control in about 60% to 80% of patients after thoracoscopic splanchnicectomy performed unilaterally or bilaterally with or without vagotomy. Poor definition of end points, limited assessment of analgesic effects, lack of appropriate control groups, and the fact that comparable results were obtained with a variety of often quite different approaches clearly demonstrate the need for appropriately designed studies. Splanchnicectomy reduces opioid requirements in comparison to systemic pharmacotherapy, but not in comparison to celiac plexus block. Overall, the current data suggest acceptable pain relief with at least a transient decrease in opioid requirements but do not support the superiority of one ablative approach over another. Thus, the choice that patients and physicians face should primarily focus on the available expertise when selecting one technique over another.
Nerve Block or Ablative Therapy for Benign Disorders
As described earlier, the pain relief after a neurolytic block is often transient, which decreases the enthusiasm to use such approaches in patients with benign disorders and long life expectancy. Therefore, less information is available about the efficacy of nerve blocks in such patients. Several case series combining about 400 patients with chronic pancreatitis have been published, thus making this the largest patient group with benign diseases. Initial response rates varied between 30% and 90%, with limited follow-up in the majority of cases. Because of concerns about the use of neurolytic agents, the largest series combined bupivacaine with triamcinolone. Although about half the patients reported an initial benefit, sustained responses after 24 weeks were seen in only 10%. Though commonly practiced, a recently completed randomized controlled trial showed no additional benefit of triamcinolone, with response rates of about 15% 1 month after celiac plexus block with bupivacaine. Moore and coauthors reported the results of stellate ganglion and paravertebral blocks with bupivacaine in 59 patients with refractory angina, a thoracic visceral pain, secondary to coronary artery disease. About 60% of the patients experienced pain relief for more than 2 weeks. However, the benefit was transient and most patients required multiple interventions to maintain some benefit. Interestingly, the limited experience with differential neuraxial blocks suggests that primary visceral pain was present only in about a fifth of patients with chronic pancreatitis, with the majority experiencing “central pain,” defined as persistent pain despite surgical anesthesia through the epidural administration of lidocaine. Overall, the data highlight the complex etiology of chronic pain syndromes in patients with benign disorders and do not support the routine use of ablative therapies in these patients.
Adverse Effects of Neurolytic Blocks
Transient pain is the most common side effect reported when ethanol is used as the neurolytic agent. Although the use of local anesthetic before the ethanol injection decreases the incidence of this adverse effect, at least 10% to 30% of patients experience significant pain within the first hours after the procedure. The destruction of sympathetic efferent pathways causes vasodilation of the splanchnic vessels, which results in hypotension in up to 20% of patients. Similarly, the unopposed parasympathetic drive can lead to diarrhea, which again is reported by about a fifth of patients. All these problems can generally be managed medically with appropriate premedication, hydration, postprocedural observation, and appropriate symptomatic therapy. More significant adverse effects are rare but have been reported for all the different approaches.
The posterior approach may traverse the kidney, as well as the pleural space, thereby potentially leading to hematuria or pneumothorax. The anterior approach requires advancing the needle through the liver, stomach, and colon. Although this is largely inconsequential, clinically relevant perforation can occur and may be difficult to diagnose because of the neurolytic block and often concomitantly administered analgesia. Therefore, all patients should be well hydrated, receive an intravenous fluid bolus before performance of the nerve block, and remain under observation for at least 2 to 4 hours after the procedure. Delivery of neurolytic agents in close proximity to major vessels is another potential source of complications. In all cases, aspiration should ensure that the needle is not inside a vessel before the neurolytic agent is injected. However, indirect effects, such as mesenteric venous thrombosis, may occur rarely despite these precautions. Spread of the neurolytic agent within the retroperitoneal space with injury to the lumbar nerves has been reported. Paraparesis is the most feared and mostly irreversible complication that is thought to be caused by injury to the nutrient vessels supplying the spinal cord. Even though sufficiently powered studies addressing the likelihood of these adverse effects are not available, current evidence suggests that the incidence does not differ between the different approaches.
