Cost-Effectiveness of New Surgical Treatments for Hemorrhoidal Disease

A Multicentre Randomized Controlled Trial Comparing Transanal Doppler-Guided Hemorrhoidal Artery Ligation With Mucopexy and Circular Stapled Hemorrhoidopexy

Paul A. Lehur, MD, PhD; Anne S. Didnée, MD; Jean-Luc Faucheron, MD, PhD; Guillaume Meurette, MD, PhD; Philippe Zerbib, MD, PhD; Laurent Siproudhis, MD, PhD; Béatrice Vinson-Bonnet, MD; Anne Dubois, MD; Christine Casa, MD; Jean-Benoit Hardouin, PhD; Isabelle Durand-Zaleski, MD, PhD


Annals of Surgery. 2016;264(5):710-716. 

In This Article


Patient Selection

Adult patients with bleeding and/or prolapsing HD classified as GII (spontaneously reduced after defecation) or GIII (requiring digital reduction) hemorrhoids were given information about the trial.[1] Those who accepted trial entry were randomly assigned treatment by either DGHAL or SH. Common exclusion criteria were applied (see SDC1, Text 1,

Baseline and follow-up assessment included a past medical history, clinical examination, and HD grading, symptom and quality-of-life (QoL) [The Short Form (36) Health Survey (SF36)] questionnaires, pain level, and work activity.[1,7,15]


Randomization was carried out online and stratified according to HD grade (see SDC2, Text 2, Investigators did not vary from their normal practice for patient care. Both procedures have already been described.[5–7] DGHAL was carried out using THD (THD, Correggio, Italy; 106 patients) and AMI HAL-RAR (A.M.I.GmbH, Feldkirch, Austria) devices (96 patients; unrecorded: 1). After anal dilation and proctoscopy, a series of Doppler-guided absorbable sutures were placed to interrupt hemorrhoidal artery blood flow. Vertical mucopexy was carried out on hemorrhoidal prolapses upon request. SH was carried out using a PPH-03 (Ethicon Endo-Surgery, Cincinnati, OH; 106 patients) and a HEM stapler (Covidien, Inc.; 79 patients, unrecorded: 5).


Patient data were collected over 13 months. Patients were evaluated 1 month before surgery and were reviewed at Day 15 (D.15), 3 months (D.90), Month 6 (M.6), and Month 12 (M.12). Unscheduled visits were also recorded. The reporting of AE was done using the Clavien-Dindo grading system adapted to HD surgery.[16,17] Pain levels, analgesics consumption, hospital stay, and sick leave were all recorded.[15]

Primary Endpoint

The trial primary endpoint was defined as the morbidity rate at D.90 postoperatively for both procedures, and computed as "the percentage of patients suffering 1 or more AE according to the procedure-related complication score," whatever the grade of complication.[16]

Secondary Endpoints

Cost-effectiveness. The prospective economic evaluation was concurrent to the RCT as per the Consolidated Health Economic Evaluation Reporting Standards recommendations.[18] The analysis was conducted from a healthcare perspective to determine the cost per averted AE, with DGHAL compared with SH over a 90-day and a 1-year period. Both hospital and nonhospital resources were considered and valued using actual costs (hospital resources), or tariffs and compensation (days off work). Health outcomes for the economic evaluation were measured by the primary clinical endpoint at D.90 and M.12. Costs are expressed as 2015 Euro (€) and not discounted.

Clinical Endpoints. The comparison between the procedures was made for the number/type of intraoperative AEs, number/severity of postoperative AEs at D.90 after surgery, and the success rate in the reduction of HD at M.12. Persisting or recurrent symptoms were recorded. In cases of multiple AEs, the most serious was recorded.

Sample Size Calculation. The superiority of DGHAL compared with SH was expected with respect to the primary endpoint. We foresaw a possible switch to a noninferiority study, should the superiority of DGHAL not be proven. Therefore, the trial was planned to demonstrate the noninferiority of DGHAL, defined as a complication rate lesser or equal to the complication rate of SH (equal to 15% from a literature search) proportionally increased by 20% (ie, 15% × 20% = 3% for the planning step). This increase is considered clinically insignificant. Power and type I error were defined, respectively, at 80% and 5%, the required sample size was computed at 438 and 420 patients, respectively, for a superiority or a noninferiority analysis.

Statistical Analysis

Continuous and qualitative variables were, respectively, described as a mean ± standard deviation (SD) and percentages. Intergroup comparisons of complications and costs were performed using statistical tests adapted to group sizes (Student, chi-square, Mann-Whitney, and Fisher exact tests).

Primary Endpoint. To switch to a noninferiority analysis, we used the superior limit of the unilateral 95% confidence interval (CI) of the complication rate in the DGHAL group. The demonstration of DGHAL noninferiority in terms of complications is defined by a limit inferior to that of the SH group, proportionally increased by 20% (noninferiority margin).[19]

Economic Evaluation. Univariate sensitivity cost-analyses were performed and represented on tornado diagrams. A joint comparison of costs and effects was performed by bootstrapping with 1000 resamples, and the result of the booststrap replications presented on cost-effectiveness planes.

The significance threshold was <0.05.

Ethical Committee Approval

The study protocol was approved by the "Ouest-1" Ethical Committee (Tours, France; Ref: 2010-R26) for all investigating centers. Informed consent was obtained from each patient (except for 3 patients subsequently excluded from the analysis). A detailed information leaflet was provided to patients and they were informed of the assigned procedure when entering the hospital for surgery.