Chronic pancreatitis (CP) is characterized by recurrent or chronic abdominal pain with late progression to exocrine and endocrine insufficiency. The pain can be so severe that affected patients continue to have pain despite treatment with narcotic analgesics. At that point, the objective of therapy is to interrupt or eliminate the root cause of the pain, while preserving as much exocrine and endocrine function as possible. Because intraductal pressure may be increased in some cases of painful CP, the first-line treatment is usually endoscopic sphincterotomy and stenting procedures (now preferred over surgical drainage) (1). Celiac nerve blocks also can be administered, although any pain relief they confer is usually transient.
If the above measures fail, pancreatectomy can eliminate or significantly reduce pain and restore quality of life (2). However, total pancreatectomy alone is complicated by the development of diabetes due to loss of β cell mass. Islet autotransplantation (IAT) performed at the time of pancreatectomy can preserve β cell mass, minimizing the severity of diabetes, or, in some patients, preventing it altogether (3).
Islet autotransplantation after pancreatectomy for CP was first performed at the University of Minnesota (UM) Medical Center in 1977 (4). Since then, more than 300 such transplants, nearly all in adults, have been reported in the medical literature (2). The probability of pain relief with total pancreatectomy is relatively high, and most patients are able to discontinue narcotic analgesics postoperatively (5–8). However, some continue taking narcotics even if pain is relieved because of intractable addiction or narcotic-induced hyperalgesia induced by long-term use.
There is no guarantee that IAT will be successful, so recipients must be willing to accept diabetes as a tradeoff for the chance to withdraw from narcotics. Thus, insulin independence is a desirable but secondary outcome for CP patients with severe pain. Nevertheless, more than 40% of adult patients have achieved insulin independence after pancreatectomy with IAT (5,7–9), for more than 13 years in the recipient with the longest reported follow-up time (10).
The prevalence of CP in pediatric patients is unknown, but it is rare. In children without predisposing genetic mutations, the cause is usually unknown (11). In contrast, in adults with nonhereditary CP, the most common identifiable cause is alcoholism, followed by idiopathic causes (12). In children, as in adults, the pain and inflammation associated with CP may result in narcotic dependence, may interfere with routine childhood activities (including school attendance), and may impair normal growth and development. The rarity of this disorder in the pediatric population can delay diagnosis and proper medical treatment.
A small number of children with intractable pancreatic disease have undergone pancreatectomy with IAT. However, only 1 previous pediatric case report has been published in the medical literature (13). To better characterize outcome after pancreatectomy and IAT in pediatric patients, we retrospectively reviewed the medical records of all 24 patients younger than 19 years who underwent this combined procedure at the UM Medical Center from 1989 to mid-2006. We hypothesized that in children, as in adults, pancreatectomy would result in significant pain relief with reduction in incidence of narcotics dependence, and that IAT would prevent the development of diabetes, or preserve at least some β cell function, in a proportion of patients. We further sought to compare outcomes in patients transplanted during preadolescence (<13 years) vs those transplanted during adolescence (13–18 years).
PATIENTS AND METHODS
The study protocol was reviewed and approved by the UM institutional review board, Human Patients Committee. We reviewed records from the UM Department of Surgery's Division of Transplantation to select all pediatric patients who underwent a pancreatectomy with IAT and to analyze their medical and surgical characteristics.
Data obtained from medical records included the etiology of pancreatic disease, age at surgery, presurgical metabolic status, and prior treatment history including endoscopic retrograde cholangiopancreatography or pancreatic surgery. Past surgical procedures included partial pancreatectomy and lateral pancreaticojejunostomy (Puestow or Frey). Either procedure may result in loss of endocrine pancreatic tissue. In partial (distal) pancreatectomy, as much as half to two thirds of the pancreas (body and tail) is removed. In lateral pancreaticojejunostomy, a longitudinal incision is made through the pancreas and the pancreatomy is anastomosed to the jejunum to allow pancreatic drainage; depending on surgical technique, this procedure may involve limited resection of tissue in the pancreatic head (Frey) or tail (classic Puestow).
The technique of pancreatectomy and IAT in children is identical to the procedure performed in adults and is described in detail elsewhere (2,14). Briefly, during resection of the pancreas, the pylorus and spleen are spared whenever possible. The excised pancreas is distended intraductally with collagenase (15), then digested using an automated method (16). Islets are suspended in approximately 500 mL of culture medium and infused into the portal vein. Before islet infusion, patients undergo anticoagulation with heparin, as prophylaxis against portal vein thrombosis. Portal pressures are monitored throughout islet infusion. If elevated portal pressures (≥25–30 mmHg) prevent infusion of all islets intraportally, the remaining islets are transplanted in the intraperitoneal cavity, stomach subserosa, or Gerota fascia of the kidney (at the surgeon's discretion).
Clinical Outcomes Follow-up
Using basic information maintained by the Department of Surgery, we attempted to contact all 24 patients by telephone or mail. Informed consent was given by the parents of pediatric patients and consent or assent, as appropriate, by the patients themselves.
A telephone survey of patients or their parents included questions regarding each patient's current height and weight, the presence or absence of diabetes, the use of insulin or other diabetes medications, the results of blood glucose or hemoglobin A1c (HbA1c) tests performed in their physician's office or at home, the presence of gastrointestinal or digestive problems, any pain symptoms, the need for narcotic pain medications, their quality of life (assessed by 5-point Likert scale), their ability to work or attend school, and the frequency of hospitalizations. Narcotics use was classified as chronic if narcotic medications were required on a daily basis and as intermittent if they were required only during episodes of disease exacerbation. We classified islet graft function as full for patients who were insulin independent, partial for those requiring once-daily basal insulin (glargine or Detrimer), or graft failure for those on a basal-bolus (typical diabetes) insulin regimen. Using Centers for Disease Control and Prevention standards, we calculated z scores (standard deviation [SD] scores) for weight by age and sex to the nearest 0.25 SDs.
Patients or parents contacted by telephone were mailed HbA1c kits to complete at home themselves and then return in provided envelopes; verbal and simple standardized written instructions were provided. All of the samples were analyzed in our UM laboratory using high-performance liquid chromatography. In addition, patients were asked to record fasting, premeal, 2-hour postprandial, and bedtime blood glucoses for 2 consecutive days, then to mail us the recorded values. Before discharge from our institution, all pediatric patients were educated in the use of home glucometers for blood glucose monitoring regardless of insulin needs.
We used descriptive statistics to examine the patients' baseline characteristics and the correlation between the observed variables. To compare categorical variables, we used the Pearson chi-square test; to compare differences between continuous variables, we used a Wilcoxon test. We also categorized continuous variables for use in multivariate models. We used the Kaplan-Meier method to calculate graft survival rates. To estimated group differences, we used the generalized Wilcoxon test for short-term differences and the log-rank test for long-term outcome. We used a stepwise Cox proportional hazard model to independently estimate the impact of potential risk factors on outcome.
A P value of ≤0.05 was considered to be statistically significant. Statistical analysis was performed using SPSS 14.0 (SPSS, Chicago, IL).
From July 1989 through June 2006, 23 patients younger than 19 years underwent total pancreatectomy and 1 patient underwent subtotal pancreatectomy (body and tail resection) with IAT. Patient ages ranged from 5.8 to 18.9 years. The most common etiology of underlying pancreatic disease was hereditary (9 patients) or idiopathic pancreatitis (8 patients). The median duration of presurgical disease was 4 years (range 1–16 years), although it was not uncommon for patients to report recurrent abdominal pain before establishing the diagnosis of pancreatitis.
A total of 18 patients (75%) required chronic narcotics for pain control preoperatively; the remaining 6 patients (25%) required intermittent narcotic medications. All 24 patients experienced multiple or prolonged hospitalizations in the year before their pancreatectomy, with a median of 5 hospitalizations in the year before surgery. Before pancreatectomy, 16 patients (67%) underwent endoscopic retrograde cholangiopancreatography with sphincterotomy and/or stenting procedures. Eleven (46%) had a history of prior pancreatic surgery—partial or distal pancreatectomy in 4 (17%), lateral pancreaticojejunostomy (Puestow or Frey) in 5 (21%), and both partial pancreatectomy and lateral pancreaticojejunostomy in 2 (8%). At the time of presurgical admission, 6 patients (25%) were receiving total parenteral nutrition because of an inability to tolerate oral intake. Among the 18 patients who were not dependent on total parenteral nutrition, 11 (61%) were being treated with oral pancreatic enzymes, in the hope that feedback inhibition of the pancreas would result in resolution of the pain. These measures were judged ineffective, prompting surgical consultation as a candidate for pancreatectomy.
For most patients, the primary indication for pancreatectomy was persistent pain unresponsive to conservative measures, including endoscopic retrograde cholangiopancreatography with stenting or sphincterotomy. However, 2 patients had a documented impairment in the acute insulin response to glucose on intravenous glucose tolerance testing, which contributed to the decision to proceed to pancreatectomy in an attempt to prevent further injury to their islet cells and to avoid progression to diabetes. Only 1 subject had diabetes before pancreatectomy. Preoperative HbA1c levels were documented for 8 patients, with a median of 5.0% (range 4.2%–6.0%).
Islet yields were quantified in terms of islet equivalents (IE), in which the islet mass transplanted is standardized to an islet size of 150 μm in diameter (17). After pancreatectomy, patients received a mean of 4392 IE/kg body weight (range 25–12,576 IE/kg). The patient with the lowest islet yield (25 IE/kg) was the 1 patient in our series with a diagnosis of diabetes preoperatively. Each patient had all or most of their islets transplanted intraportally.
The median length of postoperative stay was 17.5 days (range 8–89 days). Postoperatively, 13 patients (54%) had no complications, 6 patients (25%) had 1 complication, and 5 (21%) had multiple complications, including urinary tract infection (4/24), bacteremia or line infection (3/24), fungemia (3/24), intraabdominal abscess (3/24), pneumonia (2/24), deep venous thrombosis (1/24), wound dehiscence (1/24), omental infarction (1/24), and Clostridium difficile infection (1/24). In 1 case, the islet culture grew Escherichia coli, but this was without clinical sequelae. Only 3 patients (13%) had 1 or more major complications requiring procedural or surgical intervention. In all 3 patients, an intraabdominal abscess required percutaneous drainage or reoperation. Additionally, 1 of these patients required omental resection; another required multiple surgeries related to intraabdominal infection, including 3 incision and drainage procedures, externalization of a ventroperitoneal shunt, and subsequent shunt revisions. No perioperative deaths occurred.
Six patients (25%) were lost to follow-up, all 3 to 8 years out from surgery, including the patient with preoperative diabetes who also had the lowest islet yield. We obtained follow-up data for the other 18 patients, including direct telephone contact with 17 patients or parents and medical record information on 1 additional patient (Table 1). Of these 18 patients, 2 were reported to have died: 1 from narcotic overdose and the other from postsplenectomy sepsis (splenectomy had been performed before referral for pancreatectomy); for both, we obtained medical history at the time of death. The median time since pancreatectomy and IAT at follow-up was 2.5 years (range 0.2–7.5 years, except for 1 outlier at 17.1 years). The median age at follow-up was 15.9 years.
Gastrointestinal and Nutritional Outcome
Of the 15 living patients available for the telephone survey, 14 took pancreatic enzymes on a regular basis, 13 of whom felt the enzymes were adequate in controlling symptoms of malabsorption. Gastrointestinal symptoms in these patients did not appear to adversely affect weight gain; the median weight for age z score (SD) was 1.25 both preoperatively and at follow-up.
Of the 18 patients, 11 (61%) had discontinued all of the narcotic pain medications at follow-up or time of death. Of the 7 patients taking chronic narcotics, 5 were taking traditional narcotics (methadone, fentanyl, or hydromorphone), and 2 were taking relatively mild narcotics, tramadol (a mild central mu-opioid receptor agonist) in 1 case and buprenorphine (an opioid agonist-antagonist) in the other.
Of the 15 living patients available for the telephone survey, 10 (67%) reported no pain symptoms (1 of whom was taking narcotics); 4 (27%) reported pain improvement as compared with preoperatively (3 taking narcotics); and only 1 reported pain similar in severity to preoperative pain (taking narcotics). The last patient, 17 years after surgery, reported a pain-free interval of 12 to 14 years before the recurrence of pain symptoms; because this patent underwent a total pancreatectomy, the cause of his current pain is unclear. Of the 15 patients, 12 (80%) were able to attend school or work, and 11 (73%) reported their overall quality of life as excellent or good. Factors identified as interfering with work or school attendance or as impairing quality of life varied but included persistent pain or other medical or psychosocial issues unrelated to CP.
We had follow-up data on insulin requirements for 14 patients at 1 year or more after transplant. Seventy-eight percent had partial or full graft function at 1 year posttransplant, and 56% were insulin independent at 1 year posttransplant.
Graft function at the time of follow-up is summarized in Table 1. Some patients exhibited a change in graft function over time. In the graft failure group, 2 patients had originally achieved insulin independence, but then required initiation of insulin treatment at about 1 and 1.5 years posttransplant. Graft failure was precipitated by a major illness in 1 patient; in the other, graft function progressively declined as the patient grew and progressed through puberty. Another 2 patients with partial graft function initially were insulin independent for several months posttransplant.
Good metabolic control in patients with full or partial graft function was evidenced by self-reported normal or near-normal HbA1c levels and blood glucose profiles. We were able to verify this self-assessment in the 8 patients who mailed us HbA1c samples and blood glucose records. (The others did not because they were too busy or did not want to do more fingersticks.) HbA1c levels were normal (<6.0%) in the 5 patients with full graft function. In 2 patients with partial graft function, HbA1c levels were 5.1% and 6.3%; the latter value was from a patient who had not taken prescribed insulin glargine for several weeks beforehand. The single HbA1c level from a patient with graft failure was 12.5%, indicating that this young man's diabetes was controlled poorly. The 2-day mean fasting blood glucose levels in the patients with full and partial graft function ranged from 74 to 96 mg/dL and 89 to 103 mg/dL, respectively. The mean 2-day premeal, 2-hour postprandial, and bedtime blood glucose levels ranged from 93 to 119 mg/dL for patients with full graft function and 91 to 106 mg/dL for patients with partial graft function.
Predictors of Graft Function
In general, higher islet yields were associated with better graft function. The mean IE/kg was 7467 for patients with full graft function, 4066 for partial graft function, and 2890 for graft failure. There was considerable overlap in IE/kg between groups, and no single islet yield was fully predictive of graft function. However, patients with the lowest islet yields (<2000 IE/kg) were unlikely to achieve sustained insulin independence. Patients with islet yields >2000 IE/kg were significantly more likely to achieve insulin independence (P ≤ 0.049). All 5 of the patients with >6600 IE/kg had partial or full graft function.
A history of pancreatic surgery was inversely correlated with insulin independence. The likelihood of insulin independence at 1 year posttransplant was 67% in patients without a history of pancreatic surgery versus 33% in patients with prior pancreatic surgery (P ≤ 0.021). Our small sample of patients did not permit comparison of outcome by type of previous surgery (distal pancreatectomy vs lateral pancreaticojejunostomy). No patient in our series had undergone a Whipple procedure (pancreatic head resection).
We performed a Cox regression analysis for insulin independence, considering multiple potential predictor values including age category (preadolescent vs adolescent), surgical history, and islet yield. Likelihood of ever achieving insulin independence was correlated most strongly with a lack of prior pancreatic surgery and with an islet yield >2000 IE/kg (P = 0.011). In this model, the hazard ratio of insulin dependence was 5.0 (P = 0.019) for patients with prior pancreatic surgery and 0.264 (P = 0.057) for IE/kg >2000.
Outcomes in Preadolescent Versus Adolescent Patients
We compared outcomes in preadolescent (<13 years old) versus adolescent (13–18 years old) patients at the time of pancreatectomy and IAT (Table 2). Preadolescents were more likely to be free of narcotic pain medications and to have full or partial graft function at follow-up. Of 9 preadolescent patients, only 1 (11%) required chronic narcotics postoperatively vs 6 (67%) of 9 adolescent patients (P = 0.05). One (11%) preadolescent had islet graft failure versus 5 (56%) adolescents (P = 0.02). Preadolescents tended to have a shorter duration of disease (Fig. 1A) and a greater IE/kg yield (Fig. 1B), but these differences were not statistically significant. Fewer preadolescents had a history of pancreatic surgery, although this difference was not statistically significant.
When the pain of CP cannot be relieved by endoscopic or surgical drainage procedures, pancreatectomy may be indicated to provide pain relief, allow withdrawal from narcotics, and restore quality of life. Islet autotransplantation can prevent or reduce the severity of postoperative diabetes. To date, nearly all of the medical literature on pancreatectomy and IAT focuses on these advantages in adult recipients (5–8). This report is the first of a case series of IAT in the pediatric population. We show here that pancreatectomy with IAT can relieve pain in children with chronic pancreatitis and can prevent overt diabetes, especially when performed in the preadolescent period.
Pain relief is the primary indication for pancreatectomy. More than 90% of our study patients noted that their pain resolved or improved after pancreatectomy; 61% had discontinued all narcotic pain medications at follow-up, whereas the majority of those patients still taking narcotics reported better control of pain symptoms. This finding is comparable to adult series, in which 50% to 80% become narcotic independent (5–7,18). Adult patients with chronic pancreatitis are more likely to require reinitiation of narcotic therapy after pancreatectomy if narcotics were used on a daily basis for at least 3 months preoperatively (19).
In our present series, preadolescents were more likely than adolescents to have discontinued narcotic pain medications at follow-up, perhaps because preadolescents were simply treated earlier, with a shorter duration of the primary disease or of narcotics use preoperatively. In addition to addiction, prolonged narcotics use can result in an opiate-induced hyperalgesia, characterized by a decreased threshold for pain, in turn perpetuating narcotics use (20,21). This may explain why some patients in our series continue taking narcotic pain medications despite resection of the pancreas.
At 1 year after IAT, insulin independence was sustained in 56% of our study patients and another 22% required only small doses of basal insulin. Those without a history of partial pancreatectomy or surgical drainage procedure (Puestow or Frey) were more likely to be insulin independent, likely related to a higher islet yield. Although no pediatric patients had a history of a prior Whipple operation, in adult case series, Whipple procedures have had less impact on metabolic outcome than distal pancreatectomy or lateral pancreaticojejunostomy, presumably because of relative sparing of the islets when the body/tail is both retained and not violated (a prior Puestow or Frey procedure interferes with the ability to cannulate the pancreatic duct for collagenase infusion and exocrine pancreatic disruption as part of the islet isolation process) (22). Patients who received a greater number of islets tended to have better graft function, a finding that is consistent with previous reports in adult patients (5,14,18). However, no specific islet yield was predictive of insulin independence, a finding that is possibly related to islet quality. The viability of the islets may be influenced by the original disease state; in addition, the ability of islets to engraft may be determined by the implantation site (portal vein vs portal with additional site). Furthermore, patient-related factors such as insulin resistance may play a role. This latter factor may be particularly relevant in adolescents.
In our series, islet autotransplants were more successful in preadolescents than in adolescents. There are many potential reasons for this finding. Fewer preadolescents had a history of partial pancreatectomy or surgical drainage procedures, which we found to be associated with a lower likelihood of successful IAT, presumably because of surgical loss of healthy islets. Preadolescents, for the most part, received higher islet yields and may have had less inflammatory damage to their pancreatic β cells. However, even adolescents with a short duration of preoperative disease were less likely to have a favorable clinical course, suggesting that the success of treatment at an earlier age is not attributable simply to less time for damage to islet cells. In fact, preadolescents with lower islet yields had better metabolic outcomes than adolescents with a greater number of recovered islets. This suggests that islet autografts in adolescents may be subject to poorer engraftment, increased loss, or decreased function, perhaps due to the metabolic demands of the increased insulin resistance that is associated with normal puberty. Further prospective studies are warranted in preadolescent and adolescent patients to better assess their metabolic outcomes.
Our study has several limitations. These data represent a case series with a relatively small number of patients. Without a control group of pediatric patients similarly affected with severe chronic pancreatitis who did not undergo surgery, we are unable to predict what proportion of patients, if any, would have experienced spontaneous resolution of pain over time. Because our patients are referred from around the country, we were unable to reach them all, so ascertainment of outcomes was incomplete. Only about half of the patients were willing to send HbA1c samples or blood glucose records. Detectable c-peptide levels, typically used in islet allograft recipients to confirm residual β cell function, were not documented in the patients we classified as partial graft function. However, these 4 patients were receiving minimal insulin doses (0.09–0.18 U · kg−1 · day−1) with normal blood glucoses and normal or near-normal HbA1c levels documented in 2 patients and self-reported normal or near-normal blood glucoses and/or HbA1c levels in the remaining 2 patients (data not presented). Duration of follow-up for all except 1 patient was ≤7 years, and many patients have not yet reached adulthood. Longer follow-up of these patients will help determine whether their graft function changes over time.
In conclusion, in most of our study patients with CP refractory to medical treatment or to lesser surgical procedures, pancreatectomy alleviated pain symptoms, and IAT prevented or lessened the severity of postoperative diabetes in most. Preadolescent IAT recipients were more likely to sustain insulin independence, perhaps because of less prior surgical intervention, less chronic damage to the pancreas, and the protective effect of a less-insulin-resistant metabolic milieu. Pancreatectomy with IAT should be considered early in the surgical care of pediatric patients with intractable pain from CP.
1. Ahmad SA, Wray C, Rilo HL, et al
. Chronic pancreatitis: recent advances and ongoing challenges. Curr Probl Surg 2006; 43:127–238.
2. Carlson AM, Kobayashi T, Sutherland DER. Islet autotransplantation to prevent or minimize diabetes after pancreatectomy. Curr Opin Organ Transplant 2007; 12:82–88.
3. Robertson GS, Dennison AR, Johnson PR, et al
. A review of pancreatic islet autotransplantation. Hepatogastroenterology 1998; 45:226–235.
4. Sutherland DE, Matas AJ, Najarian JS. Pancreatic islet cell transplantation. Surg Clin North Am 1978; 58:365–382.
5. Ahmad SA, Lowy AM, Wray CJ, et al
. Factors associated with insulin and narcotic independence after islet autotransplantation in patients with severe chronic pancreatitis. J Am Coll Surg 2005; 201:680–687.
6. Clayton HA, Davies JE, Pollard CA, et al
. Pancreatectomy with islet autotransplantation for the treatment of severe chronic pancreatitis: the first 40 patients at the Leicester General Hospital. Transplantation 2003; 76:92–98.
7. Rodriguez Rilo HL, Ahmad SA, D'Alessio D, et al
. Total pancreatectomy and autologous islet cell transplantation as a means to treat severe chronic pancreatitis. J Gastrointest Surg 2003; 7:978–989.
8. Sutherland DER, Gruessner RWG, Tun J, et al
. Pancreatic islet autotransplantation for chronic pancreatitis. Clin Transplant 2004; 18(Suppl 13):17.
9. Farney AC, Hering BJ, Nelson L, et al
. No late failures of intraportal human islet autografts beyond 2 years. Transplant Proc 1998; 30:420.
10. Robertson RP, Lanz KJ, Sutherland DE, et al
. Prevention of diabetes for up to 13 years by autoislet transplantation after pancreatectomy for chronic pancreatitis. Diabetes 2001; 50:47–50.
11. Lowe ME. Pancreatitis in childhood. Curr Gastroenterol Rep 2004; 6:240–246.
12. Ammann RW. Diagnosis and management of chronic pancreatitis: current knowledge. Swiss Med Wkly 2006; 136:166–174.
13. Wahoff DC, Papalois BE, Najarian JS, et al
. Islet autotransplantation after total pancreatectomy in a child. J Pediatr Surg 1996; 31:132–136.
14. Farney AC, Najarian JS, Nakhleh RE, et al
. Autotransplantation of dispersed pancreatic islet tissue combined with total or near-total pancreatectomy for treatment of chronic pancreatitis. Surgery 1991; 110:427–439.
15. Lakey JR, Warnock GL, Shapiro AM, et al
. Intraductal collagenase delivery into the human pancreas using syringe loading or controlled perfusion. Cell Transplant 1999; 8:285–292.
16. Ricordi C, Lacy PE, Scharp DW. Automated islet isolation from human pancreas. Diabetes 1989; 38(Suppl 1):140–142.
17. Ricordi C, Gray DWR, Hering BJ, et al
. Islet isolation assessment in man and large animals. Acta Diabetol Lat 1990; 27:185–195.
18. Wahoff DC, Papalois BE, Najarian JS, et al
. Autologous islet transplantation to prevent diabetes after pancreatic resection. Ann Surg 1995; 222:562–579.
19. Alexakis N, Connor S, Ghaneh P, et al
. Influence of opioid use on surgical and long-term outcome after resection for chronic pancreatitis. Surgery 2004; 136:600–608.
20. Angst MS, Clark JD. Opioid-induced hyperalgesia: a qualitative systematic review. Anesthesiology 2006; 104:570–587.
21. Chu LF, Clark DJ, Angst MS. Opioid tolerance and hyperalgesia in chronic pain patients after one month of oral morphine therapy: a preliminary prospective study. J Pain 2006; 7:43–48.
22. Gruessner RW, Sutherland DE, Dunn DL, et al
. Transplant options for patients undergoing total pancreatectomy for chronic pancreatitis. J Am Coll Surg 2004; 198:559–569.
© 2008 Lippincott Williams & Wilkins, Inc.