Gastric Phase 3 Motility After Pylorus-Preserving Pancreatoduodenectomy

Abstract

Objective

To analyze factors affecting the recovery course of phase 3 activity of the gastric migrating motor complex after pylorus-preserving pancreatoduodenectomy (PPPD) and investigate effects of the recovery of gastric phase 3 on gastric emptying after feeding.

Summary Background Data

Whether early recovery of gastric phase 3 during fasting would predict early recovery of the fed-state gastric emptying function after PPPD has not been well documented.

Methods

Manometric recording from the gastric antrum was repeated at a weekly interval until the first appearance of gastric phase 3 in 57 patients after PPPD. Twenty-three clinical parameters were assessed as possible factors affecting the recovery course of gastric phase 3 by simple and multiple regression analyses. A gastric emptying study after feeding of a test meal was performed by the acetaminophen method and the values were compared between patients with and without gastric phase 3 after PPPD.

Results

The mean period before the first appearance of gastric phase 3 was 38 days. Among 23 parameters, only lymph node dissection along the hepatoduodenal ligament significantly delayed recovery of gastric phase 3 after PPPD by univariate and multivariate analyses. The presence or absence of gastric phase 3 in the early postoperative period did not influence gastric emptying after feeding in the intermediate period after PPPD.

Conclusions

Avoiding lymph node dissection along the hepatoduodenal ligament, if applicable, may contribute to early recovery of gastric phase 3 after PPPD. The recovery state of gastric phase 3 during fasting, however, is not necessarily consistent with the degree of improvement of gastric emptying after feeding.

During fasting, the upper gastrointestinal tract of most mammals, including humans, exhibits a cyclic change in motor activity called the migrating motor complex (MMC). ¹ Phase 3 is the most characteristic activity of the MMC, during which the upper gastrointestinal tract shows strong regular contractions of maximal frequency and amplitude. ¹ The occurrence of gastroduodenal phase 3 is associated with peaks in the plasma motilin concentration. ² Tanaka and Sarr ³ showed that resection of the canine duodenum as the principal origin of motilin secretion abolished the cyclic increase in the plasma motilin concentration and resulted in a loss of gastric phase 3.

It is well known that abdominal surgery temporarily abolishes gastrointestinal phase 3, and the period until recovery of phase 3 depends on the type of operation. ^4,5 Some previous reports showed that although gastrointestinal phase 3 returned immediately after cholecystectomy, colectomy impaired phase 3 for 4 to 6 days. ^4,5 On the other hand, we previously reported that the mean period until recovery of gastric phase 3 after pylorus-preserving pancreatoduodenectomy (PPPD) was 30 to 40 days. ^6–8 One possible explanation for the delayed recovery of gastric phase 3 after PPPD is considered to be the low plasma motilin concentrations caused by resection of the duodenum. ^3,6 However, other causes except for motilin influencing delayed recovery of gastric phase 3 after PPPD have not been well documented.

The gastrointestinal motor pattern in the fed state is completely different from that during fasting. ⁹ The gastrointestinal MMC disappears immediately after feeding, and the fed pattern, characterized by intermittent, irregular contractions like phase 2, starts and lasts for several hours. ⁹ Although recovery of phase 3 after abdominal surgery is often used to evaluate postoperative gastrointestinal activity, ^4,5 whether recovery of the fasting gastrointestinal motor pattern would also predict recovery of fed-state motor activity has not been documented. Even in the field of PPPD, there have been no reports describing the relationship between the recovery of gastric phase 3 during fasting and fed-state gastric motor activity. The recovery course of gastric phase 3 varies from 3 weeks to more than 2 months after PPPD. ^6–8 Early recovery of gastric phase 3 might also predict early recovery of the fed-state gastric emptying function after PPPD; however, this hypothesis has not been confirmed. The aims of this study were to analyze the factors influencing the recovery course of gastric phase 3 after PPPD by univariate and multivariate analyses, and to compare gastric emptying measured by the acetaminophen method after feeding of a test meal between the patients with gastric phase 3 and those without phase 3.

PATIENTS AND METHODS

Patients

This series comprises 57 Japanese patients who underwent PPPD at the Department of Surgery 1, Kyushu University Hospital, from January 1992 through August 2000. They consisted of 31 men and 26 women, with a mean age of 61.0 years (range 33–81). The first 28 patients have been described in our previous articles on manometric and emptying studies of gastric motility after PPPD. ^6–8,10 Patients who had major postoperative complications such as anastomotic leakage, abscess, bleeding, or cardiopulmonary diseases were excluded. All the patients underwent preoperative endoscopic retrograde cholangiopancreatography. Each patient was fully informed of the aims and methods of this study and gave written consent.

PPPD was performed for 35 malignant and 22 benign periampullary diseases; 16 patients with pancreatic carcinoma, 10 with ampullary carcinoma, 6 with bile duct carcinoma, 3 with intraductal papillary adenocarcinoma, 15 with intraductal papillary adenoma, 5 with chronic pancreatitis, and 2 with benign endocrine tumor. Seventeen of the 35 patients with malignant diseases had metastasis to regional lymph nodes.

Manometric Apparatus

Four polyethylene tubes were bundled for gastrointestinal pressure recording as previously reported. ^6–8 The outer diameter of each tube was 2.0 mm; the inner diameter 1.4 mm. A 0.5-mm side hole was drilled just before the end sealed with a radiopaque material, thus enabling us to confirm the correct position of the recording side hole by radiography. The side holes of the tube assembly were positioned 0.5, 20, 35, and 38 cm from the tip.

Surgical Procedures

The proximal duodenum was transected 2 to 6 cm distal to the pylorus, preserving (16 patients) or dividing (41 patients) the right gastric artery. Lymph nodes along the hepatoduodenal ligament ¹¹ were dissected in 39 patients. We performed end-to-end duodenojejunostomy as the first choice from 1992 to date. We consecutively performed end-to-side duodenojejunostomy from January 1998 through May 1999 to compare the postoperative clinical course between the two procedures (unpublished data). There were 46 patients with end-to-end duodenojejunostomy and 11 with end-to-side. All the patients except for one with total pancreatectomy underwent end-to-side pancreaticojejunostomy. Intraoperative radiotherapy (15–25 Gy) was performed in 16 patients. The manometric tube assembly was inserted into the stomach via a small incision on the anterior wall and down into the gastric antrum and jejunum, and the recording side holes were positioned in the gastric antrum (two holes, 3 cm apart) and jejunum (two holes, 34.5 and 15 cm apart from the distal gastric hole), as described in our previous reports. ^6–8 In the patients with end-to-side duodenojejunostomy, the tube assembly was inserted into the afferent jejunum through the stomach. The jejunal tube was also used for enteral feeding when necessary. A gastrostomy tube was routinely inserted into the gastric fundus through the anterior gastric wall for drainage of the gastric juice, negating the need for a nasogastric tube after surgery.

Manometric Recording

Recording of gastrointestinal motility for 3 hours a day was begun 7 to 14 days after surgery and repeated at a weekly interval until the first appearance of gastric phase 3. Figure 1 shows manometric recordings obtained in one patient on postoperative day 21. Phase 3 was identified as regular and high-amplitude contractions lasting at least 3 minutes at maximal frequency (3 per minute in the stomach and 11 or 12 per minute in the jejunum). ¹² The use of H₂ receptor antagonists and gastrokinetic drugs was avoided from 2 days before recording. Neither morphine nor epidural anesthesia was administered. The patients lay in bed in the supine position. The manometric tubes were infused with sterile distilled water at a rate of 0.25 mL/min using a low-compliance pneumohydraulic capillary infusion system connected to external pressure transducers (DX-360, Spectramed Medical Products, Singapore). The changes in pressure were recorded using a thermal-array multichannel recorder (RTA-1200 mol/L; Nihon Kohden Co., Tokyo, Japan) and Macintosh software (Mac Lab Chart 3.5.2, AD Instruments, Castle Hill, Australia) at a count speed of 40 samples/min. Zero reference was set by placing the side holes of an identical tube assembly connected to the recording system at the level of the patient’s midaxillary line at atmospheric pressure. The venous blood glucose level was checked to ensure that it was less than 150 mg/dL before the test. Manometric tracings from the stomach and jejunum were analyzed by visual inspection.

Figure 1. Manometric recordings from the gastric antrum (S-1, S-2) and jejunum (J-1, J-2) in a patient after pylorus-preserving pancreatoduodenectomy with reconstruction by end-to-end duodenojejunostomy obtained on postoperative day 21. Closed and open circles indicate gastric and jejunal phase 3, respectively. There seems to be no temporal association between gastric and jejunal migrating motor complexes. One jejunal phase 3 (second phase 3 in J-1) did not propagate distally.

Analysis of Factors Affecting Delayed Recovery of Gastric Phase 3

The following 23 clinical parameters, 12 preoperative and 11 perioperative, were assessed as factors possibly affecting the recovery course of gastric phase 3. Preoperative parameters were patient age (younger than 60 years, or 60 or older), sex, presence or absence of jaundice, internal biliary drainage by endoscopic stenting or external biliary drainage through a percutaneous transhepatic tube, diabetes mellitus, hemoglobin concentration, serum albumin level, blood urea nitrogen level, total bilirubin level, pancreatic exocrine function, and American Society of Anesthesiologists physical status score (I vs. II and III). Perioperative parameters included operative time (≤540 minutes or >540), blood loss during operation (≤1,500 mL or >1,500), perioperative blood transfusion, division of the right gastric artery, lymph node dissection, length of the preserved duodenum (<4 cm or ≥4), intraoperative radiotherapy, type of gut reconstruction, pathologic results of resected specimens (malignant or benign), histopathologic diagnosis, and presence or absence of lymph node metastasis. Normal ranges of hemoglobin were 14 g/dL or more in men and 12 g/dL or more in women, albumin 4.0 to 5.0 g/dL, blood urea nitrogen less than 20 IU/L, and total bilirubin less than 1.5 mg/dL in our institution. The N-benzoyl-l-tyrosyl-p-aminobenzoic acid excretion test, with a normal value of 70% or more, was used to assess the pancreatic exocrine function. This test was described in our previous report. ¹³ The diagnosis of diabetes mellitus was made based on the criteria set by the 1985 World Health Organization study group on diabetes mellitus by the use of the oral glucose tolerance test. ¹⁴

Gastric Emptying Study

Gastric emptying was measured by the acetaminophen method, as described in the previous report, ¹⁰ before surgery and 1 month after surgery. Patients received no medication known to influence gastrointestinal motility for at least 2 days before the examination. All patients remained seated upright during the study. After an overnight fast, each patient ingested 20 mg/kg acetaminophen (Kalonal; Showa Yakuhin Co., Ltd., Tokyo, Japan) dissolved in 5 mL/kg of a nutrient pasty liquid test meal (Ensure Liquid; Dinabot Co., Ltd., Osaka, Japan) within 2 minutes. The test meal has a caloric value of 1 kcal/mL and contains 3.5% protein, 3.5% fat, and 13.6% carbohydrate. Blood was sampled through an indwelling catheter inserted into an antecubital vein before and 15, 30, 45, 60, 75, 90, 120, 150, and 180 minutes after the administration of the test meal. All the plasma samples were stored at −80°C until assay. The concentrations of unconjugated acetaminophen in the plasma were measured by high-performance liquid chromatography. The venous blood glucose level was checked to ensure that it was less than 150 mg/dL before the test.

The following parameters were calculated from the data: area under the curve of the acetaminophen concentration against time until 60 minutes (AUC₆₀) and 90 minutes (AUC₉₀) after the meal, maximum plasma acetaminophen concentration (C_max), and time to reach the maximum acetaminophen concentration (T_max). These parameters were compared between patients with and without gastric phase 3 to investigate the effect of the recovery of gastric phase 3 on fed-state gastric emptying. Twenty-five healthy male volunteers (mean age 31.2 years, mean body weight 64.4 kg) without any digestive symptoms were recruited as a reference group.

Statistical Analysis

Values were expressed as a mean ± standard deviation. The difference between two groups was analyzed using the unpaired t test. A univariate analysis regarding the recovery course of gastric phase 3 was performed using simple regression analysis. A stepwise multiple regression analysis was used to determine the effect of 23 factors on delayed recovery of gastric phase 3. P < .05 was considered statistically significant.

RESULTS

Manometric Recording

The mean period until the first appearance of gastric phase 3 was 38.5 ± 15.3 days (range 7–85). Jejunal phase 3 occurred at the time of the first recording (7–14 days) in 54 patients and the second recording (14–21 days) in the other 3. There seemed to be no temporal association between the gastric and jejunal MMC in any patients. Two patients (4%) started oral intake after the recovery of gastric phase 3, but the other 55 (96%) could resume oral intake before the recovery of gastric phase 3.

Analysis of Factors Affecting Recovery Course of Gastric Phase 3

Of the 12 preoperative factors, none was related with recovery of gastric phase 3 (Table 1). Of the 11 perioperative factors, only dissection of the lymph nodes along the hepatoduodenal ligament significantly delayed recovery of gastric phase 3 after PPPD (P = .007) (Table 2). The stepwise multiple regression analysis of the 23 factors also revealed that lymph node dissection along the hepatoduodenal ligament significantly led to delayed recovery of gastric phase 3 (regression coefficient = 9.595, standard error = 3.692, F = 6.754, P = .01).

Table 1. COMPARISON OF PERIOD UNTIL THE FIRST APPEARANCE OF GASTRIC PHASE 3 AFTER SURGERY: 12 PREOPERATIVE FACTORS

* 39 patients did not undergo preoperative biliary drainage.

BT-PABA, N-benzoyl-l-tyrosyl-p-aminobenzoic acid; ASA, American Society of Anesthesiologists.

Table 2. COMPARISON OF PERIOD UNTIL THE FIRST APPEARANCE OF GASTRIC PHASE III AFTER SURGERY: 11 PERIOPERATIVE FACTORS

* Lymph nodes along the hepatoduodenal ligament. ¹¹

Gastric Emptying

The gastric emptying study by the acetaminophen method was performed in 25 patients. Gastric phase 3 was recovered in 14 of them but not in the other 11 at the time of the postoperative study. The preoperative values of AUC₆₀ and AUC₉₀ were significantly smaller in all the patients than in healthy controls (Table 3). C_max tended to be lower and T_max longer in the patients than in the controls, but the differences were not statistically significant. Compared with their own preoperative parameters, postoperative AUC₆₀ and AUC₉₀ in the 25 patients were significantly smaller, C_max was lower, and T_max was longer. There was no significant difference in all those parameters between those with phase 3 and those without phase 3 (Table 4).

Table 3. PARAMETERS OF GASTRIC EMPTYING IN HEALTHY CONTROLS AND PATIENTS

AUC₆₀, area under the curve of the acetaminophen concentration against time until 60 minutes; AUC₉₀, area under the curve of the acetaminophen concentration against time until 90 minutes; C_max, maximum plasma acetaminophen concentration; T_max, time to reach maximum acetaminophen concentration.

*P < .05 vs. healthy controls.

†P < .01 vs. preoperative value in same patients.

‡P < .05 vs. preoperative value in same patients.

Table 4. PARAMETERS OF GASTRIC EMPTYING IN PATIENTS WITH AND WITHOUT RECOVERY OF GASTRIC PHASE 3

AUC₆₀, area under the curve of the acetaminophen concentration against time until 60 minutes; AUC₉₀, area under the curve of the acetaminophen concentration against time until 90 minutes; C_max, maximum plasma acetaminophen concentration; T_max, time to reach maximum acetaminophen concentration.

DISCUSSION

The role of phase 3 activity during fasting is considered to clean up meal residue, basal secretions, and cellular debris from the gastrointestinal tract, thus the name of “housekeeper.”¹ Although gastric emptying literally means the ability of the stomach to propel out liquid and/or solid food distally, gastric phase 3 is responsible for gastric emptying during the immediate postoperative period before resumption of food intake. It would be ideal if we could also assess the emptying function of gastric phase 3 during fasting quantitatively, but this methodology has not been established. Thus, we evaluated gastric emptying only after feeding. There have been few reports describing the relationship between the recovery of fasted and fed-state gastric motility after surgery. Only two previous articles by Hotokezaka et al ^15,16 showed that the presence of gastric myoelectric dysrhythmias during fasting did not accurately predict fed-state gastric emptying after abdominal surgery in dogs. This is the first study to analyze factors affecting the recovery course of gastric phase 3 after PPPD and to elucidate whether early recovery of gastric phase 3 would reflect the early improvement of the fed-state gastric emptying function to the normal level in patients after PPPD.

Lymph node dissection along the hepatoduodenal ligament significantly delayed recovery of gastric phase 3 after PPPD. The most important factor contributing to the effect of lymph node dissection along the hepatoduodenal ligament on gastric motility is considered to be damage to vagal nerve branches. ¹⁰ Braasch et al. ¹⁷ and Williamson et al. ¹⁸ stated that damage to the vagal branches would cause transient impairment of gastric motility after PPPD. On the other hand, Shimada et al. ¹⁹ reported that preservation of pyloric branches of the vagal nerves did not influence gastric motility after PPPD. The present study supports the former statements. Damage to the vagal nerves is reported to induce gastric dysrhythmia, in particular tachygastria, which is considered to be associated with the impaired gastric MMC. ^20,21 We and others showed that gastric dysrhythmia was often observed in the early period after PPPD. ^10,22 Thus, we suggest that gastric dysrhythmia caused by damage to the vagal branches may play a role in the delayed recovery of the gastric phase 3 after PPPD. Avoiding lymph node dissection along the hepatoduodenal ligament whenever possible would contribute to early recovery of the gastric phase 3 after PPPD.

Prolonged gastric stasis, defined as the need for gastric decompression for more than 7 days after surgery, ^17,18 is a common complication after PPPD, occurring in 30% to 50% of the patients in the early postoperative period. It results in a prolonged loss of a large amount of gastric juice, a delay of food intake, and a prolonged hospital stay. ^23–25 We previously showed slower recovery of gastric phase 3 and lower plasma motilin concentrations in patients after PPPD than in those after duodenum-preserving pancreas head resection, and we speculated that prolonged gastric stasis after PPPD might be attributable at least in part to delayed recovery of the housekeeping function of gastric phase 3 as a result of the low plasma motilin concentrations caused by resection of the duodenum. ⁶ In fact, exogenous administrations of motilin and erythromycin as a motilin agonist produced phase 3-like gastric contractions and reduced gastric juice output in the early period after PPPD. ^7,8 Yeo et al ²⁶ also showed that exogenous erythromycin accelerated gastric emptying and reduced the incidence of prolonged gastric stasis after PPPD. On the other hand, resolution of gastric stasis after PPPD does not necessarily need recovery of gastric phase 3, because most of our patients (96%) after PPPD could resume oral intake before the recovery of gastric phase 3. We could not explain the discrepancy between these clinical points of view. Further investigations are necessary to elucidate the precise mechanism of prolonged gastric stasis after PPPD.

There are several types of gastric emptying studies, such as the radionuclide method, the ultrasonographic method, the acetaminophen absorption test, and the ¹³C acetate breath test; the radionuclide method is most frequently used. ²⁷ All these studies are designed as those after feeding of a test meal. ²⁷ We have used the acetaminophen method to examine the recovery course of gastric emptying after PPPD from 1992 to date because the radionuclide method has several technical problems, such as correction for attenuation, septal penetration, and down-scatter;²⁷ moreover, the repetitive use of the radioactive material is prohibitively expensive and impractical in an outpatient unit. The rationale of the use of acetaminophen for gastric emptying tests is that it is absorbed only from the small intestine. ²⁸ Some previous reports showed that the result of the acetaminophen study was correlated well with those of the radionuclide test and ultrasonographic test. ^28–30 We could use the acetaminophen method easily, safely, and repeatedly.

Gastric emptying after feeding has been reported to be delayed in patients of advanced age and those with diabetes mellitus, insufficient pancreatic exocrine function, and malignant diseases. ^20,31,32 One explanation for delayed gastric emptying measured by the acetaminophen method after feeding of a test meal in preoperative patients in the present study compared with that in healthy controls may be that the patients undergoing PPPD often have these problems.

Contrary to our expectations, we could not find any advantage of recovered gastric phase 3 on fed-state gastric emptying in this study. The rate of gastric emptying is considered to be determined by the degree of antroduodenal coordination and the contractile power of the stomach. ³³ Thus, one of the reasons for the lack of a positive relationship between recovery of gastric phase 3 and the gastric emptying function may be the absence of coordination of the gastric and jejunal MMCs caused by disturbed propagation across the duodenojejunostomy. In healthy humans, gastric phase 3 propagates to duodenum and then jejunum, which is important to transport the gastric content into the jejunum. The lack of this coordination might occur even after feeding. It has been reported that it can be more than 2 months before phase 3 begins to propagate across the anastomosis after partial resection of the canine jejunum. ³⁴ Another possible explanation is the reduced gastric contractile power. This is conceivable because the use of a gastrokinetic agent known to increase gastric contractile amplitude accelerates the impaired gastric emptying during the early period after PPPD. ¹⁰ The recovery of gastric phase 3 would not reflect the improvement of contractile power of the stomach after feeding, and gastric phase 3 itself might not have sufficient contractile amplitude compared with that in the normal condition.

We previously showed that acetaminophen gastric emptying after feeding was delayed in the early period after PPPD and needs 6 months to recover to the preoperative level. ¹⁰ If the patients in this series were observed for more than 6 months, the coordination of gastric and jejunal MMCs and contractile power of the stomach would return to normal. Although the early recovery of gastric phase 3 does not predict the early improvement of fed-state gastric emptying, this should be one of the important steps of the long-term recovery process of gastric motility after PPPD. Unfortunately, we could not determine the emptying function of gastric phase 3 itself during fasting after PPPD because of technical difficulty. Further studies are needed to clarify the clinical significance of the recovery of gastric phase 3 and the recovery process of gastric motility, including both the fasted and fed state after PPPD.

In conclusion, avoiding lymph node dissection in the hepatoduodenal ligament, if possible, may contribute to early recovery of the gastric phase 3 in patients after PPPD. However, early recovery of gastric phase 3 during fasting does not necessarily lead to early improvement of gastric emptying after feeding.

Acknowledgment

The authors thank Ms. Naoko Kinukawa, Department of Medical Informatics, Graduate School of Medical Sciences, Kyushu University, for her assistance in the statistical analysis of this study.