Design

The systematic review was performed in accordance to the PRISMA guidelines (12) and the protocol registered with the PROSPERO database (13). Two independent authors (MJ, SG) performed all aspects of the literature retrieval and quality assessment. When results were compared, disagreement was reviewed by a third independent author (JH).

Included Study Designs

Included study designs were randomised controlled trials (RCTs), case-control and retrospective cohort studies of patients undergoing EL. Excluded studies included those conducted in an elective setting, studies with cohorts of emergency abdominal surgery patients where EL was not the sole procedure investigated, review studies and case-reports. Non-English language studies were excluded.

EL was the only procedure of interest and the definition for eligibility was consistent with that of National Confidential Enquiry into Patient Outcome and Death (NCEPOD); “An expedited, urgent or emergency unscheduled abdominal surgery via a midline abdominal incision or laparoscopic approach including laparoscopic converted to open or laparoscopic assisted procedures.” This excluded elective procedures and surgeries defined as uncomplicated cholecystectomy, pancreatectomy and appendectomy, as well as gynaecological laparotomy, vascular surgery, and organ transplantation (14, 15). The inclusion criteria was therefore patients aged over 16 undergoing an EL.

Definition of exposure

The diagnosis of sarcopenia compared to no sarcopenia was considered the exposure. Diagnosis was assessed using measurement of psoas muscle area (PMA) at level L3 or L4 on preoperative CT scan.

Search strategy and study selection

Literature searches were conducted across five multi-disciplinary electronic databases: MEDLINE (Ovid), EMBASE (Ovid), CINAHL (Embase), Cochrane Central Register of Controlled Trials (CENTRAL), and Web of Science (Thompson Reuters). The last search was conducted on the 14th February 2022. Databases were searched using appropriate Medical Subject Headings (MeSH) and keyword searches. Terms relating to ‘sarcopenia’, ‘emergency laparotomy’ and ‘association’ allowed us to utilize a concept-based approach. All types of emergency abdominal surgery were incorporated into our initial search terms ensuring the search was sufficiently comprehensive. In the same way, we deemed it inappropriate to use a filtering term for mortality or morbidity at the risk of excluding relevant literature associated with other relevant outcomes. The full search strategy conducted in Ovid MEDLINE is shown in Appendix. 1.

Citation lists of all included studies and relevant reviews were investigated alongside the contents lists of relevant journals within the field; Age and Ageing and British Journal of Surgery (BJS). Once duplicates were removed, the retrieved studies underwent title and abstract screening to identify studies to undergo full-text review. Here the inclusion and exclusion criteria were applied, and the articles were selected for inclusion.

Risk of Bias (RoB) and Quality Assessment (QA)

Included studies were quality assessed using the Newcastle-Ottawa Scale (NOS) (16). This assessment tool for observational studies evaluates quality over the three domains of selection, comparability and either exposure (case-control) or outcome (cohort); overall comprising eight distinct categories. We considered each domain separately to conclude whether studies were of good, fair or poor quality. A risk of bias table was produced to illustrate the implementation of the NOS to the included studies, see Appendix 2.

Included Studies Characteristics

Information relating to the characteristics of the included studies were extracted by both authors. This included data regarding the country of population origin, journal of publication, study design, sample size, sarcopenia criteria applied, mortality assessment timeframe, mortality rates, morbidity outcomes identified and associated data, and characteristics of populations where available (male: female ratio and age range).

Outcomes

The primary outcome was mortality following the EL procedure at 30, or 90-days.

Secondary outcomes were: LOS and major complications following EL reported as ≥IIIb on the Clavien-Dindo Classification (CDC) - a widely adopted measure of surgical outcomes, categorising outcomes based on the type of therapy required to treat the complication (18).

Data Synthesis & Analysis

Due to the contextual heterogeneity of non-randomised designs, each study was described narratively (Table. 1) in terms of their design. Only studies that were clinically homogeneous with respect to population and sarcopenic definition were considered for pooling into meta-analysis (18). Meta-analysis was undertaken when two or more studies examined the same outcome. Outcome variables that were categorical and dichotomous in nature, were pooled as estimated odd ratios (ORs). LOS was summarised as a mean difference (MD) between those with sarcopenia and those without. Pooled results were estimated using the software package Review Manager 5.3; the Mantel-Haenszel estimator calculated relative risks (RR) and the inverse-generic variance method calculated pooled MD estimates. All meta-analyses were subsequently represented visually using forest plots with 95% confidence intervals (CIs).

Table 1 Main characteristics of included studiesStatistical heterogeneity and Subgroup analysis

Statistical heterogeneity was evaluated amongst the studies by visual inspection and calculation of the Higgins I2 statistic. Heterogeneity measured above 75% was considered high and would undergo further subgroup analysis. Subgroups used to explore heterogeneity were: study design (surgical procedure of interest), median age, and sarcopenic definition utilised in study. Repeating the primary analysis when we removed one study at a time allowed us to further assess the robustness of our results (18).