Br. J. Surg. Vol. 66 (1979) 772-775
Intestinal bacterial metabolism of protein and bile acids: role in pathogenesis of hepatic disease after jejuno-ileal bypass surgery P . R . POWELL-JACKSON, D . P. M A U D G A L , D O R O T H Y S H A R P , A L I S O N G O L D I E A N D J . D . MAXWELL* SUMMARY
Jejunal bacterial colonization and intestinal metabolism of bile acids and protein by bacteria have been investigated in 12 patients with abnormal liver histology following jeujno-ileal bypass surgery for obesity. Aerobic and/or anaerobic colonic flora was present in jejunal aspirates from 8 of 12 bypass patients, but in none of the controls. Intestinal protein metabolism and bile acid deconjugation (measured by urinary indican excretion and 14C-glycocholic acid breath test ) was significantly enhanced in bypass patients. Intestinal bacterial overgrowth, with abnormal intestinal metabolism by bacteria of ingested nutrients and bile acids, could contribute to hepatic disease after bypass surgery via the production of endogenous hepatotoxins. JEJUNO-ILEAL bypass is a highly effective treatment for selected patients with gross refractory obesity, but this operation may be followed by a variety of complications. In addition to anticipated nutritional disturbances, there are a number of unexplained sequelae including arthralgia, colonic pseudo-obstruction and liver disease. I n most cases the hepatic disturbance, characterized by hepatic steatosis and patchy hepatocellular necrosis, is subclinical. However, death from hepatic failure has been reported in between 2 and 8 per cent of patients (Maxwell et al., 1977). While bacteria have been implicated in the arthralgia and colonic pseudo-obstruction following bypass surgery (Wands et al., 1976; Barry et al., 1977), the pathogenesis of liver disease, the most serious complication of this operation, remainds unexplained. In previous studies neither hepatitis B infection nor halothane anaesthesia could be implicated (Maxwell et al., 1977). Although nutritional disturbances may contribute, there is no simple relationship between altered liver function and weight loss following bypass (Salmon and Reedyk, 1975). As abnormal intestinal bacterial metabolism of ingested nutrients and bile acids might be relevant t o post-bypass hepatic dysfunction, we studied the effects of jejuno-ileal bypass on the bacterial flora in the jejunum proximal t o the excluded segment, and on the bacterial metabolism of tryptophan and bile acids within the intestine.
Patients and methods Twelve unselected patients (all female, aged 28-64 years) were investigated a t least 3 years after jejuno-ileal bypass surgery as part of a routine postoperative evaluation. The operation performed was a modification of the standard procedure. Four inches (10 cm) of jejunum measured from the duodenojejunal junction were anastomosed end to side to the terminal 10 in (25 cm) of ileum, thus producing a n isoperistaltic excluded segment of jejuno-ileum several feet in length. None of the bypass patients had any clinical o r biochemical features of liver disease a t the time of study. However, liver biopsy revealed extensive fatty change, together with variable degrees
of patchy focal necrosis, portal tract hypercellularity and parenchymal fibrosis, in all patients. Five control subjects, matched for age and sex, were selected from obese patients awaiting jejuno-ileal bypass surgery and from hospital inpatients. None of the controls had undergone bowel surgery, nor had any of the patients o r controls recently received any antibiotic therapy. The nature of the study was explained, and all subjects gave informed consent before investigation. The following investigations were performed: Jejunal aspiration: A jejunal aspirate was obtained from the fasting subject by passing a sterile Andersen radio-opaque long weighted sump tube transnasally and positioning it, under radiological control, just distal t o the duodenojejunal flexure. Two millilitres of jejunal contents were aspirated into a sterile syringe, any air was excluded and the syringe was made airtight with a cap. Samples were transported to the laboratory where they were either cultured immediately or frozen for culture at a later date. Aliquots (1 ml) of samples t o be frozen were injected into a carbon dioxide gassed vial containing 9 ml of 10 per cent glycerol broth. The vial was then immersed in a methanol/acetone/solid CO, mixture in a vacuum flask. Frozen samples were stored at -20 "C and cultured within 3 weeks. Aerobiclanaerobic bacterial culture : Serial logarithmic dilutions of the aspirate were made in brain-heart infusion broth up to lo-*. Equal volumes of each dilution (0.1 ml) were spread o n t o a series of selective and non-selective media and incubated under appropriate atmospheric conditions for up to 5 days. Bacterial counts and identifications of species were carried out according to standard methods as described by Holderman and Moore (1972) and Cowan and Steel (1974). Bile acids in postprandial bile: After obtaining the initial fasting jejunal aspirate for bacterial culture, the patient was given a standard Lundh test meal t o stimulate bile secretion. Bile was siphoned off for 2 h into a sterile bottle kept on ice. Samples were stored at -20 "C before analysis for deconjugated bile acids by thin-layer chromatography. This was carried out on 0.25 mm silica gel plates developed with solvent systems using trimethylpentane : ethylacetate : acetic acid (5 : 5 : 1) and benzene :isopropanol :acetic acid (30 : 10 : 1) (Eneroth, 1963). Bile acid breath test: Hourly breath samples were collected into hyamine hydroxide for 8 h after oral administration of 5 VCi 14C-glycocholatewith a Lundh test meal after an overnight fast. No jejunal fluid was aspirated on this occasion. 14C radioactivity was measured by liquid scintillation counter and expressed as a percentage of the dose given per mmol C 0 , x body weight (kilogrammes) (James et al., 1973). Urinary indican excretion: Urine was collected for 24 h into bottles containing thymol as preservative. All subjects were on a normal diet, taking 60-100 g protein daily. The indican content was measured by the spectrophotometric method of Curzon and Walsh (1962). Urinary creatinine excretion was also measured t o evaluate the completeness of urine collection.
Results Bacterial culture Counts of aerobic organisms in the group of 12 postbypass patients were higher than in the group of 5
* St George's Hospital Medical School, Cranmer Terrace, London. Correspondence to: J. D. Maxwell.
Intestinal bacterial metabolism
I80 l
i
2000
773
0
I500
A
0 0
2
7 4 : 0-I
-
Subjects
Controls
-
0
5
v 8 a
Subjects
Controls
.-8 1000
0
.-
-0 c
x
L .o
Aerobes
Anaerobes
I
Fig. 1. Log counts of aerobic and anaerobic organisms from proximal jejunal aspirates in post-bypass patients and controls. 0 ,Small intestinal flora. A, Colonic flora.
3 0O0
T
500 0
..
0
€
0.3
-
Liver disease
-
Y z M
Fig. 3. Urinary indican excretion (pmo1/24 h) in post-bypass patients (O), normal controls and liver disease controls ( 0 ) . All results given as meanfs.e. mean.
h
V
x, 8-
Bypass
Controls
I-
D
Normal
0.2
2 E . ! s 0-I V
0 2
4
6
8
Time (h)
Fig. 2. Bile acid breath test in post-bypass patients (0) and . results given as mean&s.e. mean. controls ( 0 ) All
control subjects (P0.05, MannWhitney U test) (Fig. 1). A separate estimation was made regarding the presence or absence of colonic flora in jejunal aspirates. Aerobic colonic flora were present in 6 of the post-bypass patients but in none of the controls (P= 0.075, Fisher’s exact test). Anaerobic colonic
flora were present in 7 of the post-bypass patients but in none of the controls (P= 0.04, Fisher’s exact test). Aerobic and/or anaerobic colonic flora was present in 8 of the post-bypass patients but in none of the controls (P= 0.02, Fisher’s exact test). Neither subjects nor controls had aerobic bacteria which could deconjugate bile salts. Ten out of 12 subjects and 4 out of five controls had anaerobic bacteria which could deconjugate bile salts. In the control group only one species was present which could deconjugate bile salts, i.e. veillonella. However, in the group of subjects, a variety of deconjugating bacteria was present. These included bacteroides, veillonellae, bifidobacteria and other anaerobic Grampositive rods. More than one species was often present in each of the subjects. Thin-layer chromatography Thin-layer chromatography using two different solvent systems showed the presence of deconjugated bile acid in jejunal aspirate from 2 of the 12 post-bypass patients. One patient showed the presence of deconjugated deoxycholic acid and the other showed deconjugated cholic acid, chenodeoxycholic acid and deoxycholic acid. Bile acid breath test There was increased deconjugation of bile acids in the group of 11 post-bypass patients compared with the
774
P. R. Powell-Jackson et al.
group of 1 1 normal control subjects (P
Intestinal bacterial metabolism of protein and bile acids: role in pathogenesis of hepatic disease after jejuno-ileal bypass surgery P . R . POWELL-JACKSON, D . P. M A U D G A L , D O R O T H Y S H A R P , A L I S O N G O L D I E A N D J . D . MAXWELL* SUMMARY
Jejunal bacterial colonization and intestinal metabolism of bile acids and protein by bacteria have been investigated in 12 patients with abnormal liver histology following jeujno-ileal bypass surgery for obesity. Aerobic and/or anaerobic colonic flora was present in jejunal aspirates from 8 of 12 bypass patients, but in none of the controls. Intestinal protein metabolism and bile acid deconjugation (measured by urinary indican excretion and 14C-glycocholic acid breath test ) was significantly enhanced in bypass patients. Intestinal bacterial overgrowth, with abnormal intestinal metabolism by bacteria of ingested nutrients and bile acids, could contribute to hepatic disease after bypass surgery via the production of endogenous hepatotoxins. JEJUNO-ILEAL bypass is a highly effective treatment for selected patients with gross refractory obesity, but this operation may be followed by a variety of complications. In addition to anticipated nutritional disturbances, there are a number of unexplained sequelae including arthralgia, colonic pseudo-obstruction and liver disease. I n most cases the hepatic disturbance, characterized by hepatic steatosis and patchy hepatocellular necrosis, is subclinical. However, death from hepatic failure has been reported in between 2 and 8 per cent of patients (Maxwell et al., 1977). While bacteria have been implicated in the arthralgia and colonic pseudo-obstruction following bypass surgery (Wands et al., 1976; Barry et al., 1977), the pathogenesis of liver disease, the most serious complication of this operation, remainds unexplained. In previous studies neither hepatitis B infection nor halothane anaesthesia could be implicated (Maxwell et al., 1977). Although nutritional disturbances may contribute, there is no simple relationship between altered liver function and weight loss following bypass (Salmon and Reedyk, 1975). As abnormal intestinal bacterial metabolism of ingested nutrients and bile acids might be relevant t o post-bypass hepatic dysfunction, we studied the effects of jejuno-ileal bypass on the bacterial flora in the jejunum proximal t o the excluded segment, and on the bacterial metabolism of tryptophan and bile acids within the intestine.
Patients and methods Twelve unselected patients (all female, aged 28-64 years) were investigated a t least 3 years after jejuno-ileal bypass surgery as part of a routine postoperative evaluation. The operation performed was a modification of the standard procedure. Four inches (10 cm) of jejunum measured from the duodenojejunal junction were anastomosed end to side to the terminal 10 in (25 cm) of ileum, thus producing a n isoperistaltic excluded segment of jejuno-ileum several feet in length. None of the bypass patients had any clinical o r biochemical features of liver disease a t the time of study. However, liver biopsy revealed extensive fatty change, together with variable degrees
of patchy focal necrosis, portal tract hypercellularity and parenchymal fibrosis, in all patients. Five control subjects, matched for age and sex, were selected from obese patients awaiting jejuno-ileal bypass surgery and from hospital inpatients. None of the controls had undergone bowel surgery, nor had any of the patients o r controls recently received any antibiotic therapy. The nature of the study was explained, and all subjects gave informed consent before investigation. The following investigations were performed: Jejunal aspiration: A jejunal aspirate was obtained from the fasting subject by passing a sterile Andersen radio-opaque long weighted sump tube transnasally and positioning it, under radiological control, just distal t o the duodenojejunal flexure. Two millilitres of jejunal contents were aspirated into a sterile syringe, any air was excluded and the syringe was made airtight with a cap. Samples were transported to the laboratory where they were either cultured immediately or frozen for culture at a later date. Aliquots (1 ml) of samples t o be frozen were injected into a carbon dioxide gassed vial containing 9 ml of 10 per cent glycerol broth. The vial was then immersed in a methanol/acetone/solid CO, mixture in a vacuum flask. Frozen samples were stored at -20 "C and cultured within 3 weeks. Aerobiclanaerobic bacterial culture : Serial logarithmic dilutions of the aspirate were made in brain-heart infusion broth up to lo-*. Equal volumes of each dilution (0.1 ml) were spread o n t o a series of selective and non-selective media and incubated under appropriate atmospheric conditions for up to 5 days. Bacterial counts and identifications of species were carried out according to standard methods as described by Holderman and Moore (1972) and Cowan and Steel (1974). Bile acids in postprandial bile: After obtaining the initial fasting jejunal aspirate for bacterial culture, the patient was given a standard Lundh test meal t o stimulate bile secretion. Bile was siphoned off for 2 h into a sterile bottle kept on ice. Samples were stored at -20 "C before analysis for deconjugated bile acids by thin-layer chromatography. This was carried out on 0.25 mm silica gel plates developed with solvent systems using trimethylpentane : ethylacetate : acetic acid (5 : 5 : 1) and benzene :isopropanol :acetic acid (30 : 10 : 1) (Eneroth, 1963). Bile acid breath test: Hourly breath samples were collected into hyamine hydroxide for 8 h after oral administration of 5 VCi 14C-glycocholatewith a Lundh test meal after an overnight fast. No jejunal fluid was aspirated on this occasion. 14C radioactivity was measured by liquid scintillation counter and expressed as a percentage of the dose given per mmol C 0 , x body weight (kilogrammes) (James et al., 1973). Urinary indican excretion: Urine was collected for 24 h into bottles containing thymol as preservative. All subjects were on a normal diet, taking 60-100 g protein daily. The indican content was measured by the spectrophotometric method of Curzon and Walsh (1962). Urinary creatinine excretion was also measured t o evaluate the completeness of urine collection.
Results Bacterial culture Counts of aerobic organisms in the group of 12 postbypass patients were higher than in the group of 5
* St George's Hospital Medical School, Cranmer Terrace, London. Correspondence to: J. D. Maxwell.
Intestinal bacterial metabolism
I80 l
i
2000
773
0
I500
A
0 0
2
7 4 : 0-I
-
Subjects
Controls
-
0
5
v 8 a
Subjects
Controls
.-8 1000
0
.-
-0 c
x
L .o
Aerobes
Anaerobes
I
Fig. 1. Log counts of aerobic and anaerobic organisms from proximal jejunal aspirates in post-bypass patients and controls. 0 ,Small intestinal flora. A, Colonic flora.
3 0O0
T
500 0
..
0
€
0.3
-
Liver disease
-
Y z M
Fig. 3. Urinary indican excretion (pmo1/24 h) in post-bypass patients (O), normal controls and liver disease controls ( 0 ) . All results given as meanfs.e. mean.
h
V
x, 8-
Bypass
Controls
I-
D
Normal
0.2
2 E . ! s 0-I V
0 2
4
6
8
Time (h)
Fig. 2. Bile acid breath test in post-bypass patients (0) and . results given as mean&s.e. mean. controls ( 0 ) All
control subjects (P0.05, MannWhitney U test) (Fig. 1). A separate estimation was made regarding the presence or absence of colonic flora in jejunal aspirates. Aerobic colonic flora were present in 6 of the post-bypass patients but in none of the controls (P= 0.075, Fisher’s exact test). Anaerobic colonic
flora were present in 7 of the post-bypass patients but in none of the controls (P= 0.04, Fisher’s exact test). Aerobic and/or anaerobic colonic flora was present in 8 of the post-bypass patients but in none of the controls (P= 0.02, Fisher’s exact test). Neither subjects nor controls had aerobic bacteria which could deconjugate bile salts. Ten out of 12 subjects and 4 out of five controls had anaerobic bacteria which could deconjugate bile salts. In the control group only one species was present which could deconjugate bile salts, i.e. veillonella. However, in the group of subjects, a variety of deconjugating bacteria was present. These included bacteroides, veillonellae, bifidobacteria and other anaerobic Grampositive rods. More than one species was often present in each of the subjects. Thin-layer chromatography Thin-layer chromatography using two different solvent systems showed the presence of deconjugated bile acid in jejunal aspirate from 2 of the 12 post-bypass patients. One patient showed the presence of deconjugated deoxycholic acid and the other showed deconjugated cholic acid, chenodeoxycholic acid and deoxycholic acid. Bile acid breath test There was increased deconjugation of bile acids in the group of 11 post-bypass patients compared with the
774
P. R. Powell-Jackson et al.
group of 1 1 normal control subjects (P
