This is a retrospective cohort study conducted at the Mercy Hospital for Women, a tertiary women’s hospital in Heidelberg, Victoria, Australia, caring for approximately 55 women per year with newly diagnosed with ovarian cancer. The study included all women treated with dexamethasone who had a history of epithelial ovarian cancer (all histopathological subtypes and stages) with advanced disease and were admitted with a MSBO or combined small and large bowel obstruction between 1 January 2005 and 1 December 2016.
The primary outcome was to evaluate the efficacy of dexamethasone in achieving full or partial symptoms control in patients with advanced epithelial ovarian cancer presenting with MSBO. The secondary outcomes were: (1) to evaluate whether the efficacy of dexamethasone declines during subsequent admissions with recurrent MSBO; (2) to compare dexamethasone dosage required to achieve symptom control in the first versus recurrent episodes of MSBO; (3) to compare the efficacy of dexamethasone in chemotherapy refractory, platinum resistant and platinum sensitive disease; and (4) to assess the prognostic significance of MSBO in patients with advanced ovarian cancer.
Data was obtained from review of the patient medical record. The diagnosis of bowel obstruction was based upon a patient’s signs, symptoms and imaging during admission (abdominal X-ray or Computerized Tomography (CT) of the abdomen pelvis). A diagnosis of partial bowel obstruction was made when there was ongoing passage of flatus and liquid stool, whereas complete bowel obstruction was diagnosed in the setting of failure to pass flatus and liquid stool and/or an empty rectum on radiography. The medical record was reviewed to determine patient demographics including age, 2014 International Federation of Gynaecology and Obstetrics (FIGO) stage at diagnosis, histopathological subtype, site of primary disease, date of diagnosis, date of death; primary treatment; recurrence data; number of presentations with MSBO; MSBO symptoms, dexamethasone dosage and duration; alternate management (including surgery); resolution of MSBO with symptom control and clinical outcome.
Complete symptom control was defined as discharge (to home or other, including palliative care facility) on full diet with resolution of all symptoms. Partial symptom control was defined as symptom improvement as reported by the patient and discharge tolerating clear or free fluids and oral medication with possible mild persistent symptoms such as nausea and or vomiting.
Conservative management of MSBO included nil by mouth (ice for comfort), intravenous fluids, electrolyte optimisation, analgesics, antiemetics and anti-reflux medication (usually a proton pump inhibitor). Other interventions in selected cases only included the use of a nasogastric tube (NGT), gastrograffin enema, percutaneous endoscopic gastrostomy (PEG) tube, ascitic tap or surgery. Usual dose of Dexamethasone was 8mg/day with a reducing dose of 4mg/day following resolution of symptoms.
Platinum resistant disease was defined as evidence of progressive disease (radiological, clinical or biochemical) within six months of completing first line platinum based chemotherapy.
Women who had evidence of isolated large bowel obstruction (where surgery is more often required) and women who had no evidence of advanced ovarian cancer as a cause of their obstruction were excluded, as were women with MSBO occurring during primary therapy (either perioperatively or during first line chemotherapy). Analysis was truncated for the first six admissions with MSBO.
Patient characteristics and outcome data were summarized as mean (SD), median [25th to 75th percentile] and number (%) according to type and distribution. For outcomes the precision of percentage point estimates was presented as 95% confidence intervals calculated using the Wilson score test. Assessment of longitudinal change across presentations for responsiveness and dosage per was performed using a marginal mean model using generalized estimation equations with robust standard errors and an exchangeable correlation structure. All statistical tests were two-sided, with a p value of < 0.05 considered significant. Statistical analysis was performed using Stata v16 statistical software (StataCorp. 2019.Stata Statistical Software: Release 16. College Station, T. USA).
Ethics approval was obtained through the Mercy Health Human Research Ethics Committee (Reference number R16/50, approved 11 October 2016).