Cost- Effectiveness and Cost-utility of Laparoscopic Versus Open Repair of Ventral Hernia

Population: 130 patients, diagnosed with clinic and radiologic (abdominal wall CT scan) ventral hernia, were recruited, after expressing their willingness, for either laparoscopic or laparotomy repair to be conducted at Infanta Elena Hospital, Huelva (Spain), between January 2005 and October 2008. Recruitment criteria were: patients above 18 years old, diagnosed with primary or incisional ventral hernia, including recurrent ones, featuring a prior-to-operation estimated size from 20 to 225 cm2. Exclusion criteria included type 4 or 5 of ASA (American Society of Anesthesiologist), to suffer from a disease limiting lifespan to less than 2 years, cirrhotic ascites, emergency surgery or any complication as intestinal obstruction, strangulated hernia, peritonitis, local or systemic infection, failure of patient to ensure an at least 2 years follow-up or abandonment protocol. Every patient was fully informed and signed his willingness. Defect sizes were estimated and classified following Chevrel Classification. Patients were assigned a randomly generated number (see validated and published tables), meaning both laparoscopic or open repair, and informed same morning of operation. All operations were conducted by experienced surgeons. Protocol was approved by Local Ethics Committee.

General techniques

All patients were given an heparin prophylactic dose to prevent thromboembolic events evening prior to operation-day, as well as 1500 mg cefuroxime intravenous (IV) during anesthetic induction (in case of allergy, 600 mg of clindamycin IV + 80 mg of gentamicin IV). Also, evening before intervention-day all patients filled in survey (SF-36 questionnaire). Skin was prepared with an alcoholic solution of clorhexidine. Surgical pain was treated with magnesium metamizole (2gr IV each 6 hours). Upon patient request, 100 mg of tramadol IV was added (a maximum amount of 400 mg). Every patient was encouraged to hang around as soon as possible and be on a diet 8 hours after surgery, supplying oral treatment from then onwards. On discharge, patients were recommended wearing an abdominal binder for a month.

Laparoscopic repair: Pneumoperitoneum was performed to 12 mmHg after placement first 12 mm trocar at level of umbilicus in anterior axillary line. Subsequently are placed two 5 mm auxiliary trocars. In all cases we explore entire abdominal wall for other defects not detected preoperatively. Herniary contents and sac are reduced releasing adhesions with diathermy or harmonic scalpel. Defects are repaired with a polytetrafluoroethylene (PTFE) patch (Dual Mesh; W.L Gore and Associates, FlagstaV, AZ USA) with double crown fixation as technique of Carbajo et al, ensuring exceed 3 cm edge of defect, using 5mm tackers (Protack, Autosuture; Tyco Healthcare, USA), reducing intrabdominal pressure to 8 mmHg. Not abdominal drainage was used. Skin was closed with metal staples. All operations were performed by experienced surgeons, over than 40 laparoscopic ventral hernia repairs.

Open repair: Incision was made over hernia defect, reducing herniary contents by opening sac if it is necessary. We made 4 cm soft tissue flaps around edge of defect depending on available healthy fascial tissue. In all cases we use Chevrel technique with fascial closure using anterior rectus sheath with continuous and absorbable suture and placement of polypropylene mesh (Parietene, standart polypropylene mesh, Covidien, Norwalk, CT) in an onlay position fixed with polypropylene suture. We used one or two suction drains below subcutaneous flap which was sutured with absorbable suture, and metallic skin staples.

Follow-up

A given follow-up was always conducted by same researcher before hospital discharge, after 30 days, 60 days, a year and once a year thereafter. All postoperative reports were analyzed with aim of finding protocol deviations or intraoperative complications. At hospital follow-up, before discharge, complications were actively looked (seroma, hematoma, wound infection, prolonged ileus, urinary retention and medical complications such as stroke, respiratory failure, arrhythmia, deep vein thrombosis, adverse drug reaction or pneumonia), hospital stay was achieved and onset of walking. Patient-centered outcomes analyzed were pain measured by visual analogue scale (VAS), addition to daily physical activity, analysis of quality of life related to health by surveys after 2 and 5 days (CVP-CG questionnaire) and difference between preoperative values, at 3 months and 12 months of physical and mental construct by survey (SF-36 questionnaire). At follow-up after discharge several facts were taken as exhibiting a satisfactory evolution, among them absence of hernia recurrence (defined as clinic examination and radiologic non-defect abdominal wall), absence of symptoms related to disability (chronic pain, intermittent ileus, trocar site hernia, fistula, abscess or prosthesis extrusion) and also absence of readmission or reoperation due to hernia repair.

Cost analysis

Expenses spent in ventral hernia repair were calculated from viewpoint of funder (health costs), taking into account only direct expenses and discarding indirect and intangible ones, as nearly all patients were economically assisted during convalescence period. Euro currency was used and a prospective approach to costs analysis was employed, following cost-opportunity method. As in-hospital expenses following variables were used materials, such as prosthesis, sutures, trocars, drainages and other surgical instruments; preoperative, anesthetics and postoperative medication (included recommended at discharge); operation room personnel cost and hospital wards; cost of diagnostic tests performed and consumption of hospitality. Installation costs, equipment inventoried and cost of warehousing and inventory management of consumables were neglected. As extra-hospital costs following items were included: personnel and materials used in medical visits and cures, drugs not prescribed, travel costs in medical transports, totality of which generated as a consequence of convalescence period (up to 60 days). Finally, expenses produced by a new hospital stay or reoperation because of complicated issues within two years after hernia repair were estimated, according to direct hospital costs at the beginning of study and further added to hospital expenses. As expenditure variable we use average patient cost.

Statistical analysis

A sample size of 130 patients was considered in order to detect differences in morbidity or an increase in hospital stays of 20%, assuming a type I error of 5%, powder of 80%, with losses of 10% and 2% crosslinking. To confirm normality Kolmogorov-Smirnov test was employed. Discrete variables are presented as percentage and confidence intervals of 95% and were compared through either Pearson test or Fisher exact test. Quantitative variables are depicted as mean values and standard deviations confidence intervals 95% using Student's t test. In all cases analysis was by intention of treat including in laparoscopic group those patients that required conversion to laparotomy. Also, an incremental cost-effectiveness analysis with confidence intervals of 95% was conducted, following Fieller theorem, and for sensibility analysis inversion cross-point was taken (threshold analysis). As efficiency measure a clinic profit analysis was carried out by means of a composite analysis that integrates absence of recurrence, readmission, reoperation or disability. As an indirect measure of utility is used values obtained from questionnaires (HRQOL -Health Related Quality Of Life), assuming standard mean response in case of SF-36 and mean response standardized by standard deviation of difference in case of CVP-CG instruments between two types of repairs as well as effect size between preoperative value and postoperative one in case of SF-36. choice of measurement of HRQOL as a measure of utility was made under requirement that in absence of clinically relevant differences in mortality, recurrence or morbidity, a clinically relevant difference in HRQOL is a preference by patient and thus a choice between treatment alternatives. A minimum significance level of 0.05 was assumed. Every analysis was conducted by blinded statistician through statistical package (SPSS statistical program, version 11.5, SPSS Inc, Chicago).