The objective of the HULC trial is to investigate whether prophylactic OMA in addition to abdominal wall closure in SST reduces the risk of IH formation in patients undergoing elective midline laparotomy compared to SST alone.
HULC is a multicentre randomized controlled, observer and patient blinded surgical effectiveness trial with two parallel study groups.
Patients and trial centres
To enrol the required number of patients in the planned recruitment period, 12 trial sites of the Clinical Trials Network of the German Surgical Society (CHIR-Net) will participate in this trial (www.chir-net.de). These 12 centres will be high-volume centres committing to include at least 50 patients each. To improve recruitment at all centres, brochures and flyers for patients including information about the trial will be available.
Patient inclusion criteria
All patients scheduled for elective clean or clean-contaminated (25) abdominal surgery as defined by the Centers for Disease Control and Prevention (CDC) via a midline laparotomy for any indication will be screened consecutively for eligibility and will be informed about the HULC trial. All subjects must be able to understand the nature and extent of the trial and only adult patients (> 18 years of age) with a life expectancy of at least two years who provide written informed consent will be included in the trial. The participant informed consent form includes general information about the trial (indication, clinical data about incisional hernias), details about the experimental and control intervention, randomisation, benefits and risks of participation in the trial, and trial visits. Furthermore, it includes general study information like the voluntariness of participation, possibility of termination of the trial, organisation and financing of the trial, data protection, and important contact details for further questions. As the intervention of the trial includes two well established surgical standard therapies for abdominal wall closure, there is no increased risk of any harm to be expected through participation in the trial. Thus, no additional compensation for any postoperative complications or harm is planned. Patients will be insured against travel accidents for their follow-up visits.
Patient exclusion criteria
Patients with planned relaparotomy via the midline incision within 2 years after trial intervention, midline laparotomy within the last 60 days prior to trial intervention or previous IHs or fascial dehiscences will be excluded from the trial. Moreover, patients with concurrent abdominal wall infections will not be included in the trial, in order to reduce the risk of SSI and potential mesh infections. Furthermore, patients with an ASA grade > 3 according to the American Society of Anesthesiologists (ASA) classification, pregnant or lactating women and patients that participate in another intervention-trial with interference of the intervention and/or outcome of the HULC trial will be excluded.
Patient withdrawal criteria
Patients are free to stop their trial participation at any time and without giving reasons for their decision. When a trial participant withdraws his/her informed consent, he/she is asked to decide whether his/her data captured so far may be analysed or if it should be discarded. In addition, if, in the surgeon’s opinion at the end of the operation, the trial intervention will be detrimental to the subject’s well-being, the trial participation can be stopped for this patient. In this case, the patient will not be randomized and the reason for screening failure must be recorded in the screening log. All randomized patients, including those with premature trial termination, will be included in the final analysis.
Patients in both groups will receive closure of the midline incision with a slowly absorbable monofilament suture (USP 2-0, PDS Plus, Ethicon, Somerville, NJ, USA) in SST as in previous trials (1, 10). Tissue bites of 5mm and intersuture spacing of 5mm are applied exclusively to the fascia within the linea alba (omitting subcutaneous fat and muscle tissue). Suturing will be initiated at both ends of the median laparotomy towards the centre. An overlap of up to 2 cm may be created. Both sutures should be knotted independently. The suture length to wound length ratio (SL:WL) must be ≥ 4:1. The SL:WL ratio is recorded intraoperatively and is calculated as follows (16): SL:WL = (A-(B+C)) / D (A = total length of suture used (in cm); B = Length of suture remnants at starting knots (in cm); C = Length of suture remnants at finishing knots (in cm); D = Length of fascial incision (in cm))
In the experimental group, but not in the control group, the abdominal wall closure is augmented with a light-weight polypropylen mesh in onlay technique (OMA). To this end an anterior plane will be created between the anterior rectus fascia and the subcutis. The mesh should overlap the fascial midline incision by 3-4 cm on all sides (18, 24) and must be fixed to the fascia tension-free with USP 2-0 Prolene single knots (Ethicon, Somerville, NJ, USA). The mesh material will be standardized and an Optilene® Mesh (B. Braun, Melsungen, Germany) will be used. The interventional procedure will prolong the operation by approximately 20 minutes.
Materials and surgical technique will be standardized. HULC will use the same materials and surgical technique as the previous PRIMA and STITCH trials (1, 24) to ensure comparability of results and to avoid potential bias.
Closure technique of the skin and the subcutaneous tissue will be the same in both groups and will be standardized to reduce dead space and seroma formation. The subcutaneous tissue should be closed with mono- or polyfilament absorbable sutures. No subcutaneous drains should be placed. The subcutaneous sutures in the experimental group will include the mesh in the midline in order to reduce seroma formation as the latter was increased in previous OMA trials without subcutaneous sutures (18, 24), but not in OMA trials with subcutaneous sutures (19). The skin will be closed using staples.
Assignment of intervention and randomization
In order to ensure equal distribution of patient characteristics randomization will be used. Allocation of treatments will be performed using a web-based randomization tool (www.randomizer.at) by means of block-wise randomization. Randomization will be performed intraoperatively at the end of surgery, after closure of the fascia. This prevents potential bias by different intraoperative techniques. Before randomization, the surgeon needs to confirm a clean or clean-contaminated operation according to CDC definition (25). Randomization will be stratified by centre and by IH risk (low vs. high risk patients, defined as patients with BMI ≥ 27 and/or surgery for abdominal aortic aneurysm). The surgeon who will perform the closing technique must be chosen before abdominal wall closure.
Patients, observers, and data analysists will be blinded to the intervention in order to guarantee unbiased assessment of the primary outcome. The person performing randomization and the surgical team conducting the control/experimental intervention (“unblinded” study members) will be documented and will not be part of further outcome assessment. Moreover, neither the operation report nor the discharge letter will contain information regarding group allocation. In any case of an emergency including possible reoperation or a clinical situation that necessitates the knowledge of the trial group of the participant, patients can be unblinded.
Other methods against bias
To minimize performance bias, the intervention will be standardized in both groups and the suture-to-wound length ratio must be recorded intra-operatively and will be monitored. Furthermore, to minimize training effects all participating surgeons must pass an obligatory eLearning tutorial demonstrating the SST before participation in the trial. Only surgeons having performed a minimum of 10 SST abdominal wall closures are allowed to perform interventions in the HULC trial. In addition, only centres committing to include at least 50 patients will participate in the trial.
Primary outcome measure of the trial will be the occurrence of IHs within 24 months after surgery as defined by the European Hernia Society (EHS) (26). Consequently, “any abdominal wall gap with or without a bulge in the area of a postoperative scar perceptible or palpable by clinical examination or imaging” is regarded as an IH. Occurrence of a burst abdomen will not be counted as primary endpoint, but as a secondary endpoint by consensus (1, 3). Follow-up time will be 24 months as has been recommended by the EHS (26) since IH incidence increase over time (15). Patients will be assessed for the primary endpoint at 6, 12, and 24 months after trial intervention. At these time points patients will be examined by a clinician blinded for the trial intervention and by a radiologic examination performed by a blinded assessor. Radiologic exams allowed in the trial are sonography, CT or MRI scans. In case of conflicting results between clinical and radiologic exams, the radiologic imaging is decisive to increase sensitivity (26). If only one of the two examinations is performed (i.e. either clinical or imaging) the result of this assessment will be used for analysis. Possible results are listed in table 1. As many patients included in this trial are expected to have an oncological indication for laparotomy and the included centres perform their oncological follow-up themselves, the loss to follow-up of patients is expected to be low. For patients who are unable or unwilling to attend the follow-up visits a telephone follow-up is incorporated. The patient-reported outcome questionnaire developed by Jairam et al. (27) will be used as it exhibits a high reliability. It will be used as a screening tool, i.e. patients who are suspected to have an IH based on the questionnaire, might be convinced to attend an outpatient visit even if they were reluctant to do so before.
In the recently released addendum to the ICH E9 guideline (draft version) (28), the estimands framework is recommended as clear and transparent definition of “what needs to be estimated to address a specific scientific question of interest”. Such an estimand can be defined through the population of interest, variable of interest, specification of how intercurrent events are handled, and summary measure. The specification of how intercurrent events are handled is referred to as intervention effect in the following.
Population: The population is defined as all patients fulfilling the in- and exclusion criteria.
Variable: The variable is the occurrence of IHs as defined by the European Hernia Society (EHS) within 24 months after intervention.
Intervention effect: Possible intercurrent events and the strategies to handle them are as follows: missing values due to death, drop-out, loss to follow-up and re-laparotomy will be replaced by using multiple imputation. Since re-laparotomy changes the probability of occurrence of an IH, information of occurrence or non-occurrence of IH after re-laparotomy will not be considered for primary analysis. This represents a hypothetical strategy for the post randomization events re-laparotomy, drop-out, loss to follow-up, and death. Besides these events, other post randomization events will not be considered, thus reflecting a treatment policy approach, which means that the effect of randomized treatment is estimated irrespectively of other post-randomization events not captured in the primary endpoint definition.
Summary Measure: The summary measure is the odds ratio. The odds ratio will be calculated by a two-level binary logistic regression analysis including the fixed factors treatment group and IH risk (low vs. high), the latter being deemed as the by far most important confounder and being also used for stratification in the randomization procedure, and the random factor centre. Confounding by other less important prognostic and predictive factors can assumed to be controlled by the randomized study design. The model will be fitted using the variance-components covariance matrix. The level of significance is set to 5% (two-sided). The p-value for judging the primary hypothesis will result from the two-level binary logistic regression model, where the coefficient of the factor treatment effect is tested against zero using the Wald test.
Additionally, sensitivity and supplementary estimands will be considered, but not described in further detail in this publication.
Key secondary endpoints
The secondary measurements chosen in the HULC trial have been proposed by international guidelines (26). For an adequate evaluation of the secondary endpoints follow-up visits on postoperative day 5 to 7, 10 to 14 and 30 to 35 will be performed (see table 2) in addition to the follow-up visits for the primary endpoint described above (at 6, 12 and 24 months postoperatively). The key secondary endpoints are as follows:
- Risk of superficial and deep surgical site infections (SSIs) within one year in both groups (25). SSIs will be assessed by clinical examination as defined by the CDC (25). Organ-space SSIs are excluded in this measurement as they are independent of abdominal wall closure technique, but rather depend on the underlying surgery. Consequently, organ-space SSIs will be recorded in the overall complication rate and as serious adverse events if applicable. Follow-up for SSI is one year in line with CDC guidelines as patients in the experimental group undergo implantation of alloplastic material (mesh).
- Postoperative 30-day morbidity. Complications will be recorded and classified according to the Dindo-Clavien classification (29).
- Occurrence of non-infectious wound complications (hematoma, seroma) within 30 days. Seroma is defined as a collection of serous fluid in a dead space, which can either be in situ or leaking through a wound. Hematoma is defined as an accumulation of blood in the wound area, which warrants (bedside) surgical exploration and intervention.
- Occurrence of postoperative burst abdomen within 30 days. Postoperative burst abdomen will be defined as missing continuity of the fascia in combination with wound dehiscence with consecutive re-operation.
- Postoperative wound pain at rest and during movement. Assessment will be performed using the well-established numeric pain rating scale. Pain is an important patient-reported outcome measure and is influenced by hernia occurrence and by surgery. Pain will be assessed as visits 3-5 (see table 2) and the mean pain will be compared between the groups at these time points.
- HRQoL measured with the SF-36 and EQ-5D questionnaires. As HRQoL is an important patient-reported outcome measure and is influenced by hernia occurrence and by surgery (2), it will be recorded both preoperatively (visit 1) and during follow-up (visit 6-8) (see table 2). The median HRQoL will be compared at these time points between the groups as well as the change from baseline.
- Length of primary hospital stay in days from index operation.
Patient timeline and trial visits
Patients scheduled for elective abdominal surgery via a midline incision are screened preoperatively at day 0 (visit 1). Patients are enrolled given their ability to understand the extent and nature of the trial as well as their written informed consent after detailed patient information. All inclusion criteria and no exclusion criteria must be fulfilled. Baseline data are collected during the screening/baseline visit. The duration of visit 1 will be approximately 25 minutes. Included patients are randomized during surgery (visit 2) after closure of the fascia in SST. Follow-up visits will be on postoperative day 5 to 7, 10 to 14 and 30 to 35 (visits 3-5) for evaluation of secondary endpoints (time expenditure approximately 15 minutes). In addition, 6, 12 and 24 months (visits 6-8) after surgery patients are planned for follow-up to evaluate primary and secondary outcome parameters. The expenditure of time for each visit will be approximately 30 minutes per patient. An overview of trial visits and items captured during the trial visits is presented in table 2 according to the guidelines of the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) (30).
All protocol-required information collected during the trial must be entered by the investigator, or designated representative, in an electronic case report form (eCRF) implemented in the REDCapTM system (31) (www.project-redcap.org). An explanation should be given for all missing data. The completed eCRF must be reviewed and signed by the investigator named in the trial protocol or by a designated sub-investigator. The Institute of Medical Biometry and Informatics of the University of Heidelberg (IMBI) is responsible for the data management within the trial. To assure a safe and secure environment for data acquired, data transmission is encrypted with secure socket layer (SSL) technology. Only authorized users are able to enter or edit data, the access is restricted to patients’ data in the respective centre. All changes to data are logged with a computerized timestamp in an audit trail. All data will be pseudonymized. Completeness, validity and plausibility of data will be checked in time of data entry (edit-checks) and using validating programs, which will generate queries. If no further corrections are to be made in the database, eCRF data will be locked. All data management procedures will be conducted according to written defined standard operating procedures (SOPs) of the IMBI that guarantee an efficient conduct complying with good clinical practice (GCP).
Sample size calculation
The sample size calculation is based on the primary efficacy endpoint (IH risk) within 24 months after surgery. Based on the assumption that the percentage of patients developing an IH after midline laparotomy in a general surgical population closed with the SST is approximately 15% for the control group, we hypothesize a reduction of 7% in the intervention arm based on previous RCTs (13, 20). Consequently, a sample size per group of 325 patients is needed for the between-group comparison by the Chi-squared test to achieve 80% power in detecting this difference in IH risk at a two-sided level of significance of 5%. It is assumed that using a two-level logistic regression model adjusting for the random factor centre and the fixed factor IH risk (low risk vs. high risk patients: BMI ≥ 27 and/or surgery for abdominal aortic aneurysm) (20) in the primary analysis will lead to less unexplained variance and thus to an increased power. Assuming a drop-out rate of up to 20% based on previous trials (1, 3, 10, 18), a total of 812 patients (406 per group) will be randomized in the study (figure 1). The potential occurrence of missing values for the primary outcome is partially addressed by the predefined multiple imputation strategy. Sample size calculation was performed using ADDPLAN v6.1.
Analysis variables and statistical methods
The primary efficacy analysis will be based on the full analysis set (FAS) built according to the intention-to-treat (ITT) principle thus reflecting the recommendations given in guidelines (32). As a sensitivity analysis, an evaluation based on the per-protocol (PP) population (based on those patients without major protocol violation and excluding patients that receive a fascial closure not predefined in the randomization scheme) will be performed. The risk of IH will be analysed via a two-level binary logistic regression model including the fixed factors treatment group and IH risk (high vs. low), the latter being deemed as the by far most important confounder and being also used for stratification in the randomization procedure, and the random factor centre. Confounding by other less important prognostic and predictive factors can assumed to be controlled by the randomized study design. The model will be fitted using the variance-components covariance matrix. The level of significance is set to 5% (two-sided). All secondary outcomes will be evaluated descriptively, and descriptive p-values will be reported together with 95% confidence intervals for the corresponding effects.
Interim analyses during the trial are not predefined or planned but depending on the frequency of serious adverse events (SAEs) in both groups, safety analyses can be performed as recommended by the independent data safety and monitoring board. Further sensitivity analyses will be performed with the per protocol set and the results will be compared with those of the ITT analysis. Moreover, for missing data in the ITT population set further sensitivity analyses will be conducted by a worst-case scenario for the intervention, a minimal and a maximal IH risk imputation and by another alternative method of dealing with missing data as described by Higgins et al. (33). Furthermore, a time-to-event analysis for the outcome “time from randomization to occurrence of IH” will be performed in the ITT population according to Kaplan-Meier. Additionally, pre-specified subgroup analyses will be performed in the ITT population for the risk of incisional hernias in the subgroups of different types of surgery (colorectal, small bowel, hepatobiliary-pancreatic, upper GI (oesophageal and gastric), vascular, others), adipose vs. non-adipose patients and the presence or absence of neoadjuvant therapy, previous laparotomy or chronic obstructive pulmonary disease. All secondary outcomes will be evaluated descriptively, and descriptive p-values will be reported together with 95% confidence intervals for the corresponding effects.
The assessment of safety will be based on the frequency of SAEs in both groups, which will be analysed via descriptive statistical methods in the study population. For comparisons of frequencies between groups the Chi-squared test will be used. All analyses will be done using SAS version 9.4 or higher.
Clinical data monitoring
Clinical monitoring will be performed by independent monitors of the SDGC according to its standard operating procedures in line with the ICH-GCP guideline (E6) (34). A risk-based monitoring strategy will be conducted based on patient safety, patient rights, protocol adherence and data. The frequency of monitoring visits will be determined depending on recruitment numbers and individual performance of each centre based on feedback from project and data management.
Premature closure of the trial
The trial may be prematurely closed by the coordinating investigator in consultation with the Steering Committee including the responsible biometrician. If the termination of the trial becomes necessary, the Steering Committee of the trial will discuss this issue with the independent Data Safety Monitoring Board (DSMB). Similarly, the DSMB can recommend closing the trial based on the safety reports; however, the decision remains with the Steering Committee. Reasons that may necessitate the termination of the trial include the following: The incidence or severity of SAEs, morbidity or complications in this trial indicate a potential health hazard caused by the study treatment. Furthermore, the trial should be terminated if it appears that patients’ enrolment is unsatisfactory with respect to quality and/or quantity or data recording is severely inaccurate and/or incomplete. Another case in which termination of the trial is necessary is, if external evidence demands a termination of the trial.