Celiac Plexus Block for Pancreatic Cancer Pain: Factors Influencing Pain, Symptoms and Quality of Life

Transcription

1 1140 Journal of Pain and Symptom Management Vol. 26 No. 6 December 2003 Original Article Celiac Plexus Block for Pancreatic Cancer Pain: Factors Influencing Pain, Symptoms and Quality of Life Sebastiano Mercadante, MD, Elena Catala, MD, Edoardo Arcuri, MD, and Alessandra Casuccio, BS Pain Relief & Palliative Care Unit (S.M.), La Maddalena Cancer Center, Palermo, Italy; Anesthesia and Pain Therapy Department (E.C.), Hospital de la Santa Creu I Sant Pau, Barcelona, Spain; Pain Relief & Intensive Care (E.A.), National Cancer Institute Regina Elena, Rome, Italy; and Department of Microbiology and Hygiene (A.C.), University of Palermo, Palermo, Italy Abstract Neurolytic celiac plexus block (NCPB) is claimed to be an effective method of pain control for pancreatic cancer pain. However, the factors that may influence long-term analgesia, adverse effects, and quality of life after performing NCPB have never been determined. In a prospective multicenter study, 22 patients who underwent NCPB were followed until death. Numerous parameters other than pain and symptom intensity were evaluated, including age, gender, initial site of cancer, sites of pain, possible peritoneal involvement, technique, and oncologic interventions. Indices were calculated to determine the opioid consumption ratio (EAS) and the trend of opioid escalation (OEI). NCPB was effective in reducing opioid consumption and gastrointestinal adverse effects for at least 4 weeks. In the last four weeks prior to death, there was the typical trend of increasing symptom intensity common to the terminal cancer population. None of the factors studied influenced the analgesic effectiveness of NPCB. NPCB, performed by skilled clinicians, regardless of the technique chosen, is a safe and useful means that should be considered as an adjuvant to common analgesic regimens at any stage, as it may allow the reduction of the visceral component of pancreatic pain that may prevail in certain phases of the illness. The analgesic and symptomatic effect of NCPB is presumably advantageous for about four weeks. A possible factor interfering with long-term outcome includes the capacity of cancer to involve the celiac axis, which can distort the anatomy and prevent neurolytic spread, or modify the pain mechanisms. Outcomes are strongly based on individual variation. J Pain Symptom Manage 2003;26: U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. Key Words Pancreatic cancer pain, neurolytic celiac plexus block, opioids, cancer epidemiology Address reprint requests to: Sebastiano Mercadante, MD, Director, Anesthesia and Intensive Care Unit & Pain Relief of Palliative Care Unit, La Maddalena Cancer Center, Via S. Lorenzo 312, Palermo, Italy. Accepted for publication: April 11, U.S. Cancer Pain Relief Committee Published by Elsevier Inc. All rights reserved. Introduction Pancreatic cancer is usually very painful and highly lethal, and the major focus of care is symptom relief and supportive care. 1,2 Neurolytic celiac plexus block (NCPB) is claimed to be /03/$ see front matter doi: /j.painsymman

2 Vol. 26 No. 6 December 2003 Celiac Plexus Block for Pancreatic Cancer Pain 1141 an effective method of pain control, with acceptable adverse effects and limited severe complications. Controlled randomized studies have reported prolonged efficacy in terms of balance of analgesia and adverse effects. 3 5 Factors implicated in the response to this procedure, however, have been variably explored. Some authors suggest an early intervention, 6 based on less compromised anatomy and on the prevalent pain mechanism before the tumor spreads over the celiac axis. Some authors have even postulated that NPCB may have pre-emptive effects, 7 but there is no empirical evidence for this. Apart from the technical problems that may have importance for the immediate effect, tumor spread will be inevitable, sooner or later, in the course of the illness. Tumor may injure nonvisceral structures and the underlying pain mechanism may change, due to involvement of neural and somatic structures. Tumor progression may be one reason that the pain-relieving effects of NPCB are rarely complete. Continuing use of analgesics at an appropriately reduced dose is usually required, post-block. Other than variations in technique, no study has addressed possible factors influencing the success rate of NCPB with regard to analgesia, analgesic consumption, adverse effects, and quality of life measure. 3 5,8 10 The aim of this study was to determine possible factors that may influence the block response during the course of the illness in a prospective multicenter study. Methods Twenty-two patients with pancreatic cancer pain were selected for this prospective multicenter study. Patients had to provide their informed consent after clear explanation about the possible advantages and risks of the procedure. Patient characteristics are given in Table 1. Almost all the patients were receiving opioids at the time of performing the block. Each center could use their preferred technique, including the anterior ultrasound, the anterior computerized tomographic-guided, or the classical posterior approach. Similarly, dosages and concentration of alcohol were left to the local preference. Demographics, performance status, the opioid starting dose (OSD) in mg at referral, the maximum dose of opioids (OMD) in mg, previous surgery and chemotherapy, survival (from NPCB to death), and adjuvant medications including nonopioid analgesics were recorded. The following parameters were monitored before performing the block (W0) and at week intervals until death: Pain intensity was measured using the patient s self report on a 0 to 10 numerical rating scale. In a few cases, when self-rating was not possible due to cognitive impairment, a physician proxy measure was used to rate apparent pain intensity or relief. Symptoms associated with opioid therapy or commonly present in advanced cancer patients, such as nausea and vomiting, drowsiness, confusion, xerostomia, and others were monitored using a scale from 0 to 3 (not at all, slight, a lot, awful). A distress score (DS), including nausea, vomiting, drowsiness, confusion, xerostomia, and constipation, was calculated as the sum of the symptom intensity scores. This score has not been validated, but should express the physical burden due to illness or opioid therapy and has been previously used. 11 The occurrence of jaundice or bowel obstruction was also recorded. Nutrition intake and level of oral hydration were measured using a scale from 0 (normal) to 3 (absent). Table 1 Patient Characteristics Age 61.4 (SE 1.7, 95% CI 57 65) Gender (M) 14/22 Clear peritoneal involvement 8/22 Pain mechanisms visceral 12, somatic 1, visceral-somatic 9 Pain site abdominal 13, addominal-back 6, abdomen-shoulder 1, epigastrium 2 Weight (Kg) 57.5 (SE 1.7, 95% CI 53 61) Height (cm) 163 (SE 0.014, 95% CI ) Survival (days) 71.9 (SE 11, 95% CI 49 94) Opioid starting dose (mg) (oral morphine equivalent) 97 (SE 13.8, 95% CI )

3 1142 Mercadante et al. Vol. 26 No. 6 December 2003 Quality of life was measured with Spitzer score (five items including activity, daily living, health, support, outlook, from 0 to 2, for a maximum score (10), which is a well validated system. Performance status was measured using the Eastern Cooperative Oncology Group Scale. Sites of cancer (head, body, tail), pain and possible peritoneal involvement, according to clinical judgment and imaging studies at time of referral, were recorded. To follow the patient s course over time, the following indices were calculated. 12 The opioid escalation index percentage (OEI%) is the mean increase in the percentage of opioid dosage from OSD, using the following formula: ((OMD OSD)/ OSD)/days 100, where OMD is the maximum opioid dose, expressed as oral morphine equivalents, and OSD is the starting dose at time of NPCB. The opioid escalation index in mg (OEI mg) is the mean increase of opioid dosage in mg, using the formula (OMD OSD)/ days. The effective analgesic score (EAS) was calculated at fixed weekly intervals on the basis of the following formula: 1 O/10 (pain intensity), where 1 indicates the administration of NSAIDs at fixed times and at full dosage, O indicates the dosage in oral morphine equivalent of the opioid used (mg), pain intensity by a numerical scale (see above). This score monitors the analgesic consumption/pain relief ratio and fits very well to monitor the course of an analgesic intervention after a procedure. 3 The patients were continuously assessed for changes in pain and symptom intensity with frequent contacts or visits at home until their death. The doses of analgesics and adjuvants were tailored for each patient according to their needs and kept as low as possible and titrated to achieve an acceptable analgesia by the patient s report, with minimal adverse effects. The use of other drugs was allowed, including those generally administered in palliative care to control symptoms due to illness or treatment. Statistical Analysis Frequency analysis was performed with chisquare test. The paired Wilcoxon signed-rank test and paired samples Student s t test were used to compare the scores of non-parametric and parametric variables, respectively, at the different time intervals. The one-way analysis of variance (ANOVA) and Kruskal Wallis statistic test were used to evaluate differences between the parameters. All P values were two-sided and P values less than 0.05 were considered statistically significant. To normalize the data, measures were analyzed not only prospectively from the time of admission and NCPB, but also retrospectively for the eight-week period prior to death. Results Of the twenty-two patients enrolled in this study, ten patients had previously undergone surgery; two patients had received duodenopancreatectomy, three gastrojejunostomy, four biliary bypasses, and one had an exploratory laparotomy. One patient had jaundice at the time of the block, and one patient was on parenteral nutrition. Twelve patients had received or were receiving chemotherapy. One patient was receiving paracetamol but presented with severe pain and required strong opioids immediately. Similarly, two patients were receiving codeine unsuccessfully. Two patients were receiving transdermal fentanyl, two methadone, and one subcutaneous morphine. All the other patients were receiving oral morphine. The mean equivalent doses of oral morphine were 97 mg daily. After the NCPB, all the patients survived at least three weeks; twelve survived more than eight weeks, and one subject lived for 38 weeks post-ncpb. Prolonged post-block diarrhea and hypotension were reported in three patients. No other relevant complications attributable to NCPB were found. Data regarding pain intensity, opioid doses and related parameters are shown in Table 2. Pain intensity significantly decreased at W1 (P ). Pain control remained adequate until death. Opioid doses significantly decreased for three weeks (P at W1, P at W2, P at W3), and subsequently tended to increase, reaching the same pre-blocklevel inthe lastweeksof lifein thebackward analysis (Table 3, see below). The OEI mg was 1 mg (SE 0.06, CI 95% ), and OEI% was 5% (SE 15.6, CI 95% ).

4 Table 2 Pain Intensity, Opioid Doses and Related Parameters Investigated Upon Admission and in the Following Weeks After NCPB Up to Time of Death Week T n Pain Nausea Vomiting Constipation Diarrhea Drowsiness Confusion Xerostomia Dysphagia Weakness Nutrition Hydration ECOG QOL Opioid (mg) EAS DS Data Recorded Expressed as a Mean (see text) Vol. 26 No. 6 December 2003 Celiac Plexus Block for Pancreatic Cancer Pain 1143

5 1144 Mercadante et al. Vol. 26 No. 6 December 2003 Table 3 Pain Intensity, Opioid Doses and Related Parameters Investigated in the Last Eight Weeks Before Death Backward Approach -8W -7W -6W -5W -4W -3W -2W -1W Pts Pain Nausea Vomiting Constipation Diarrhea Drowsiness Confusion Xerostomia Dysphagia Weakness Nutrition Hydration ECOG QOL Opioid mg EAS DS Data Recorded Experssed as a Mean (see text). When considering the last 8 weeks of life, pain intensity significantly decreased one week before death ( W1) when compared with W 3 (P 0.013) and W 4 (P 0.019). Opioid doses showed a progressive increase in the last four weeks. The differences reported were statistically significant (see Table 3). The EAS showed a similar trend, although the large variability of values calculated, due to different stages of disease, could reach statistical significance only at certain intervals. No statistical relationship was found between pain intensity, OEI, and EAS, and all the variables studied, including age, gender, pain site, pancreatic cancer site, survival, diagnosis-block time, surgery, and known peritoneal involvement, NCPB technique, adjuvant drugs and their dose. Patients who had received or were receiving chemotherapy showed significantly higher OEI% (P 0.05) and OEI mg (P 0.016). DS dramatically decreased at W1 (Table 2) and remained significantly lower for four weeks after NCPB (range P ) until W9 (P 0.041). However, after the sixth week, this score progressively increased (W1 versus W6, P 0.028). When considering the backward approach for all the patients in the last 8 weeks of life, significant changes were observed, with a clear cut-off at W4 (Table 3). Gastrointestinal symptoms presented relevant changes in the trend of their intensity after the NCPB. Nausea was significantly improved during the initial weeks after the block and this condition was maintained up to 5 weeks (P 0.05). Then, nausea significantly increased in time. Vomiting intensity significantly decreased up to 4 weeks after the block (at W1 P 0.014, at W2 P 0.026, at W4 P 0.037). In the following weeks, symptom intensity tended to progressively increase (W3 versus W6 P 0.034, versus W7 P 0.034; W4 vs. W6 P 0.023, versus W7 P 0.038). Constipation was strongly influenced by NCPB and the effect was maintained for 10 weeks (P 0.05). Diarrhea significantly increased immediately after the block (P at W1), but progressively declined in the subsequent weeks (W1 versus W3 P 0.024, versus W4 P 0.034). Nutrient and fluid intake showed a parallel trend, with the ECOG status presenting a progressive decline after five-six weeks. The Spitzer index started to statistically decrease at W3 (P 0.005), and progressively worsened until W10. Drowsiness was significantly improved in the week subsequent to NCPB (P 0.038). However, this symptom tended to progressively increase in intensity three weeks after (P 0.05 for W2 versus W3,W4,W5,W6). When considering symptom intensity for all the patients in the last 8 weeks of life, dry mouth, dysphagia, and weakness, which did not immediately change after NPCB, and nausea, had a significant peak in intensity between W4

6 Vol. 26 No. 6 December 2003 Celiac Plexus Block for Pancreatic Cancer Pain 1145 and W5, which was maintained until death, while drowsiness and confusion significantly worsened in the last two weeks of life. Hydration and nutrition intake, ECOG, as well as quality of life showed as a similar trend as that of symptom intensity (see Table 3). Discussion An effective NCPB is able to abolish the visceral component of pain due to pancreatic cancer. It is unable to ensure complete and lasting pain relief. Efficacy data are not often available in most studies of NCPB. There is usually limited information about pain characterization, opioid consumption, pain intensity over time, adverse effects or symptoms associated with the illness, and eventual factors implicated in these outcomes. 3 5,13 In this series, NCPB, regardless of the technique used, produced immediate analgesia and allowed a reduction in opioid dose and an evident improvement of opioid-induced adverse effects or symptoms associated with the illness, particularly gastrointestinal ones. This therapeutic effect lasted about four-five weeks, after which symptom intensity worsened. Significant increases in DS were observed four to five weeks before death. The analysis shows an emergent pattern of distressing symptoms foreshadowing imminent death in pancreatic cancer patients, as confirmed by the parallel worsening of weakness, food and fluid intake, dysphagia, QOL, ECOG status. 12 Regardless of timing of NCPB, opioid doses showed a progressive increase in the last four weeks, despite improved analgesia provided by NCPB in the preceding weeks. This might be attributable to opioid use for control of other non-pain-related distress. As previously observed, 3,4 NCPB may have direct and indirect therapeutic effects on gastrointestinal symptoms and improve quality of life for an average of about four weeks, probably due to a concomitant decrease in opioid doses and an improvement of some gastrointestinal adverse effects. No specific factor, such survival, age, gender, site of pain, and other parameters taken into consideration has been identified to influence the outcome of NCPB. Controversy exists on the role of NCPB in advanced cancer, when the pain syndrome may assume other characteristics, with a possible involvement of structures other than viscera. NCPB has been considered to be more effective if performed early after pain onset, when pain is still or mainly of celiac type and may respond to nonopioid analgesics In Polati et al. s 5 study, for example, patients selected were receiving nonopioid analgesics or minimal doses of opioids (mean oral morphine equivalents 10 mg), raising the problem of timing the procedure. However, the probability of patients remaining pain-free diminishes with increased survival time, as progression of disease tends to involve more anatomical structures, independently of performing the NCPB or not. Considered from another point of view, an early block means an immediate effect but not a guarantee for the future in patients who have long survival, with more possibility of peritoneal, somatic or neural involvement. The assertion that NCPB performed in patients without pain may prevent the subsequent onset of pain as well as improve survival 7 has been not confirmed in randomized double-blind studies. 5 The origin of pancreatic cancer pain is still disputed. Typically, pain has been described as a result of the tumor infiltrating tissues; obstructing blood vessels, ducts, or viscera; stretching a capsule; or causing necrosis, inflammation or ulceration. 15 Visceral pain usually is quite responsive to common analgesics, as well as NCPB. Subsequently, pain could be due to the progressive infiltration and then to destruction of the celiac ganglia. As a consequence, different degrees of neural involvement may cause fluctuation of the opioid response. Profound neural infiltration may cause a progressive suppression of this kind of input, similarly to that produced by a neurolytic intervention. In the meantime, somatic structures may be progressively involved due to invasion of peritoneum and/or abdominal wall or diaphragm. 16 Thus, the evolution is unpredictable, and based on individual anatomical factors due to the characteristics of local progression of disease. In a study focusing on the efficacy of NCPB in varying locations of pancreatic cancer, neurolysis was more effective in cases with tumor involving the head of the pancreas than in patients with cancer of the body and tail of the pancreas. 17 This observation was not confirmed in the present study. In another

7 1146 Mercadante et al. Vol. 26 No. 6 December 2003 survey, patients with tumors in the head of the pancreas had less pain than patients with cancer in the body or tail of the pancreas, and this could not be explained by stage or size of the tumor. 18 Simulated needle placement was more likely to fail in patients with pancreatic cancer than in patients without cancer, due to a reduction in the right retrocrural space 8 or distortion of the celiac area preventing appropriate neurolytic spread. 9,10 Thus, patients with unsatisfactory pain relief after NCPB may show massive growth of the tumor around the celiac axis with metastases. 17 Additional intermittent administration of bupivacaine through a catheter previously placed near the celiac plexus provided prolonged pain relief in patients who had a substantial analgesic effect lasting three-four weeks after NCPB. 19 The reason why local anesthetic may still have an effect even in advanced conditions, despite potentially compromised access to the celiac area, could be an action of the large volumes used on somatic structures involved later in the progress of disease. This should be better tested in appropriate studies. It is unclear if the use of chemotherapy provides any palliation, although new treatments seem to improve survival. 1 In this study, patients on chemotherapy had higher values of OEI% and OEImg when compared with patients not receiving chemotherapy. Chemotherapy could induce more neural damage, and would require higher opioid doses, but this hypothesis is difficult to prove. However, in a previous experience in a general oncological population, it has been reported that older patients, who have received less oncologic treatment, reported less or similar adverse effects but an equally effective pain relief, despite reporting a lower OEI. Previous chemotherapy may affect the progression of the disease or, alternatively, influence the pain syndrome and, as a consequence, the trend in opioid consumption. 20,21 Although this relationship is difficult to demonstrate, the putative palliative effects of chemotherapy should be regarded more cautiously. Different techniques have been proposed in an attempt to improve the analgesic effects and reduce the risk of complications. However, the technique does not seem important in terms of immediate or up-to-death results, or complications reported, 14 and the operator s experience remains the more important factor. In cases with advanced tumor proliferation, regardless of the technique used, the analgesic effect of celiac plexus block are not satisfactory. 19 It was for this reason that it was decided to let clinicians use their own preferred technique in this study. This was confirmed by the absence of evident differences in results reported in the different centers. Sometimes, repeated blocks are performed when symptoms re-emerge. However, this practice has never assessed in appropriate studies, and the distorted anatomy in advanced phases of the illness would make this approach less useful. Conclusion NCPB, performed by skilled clinicians, regardless of the technique chosen, is a safe and useful means that should be considered as an adjuvant to common analgesic regimens at any stage, as it may allow the reduction of the visceral component of pancreatic pain, which may prevail in certain phases of the illness. It is less useful when death is near. No factors surveyed in this study influenced the outcome of the NCPB. Outcomes were probably influenced only by the local spread of cancer. A clear effect of the NCPB is observed for at least four weeks. The global response over time, however, probably depends on the capacity of cancer to involve the celiac axis, distorting the anatomy, and preventing neurolytic spread. This evolution will be strongly based on individual local variation, and unfortunately, cannot be foreseen. References 1. Van Hoff DD, Goodwin AL, Garcia L. Advances in the treatment of patients with pancreatic cancer: improvement in symptoms and survival time. The San Antonio Drug Development Team. Br J Cancer 1998;78(3): Caraceni A, Portenoy RK. Pain management in patients with pancreatic carcinoma. Cancer 1996; 78: Mercadante S. Celiac plexus block versus analgesics in pancreatic cancer pain. Pain 1993;52: Kawamata M, Ishitani K, Ishikawa K, et al. Comparison between celiac plexus block and morphine treatment on quality of life in patients with pancreatic cancer pain. Pain 1996;64:

8 Vol. 26 No. 6 December 2003 Celiac Plexus Block for Pancreatic Cancer Pain Polati E, Finco G, Gottin L, et al. Prospective randomized double-blind trial of neurolytic coeliac plexus block in patients with pancreatic cancer. Br J Surg 1998;85: Ischia S, Polati E, Finco G, et al. The role of the neurolytic celiac plexus block in pancreatic cancer pain management: do we have the answers? Reg Anesth Pain Med. 1998;23: Lillemoe K, Cameron J, Kaufman H, et al. Chemical splanchnicectomy in patients with unresectable pancreatic cancer. A prospective randomized trial. Ann Surg 1993;217: Weber JG, Brown DL, Stephens DH, et al. Celiac plexus block. Retrocrural computed tomographic anatomy in patients with and without pancreatic cancer. Reg Anesth 1996;21: Di Cicco M, Matovic M, Balestreri L, et al. Singleneedle celiac plexus block: is needle tip position critical in patients with no regional anatomic distortion? Anesthesiology. 1997;87: Di Cicco M, Matovic M, Bortolussi R, et al. Celiac plexus block: injectate spread and pain relief in patients with regional anatomic distortions. Anesthesiology 2001;94: Mercadante S, Casuccio A, Groff L, et al. Switching from morphine to methadone to improve analgesia and tolerability in cancer patients a prospective study. J Clin Oncol 2001;19: Mercadante S, Casuccio A, Fulfaro F, et al. The course of symptom frequency and intensity in advanced cancer patients followed at home. J Pain Symptom Manage 2000;20: Eisenberg E, Carr D, Chalmers TC. Neurolytic celiac plexus block for treatment of cancer pain: a meta-analysis. Anesth Analg 1995;80: Ischia A, Ischia A, Polati E, et al. Three posterior percutaneous celiac plexus block techniques. A prospective, randomized study in 61 patients with pancreatic cancer pain. Anesthesiology 1992;76: Alter CL. Palliative and supportive care of patients with pancreatic cancer. Semin Oncol 1996;23: Arcuri E, Mercadante S, Laurenzi L, et al. Opioid nonresponsiveness in cancer can be reversibile. A serendipitous conclusion of a retrospective analysis. J Pain Symptom Manage 2000;20: Rykowski JJ, Hilgier M. Efficacy of neurolytic plexus block in varying locations of pancreatic cancer: influence on pain relief. Anesthesiology 2000;92: Graham AL, Andren-Sandberg A. Prospective evaluation of pain in exocrine pancreatic cancer. Digestion 1997;58: Vranken JH, Zuurmond WWA. Increasing the efficacy of a celiac plexus block in patients with severe pancreatic cancer pain. J Pain Symptom Manage 2001;22: Mercadante S, Dardanoni G, Salvaggio L, et al. A Monitoring of opioid therapy in advanced cancer pain patients. J Pain Symptom Manage 1997;13: Mercadante S, Casuccio A, Pumo G, et al. Factors influencing the opioid response in advanced cancer patients with pain followed at home: the effects of age and gender. Support Care Cancer 2000;8: