Effect of cholecystectomy on bile acid diarrhoea biomarkers: A prospective clinical study

The pathophysiological mechanisms of bile acid diarrhoea after cholecystectomy are unknown. Therefore, we aimed to explore the effects of cholecystectomy on the plasma biomarkers of bile acid diarrhoea: fibroblast growth factor 19 and 7α‐hydroxy‐4‐cholesten‐3‐one.


| INTRODUC TI ON
Chronic watery diarrhoea is a common complaint after cholecystectomy, with an estimated cumulative incidence of 9%, 1 and bile acid diarrhoea may account for 66% of this. [1][2][3] Bile acid diarrhoea is caused by excess bile acids spilling over from the small intestine to the large intestine, triggering secretion and peristalsis. 4 Ultimately this leads to diarrhoea that may be socially debilitating. 5 Fort et al found that cholecystectomy generally accelerated colonic transit time, which correlated clinically with increased stool frequency and looser consistency. 6 However, in most patients this did not result in diarrhoea. 6 It is unclear why some patients develop diarrhoea after cholecystectomy and whether this involves an altered bile acid homeostasis. 1 Both faster enterohepatic cycling of bile acids with increased bile acid synthesis and a changed composition of the bile acid pool 7 could play a role. 1 The bile synthesis rate may be assessed with plasma 7α-hydroxy-4-cholesten-3-one (C4), 8,9 whereas ileal reabsorption of bile acids may be assessed with plasma fibroblast growth factor 19 (FGF19). 10,11 Patients with bile acid diarrhoea diagnosed with the gold standard 75-seleno Tauro-homocholic acid (SeHCAT) retention test have lower fasting FGF19 12,13 and higher fasting C4 [13][14][15][16] than diarrhoea controls as does patients with postcholecystectomy bile acid diarrhoea. [12][13][14] Reports on the specific effects of cholecystectomy on the biomarkers are limited-however, two studies found substantial changes in FGF19 and C4 but only minor impact on bowel habits. 17

| Procedures
The study specified a preoperative visit and a visit to be scheduled 3-5 months postoperatively. The patients registered bowel habits in a Bristol stool scale diary for 7 days before each visit and answered the short health scale 19 and the gastrointestinal quality of life index 20 questionnaires, both with a 2-week recall. We defined diarrhoea as more than three bowel movements per day or more than one liquid stool (Bristol stool scale type 6 or 7) per day. 21 The patients attended each visit after an overnight fast and ingested 1250 mg unconjugated chenodeoxycholic acid (Xenbilox®) with a solid study meal consisting of two slices of toast and two boiled eggs as previously described. 22 Blood was collected before and 60, 90, 120 and 150 minutes after ingestion of the chenodeoxycholic acid capsules.
The results of the plasma bile acids are presented as sums of unconjugated and conjugated forms (taurine, glycine or sulphate) of the primary bile acids (chenodeoxycholic acid and cholic acid) and the secondary bile acids (deoxycholic acid, lithocholic acid, ursodeoxycholic acid and hyodeoxycholic acid) unless else is specified.

| Endpoints
All endpoints compared preoperative with postoperative measurements. The primary endpoint was the increment in FGF19 from fast-

| Power calculation
Stimulated FGF19 peak values at 150 minutes (Δ 0-150 FGF19) were estimated to 350 pg/mL 22,26 before and 250-300 pg/mL after cholecystectomy, 18 both with a standard deviation of 100. We expected no significant change in fasting FGF19 relative to the peak levels.

| Statistical analysis
The normality of data distribution was assessed with probabilityprobability plots. FGF19 and C4 were lognormally distributed and reported as means with 95% CIs and analysed with paired t-tests.
Other data were largely non-parametrically distributed. These data were analysed with the Wilcoxon signed-rank test and are reported as medians with interquartile (Q1-Q3) ranges. Confidence intervals for proportions were calculated with Wilson's method. The total AUCs were calculated with the trapezoid rule. Associations between continuous parameters were assessed with Spearman correlations. Two-sided P < .05 were considered significant; no adjustment for multiple statistical testing was made. Statistical analyses were performed in SPSS version 27.

| Recruitment
Invitations were sent to 198 patients referred to the surgical outpatient clinic for evaluation of a possible cholecystectomy. Sixty patients (60% female) were scheduled for cholecystectomy; of these, 23 were enrolled, and 18 completed the study (see Figure S1 for a detailed recruitment flow chart).

| Demographic characteristics
The median participant age was 55 (interquartile range, IQR 42-63) years; 50% were female. Baseline biochemistry was normal. The median time from the operation date to the postoperative visit was 4.6 months (3.9-5.8; minimum 3.2, maximum 10.2) ( Table 1). Alcohol and lipid-lowering medications may influence FGF19 and C4. 27,28 One patient took atorvastatin during the study; none reported excessive alcohol intake.

| Clinical outcomes
There were no surgical complications. Diary results were available for 17 patients as one diary was lost. Cholecystectomy did not change diary-recorded stool frequency and consistency (Table 2) or plasma lipid levels (Table S1). One patient fulfilled the diarrhoea criteria both at baseline with 3.3 daily bowel movements and after cholecystectomy with 3.7 daily bowel movements. During the 7-day diary registration this patient had one loose stool (Bristol type 6) before cholecystectomy, and after cholecystectomy the patient had six loose stools (Bristol type 6) with five being in 1 day.

| Plasma bile acid species
The serial plasma measurements of unconjugated chenodeoxycholic acid reflected the ingestion of this bile acid with the meal (Table 4).
Peak levels (P = .59) and total AUCs (P = .59) were similar before and after cholecystectomy. There were no significant changes in the bile acid composition measured in plasma at fasting. The total AUC for the summed secondary bile acids increased from median 269 µM · minutes (186-349) to 347 µM · minutes (187-432) (P = .02) after cholecystectomy, driven by deoxycholic acid that increased from median 207 µM · minutes (157-300) to 283 µM · minutes (156-387) (P = .02). There were no significant changes in the AUC for the total bile acids or the summed primary bile acids. After cholecystectomy, the total AUC for deoxycholic acid correlated strongly with the total AUC of FGF19 (r = .57, P = .014), but before cholecystectomy it did not (r = .28, P = .26).
There were no differences in the reported frequency of the gastrointestinal quality of life index items "frequent bowel movements", "diarrhoea", "urgent defecation", or "sudden uncontrollable bowel movements" before and after the cholecystectomy (P > .10).

| Adverse events
The 23 enrolled patients in total completed 40 visits, which included ingestion of chenodeoxycholic acid. There were no serious adverse events. Non-serious adverse events were acute diarrhoea (70%),   We found no significant change in the fasting plasma biomarkers 3-6 months after cholecystectomy. Three patients had FGF19 less than 60 pg/mL (Figure 1), suggestive of bile acid diarrhoea, 13 both before and after the operation. These patients' stool diaries were normal, reflecting the poor positive predictive value of FGF19. 12,13 By contrast, no patients had elevated C4 levels above either of the thresholds of 30 ng/mL 13,14 and 48-52 ng/mL 13,31,32 that are verified for the biochemical diagnosis of bile acid diarrhoea. This consistency between normal C4 measurements and no diarrhoea symptoms supports the specificity of C4 as a biomarker for bile acid diarrhoea.

| D ISCUSS I ON
Furthermore, it suggests that elevated C4 in a patient with postcholecystectomy diarrhoea is not a clinically insignificant consequence of the cholecystectomy. However, these results differ from the findings of two previous studies where cholecystectomy profoundly changed the fasting biomarkers. One study found a two-fold increase in C4 three months after cholecystectomy, with five of 10 patients having values above 48 ng/mL, 18 and mean FGF19 decreased by 70 pg/mL. 18 The other study also found a two-fold increase in C4 from a mean of 25 ng/mL before to 47 and 53 ng/mL one and three months after cholecystectomy respectively. 17 Only three of 51 patients had intermittent diarrhoea after three months. 17 Both the frequency of approximately 50% of participants having a substantial C4 increase and the lack of associated diarrhoea symptoms are puzzling. It seems disproportionate compared with an estimated 9% incidence of postcholecystectomy diarrhoea found in epidemiological studies 1 and that elevated C4 is a strong indicator of bile acid diarrhoea. 31 The timing of the postcholecystectomy assessment does not explain the discrepancy since it was three months in all studies.
Differences in diets could play a role because diet affects bile acid pool size and composition. 33 A high-fat diet is a proposed predictor for developing postcholecystectomy diarrhoea, 29 and a low-fat diet is beneficial in bile acid diarrhoea. 34 Since no patients in this prospective cohort developed postcholecystectomy diarrhoea, extrapolation to such cases is warranted.
We included seven patients with a history of cholecystectomy in a previous study on the biochemical diagnosis of bile acid diarrhoea. 13 The seven patients had a median 1-week SeHCAT retention of 8 We conclude that cholecystectomy increased the FGF19 response, which may reflect an accelerated enterohepatic circulation of bile acids. There was no change in bowel habits nor fasting biomarkers of bile acid diarrhoea when assessed 3-6 months postoperatively. In contrast to previous works, these results do not suggest that the reference intervals for the fasting biomarkers need adjustment in cholecystectomised patients.

ACK N OWLED G EM ENTS
The authors thank for the collaboration with the surgical depart-

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1002/ygh2.489.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on reasonable request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.