Choledochal malformations in adults in the Netherlands: Results from a nationwide retrospective cohort study

Abstract Background and Aims Patients with a choledochal malformation, formerly described as cysts, are at increased risk of developing a cholangiocarcinoma and resection is recommended. Given the low incidence of choledochal malformation (CM) in Western countries, the incidence in these countries is unclear. Our aim was to assess the incidence of malignancy in CM patients and to assess postoperative outcome. Methods In a nationwide, retrospective study, all adult patients who underwent surgery for CM between 1990 and 2016 were included. Patients were identified through the Dutch Pathology Registry and local patient records and were analysed to determine the incidence of malignancy, as well as postoperative mortality and morbidity. Results A total of 123 patients with a CM were included in the study (Todani Type I, n = 71; Type II, n = 10; Type III, n = 3; Type IV, n = 27; unknown, n = 12). Median age was 40 years (range 18‐70) and 81% were female. The majority of patients (99/123) underwent extrahepatic bile duct resection, with additional liver parenchyma resections in eight patients, only exploration in two, and a local cyst resection in eight patients. Postoperative 30‐day mortality was 2% (2/123) and limited to patients who underwent liver resection. Severe morbidity occurred in 24%. In 14 of the 123 patients (11%), a malignancy was found in the resected specimen. One patient developed a periampullary malignancy 7 years later. Conclusions In a large Western series of CM patients, 11% were found to have a malignancy. This justifies resection in these patients, despite the risk of morbidity (24%) and mortality (2%).


| INTRODUC TI ON
Choledochal malformation (CM) is a rare congenital malformation of the bile ducts and is considered a premalignant condition. CM is generally characterized by fusiform or saccular dilatation of the extrahepatic (or larger intrahepatic) bile duct(s) and an alternative (now obsolete) term for these abnormalities is choledochal cysts. The exact incidence of CM is largely unknown and is race-dependent. Some authors report an estimated 100 to a 1000 higher incidence of CM in Asia, compared to Western countries, with a strong female preponderance. [1][2][3][4] In children the incidence of CM is estimated at 1 in 59 000 live births. 5,6 This is higher than previously suggested in the literature, possibly as a result of improved detection of CM with the introduction of antenatal ultrasound. 7 The maximum diameter of the normal common bile duct has been described in the Japanese population by the JSGPM in 2016, but no norm value exists for non-Asian populations. 8 Because of potential malignant degeneration, standard treatment of CM consists of prophylactic resection of the affected part of the extrahepatic bile duct, including the gallbladder, followed by restoration of the biliodigestive tract via a Roux-Y hepaticojejunostomy (HJ). With this procedure the premalignant tissue is resected to avoid the later formation of cholangiocarcinoma (CCA) or gallbladder carcinoma (GBC). It is, however, difficult to ascertain the real risk of formation of biliary malignancies because the prevalence of asymptomatic CM in the general population is unknown.
In the Western population, the incidence of CM in adults is understudied with only three studies describing cohorts larger than 36 patients and this lack of data forms the basis for our investigation. [9][10][11] Extrahepatic bile duct resection (EHBDR) is associated with considerable short-term morbidity and mortality. Long-term complications include anastomotic strictures of the HJ, often requiring radiological and/or surgical re-intervention. Nevertheless, detailed information on the occurrence and optimal treatment of these complications is lacking. Because of the impact of these complications, weighing of the risks and benefits of prophylactic resection poses a clinical problem for patients and surgeons.
The aim of this study was to analyse a nationwide cohort of adult patients with CM, with special interest for the occurrence of biliary malignancy associated with CM. Furthermore, we assessed the short-term and long-term complications of surgical therapy.

| ME THODS AND MATERIAL S
This was a multicentre retrospective cohort study. Each of the eight academic medical centres in the Netherlands participated in the patient search. In the Netherlands, major biliary surgery, as well as surgery for congenital malformations, is performed mainly in one of these centres. The study was approved by the medical ethics committee of the University Medical Center Groningen (UMCG) and the need for patients' informed consent was waived (reference number METC2016/341). The STROBE guideline for reporting of cohort studies was adhered to. 12

| Patient population and identification
A structured search of the Dutch Pathology Registry (PALGA) was performed using a selection of medical subject headings (MeSH) criteria as listed in Appendix 1. PALGA, established in 1971, is the nationwide network and registry of histo-and cytopathology. 13 It prospectively collects data from all Dutch pathology laboratories and serves as a central basis for research. Conclusions of pathology excerpts of all potential cases were read and, when appropriate, excerpts were studied in more detail. If relevant to the present study, the selected pathology data were returned to each participating centre and matched with the hospital record for further analysis.
Using the same MeSH criteria, hospital-based registries were searched for eligible patient files. Data retrieved from PALGA and from hospital-based registries were cross-checked to avoid duplicates and this patient identification was used for on-site medical file evaluation. resections in eight patients, only exploration in two, and a local cyst resection in eight patients. Postoperative 30-day mortality was 2% (2/123) and limited to patients who underwent liver resection. Severe morbidity occurred in 24%. In 14 of the 123 patients (11%), a malignancy was found in the resected specimen. One patient developed a periampullary malignancy 7 years later.

Conclusions:
In a large Western series of CM patients, 11% were found to have a malignancy. This justifies resection in these patients, despite the risk of morbidity (24%) and mortality (2%).

K E Y W O R D S
bile duct carcinoma, choledochal cyst, choledochal malformation, surgery

Lay summary
A Choledochal malformation is a very rare dilatation of the bile duct. To prevent the development of bile duct cancer, an operation is advised. We gathered data of 123 patients operated upon and described their diagnosis, surgery and outcome through a nationwide collaborative study.
To determine the incidence of CM, we aimed to include all consecutive adult (Age ≥ 18 years at the time of surgery) patients who had underwent a surgical resection for CM, with or without CCA, between 1990 and 2016. We excluded patients who were not operated upon and patients in whom the final diagnosis was not a CM as termed in the operation report or the postoperative discharge letter.
In case of an unclear diagnosis or clear secondary, acquired dilatation of the bile ducts not related to CM, patients were also excluded. We did collect data on patients with a Type V CM but we excluded these patients from the present analysis because their disease is located deep into the liver parenchyma and its surgical treatment is strikingly different.

| Data collection and outcome definitions
We extracted the data from patient charts and anonymously recorded the data using a standardised case record form (CRF, Appendix 2). This CRF was defined and maintained during the complete data acquisition. All data were collected by at least 2 researchers (either RdK, AMS and/or AtH) and disagreements between researchers were resolved by consensus. The parameters collected included patient demographics (sex, age at time of surgery and the We graded the short-term complications (≤30 days) according to the highest Clavien-Dindo classification. 15 We considered shortterm complications grade 3 or higher as major complications and specified them in more detail. These included hepaticojejunostomy leakage, intra-abdominal abscess or collection, haemorrhage (requiring an intervention) and sepsis. Postoperative mortality was defined as death by any cause within 30 days of surgery.
Long-term complications (>30 days) included recurrent cholangitis, which we defined as jaundice with fever (>38.5°C) and abdominal pain, anastomotic stricture of the HJ, bile stones (intra-hepatic stone formation confirmed by ultrasonography), liver abscesses (fever with bacterial infection proven via culture and infection of the liver confirmed by ultrasonography), incisional hernia requiring re-intervention.

| Statistics
Analytical statistics were used to compare groups within the cohort.
Continuous data were presented as median with range, and categorical data as numbers with percentages. Missing data were excluded from the analysis per item. The SPSS (IBM Corporation, Armonk, NY, USA) package 26 was used for the analysis. After combining the two sources of data, we identified a total of 152 adult patients with CM who had undergone surgery. From this number, we excluded 29 patients with Type V malformations.

| RE SULTS
Ultimately, therefore, we analysed the data pertaining to 123 patients.
This gave an overall incidence of 5.85 treated cases of CM (Type I-V) per year, or 4.73 cases of CM Type I-IV per year. In 2000, the size of the Dutch adult population was approximately 11.9 million. 16

| Patient characteristics
In Table 1, we provide a list of patients' characteristics and surgical details. Female patients formed the majority of patients (81%, 100/123). Two explorations without resection were performed because of extensive malignancy. In seven patients, a Type II cyst was resected without the need for formal bile duct resection. EHBDR with additional parenchymal resections were performed in eight patients; these included five major liver resections, one pancreatic resection and one combination of a right-sided hemihepatectomy with a pylorus preserving pancreatico-duodenectomy. One patient had a conversion of a cystoduodenostomy to a formal resection of the cyst and reconstruction with HJ with a concomitant colon resection. In one instance, in 1994, a hepatico-duodenostomy was performed because the patient lacked suitable small intestine. One cyst resection was performed and bile continuity was achieved by means of a duct-to-duct anastomosis, in a patient with substantial co-morbidity.
Open surgery was performed on all patients.

| Postoperative outcome
We list postoperative outcomes that included short-term and longterm complications in Table 2. One half of the patients did not suffer

(Recurrent) cholelithiasis 3 7
Secondary biliary cirrhosis 0 0 Length of follow-up in years (median, range) 1.7 0. 1 -15.7 related to the previous CM. Cancer was only found in type I and type IV malformations. Since a more detailed evaluation of the subtypes of CM was not recorded these data were not available to us.

| D ISCUSS I ON
CM is a rare diagnosis of which the true incidence in the Western World is unknown. Based on the present nationwide cohort of 26 years, the incidence of surgically treated CM in this Western adult population of 11.9 million was 5.85 patients per year. It is highly likely that the improved detection of congenital defects with ultrasound will bring forward the detection of choledochal malformations. 6 With improved detection and imaging, mild dilatation of the bile duct, without distal obstruction, will spark the discussion if a malformation is present. In a large cohort from the USA, the occurrence of major complications following surgery for CM was reported to be as high as 56% with a mortality rate of 7%. 11 Another American study by Nicholl et al in 51 operated patients reported a morbidity rate of 25% with no operative mortality. 17 Postoperative morbidity and mortality rates in surgery for cholangiocarcinoma are known to be lower in Asian countries compared to Western countries in the Western World. 18 Accordingly, reported morbidity and mortality rates following surgery for CM are also lower in Asian series, 3%-13% and 0%-0.2% respectively. 19,20 The incidence of biliary malignancy found in this study is in accordance with a recent meta-analysis reporting a prevalence of malignancy of 10.7%. 1 Previous publications of patient series in the Western World reported varying percentages of malignancy, ranging from 3% to 19%. [9][10][11]17,21 This may be explained by the small sample sizes of some of these studies and by the fact that these studies also included patients with Of course, one can never be 100% certain in a retrospective study.
We do, however, note limitations to the present study. First, the fact that in the past some non-academic hospitals sporadically have performed resections for CM possibly leads to underestimation of the national incidence. 22 Second the fact that not all patients with CM will become symptomatic and may therefore not be diagnosed at all, contributes to an underestimation of the true prevalence of CM. Resection of CM as a prophylactic therapy for CCA remains the pivotal recommendation to all newly detected patients with CM in view of the high risk of malignant transformation.

CO N FLI C T S O F I NTE R E S T
None reported.