Outcomes of Laparoscopic Cholecystectomy for Biliary Dyskinesia in Children

Introduction

Biliary dyskinesia often presents as recurrent epigastric or right upper quadrant (RUQ) abdominal pain, nausea, and vomiting, and is thought to result from uncoordinated contraction and relaxation of the gallbladder and sphincter of Oddi.^1–18 This disorder is most commonly diagnosed based on clinical history and cholescintigraphy with a gallbladder ejection fraction (EF) <35% and/or reproduction of symptoms with administration of cholecystokinin (CCK), in the absence of gallstones or cholecystitis on ultrasound.^1–16

The true incidence of this condition is obscured by its difficult diagnosis in relation to other common causes of abdominal discomfort, including functional abdominal pain of childhood, irritable bowel syndrome, and dyspepsia.^2,4,19–21 Despite the rising prevalence of pediatric obesity and gallstone disease, biliary dyskinesia is becoming the leading indication for cholecystectomy in children, having risen sevenfold over a 13-year period.^8,22,23

Multiple studies have demonstrated the efficacy of cholecystectomy in improving symptoms of biliary dyskinesia in 35% to 100% of children early in the postoperative course, decreasing to 40% to 70% beyond 6 months.^{3–9,13,16–18,24,25} There are conflicting reports detailing the impact of patient and disease factors on surgical results.^{3–5,7,8,10,11,16–18,24–26} Gallbladder EF <15%, <35%, and greater than 70% have correlated with improved outcomes in some studies, while others have suggested no such relation.^{3–5,7,8,10,11,16–18,25–27} Similarly, there have been equivocal conclusions on the influence of sex, body mass index (BMI), disease manifestation and duration, comorbidities, pain with CCK, and degree of histologic cholecystitis on symptom resolution after cholecystectomy.^{2,4,5,8,10,11,16–18,25,26}

Nonoperative approaches, including observation and antireflux treatment, have also been examined in retrospective studies with mixed results compared with surgery.^9,12 Given the heterogeneity of existing publications, we sought to determine variables that predict successful surgery in our large cohort of 215 pediatric patients with biliary dyskinesia treated with laparoscopic cholecystectomy.

Materials and Methods

With institutional review board approval, a retrospective chart review was performed on children < 21 years old with biliary dyskinesia who were treated with laparoscopic cholecystectomy at a single institution between March 2010 and February 2016. Biliary dyskinesia was diagnosed based on gallbladder EF < 35% and/or symptom reproduction with CCK, in the absence of gallstones or cholecystitis on ultrasound. Patients identified with gallstones or cholecystitis on imaging were excluded as all underwent cholecystectomy without a need for further investigation with cholescintigraphy. Demographic (age, sex, weight, height, BMI), medical history (frequency and duration of symptoms, comorbidities, previous abdominal surgery), diagnostic imaging (ultrasound, cholescintigraphy), operative findings, pathology, and outcome (complications, follow-up) data were collected.

Postoperative symptom resolution was determined based on chart review of follow-up visits, and documented as complete resolution of pain, partial improvement of pain, or no effect. Stratified analyses and odds ratios were calculated for a priori defined variables, including sex, age, BMI, symptomatology, comorbidities, EF (both as a continuous variable and as a dichotomous variable based on thresholds of 10% and 35%), reproduction of pain with CCK, operative findings, pathology results, and follow-up, for their effect on symptom resolution. Statistical tests (analysis of variance, chi-square, binomial logistic regression analysis) were performed using GraphPad Prism 7.0b (GraphPad, San Diego, CA) and Stata 13.1 (Stata Corp, College Station, TX) statistical software. P < .05 was deemed statistically significant.

Results

A total of 215 laparoscopic cholecystectomy procedures were performed for biliary dyskinesia. Patient characteristics, imaging results, pathology findings, surgical factors, and postoperative outcomes are listed in Table 1. Patients were predominantly female (72.6%), with a mean age of 13.8 years and BMI 22.3 kg/m². Constant abdominal pain was the most common presenting symptom (70.9%), followed by nausea/vomiting (27.1%) and postprandial pain (24.6%), lasting a mean of 9.5 months. Common comorbidities include gastroesophageal reflux disease (15.2%) and previous abdominal surgery (18.1%). EF was abnormal (<35%) in 85.6%, while CCK reproduced symptoms in 84.2%. Normokinetic biliary dyskinesia (reproduction of symptoms with CCK despite normal EF) was found in 14.5%.

The mean operative duration was 45 minutes, and 13.0% of patients had concurrent abdominal procedures performed. On pathology, 85.9% had cholecystitis while 1.9% had unexpected cholelithiasis. Follow-up time was a median of 2.7 weeks, with inadequate follow-up data in 15.8%. Of the remaining patients, 60.8% had complete resolution of pain with cholecystectomy, while a further 28.7% had partial improvement in symptoms.

Only 10.5% of patients did not respond to surgery. Six children with partial improvement in symptoms were diagnosed subsequently with choledocholithiasis (1 patient), common bile duct (CBD) stricture (1 patient), sphincter of Oddi dysfunction (1 patient), irritable bowel syndrome (1 patient), ovarian cysts (1 patient), gastroparesis (1 patient), and conversion disorder (1 patient), which could account for their ongoing abdominal discomfort. Nine children with continuing postoperative pain were later diagnosed with sphincter of Oddi dysfunction (4 patients), gastroparesis (3 patients), celiac disease (2 patients), and lactose intolerance (1 patient).

Eight children with partially improved or ongoing abdominal pain postcholecystectomy underwent further investigation with endoscopic retrograde cholangiopancreatography (ERCP). Sphincterotomy and insertion of stent was performed in each case for the following ERCP findings: sphincter of Oddi dysfunction (sphincter hypertrophy, dilated CBD, 5 patients), CBD stricture (1 patient), choledocholithiasis (1 patient), and normal CBD with persistent pain (1 patient). In patients with sphincter of Oddi dysfunction, 1 had complete resolution of pain after cholecystectomy for 1 year before recurrence of symptoms, 4 had ongoing pain ranging from 7 weeks to 2 years, and 4 improved after sphincterotomy and stent insertion. The child with the CBD stricture had complete resolution of symptoms postoperatively, with recurrence of pain at 2 years that did not improve with sphincterotomy, duct dilation, and stent insertion, but was lost to subsequent follow-up.

The child with choledocholithiasis also had complete resolution of pain, with recurrence at 2 years that did not improve with sphincterotomy, stone clearance, and stent insertion. Eventually, this patient underwent diagnostic laparoscopy for lysis of adhesions and appendectomy with partial improvement in symptoms. The remaining child with normal duct anatomy had resolution of sharp RUQ pain after cholecystectomy, but continued to complain of postprandial epigastric cramping, nausea, and bloating. Further investigation with ERCP 7 weeks later revealed normal anatomy, but a sphincterotomy and stent were inserted empirically and upsized 3 months later due to persistent symptoms. Eventually, this patient was diagnosed with conversion disorder and gastroparesis.

Variables were analyzed after stratification into three postoperative outcomes (complete resolution of pain, partial improvement of pain, no effect) as in Table 2. There were trends for a higher percentage of nonresponders being female (P = .0677) with previous abdominal surgery (P = .0747). There were no differences with respect to age, BMI, symptomatology, EF, pain with CCK, surgical factors, or pathology between the groups. Follow-up time was significantly longer in patients whose pain did not improve after cholecystectomy (P = .0219).

Odds ratios were calculated for the categorical variables in Table 3 for complete resolution or improvement of symptoms. Female sex and previous abdominal surgery were associated with continued postoperative pain. Without accounting for reporting bias, shorter follow-up time was associated with better outcomes after cholecystectomy in both analyses. Resolution of symptoms was greater in patients with < 6 months of follow-up (complete response 97/150 [64.7%], partial improvement 44/150 [29.3%]), compared with those exceeding 6 months (complete response 13/31 [41.9%], and partial improvement 8/31 [25.8%]) with P = .0255 and P = .0002, respectively. No differences in outcome were seen with remaining variables.

The effect of previous abdominal surgery (P = .029) and follow-up time <6 months (P = .003) significantly predicted postoperative pain outcomes using binomial logistic regression analysis, whereas female sex, EF, and symptom reproduction with CCK did not.

Discussion

Over the last several decades, there has been a rising prevalence of biliary dyskinesia in children.^8,22,23 With enhanced accessibility of imaging to diagnose this condition, increased awareness of the disease, and growing surgeon comfort in performing laparoscopic cholecystectomies, the frequency of operative intervention is only expected to rise further.^8,22,23,28 In this large cohort of pediatric patients, we show that cholecystectomy is safe and effective in alleviating symptoms of biliary dyskinesia. We note that 60.8% of patients responded completely, with a further 28.7% experiencing partial improvement in symptoms, consistent with previous reports.^{3–9,13,16–18,24,25} Moreover, our complication rates after surgery are comparable to historic values. ²⁹

Analysis of factors that predict success with cholecystectomy for children with biliary dyskinesia is necessary to improve outcomes, mitigate risks of nontherapeutic surgery, and reduce healthcare costs. In those who did not respond completely to surgery, several patients were subsequently diagnosed with other causes of abdominal pain, including choledocholithiasis, CBD stricture, sphincter of Oddi dysfunction, gastroparesis, celiac disease, lactose intolerance, irritable bowel syndrome, ovarian cysts, and conversion disorder. These conditions left untreated could account for their continued discomfort, highlighting the importance of keeping a wide differential diagnosis for persistent abdominal pain and ruling out medically treatable causes before surgery.

A history of laparoscopic or open abdominal surgery was correlated with worse outcomes on univariate and multivariate analyses. Previous procedures, such as appendectomy and fundoplication, may have been performed for previous unexplained abdominal pain, nausea, and vomiting, or potentially resulted in partial or chronic adhesive bowel obstruction symptoms, both of which can mimic biliary dyskinesia. The negative association between outcome and female sex was only seen on univariate analysis and may be explained by the higher incidence of abdominal pain in girls due to gynecological (ovarian cysts, dysmenorrhea), gastrointestinal, functional, and psychiatric conditions.^30–36

However, our conclusions are limited due to the retrospective nature of all publications (to our knowledge) of pediatric biliary dyskinesia outcomes, including our own. Patient stratification into those who had complete resolution, partial improvement, and no effect categories oversimplifies the gradation of symptom resolution as an outcome. However, this approach avoids the subjectivity in quantifying the degree of improvement experienced by patients. Follow-up visits at multiple prospectively defined intervals postcholecystectomy may determine if improvement in pain is sustained or transient, leading to complete resolution of pain or eventual worsening of symptoms.

Despite improvement of symptoms in 89.5% of children in our study, this likely reflects the results of mostly short-term follow-up. Persistent symptoms after surgery were associated with longer follow-up periods. When comparing results of patients with <6 months of follow-up to those exceeding that time frame, we note a decline in symptom improvement from 94.0% to 67.7%, consistent with previous reports.^{3–6,8,16,24} However, it is unclear whether this decrease is due to a bias in the types of patients who require long-term follow-up. Children who are doing well may feel that delayed follow-up is unnecessary, thus skewing outcome results to those who have not responded to therapy. Previous reports examining long-term outcomes with surveys have been able to address the issue of reporting bias, but are susceptible to recall bias instead.^7,8,16,24

Further prospective studies examining routine short- and long-term outcomes are required to eliminate bias. Although our results have exceeded the 75% response rate of expectant management reported by a smaller study with a cohort of similar age, sex distribution, and EF, prospective studies examining outcomes of nonoperative, and surgical management of biliary dyskinesia are required to evaluate true differences between these groups to determine the utility of cholecystectomy. ⁹ Despite these limitations, we recommend that cholecystectomy is safe and effective as treatment for pediatric biliary dyskinesia. Our results may facilitate discussions on consent regarding expected outcomes after cholecystectomy.