Nonspecific supportive therapy: The first line in pain management is the use of medical therapy. The initial step of medical therapy usually is nonspecific supportive treatment. Supportive therapy is aimed at treating the concurrent symptoms and not the underlying factors in pain causation. Analgesic drugs are still the most commonly adopted method for pain relief. The obvious problem with this method of treatment is that patients often become dependant on heavy narcotic use. Most patients with chronic pancreatitis have their pain treated with analgesics on an episodic or continuing basis. Although the use of narcotics for the treatment of chronic pancreatitis is widespread, there are no controlled trials testing their efficacy as compared to the other modalities. Time intervals and doses of drug application must be adapted to the individual pain pattern. Although reluctance to use of narcotics is understandable, it should not be withheld if the treatment would otherwise not lead to adequate pain control[6].

There may be significant psychiatric, psychological, or psychosomatic contributions to the pain syndrome in these patients. Many physicians and surgeons use antidepressant medications as concomitant therapy, acknowledging the difficulty in assessing the psychological contributors to patients’ pain syndrome. The benefits are anecdotal and variable in any individual experience and have never been rigorously assessed. It has been suggested that the natural path of chronic pancreatitis is toward progressive glandular insufficiency and calcification, and with the eventual ‘burnout’ would come spontaneous remission of pain[7]. There is a school of thought against conservative therapy. Pain is endured until burnout. This theory sheds light on the uncertainty regarding the duration of clinical pain, and if burnout is indeed a certainty and not solely a proposed hypothesis[89]. In conclusion, a strategy of waiting for spontaneous pain relief is not reliable and may be unreasonable advice for the patient with persistent or frequent severe pain.

Pancreatic enzymes: The presumed mechanism for pain relief after the administration of oral pancreatic enzymes is thought to involve the negative feedback inhibition to the pancreas. A cholecystokinin (CCK)-releasing peptide in the duodenum is normally denatured by pancreatic trypsin. In chronic pancreatitis, damage to acinar cells results in decreased secretion of pancreatic trypsin and consequently insufficient denaturing of the CCK-releasing peptide. This then leads to the potentiation and increased release of CCK, which causes pancreatic pain related to an increase in pancreatic enzyme output. When pancreatic enzymes are administered orally, there is more complete denaturing of the CCK-releasing peptide, thereby diminishing the release of CCK[1011]. The results of studies examining the use of pancreatic enzymes that are administered orally to treat the pain of chronic pancreatitis have been variable, in part because of a high placebo response rate of over 35%, the potential for exogenously administered digestive enzymes to be inactivated by gastric acid and pancreatic proteases, and the lack of efficacy of enteric coated preparations[12–16].

In one of the earliest double-blind randomized trials of pancreatic enzymes, Isakson et al showed the pain relieving effect of oral enzyme preparations in a proportion of patients with chronic pancreatitis[16]. They took 19 patients with chronic pancreatitis, and treated them for 1 wk with a granulated pancreatic enzyme preparation (Pankreon^®; five times daily 7.5 mL) or placebo and vice versa. Pain was evaluated using an analog scale and by questioning. A 30% pain reduction was seen after treatment with pancreatic extract compared to placebo. Fifteen of the nineteen patients had less pain during the week of treatment with pancreatic extracts. These results could not be confirmed by Halgreen, who conducted a 4-wk double-blind cross-over study with pancreatic enzymes (Pancrease^®) in 20 chronic pancreatitis patients. There was no significant pain reduction[17]. In a placebo-controlled, double-blind, crossover study, pancrelipase (Viokase), in a dose of six tablets taken four times per day for one month, significantly reduced pain in 75% of patients with mild-to-moderate disease[15]. The best response was in young women with idiopathic chronic pancreatitis, whereas patients with advanced disease, including those with steatorrhea, had no response.

Of the 6 randomized trials published to date two studies using a non-enteric coated enzyme preparation reported benefit and four studies using an enteric-coated capsule showed no effect on pain in chronic pancreatitis. The conflicting study results led to investigators questioning the mechanism of negative feedback inhibition in the proximal small bowel[18]. As noted, the presumed mechanism for pain relief with administration of oral pancreatic enzymes is thought to involve feedback inhibition of the exocrine pancreas by the degradation of CCK-releasing peptide in the duodenum. The administered enzymes would need to release activated serine proteases into the duodenum. This is much more likely with the non-enteric coated than the enteric-coated preparations, and hence the suspicion that the former are more effective. A meta-analysis of the six randomized, double-blind, placebo-controlled trials for the treatment of chronic pancreatitis with pancreatic enzymes showed no benefit in improving pain. The pooled estimate of the percentage of patients per study who preferred enzymes relative to placebo was 52% (95% confidence interval 45%-60%). This was not statistically different from 50%. Thus, this analysis demonstrates no significant benefit of pancreatic enzyme therapy to relieve chronic pancreatitis-associated pain[19]. It should be noted that this meta-analysis combines studies using enteric-coated and studies using non enteric-coated preparations. In that way, the potential benefit of non enteric-coated enzymes may have been negated by the lack of positive effect with non-enteric-coated preparation. The role of oral pancreatic enzymes in reducing pain in chronic pancreatitis, therefore, remains unclear. Additional studies are required to establish the effectiveness of this modality of treatment and to define whether certain subsets of pain: chronic versus intermittent pain; patients with or without exocrine insufficiency; alcoholic versus idiopathic pancreatitis; minimal versus extensive pancreatic duct changes; are more likely to benefit from enzyme therapy than others. Non-enteric coated enzymes are certainly safe and reasonable to try before considering more invasive or risky therapies.

Octreotide: Cholecystokinin-receptor antagonists or somatostatin analogues, such as octreotide, have been postulated to work on the negative feedback inhibition as well as hypertension of the pancreatic duct due to outflow obstruction. Inhibition of pancreatic secretion using somatostatin might, therefore, be effective in reducing pain in chronic pancreatitis. Octreotide is a synthetic somatostatin-analogue with an increased half-life, higher potency and the possibility of subcutaneous application. Experimental data suggest that octreotide increases the contractibility of the sphincter of Oddi, while somatostatin decreases it. This has, however, not consistently been demonstrated[20]. Normally, the release of cholecystokinin from specific intestinal cells is regulated by a cholecystokinin-releasing peptide in the proximal small intestine that is luminally active and trypsin-sensitive[13]. In chronic pancreatitis, exocrine insufficiency may lead to increased cholecystokinin-mediated stimulation of the pancreas. Theoretically, this process could be interrupted by the administration of cholecystokinin-receptor antagonists, or somatostatin. In a multicenter pilot study, octreotide, in a dose of 200 &mgr;g administered subcutaneously three times per day for 4 wk, reduced pain scores by 25% or more in 65% of patients with severe chronic pancreatitis[21]. On the other hand in a randomized, prospective, double-blind, placebo-controlled study conducted in Europe [100 mg subcutaneously (sc) every 8 h] administered to 10 patients for only 3 d was no more effective than placebo in relieving pain in chronic pancreatitis[22]. In a second study[23], octreotide (100 mg sc every 8 h for 3 wk) administered to six patients in a nonblinded fashion provided relief of pain in some but not all patients. In a third study[24], octreotide was administered to 84 patients for 4 wk in a randomized, prospective, double blind trial and showed a trend toward benefit at the highest dosage used (200 mg sc every 8 h). However, this effect did not reach statistical significance in this dose-ranging study. The longevity of the possible benefit was not established. Clearly further studies are needed before the use of octreotide can be widely adopted.

Antioxidant therapy: Bhardwaj et al[25] reported a decreased micronutrient intake (Vitamin E, riboflavin, choline, magnesium, copper, manganese and sulphur) in patients with chronic pancreatitis. This was due to diet modifications due to pain, as well as to a lower caloric intake. This points to the possibility that micronutrients deficiency may contribute to increased oxidative stress. In a comparison between patients with chronic pancreatitis and acute pancreatitis, the antioxidant profiles appeared to be different. Patients with chronic pancreatitis had significantly lower plasma concentrations of selenium, Vitamins A and E, beta-carotene, xanthine, beta-cryptoxanthine and lycopene in comparison with patients with recurrent acute pancreatitis[26]. Cullen et al[27] reported a decrease in antioxidant enzyme expression in pancreatic cells from normal pancreas to chronic pancreatitis to pancreatic cancer. Another observation concerning antioxidants is the altering of antioxidant status in chronic pancreatitis patients, which is worsened in patients with diabetes mellitus[28]. A 1-year clinical trial with 10 patients studied the effect of food supplementation using a complex containing l-methionine, beta-carotene, Vitamins C and E and organic selenium[29]. This resulted in a significant decrease in the intensity of pain as well as in days of hospital admission. Based on a placebo-controlled trial, followed by a retrospective cross-sectional study in 94 patients, some authors recommend antioxidant therapy consisting of supplements of methionine, Vitamin C and selenium[30].

Based on the observations that activation of oxygen free radicals can cause metabolic changes leading to pancreatic ischemia, antioxidant treatment with allopurinol seems a valid option. A trial with 13 patients with chronic pancreatitis investigated the effect of allopurinol on pain in a cross-over double-blind, randomized treatment trial[31]. Allopurinol, which is believed to reduce oxidative stress by inhibiting xanthine oxidase and thereby preventing the formation of oxygen derived free radicals, was given to 13 patients with pain occurring at least three times each week. Allopurinol was not effective in reducing pain or improving activities of daily living compared to placebo. In contrast, others showed that addition of allopurinol or dimethyl sulfoxide to intramuscular pethidine hydrochloride significantly enhanced the efficacy of the analgesic regime[32]. This report suggests that removing oxygen free radicals in chronic pancreatitis may result in a beneficial therapeutic effect. The results of the most recent randomized trial presented only in abstract form showed that the combination of selenium, Vitamin C, β-carotene, Vitamin E, and methionine was significantly better in controlling pain compared to placebo[33]. In summary, there are conflicting data about the effectiveness of antioxidant therapy. A few trials show potential benefit, but further research is needed before it can become standard of therapy.

Endoscopists have shown that they can overcome pancreatic duct obstruction caused by ampullary stenosis, strictures, or stones. However, there have been no published validated guidelines for defining significant obstruction, and methodology for assessing patients before treatment and then judging the efficacy of that treatment. It should be noted though that the alternative to endoscopy, surgical sphincterotomy and sphincteroplasty, have already proven to be less efficient[3435]. These interventions are hardly ever used now. This may also be due to the more acceptable rate of complication with endoscopic procedures, in conjunction with stent placement and stone extraction. For the present, the decision to perform endoscopy is based partially on subjective judgments that include assessment of the need for long-term narcotic therapy, marked diminution of the quality of life because of intractable pain, or major nutritional consequences of pain. When major pain episodes cannot be controlled by major, but acceptable maintenance analgesics, intervals of narcotics, or reasonable and brief periodic hospitalizations, a trial of interventional therapy can be justified. Among three recent studies involving stent therapy in 98 patients, at times associated with other interventional therapies such as lithotripsy and/or sphincterotomy[36–38], two studies[3638] reported amelioration of pain and one did not[37].

The ideal treatment for patients with pancreatic-duct stones, dilated pancreatic ducts, and pain is not known. The stones can be easily removed coincidently with the performance of a surgical-drainage procedure, such as pancreaticojejunostomy. Alternatively, however, they can be fragmented by extracorporeal shock-wave lithotripsy (ESWL) and removed endoscopically after sphincterotomy of the pancreatic duct. Stones can be cleared by this approach in roughly 80 percent of patients, and approximately 50% of these have long-term relief of their symptoms[3940]. Dumonceau et al conducted a randomized trial comparing pain relief after extracorporeal shock wave lithotripsy alone versus in combination with endoscopic drainage of the main pancreatic duct in patients with painful calcified chronic pancreatitis. Two years after trial intervention, 10 (38%) and 13 (45%) patients of the ESWL alone and ESWL combined with endoscopy group, respectively, had presented pain relapse. In both groups, a similar and significant decrease was seen after treatment in the number of pain episodes/year (mean decrease 3.7 episodes). There was no difference between the treatment groups and the treatment costs per patient were three times higher in the ESWL combined with endoscopy group compared with the ESWL alone group[41]. The claims for the efficacy of stone removal for pain relief should be considered in context with the observations that the presence or absence of stones does not necessarily correlate with the existence of pain. In the absence of randomized prospective trials comparing stone ablation either with placebo or with surgical decompression, it is difficult to assess the results of pancreatic stone removal.

An alternative involves the use of endoprostheses or stents placed in the pancreatic duct endoscopically. Reports indicate that 30%-76% of patients receiving such stents had symptomatic improvement over a period of 14 to 36 mo of observation[42–46]. Cremer et al[42], for example, noted initial improvement of symptoms in 94% of patients who were so treated for pancreatic-duct strictures and upstream ductal dilatation. In that group of patients, 53% remained free of symptoms over a mean follow-up period of 36 mo. Similarly, Grimm et al[43] showed that 57% of their patients were symptomatically improved by this treatment over a mean follow-up period of 19 mo. Although these results seem encouraging, a criticism is that most of the data reported to date were from relatively short term, nonrandomized studies. The issue is further complicated by the fact that pancreatic-duct stents may not be entirely harmless; for example, they may cause further pancreatic duct changes and potentiation of chronic pancreatitis[47–49]. Endoprosthesis occlusion and migration also seem to be relatively common.

Analyzing all the endoscopic modalities taken together it is usual to find a report of 80%-90% complete stone clearance and good immediate pain relief[47]. The long term results were not as favorable in the larger series. Delhaye et al[39] found that of 123 patients, only 60% experienced complete or partial pain relief during 14 mo follow-up. So far there are two randomized control trials comparing endoscopic therapy with surgery[5051]. The study from Dite et al randomized 72 patients with large duct chronic pancreatitis to endoscopic therapy versus surgical lateral pancreaticojejunostomy. In addition, 68 patients were treated with endoscopy or surgery based on patient preference. The results between the randomized and nonrandomized study groups are similar. After 5 years of follow-up only 14% of the patients treated by endoscopy were pain free compare with 36% in the surgery group. The latest randomized controlled study comparing endoscopy with surgery (lateral pancreaticojejunostomy) enrolled 36 patients. The results are strikingly similar to the previous study. Pain was absent in 16% of patients treated with endoscopy and 40% in patients treated with surgery. Based on these trials it appears that surgery provides better pain relief compared to endoscopy, but even surgery fails to provide substantial pain relief in more than half of the patients[51].

Endoscopic treatment may have a place in the prevention of acute relapsing pancreatitis, more so than treatment of the pain of chronic pancreatitis. To avoid this potential problem, some have suggested that endoscopically placed pancreatic-duct stents should be used only for relatively short periods. This serves as a screening procedure, to identify those patients most likely to benefit from surgical drainage[424552]. At present, endoscopically placed stents should be considered an unproved, but potentially useful approach to the treatment of chronic pancreatitis.

Kozarek and Traverso[53] have analyzed collected experiences and indicate that the likelihood of symptomatic improvement with combination endotherapy is reported to be 50%-85% at 15 to 25 mo. Successful pain relief has been correlated anecdotally with stone removal and subsequent decrease in diameter of the pancreatic duct. As a rule, the focus is on stones in the main duct and the morbidity of side-branch stones has not been defined. Better selection of patients for endotherapy may be helpful in order to maximize results. Due to its low degree of invasiveness, however, endotherapy can be offered as a first-line treatment, with surgery being performed in case of failure and/or recurrence.

Although this modality is thought to be medical management, it may be administered via endoscopic or interventional radiological means. Although widely used, there have been relatively few formally reported experiences with nerve blocks for long-term therapy of chronic pancreatitis. Leung et al[54] studied the use of celiac block in 23 patients with chronic pancreatitis. Twelve of the 23 had complete analgesia, whereas six had partial relief. There was no effect in five patients. The mean pain-free period in the chronic pancreatitis patients was only 2 mo, and the longest 4 mo. Benefit was least in patients with previous pancreatic surgery and repeat blocks were unhelpful.

Because of possible concerns about potential irreversible nerve injury, including very rare anecdotes of paraplegia from neurolytic agents, injection of steroids for the treatment of chronic pancreatitis has been recommended, instead of the use of alcohol injected into the celiac plexus (principally used in the treatment of cancer pain)[5556]. In one study, steroid injection provided relief of pain (lasting two mo) in only 4 of 16 patients[57]. Eleven of the 12 patients who did not obtain relief were narcotic dependent, whereas none of the four who obtained relief were narcotic dependent. This finding emphasizes the complexity of treating pain in a population of patients with chemical dependencies and other abnormal psychological and psychosomatic behavior. In another report[58], which investigated the mode of delivering the nerve block, only 2 of 8 patients with a CT-guided celiac plexus block experienced relief of pain compared with 6 of 14 who were treated by endoscopic ultrasonography-guided celiac plexus block with 10 mL of bupivacaine. The benefit from endoscopic ultrasonography-guided celiac plexus block seemed to persist longer that CT-guided block. More importantly, paraplegia has not been described after endoscopic ultrasonography-guided celiac plexus block, probably because of the anterior transgastric approach taken during endoscopic ultrasonography-decreasing or even eliminating the risk of nerve or spinal cord injury. The same group of investigators more recently published their prospective experience with endoscopic ultrasonography-guided celiac plexus block with steroids in 90 patients with pain resulting from chronic pancreatitis[59]. A significant improvement in pain score occurred in 55% of the patients. The benefit persisted beyond 12 wk in 26% of patients and beyond 24 wk in only 10%. Younger patients (< 45 years) and patients with previous pancreatic surgery for chronic pancreatitis did not appear to benefit from the block.

The current evidence indicates that endoscopic ultrasonography-guided celiac plexus block is safe and well tolerated, with excellent temporary results in some patients. Unfortunately, reliable predictors of success are lacking. In the absence of long-term studies with follow-up in patients with chronic pancreatitis whose pain is chronic, the role of endoscopic ultrasonography- guided celiac plexus block should be limited to treating flares of chronic pain in patients with otherwise limited therapeutic options.

Duval pioneered efforts to treat the pain of chronic pancreatitis by surgical means in the 1950s with trans-duodenal sphincteroplasty and with caudal pancreato-jejunostomy (the Duval procedure). The results of this procedure were fraught with variable and usually poor results, perhaps only helping some of those patients with true recurrent acute pancreatitis[60]. A more extensive drainage procedure, lateral pancreatojejunostomy, described by Puestow and Gillesby[61] and subsequently modified by Partington and Rochelle[62], was applied to the subset of patients with dilated main pancreatic duct and became the first surgical treatment widely considered to be effective for pain in this disease. At that time, however, its application was hampered because there was no way to determine preoperatively if a patient with chronic pancreatitis had the dilated ducts required for this procedure because neither ERCP nor CT was available until the 1970s. Thus, at exploration an intraoperative pancreatogram was used to select who would be candidates for lateral pancreatojejunostomy. In those without dilated ducts, the remaining options were to perform a sphincteroplasty (which was largely abandoned because of its failure) or to do nothing further. In the 1960s, surgeons began performing pancreatic resections for chronic pancreatitis, initially distal pancreatectomies (with poor results) and later distal subtotal (95%) resections, which were relatively more effective for pain, but rendered most patients diabetic[63]. Proximal resections of the head of the pancreas (i.e. Whipple procedures) were not widely applied until the 1980s, when the associated operative morbidity and mortality fell substantially[64–66].

Patients whose pain persists in spite of aggressive noninvasive treatment should undergo endoscopic retrograde pancreatography to define the caliber and morphologic characteristics of their pancreatic ducts. Depending on the population being studied, up to half of these patients may have dilated ducts, frequently with areas of stricture-the "chain of lakes" or "string of pearls" appearance; the remainder have either ducts of normal caliber (2 to 4 mm in diameter) or small ducts that may lack side branches-the "tree in winter" appearance[6768]. Ducts larger than 8 mm in diameter can be successfully decompressed by an internal surgical-drainage procedure, such as a longitudinal pancreaticojejunostomy (the modified Puestow procedure)[6062], but smaller ducts are not amenable to internal surgical drainage or resection.

Like most surgical procedures currently in use, those for chronic pancreatitis gradually became part of the armamentarium without undergoing rigorous testing and were never compared against medical treatment or no treatment. The vast majority of patients are still operated on when they continue to have intractable pain despite medical treatment. There are very few controlled trials in the surgical literature on this disease. The two randomized controlled studies comparing surgery with endoscopic therapy are discussed in the endoscopic therapy section. Surgical options include decompression/drainage operations, pancreatic resections, and denervation procedures. As with endoscopic interventional therapy, objective transferable criteria for the need for surgical intervention have not been developed or agreed upon.

Decompression/drainage operations: At present, the ultimate role of these various invasive approaches to the treatment of patients with large-duct, symptomatic chronic pancreatitis has not been established. Given the information available at the present time, most physicians recommend longitudinal pancreaticojejunostomy for patients with pain and dilated ducts. This operation may also retard the progression of exocrine and endocrine insufficiency[6970]. Surgical decompression of the obstructed main pancreatic duct was for a long time the gold standard[71]. Drainage procedures today are most commonly side to side pancreaticojejunostomy. This particular procedure preserves parenchymal function. Longitudinal pancreaticojejunostomy is also used based on the concept the ductal obstruction leads to distention and that this in turn gives rise to pain and should thus be favored if the duct is widened. Ebbehoj et al[72] were able to show a relationship between the degree of pain and intrapancreatic pressure. Pancreatic pressure was measured by a percutaneously placed needle preoperatively, postoperatively, and one year after pancreatic duct drainage. Patients whose pressure decreased after surgery and remained low were pain free, whereas those with recurrent pain had increased pressure.

Theoretically, any procedure that improves drainage, either by improving flow into the jejunum or stomach, might be expected to relieve pain. Pancreatic decom- pression results in immediate and lasting pain relief in a high proportion (80%-90%) of patients with non-alcoholic chronic pancreatitis[73]. These procedures have been less successful with alcoholic chronic pancreatitis with pain relief averaged at 60%[74]. Although early good results have also been reported after a lateral pancreaticojejunostomy in patients with alcoholic pancreatitis, when these patients are followed for 5 years only 38%-60% of them continue to be pain free[75]. These operations are predicated upon the presence of a widely dilated main pancreatic duct (generally taken as > 6 to 7 mm) and the presumption that the dilated ducts imply an abnormally high pressure in the duct system[75] and in the pancreatic parenchyma[7276]. The operation most commonly performed is a variant of the Puestow procedure, which is actually the Partington-Rochelle modification (lateral pancreaticojejunostomy)[62].

Many of the studies of lateral pancreatico-duodenectomy find that short-term pain relief is achieved in about 80% of patients and that the operation can be performed with a very low morbidity and mortality (0%-5%). Although the short-term studies shine a positive light on the procedure, long-term follow-up studies show that pain not uncommonly recurs. As time goes by, pain recurs, perhaps related to progression of the pancreatic injury and fibrosis. Pain relief for greater than two years is achieved in only 60% of patients[7778]. Strategies for salvage in patients with persistent or recurrent pain after drainage procedures include redoing or extending the pancreatojejunostomy and resection procedures[79]. Of patients undergoing pancreatic duct drainage procedures, 25%-66% require concomitant biliary or gastric drainage, because of functionally significant obstruction of the bile duct or duodenum[8081]. Biliary or duodenal strictures have been reported to be more likely in patients with large-duct disease than in their counterparts without dilated ducts[81].

The only reported attempt made to compare pancreatic duct drainage with no intervention in the management of pain is that of Nealon and Thompson[70]. In a series of 143 patients with chronic pancreatitis, 85% of the 87 patients who were treated by pancreatic duct decompression achieved pain relief, whereas pain abated spontaneously in only 1.3% of the 56 nonoperative patients. The study was not randomized, however, the principal criterion to determine candidacy for the operation was the presence of a dilated pancreatic duct. Thus, what the study actually reports is the outcome of pancreatojejunostomy in patients with dilated ducts versus the natural history of patients with chronic pancreatitis and no duct dilation. The study also found that deterioration of pancreatic function was slower in their patients with dilated ducts than in those with small ducts. Although this effect was ascribed by the investigators to the protection or relief afforded by the surgical drainage procedure, the cause and effect relationship is uncertain because of the differences in the patient population.

The consensus, albeit based on evidence from collected experiences, states that pancreatic duct decompression via lateral pancreatojejunostomy (a Puestow-type operation) can be accomplished with low associated morbidity and mortality and that pain relief will be achieved in the majority of patients. For most experienced pancreatic surgeons, it is the preferred surgical treatment option in patients whose main pancreatic duct measures 6 mm or more because of its simplicity, safety, and benefits, including the advantage that remaining pancreatic tissue and function are at least not compromised further by loss from resection.

Drainage of pancreatic pseudocysts provides another form of pancreatic decompression in conjunction and even in continuity with a lateral pancreatojejunostomy when the main duct is also dilated. Up to 39% of patients undergoing lateral pancreaticojejunostomy have evidence of pseudocysts disease at the time of surgery[82]. Pseudocysts are found in about 25% of patients with chronic pancreatitis and have a much lower rate of spontaneous resolution than those that are a consequence of an attack of acute pancreatitis[82–84]. They can be the source of pain indistinguishable from that of the underlying chronic pancreatitis. In one study, surgical drainage resulted in complete short-term pain relief in 96% of 55 patients, and 53% remained pain free after a median follow-up of 11 years[84]. Endoscopic drainage of pseudocysts into the stomach or duodenum may be an alternative, especially in patients who do not have associated duct dilation. Studies directly comparing surgical with endoscopic drainage of pseudocyst are lacking.

It should also be mentioned that there are numerous variations of the previously mentioned operations. Frey et al[8586] combined a coring out of the pancreatic head with a lateral pancreaticojejunostomy. In his series, the pain relief after 5 years was complete or improved in 87% of cases. There is also one randomized series of patients comparing the Beger and Frey procedure[86–90], with no difference in decrease of pain, but less morbidity with the Frey procedure.

Resection procedures: The therapeutic principle of resection is based on the assumption that pain in chronic pancreatitis is predominantly caused by inflammation. This inflammation then becomes the nidus for qualitative and quantitative changes of nerve fibers. This is especially seen in the clinical scenario of normal sized ducts and masses of the head of the pancreas. Thirty percent of patients with chronic pancreatitis develop inflammatory enlargement of the pancreatic head with subsequent obstruction of the pancreatic duct, and sometimes also of the common bile duct and duodenum. In these cases a pancreaticoduodenectomy, "Whipple procedure", has been the procedure of choice for a long time, as it provides reasonably effective pain relief. These resections, however, have both immediate postoperative morbidity and long-term morbidity. Insulin dependent diabetes mellitus has an increase in the incidence from 20% preoperatively to 60% in the years that follow[81]. Also, postgastrectomy complications detract significantly from the overall quality of life. The long-term mortality rate and quality of life after this procedure in patients with chronic pancreatitis has not always been encouraging, and in some studies disappointing[71]. Distal pancreatectomy alone had poor results unless the disease is largely confined to the body and tail of the gland, e.g. with an occlusion of the mid-pancreatic duct or with a pseudocyst in the tail. By contrast, resection of the pancreatic head by either a conventional or pylorus-preserving pancreaticoduodenectomy will provide pain relief in up to 85% of patients, even if the disease extends into the distal pancreas. In order to deal with these undesirable consequences of the Whipple procedure, surgeons turned to the pylorus preserving pancreaticoduodenectomy (PPPD) and the "Beger procedure"[88–91]. Russel[92], in studying the results of preservation of the duodenum in total pancreatectomy compared with those of standard pancreaticoduodenectomy, found no difference in pain relief between the results of the two operations. He noted that 13 (14%) of the 32 still had severe pain after duodenum preserving total pancreatectomy, and that six required major analgesics. The purported benefits of better postoperative nutritional status and glucose control in the duodenum-preserving procedure were addressed in two randomized trials[9394].

Frey and Amikura have recently reported a surgical modification that combines removing part of the anterior segment of the pancreatic head with longitudinal duct anastomosis to the jejunum[86]. A randomized trial[87] found little difference between the Frey procedure and the duodenum-preserving resection of the pancreatic head as described by Beger and Buchler[95].

Noteworthy in recent years has been the very low operative morbidity and mortality of pancreatic resection, which may be one reason for the larger numbers of patients with benign disease being referred for surgical treatment. In a recent series of 231 pancreatic resections, the most frequent indication being chronic pancreatitis, the operative mortality was 0.4%[66]. McLeod et al[96] studied the morbidity of the Whipple operation. Although the study focused on resections for neoplasms, the observations pertain as well to those for chronic pancreatitis and show satisfactory digestion, weight maintenance, and activity level in the great majority of patients. A study of quality of life after pancreatic resections found that diabetes and its complications had the greatest negative influence on everyday well-being[97].

Distal pancreatectomy[98] has a very limited role in management of pain, and only in patients with non-dilated pancreatic duct and pseudocysts involving the tail of the pancreas does this procedure seem to be associated with a good outcome[99]. Keith et al[100], analyzed the results of 80% distal pancreatectomy, pancreaticoduodenectomy and total pancreatectomy. After an average follow-up of 5 years, 9 years, and 6 years, respectively, he found that four of five patients after pancreaticoduodenectomy required narcotics. Thirteen of 32 patients had complete pain relief after 80% distal pancreatectomy. Finally total pancreatectomy is usually reserved as a last resort following a failed partial pancreatic resection.

Resection of pancreatic tissue results in the loss of some exocrine and endocrine function and increases the possibility or hastens the onset of fat malabsorption and diabetes. Whereas only 20% of normal pancreatic tissue is required for clinically adequate function, the pancreas already damaged by chronic pancreatitis may have substantially reduced reserves even before resection. Because of the complete lack of insulin and glucagon after total pancreatectomy, very brittle diabetes may ensue and can be the source of considerable morbidity and even mortality. In an attempt to lessen these adversities, autotransplantation of either part of the organ[101] or of islet tissue[102] has been described. In the latter study, Farney et al obtained insulin independence in 20% of 24 patients at a mean follow-up of 5.5 years. A more extensive experience with islet cell autotransplantation was reported by the Minnesota group in 1995 comprising 48 patients[103]. Forty-seven of the 48 patients had small duct chronic pancreatitis. Only one postoperative death resulted, but 25% of patient’s encountered complications. There were 8 deaths in the follow-up period, none apparently attributable to the operation. In follow-up, from 1 mo to 17 years, 39% of patients reported that pain was resolved, and 61% still had some degree of pain. Twenty of 39 evaluable patients (51%) had initial (less than 1 mo) insulin independence, but this dropped to 15 patients (38%) beyond 1 mo. A more recent European experience of 13 patients indicated sustained insulin independence in 5 of 9 surviving patients (4 late deaths) from 9 to 48 mo after surgery[104]. The latest studies suggest improvement in both the areas of brittle diabetes and in pain control. Rodriquez et al[105] recruited 22 patients who underwent pancreatectomy and autologous islet cell transplantation. All patients demonstrated C-peptide and insulin production indicating graft function. Forty-one percent were insulin dependent, and 27% required minimal amount of insulin or a sliding scale. Eighty-two percent no longer required analgesics postoperatively and 14% experienced a decrease in need for narcotics. Their success was attributed due to the provision of pancreatectomy and islet cell transplantation earlier in the course of the disease. Clayton et al[106] followed 40 patients who had pancreatectomy followed by islet cell transplantation. At 2 years post-transplant, 18 patients had a median HbA1c of 6.6% (5.2%-19.3%), fasting C-peptide of 0.66 ng/mL (0.26-2.65 ng/mL), and required a median of 12 (0-45) units of insulin per day. At 6 years, these figures were 8% (6.1%-11.1%), 1.68 ng/mL (0.9-2.78 ng/mL) and 43 U/d (6-86 U/d), respectively. The majority of patients (68%) no longer require opiate analgesia. Finally, Gruessner et al[107] performed 112 islet autotransplants at the time of total pancreatectomy. They found that islet autotransplants, at the time of total pancreatectomy in patients who had not had previous operations on the body and tail of the pancreas, were associated with > 70% of the recipients achieving complete insulin independence. In contrast, a previous distal pancreatectomy or a Puestow drainage procedure was associated with complete insulin independence in < 20%. Islet autotransplantation offers a valuable addition to surgical resection of the pancreas, as a treatment for chronic pancreatitis; and even in cases in which insulin independence is not achieved, the potential beneficial effects of C-peptide make the procedure worthwhile, particularly in early disease.

Many studies on pancreatic resection and even those on drainage procedures show that up to 15% of patients undergoing these surgical treatments for treatment of pain due to chronic pancreatitis will be found to have pancreatic cancer[77999108] and it has been shown that a chronic pancreatitis is in fact, a small, but real risk factor in the development of pancreatic cancer[109]. This is an important consideration to keep in mind during the diagnostic work-up and choice of operation. The morphology of the pancreas by CT imaging and by cholangiopancreatography may fail to discriminate between cancer and chronic pancreatitis. Cytological confirmation by fine-needle aspiration is helpful when positive, but the true diagnosis may become known only with resection (10% of cases). This consideration in some cases may determine the treatment strategy.

Surgical denervation: Most of the sensory nerves returning from the pancreas pass through the celiac ganglion and splanchnic nerves. It is hypothesized that interruption of these fibers may lessen pain. Mallet-Guy[110] reported an experience with 215 patients over 30 years whose principal treatment for pain was by sensory denervation. These patients first underwent abdominal exploration to document the absence of pancreatic ductal dilation or pseudocysts and to correct any associated biliary pathology; this was immediately followed by resection of the greater splanchnic nerve and celiac ganglion through a left translumbar approach. Although excellent long-term results are reported (90% of patients were pain-free, with 60% followed for more than five years), the heterogeneity of the patient population and the simultaneous use of biliary diversion procedures in many cases precludes meaningful conclusions. This treatment has not been widely accepted.

The celiac block can be done during laparotomy or percutaneously, usually from the back. The placement of the injection can be done simply by using anatomical landmarks or by checking the position with an imaging modality: fluoroscopy, scout X-ray films, ultrasonography, computed tomography, or at angiography. A nerve block with 25 mL of 50% alcohol on each side should be preceded by a positive diagnostic block with long acting local anesthesia, carried out at least 1 d earlier. The method aims at blockage of the splanchnic nerves before they reach the celiac plexus[111].

Stone and Chauvin reported on 15 patients with chronic pancreatitis who had previous unsuccessful operative procedures for pain[112]. Denervation was accomplished with a transthoracic left splanchnicectomy with concomitant vagotomy, and all 15 patients had immediate pain control. Five later suffered recurrent pain, but were successfully treated with a right splanchnicectomy. The long-term outcomes are not known. The advent of thoracoscopic surgery has made this procedure more attractive, and a few small series have reported its feasibility and early results[113114]. Maher et al recently reported on 15 patients with chronic pancreatitis, mostly idiopathic, with chronic pain measured by visual analogue pain scale[115]. Unilateral thoracoscopic splanchnic nerve resection in eight patients and bilateral in seven patients resulted in significant decreases in pain frequency and intensity, as well as in narcotic consumption. Overall, 80% of patients had good results or were improved, with a mean follow-up of 16 mo. A controlled trial comparing this procedure to other surgical options or to medical treatment is needed. Of note, pancreaticoduodenectomy and duodenum-preserving resection of the pancreatic head may well confer pain relief at least in part through denervation.