Introduction

Biliary tumors have often caused obstructive jaundice by the time of presentation, and the relationship between jaundice and operative risk is currently under discussion.1,2 The value of preoperative biliary drainage in jaundiced patients is controversial. Early studies showed decreased mortality and morbidity rates in patients with biliary drainage, but recent prospective studies have not confirmed any clear benefit of preoperative biliary drainage.3–6 In Western countries, biliary drainage for major hepatic resection is thought to be generally required, but pancreatoduodenectomy (PD) is suggested to need biliary drainage only in selected cases.2,7 In contrast, biliary drainage is routinely performed in Japan for patients with obstructive jaundice regardless of planned PD or hepatectomy.8,9

There are two different types of biliary drainage (BD): external (EBD) being percutaneous transhepatic biliary drainage (PTBD) and endoscopic nasobiliary drainage (ENBD), and internal (IBD) being endoscopic retrograde biliary drainage (ERBD). Although quality of life of patients treated with IBD appears superior to that with EBD, IBD (ERBD) is not always adequate because it quite often leads to infection (cholangitis).10,11 For bile duct cancers (especially hilar cholangiocarcinoma), EBD was recently recommended as the most suitable method for preoperative biliary drainage in jaundiced patients who are candidates for surgical resection in Asian countries,11–13 although this strategy has not been adopted worldwide.

ENBD is an external drainage method, and the absence of intestinal bile is associated with certain problems. Our experimental studies indicated that IBD improved bacterial translocation in the gut immune system via reversal of chemokine expression.14,15 In addition, IBD is superior to EBD with respect to postoperative hepatic regeneration in experimental rat models.16,17 In humans, the value of preoperative bile replacement is demonstrated mainly by its impact on the impaired intestinal barrier function.18

Thus, when EBD is performed, bile replacement is possibly a better option for the preoperative management of jaundiced patients. However, to the best of our knowledge, the literature on bile replacement in humans is scarce and detailed changes of immune function are not clear to date. In the present study, we prospectively investigated the effect of bile replacement on immune functions over the time course of bile replacement in jaundiced patients.

Methods

Patients and Protocols

This study was a prospective clinical trial at our institute. The inclusion criteria were age between 18 and 80 years in good general condition (Eastern Cooperative Oncology Group (ECOG) performance status 0–1) with obstructive jaundice (T-bil >5.0, D-bil >3.0 mg/dl). Written informed consent for participation was obtained from each patient before enrollment. This study was approved by the institutional ethics committee of the Kobe University Hospital (No. 1194, UMIN ID 000006645).

Between October 2011 and January 2013, 15 consecutive patients with obstructive jaundice were enrolled. All patients received a regular diet preoperatively. None of them received preoperative parenteral or enteral nutritional supplementation. Eleven patients underwent ENBD, three underwent PTBD, and the remaining one patient underwent both ENBD and PTBD. Bile replacement was started soon after the total serum bilirubin concentration reached 5.0 mg/dl and was continued for 14 days. Drained bile was percolated through gauze and cooled before ingestion. Each patient received 100 ml of the drained bile orally two times per day (total 200 ml/day). Surgery was performed after the serum total bilirubin concentration reached 2.0 mg/dl. Time between starting initial biliary drainage and surgery was a median of 48 days (range, 28–122 days).

The patients were composed of 11 men and 4 women, with a mean age of 65 years (Table 1). Seven patients had bile duct cancer, six had pancreatic cancer, one had both, and the remaining patient had benign biliary stricture as a final diagnosis. One patient with pancreatic cancer did not complete the study because of rapid progression, and one patient was also excluded because of benign disease. Thus, finally we analyzed data from 13 patients.

Table 1 Patient characteristics
Full size table

Blood samples for standard laboratory tests were obtained in the morning just before beginning bile replacement (day 0), and days 1, 3, 5, 7, and 14 after bile replacement. White blood cell (WBC) counts, lymphocyte fraction, prothrombin time–international normalized ratio (PT-INR), albumin, C-reactive protein (CRP), blood urea nitrogen (BUN), creatinine, total bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and alkaline phosphatase (ALP) were quantified. Measurements of serum diamine oxidase (DAO) activity, as well as T cell responses to concanavalin A (Con A) and phytohemagglutinin (PHA), were performed on day 0 and day 3, 7, and 14 after bile replacement. Blood samples were centrifuged at 3,000g for 10 min at 4 °C and the sera were stored at −80 °C until measurement.

Cellular Immunity

Cellular immunity was examined by assessing T cell proliferative responses to the plant mitogens Con A and PHA, taken as indicators of cell-mediated immunity. Stimulation indices were calculated by dividing the counts per minute (cpm) of nuclear3 H-thymidine incorporation in mitogen-stimulated cells by the cpm in cells cultured without mitogens.19

Serum DAO Activity

Serum DAO activity was used as a marker of intestinal integrity. It was determined using the modified method of Takagi et al. 20,21 In brief, a colorimetric assay based on a coupled reaction with peroxidase and the novel chromogen 10-(carboxymethyl-aminocarbonyl)-3,7-bis(dimethylamino) phenothiazine sodium salt (DA-67; Wako Pure Chemical Industries, Osaka, Japan) was performed. Previous reports showed that the serum activity averaged 4.9–6.25 U/l in healthy controls.20,22

Analysis of a Control Group of Patients Undergoing External Biliary Drainage, but No Bile Replacement

In order to clarify the role of bile replacement, we have retrospectively analyzed 20 patients with pancreatic or biliary cancers who had obstructive jaundice and were treated with external biliary drainage without bile replacement. The patients were composed of 13 men and 7 women, with a mean age of 68 years. After serum total bilirubin concentration fell below 5.0 mg/dl, we analyzed total bilirubin, AST and ALT, albumin, BUN, creatinine, and PT-INR levels at days 0, 3, 7, and 14. Con A and PHA responses on days 0 and 7 were available in 6 of these 20 patients.

Statistical Analysis

Results are expressed as mean ± standard deviation (SD). Statistical analysis was performed using Fisher’s exact test, Student’s t test, and the Wilcoxon signed rank test using the StatView program (Abacus Concepts Inc., Berkeley, CA), as appropriate. A P value <0.05 was considered statistically significant.

Results

Patients’ Characteristics

Total serum bilirubin concentration was 13.1 ± 5.9 mg/dl (range, 6.2–25.3) at the time of starting biliary drainage. The interval between initiating biliary drainage and beginning bile replacement was 12.5 ± 8.4 days (range, 3–34). No change in body weight was observed in any patient during this study. No moderate or severe cholangitis 23 occurred during this study and none of the patients required antibiotics. Bile culture was performed for 11 of 14 patients and was positive in 7 (63.6 %). The organisms cultured were Pseudomonas aeruginosa (n = 2), Enterococcus spp. (n = 2), and Bacillus cereus, Corynebacterium spp., and Aeromonas hydrophila (n = 1 of each).

Standard Laboratory Test Results

Table 2 shows changes in laboratory tests for all patients. Serum total bilirubin, AST, and ALT levels decreased over time. Serum albumin, BUN, and creatinine levels did not change significantly during bile replacement. PT-INR increased on the first day after starting bile replacement, then gradually decreased, and was significantly lower at day 14 than at day 0 (p = 0.022). The lymphocyte fraction did not change during bile replacement (p = 0.511) (Fig. 1).

Table 2 Changes in laboratory tests
Full size table
Fig. 1
figure 1

Changes of PT-INR (a) and lymphocyte % (b) over the duration of bile replacement. # P < 0.05 versus day 0

Full size image

Changes of Con A and PHA Responsiveness

T cell proliferative responses to Con A and PHA increased gradually over time from 19,318 ± 9,361 and 23,028 ± 11,604 before bile replacement to 25,028 ± 8,963 and 27,815 ± 7,042 cpm 2 weeks after bile replacement, respectively (p = 0.032, p = 0.050). Although only 4 patients had responses to these mitogens within the reference range at day 0, this increased to 8 by day 7 and 10 by day 14 after bile replacement (Fig. 2).

Fig. 2
figure 2

Correlations between Con A and PHA stimulation (days 0 (a), 3 (b), 7 (c), and 14 (d)). The reference ranges of Con A and PHA responses are from 20,300 to 65,700 and from 20,300 to 56,800 cpm, respectively

Full size image

Changes of Serum DAO Activity

Serum DAO activity before bile replacement (day 0) was 2.8 ± 1.0 U/l, tending to decrease to 3.0 ± 0.9 at day 3, 2.8 ± 0.9 at day 7, and again 2.5 ± 1.1 at day 14 after bile replacement (Fig. 3). None of these values were significantly different from the day 0 level.

Fig. 3
figure 3

Changes of serum DAO activity between day 0 and day 3 (a), day 0 and day 7 (b), and day 0 and day 14 (c). Bars show mean ± SD

Full size image

Influence of the Volume of Externally Drained Bile

The volume of externally drained bile was monitored daily. There was considerable inter-individual variation. The mean volume of drained bile for the whole patient cohort was 548 ± 170 ml (range, 314–854 ml). We calculated the bile replacement fraction (i.e., the fraction of the total drained bile represented by the 200 ml returned to each patient). The mean bile replacement fraction for all patients was 39.8 % (range 23.4–63.7 %). We stratified patients into two groups according to whether the volume of externally drained bile was more or less than the mean: Large group (drained bile ≥548 ml/day, fraction of bile replaced ≤39.8 %, five patients) and Small group (<548 ml/day, fraction of bile replaced >39.8 %, eight patients) (Fig. 4a). PT-INR in Small group patients had decreased significantly by day 14 (p = 0.014) but was not changed in the Large group patients during bile replacement (Fig. 4b). Con A responses tended to increase more after 14 days in Small than in Large group patients, but this difference failed to achieve statistical significance (p = 0.097, Fig. 4c).

Fig. 4
figure 4

Volume of externally drained bile of each patient (a, bar shows the mean). Changes of PT-INR (b), Con A (c), and PHA (d) from day 0 to day 14 are shown. Black circles and boxes are patients with ≥548 ml/day externally drained bile (Large group); white circles and boxes are patients with <548 ml/day (Small group). Bars show SD. # P < 0.05 versus day 0

Full size image

Influence of the Interval Between Starting Biliary Drainage and Bile Replacement

The interval between starting biliary drainage and beginning bile replacement was 12.5 ± 8.4 days (range, 3–34). We stratified patients into two groups according to the interval between the start of biliary drainage and starting bile replacement. These we designated Long group (≥14 days) and Short group (<14 days) (Fig. 5a). PT-INR in Long group patients decreased gradually relative to the Short group, but differences between them were not statistically significant (Fig. 5b). The PHA response after 14 days of bile replacement in Short group patients tended to increase more than in the Long group, but again, the difference did not quite achieve statistical significance (p = 0.060, Fig. 5d).

Fig. 5
figure 5

Duration of biliary drainage of each patient (a, bar shows the mean). Changes of PT-INR (b), Con A (c), and PHA (d) from day 0 to day 14 are shown. Black circles and boxes show patients receiving externally biliary drainage for ≥14 days (Long group), and white circles and boxes <14 days (Short group). Bars show SD

Full size image

Surgical Outcomes

Eight patients underwent tumor resection, but this was not possible in five due to liver metastasis in two and being too locally advanced in three. No surgical site infections were seen in the eight resected patients. No infectious complications were seen in the remaining five patients who did not undergo tumor resection.

Comparison of Patients With or Without Bile Replacement

In the control group, total bilirubin gradually decreased (p < 0.05), and AST and ALT tended to gradually decrease, whereas albumin, BUN, creatinine, and PT-INR levels did not change significantly (Fig. 6a). Also, in the control group, mean Con A and PHA responses did not increase between day 0 and day 7 (Con A, 18,500 and 19,816 cpm, respectively, p = 0.75; PHA, 23,250 and 20,941 cpm, p = 0.46) (Fig. 6b, c). These differences were minor compared to those of the patients who did receive bile replacement (day 0 and day 7, Con A, 19,318 and 22,439 cpm, p = 0.10; PHA, 23,028 and 25,362 cpm, p = 0.16).

Fig. 6
figure 6

Changes of PT-INR (a) and Con A and PHA responses (day 0 (b), day 7 (c)) of patients receiving external biliary drainage but no bile replacement

Full size image

Discussion

The current study showed that general immune function was improved by bile replacement in jaundiced patients with EBD, as reflected in Con A- and PHA-stimulated lymphocyte proliferation. This is the first prospective study to provide clinical data on general and intestinal immune function over the bile replacement period. Obstructive jaundice is associated with T lymphocyte dysfunction,24 and Con A and PHA responsiveness is a marker of T lymphocyte function. Con A preferentially activates CD8+ T cells rather than CD4+ T cells, while PHA preferentially activates the latter.25,26 Suzuki et al. proposed that Con A and PHA were useful for evaluating postoperative immunosuppression in patients undergoing major surgical resection.19 Christou et al. also reported that the suppression of cell-mediated immunity influences postoperative complications.27 In the present study, the fraction of lymphocytes did not change during bile replacement, but Con A- and PHA-stimulated lymphocyte proliferation in most patients with biliary drainage tended to be improved at 7 days and was significantly improved at 14 days by bile replacement. In contrast, patients without bile replacement did not show obvious changes of Con A and PHA responses during 7 days of external biliary drainage. These findings are in line with the concept that bile replacement can improve immune function via improvement of T cell function. Our previous studies demonstrated that changes of the number of gut mucosal T lymphocytes were associated with the presence of bile in the gut in rats and humans.14,28 In addition, another previous clinical study showed that preoperative immunonutrition modulated both Con A- and PHA-stimulated lymphocyte proliferation and that this may protect against the aggravation of surgical complications in patients undergoing pancreatoduodenectomy.29 Collectively, examining T lymphocyte function using data on both Con A and PHA responsiveness may predict occurrence of postoperative complications.

Here, we have also investigated DAO activity to evaluate changes of intestinal immune function associated with bile replacement in these patients. However, DAO activity did not uniformly improve within 14 days. A previous report by Kamiya et al. indicated that DAO activity gradually increased over the duration of bile replacement.18 However, they also reported that DAO activity had not increased in about half of the patients within 14 days. In addition, in the present study, many jaundiced patients had a low level of DAO activity compared to healthy subjects, and almost all patients failed to reach normal levels in spite of bile replacement for 14 days. These results indicate that intestinal immune function in patients with biliary obstruction tends to deteriorate in spite of biliary drainage. DAO activity is reported to be influenced by intestinal stresses such as surgery, chemotherapy, or nutritional management.21 Because the duration of our study was limited to 14 days, longer-term observation of DAO activity may be required to accurately establish the relationship between DAO activity and bile replacement.

Preoperative bile replacement is recognized as a useful method benefitting impaired immune function. However, the optimal duration or appropriate volume of bile intake remains unclear. The volume of bile varied widely from patient to patient in the present study. Bile is generated to the amount of about 250–1,000 ml in humans.30 We found that the mean volume of drained bile in the patients reported here was 548 ml. We stratified the patients into those above and those below the mean to examine the influence of the replacement volume of drained bile. When we set the amount of bile replacement as 200 ml per day, a volume which we chose in order to make intake as easy as possible for the patients, those with a high proportion of the volume of drained bile being replaced showed better recovery of PT-INR and Con A responsiveness compared to those with a lower fraction of bile being replaced. This suggested benefits when a large fraction of drained bile is replaced, but this could be difficult in patients without a gastric or jejunal tube. From this point of view, internal biliary drainage may be the preferred choice of preoperative biliary drainage for jaundiced patients despite the risk of infection.

We started bile replacement when total serum bilirubin reached 5.0 mg/dl because our study focused on investigating changes of immune function during bile replacement.14,18,31 There are several reports of immune dysfunction in obstructive jaundice in both experimental models and humans. Obstructive jaundice leads to increased concentrations of cytokines such as IL-6 and tumor necrosis factor,31 and depressed natural killer activity of hepatic cells in experimental models which can be reversed by relief of the biliary obstruction.32 In addition, intestinal barrier function is impaired in obstructive jaundice, which can be reversed by returning bile into the gut.33 Kamiya et al. reported that bile replacement may be started as soon as possible after beginning external biliary drainage, or after hospital admission.18 Their report showed that biliary concentrations of bile acids and phospholipids were increased by bile replacement. Taking the data together, although we started bile replacement when total bilirubin fell below 5.0 mg/dl, it may be clinically better to begin bile replacement as soon as possible after starting external biliary drainage.

Concerning the relationship between the duration of biliary drainage and the recovery of immune function on bile replacement, we found that short-term biliary drainage tended to be better than long-term, but this may mainly reflect liver function that is compromised by biliary obstruction. A previous paper reported that hepatic impairment is associated with decreased expression of multidrug resistance protein 2 (MRP2), a bilirubin transporter, which is improved by biliary drainage.34 The decrease of MRP2 expression observed in jaundiced patients requiring long-term biliary drainage influences the occurrence of postoperative liver failure.35 Bile replacement may recover hepatic function to some degree, consistent with our findings here that liver enzymes gradually decreased over the 14-day course of bile replacement.

Our study has several limitations. First, the duration of bile replacement may have been too short. Kamiya et al. reported that a mean length of bile replacement of 20.7 days was associated with improved DAO activity.18 Nonetheless, none of the patients in the present study suffered any surgical site infection, suggesting that 14 days of bile replacement may have improved general immune function to a degree sufficient to decrease infectious complications. However, longer-duration bile replacement could have increased immune function still more markedly. Second, we could measure Con A and PHA responses after 7 days but not 14 days in the control group (no bile replacement). The data for Con A and PHA responses at 14 days were not available in the control group of the current study. Therefore, further studies assessing the immunological data over a longer duration in patients with biliary drainage and no bile replacement may be required. Third, the numbers of patients in our study were small, as it was a prospective study of limited duration. Further work is still required to confirm the value of bile replacement.

In conclusion, EBD plus bile replacement for 14 days was able to improve liver and general immune function. However, intestinal immune function could not be improved even by bile replacement of 14 days. Although further randomized studies are required, we propose that preoperative bile replacement for patients with EBD should be performed in patients undergoing major surgery, or in high-risk patients, in order to improve general immune function.