Introduction
The term ‘sarcopenia’ was introduced by Rosenberg in 1997 to describe an age-related decrease in skeletal muscle mass [
1]. According to the recent recommendation of the European Working Group on Sarcopenia in Older People (EWGSOP), sarcopenia can be diagnosed if low lean muscle mass stands together with either low muscle strength or low physical performance [
2]. In clinical practice, sarcopenia often remains unrecognized if only body mass index (BMI) is used to determine nutritional status [
3]. Several studies have shown that BMI does not predict low lean muscle mass accurately and the entity of sarcopenic obesity also exists [
3‐
5].
Sarcopenia can be assessed anatomically and functionally. Anatomically, a variety of methods and measures are used to evaluate muscle mass, including computed tomography (CT), magnetic resonance imaging (MRI), bioelectrical impedance analysis (BIA), dual-energy X-ray absorptiometry (DXA) and skeletal muscle index (SMI), appendicular SMI (ASMI), and total psoas muscle area (TPA) [
2]. However, there is a considerable variance in the cut-off thresholds for sarcopenia in these diagnostic modalities, while the characteristics of the (normal) reference populations also vary [
6]. Functionally, the handgrip strength test with a standard dynamometer or a physical performance test such as that using the Short Physical Performance Battery are the gold standards [
2]. Sarcopenia in the elderly is often related to adverse health outcomes including a higher risk of hospitalization and mortality with associated increased healthcare costs [
7,
8]. Recently, sarcopenia was implicated as a prognostic factor in a wide range of diseases, such as cancer [
9,
10], chronic liver diseases [
11], chronic pancreatitis [
12], rheumatic diseases [
13] and inflammatory bowel disease (IBD) [
14].
Crohn’s disease (CD) and ulcerative colitis (UC) are the two main forms of IBD, and both are characterized by chronic inflammation of the digestive tract. Bryant et al. reported low lean muscle mass and sarcopenia in 21% and 12% of adult IBD patients, respectively [
3]. In a recent systematic review, the incidence of sarcopenia was as high as 52% in CD and 37% in UC, when anatomical criteria were considered without functional strength assessment [
14]. Undesired consequences of the altered body composition in IBD include bone demineralization (osteopenia and osteoporosis), inadequate response to therapy, and poor quality of life [
3,
15]. Despite the associations supporting the adverse effect of sarcopenia in many diseases other than IBD, there have been only a few low-volume trials addressing this problem. Thus, we performed this meta-analysis to summarize and synthesize the results of the most up-to-date literature investigating the effect of sarcopenia, as a prognostic factor, on the need for disease-related surgery and on the characteristics of postoperative complications in patients with IBD.
Methods
This meta-analysis was reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement (Supplementary Table 1) [
16]. The protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) a priori under registration number CRD42018118517.
Data sources and search strategy
Our search was conducted in four electronic databases using PubMed (
http://www.ncbi.nlm.nih.gov/pubmed), EMBASE (
https://www.embase.com), Central Cochrane Register of Controlled Trials (CENTRAL) (
http://www.cochranelibrary.com) and Web of Science (
www.webofknowledge.com), last updated 13 March 2019. ‘English-language’ and ‘human’ filters were applied to the search. A manual search was also done by browsing the reference lists of relevant papers and review articles to identify additional studies. The PECO items of our prognostic meta-analysis were as follows: (P) adult patients with IBD, with available body composition assessment results; (E) those who had diagnosed sarcopenia; and (C) those who did not have sarcopenia. Our outcomes (O) included the number of surgical interventions and postoperative complications. The standardized Clavien-Dindo classification tool was applied to categorize minor and major postoperative complications [
17]. Major complications were defined as grade ≥ III on the Clavien-Dindo scale. Studies were identified by entering (
‘body composition’ OR sarcopenia) AND (inflammatory bowel disease or Crohn or ‘ ulcerative colitis’) combining Medical Subject Headings (MeSH) and free-text terms.
Study selection
After importing all references into a reference management software (EndNote X8, Clarivate Analytics, Philadelphia, PA, US), duplicates and database overlaps were removed by one of the authors (AE). To maximize the precision of selection, the remaining records were screened based on title and abstract, independently, by two of the authors (AE and PS). Finally, the same two authors verified whether the remaining full-text articles or abstracts truly fit the inclusion criteria. If there was disagreement at any stage of the selection, the opinion of a third author (PH) was sought to reach a consensus. English-language full-text articles and conference proceedings were eligible for inclusion if they met our inclusion criteria. Uncontrolled studies were excluded.
Data extraction and quality assessment
From the included studies, two authors extracted the data according to a predefined data abstraction form: first author, year and form of publication (full-text article/abstract only), study design, sample size, and outcome (rate of patients requiring IBD-related surgery and postoperative complications). The definition of sarcopenia and the type of surgery were also recorded (Table
1). Adjusted results generated from, and covariates imputed in, multivariate logistic regression models were also collected (Supplementary Table 2).
Table 1
Study characteristics and measured outcomes
| 90 (76/14) | Men: SMI < 52.4 cm2/m2 Women: SMI < 38.5 cm2/m2 | Previous surgery, not specified | Sarcopenic (41) | 21 | NA | NA | NA |
Non-sarcopenic (49) | 19 | NA | NA | NA |
| 72 (43/29) | Men: SMI < 42 cm2/m2 Women: SMI < 38 cm2/m2 | Intestinal resection, not specified | Sarcopenic (30) | 16 | NA | NA | NA |
Non-sarcopenic (42) | 9 | NA | NA | NA |
| 58 (58/0) | Men: SMI < 52.4 cm2/m2 Women: SMI < 38.5 cm2/m2 | Surgical resection, not specified | Sarcopenic (24) | 7 | 7 | NA | NA |
Non-sarcopenic (34) | 17 | 2 |
| 82 (0/82) | Men: SMI < 55 cm2/m2 Women: SMI < 39 cm2/m2 | Colectomy | Sarcopenic (57) | 16 | NA | NA | NA |
Non-sarcopenic (25) | 3 | NA | NA | NA |
| 69 (0/69) | Men: TPA < 567.4 mm2/m2 Women: TPA < 355.8 mm2/m2 | Two- or three-stage ileal J-pouch–anal anastomosis (IPPA) | Sarcopenic (18) | NA | 8 | 4 | 4 |
Non-sarcopenic (51) | NA | 5 | 3 | 2 |
| 77 (52/21)b | In case of BMI < 25 kg/m2: men: SMI < 43 cm2/m2 women: SMI < 41 cm2/m2; In case of BMI ≥ 25 kg/m2: men: SMI < 53 cm2/m2 | IPPA, ileocecectomy, hemicolectomy, colectomy, panproctocolectomy with or without ileostomy formation | Sarcopenic (30) | 77 | 6 | NA | 6 |
Non-sarcopenic (47) | 10 | 10 |
| 79 (79/0) | Men: SMI < 55 cm2/m2 Women: SMI < 39 cm2/m2 | Surgery, not specified | Sarcopenic (64) | 11 | NA | NA | NA |
Non-sarcopenic (15) | 2 | NA | NA | NA |
| 149 (149/0) | Men: SMI < 55.4 cm2/m2 Men: SMI < 38.9 cm2/m2 | Stricturoplasty, bowel resection, stoma without resection, perianal/abdominal abscess drainage, viscerolysis | Sarcopenic (50) | OR: 2.03a, (CI 0.98–4.26), p = 0.056 | NA | NA | NA |
Non-sarcopenic (99) | NA | NA | NA |
| 114 (114/0) | Men: SMI < 55 cm2/m2 Women: SMI < 39 cm2/m2 | Segmental/total colectomy, ileocecal/small bowel resection | Sarcopenic (70) | NA | 32 | 21 | 11 |
Non-sarcopenic (44) | NA | 26 | 25 | 1 |
| 99 (0/99) | Men: SMI < 55 cm2/m2 Women: SMI < 39 cm2/m2 | Colectomy | Sarcopenic (27) | 7 | NA | NA | NA |
Non-sarcopenic (72) | 7 | NA | NA | NA |
We also collected data on patient characteristics such as age, sex, disease duration, BMI, smoking, prior use of immunomodulatory or biological therapies, and preoperatively measured laboratory parameters including hemoglobin (Hb), serum albumin level, C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) (Table
2). Two independent investigators (AE and PS) performed the quality assessment separately and any disagreements were resolved by discussion. A critical appraisal tool for prognostic studies, the Quality in Prognosis Studies (QUIPS), was used to assess the methodological quality of the studies included [
18]. QUIPS covers six main domains, namely, study participation, study attrition, prognostic factor and outcome measurement, study confounding and statistical analysis, and reporting. For each item of the six domains, we used ‘yes’, ‘no’, or ‘unclear’ to assess the risk of bias. Each domain was then judged as carrying ‘low’, ‘moderate’ or ‘high’ risk of bias (Supplementary Table 3).
Table 2
Patient characteristics and preoperative laboratory parameters of the sarcopenic and non-sarcopenic study groups
| Sarcopenic (41) | 35 (26, 50)a | 28 (68) | 4 (2, 13)a | 19 (18, 24)a | 11 (27) | NA | 9 (20) | NA | 3.8 (3.5, 4.2)a | 16.8 (5.1, 37.4)a | 25 (12, 49)a |
Non-sarcopenic (49) | 35 (26, 50)a | 10 (20) | 6 (2, 12)a | 24 (22, 30)a | 9 (18) | NA | 6 (15) | NA | 4.1 (3.8, 4.3)a | 2.9 (1, 13.3)a | 20 (10, 35)a |
| Sarcopenic (24) | NA | 13 (54.2) | NA | NA | NA | 19 (79.2) | 15 (62.5) | NA | 3.49 ± 0.60b | NA | NA |
Non-sarcopenic (34) | NA | 14 (41.2) | NA | NA | NA | 26 (76.5) | 19 (55.9) | NA | 3.52 ± 0.60b | NA | NA |
| Sarcopenic (57) | 41 ± 28b | 43 (75.4) | 2 ± 7b | 23 ± 6b | 2 (12.5) | 20 (35.1) | 17 (29.8) | 11.4 ± 2.5b | 3.4 ± 0.6b | 50 ± 76b | 37 ± 42b |
Non-sarcopenic (25) | 32 ± 16b | 10 (40) | 3.5 ± 4.5b | 26 ± 8b | 1 (11.1) | 13 (52) | 9 (36) | 11.3 ± 2.4b | 3.5 ± 0.6b | 34 ± 63.2b | 42.5 ± 29.1b |
| Sarcopenic (18) | 36.0 ± 17.4b | 12 (66.6) | 7.2 ± 7.9b, g | 18.0 ± 2.9b | NA | NA | NA | NA | 3.72 ± 0.59b | 16 ± 25b | NA |
Non-sarcopenic (51) | 41.2 ± 13.4b | 33 (64.7) | 8.4 ± 7.3b, g | 21.3 ± 3.5b | NA | NA | NA | NA | 3.87 ± 0.47b | 7.0 ± 15b | NA |
| Sarcopenic (30) | 43 (20–80)c | 15 (50) | NA | 21 (16–37)f | 6 (20) | 20 (67) | NA | 11.0 (9.8–12)d 10.26 (3.7–19.6)e | 3.35 (1.7–5.1)c | 63.26 (0–547)c | NA |
Non-sarcopenic (47) | 41 (21–74)c | 31 (66) | NA | 24 (19–33)f | 7 (15) | 23 (49) | NA | 12.6 (10.5–13.7)d 12.8 (11.4–13.8)e | 3.53 (2.0–4.8)c | 46.16 (0–374.7)c | NA |
| Sarcopenic (70) | 28.8 ± 10.2b | 55 (78.6) | NA | 17.1 ± 2.9b | NA | NA | NA | 10.7 ± 2.1b | 3.6 ± 0.5b | 29.1 ± 54.1b | NA |
Non-sarcopenic (44) | 37.1 ± 11.6b | 20 (45.5) | NA | 19.1 ± 1.9b | NA | NA | NA | 11.2 ± 1.7b | 3.8 ± 0.3b | 17.2 ± 37.4b | NA |
Statistical analysis
The Comprehensive MetaAnalysis software Version 3 (Biostat, Inc., Englewood, NJ, USA) was applied to perform meta-analytical calculations with the random effects model [
19]. We computed relative measures [odds ratios (ORs) and weighted mean differences (WMDs)] and event rates with 95% confidence intervals (CIs). The main outcomes, namely, surgical interventions and postoperative complications, were handled as binary variables. First, ORs from raw 2 × 2 contingency tables were pooled [
19]. Peto’s OR was calculated in the case of the study of Carvalho et al. due to rare events [
20]. In addition, covariate-adjusted ORs computed with multivariable logistic regression models in the individual studies were pooled [
19]. For numerical variables, namely, age, disease duration, BMI, and laboratory parameters, WMDs were calculated. For differences regarding sex, smoking, prior immunomodulator and biologics use between groups, event rates were calculated.
The results of our statistical analysis are shown in tables and forest plots. All analyses were two-tailed and
p < 0.05 was considered significant. For assessing heterogeneity, Cochrane’s
Q and the
I2 statistics were used. In the case of the
Q statistic,
Q exceeds the upper-tail critical value of Chi-square with the
k − 1 degree of freedom.
I2 represents the percentage of effect size heterogeneity, which cannot be explained by random chance. Based on the cochrane handbook for systematic reviews of interventions, heterogeneity was interpreted as moderate if it was between 30 and 60%, as substantial if it was between 50 and 90%, and as considerable if it was above 75% [
20]. Publication bias was assessed by the visual inspection of the funnel plot, which was complemented with the Egger’s test for analysis of the need for surgical interventions [
20].
Discussion
Sarcopenia is relevant for patients with IBD as it can lead to poor outcomes, such as bone demineralization with consequential pathological fractures, hospitalization, reduced mobility, and compromised quality of life [
31]. Apart from the increased level of pro-inflammatory mediators, inadequate calorie intake, malabsorption, and protein-losing enteropathy, different pharmacological and surgical treatments may also impair nutritional status in IBD [
15]. However, few studies have evaluated the effect of sarcopenia as a prognostic factor of surgical outcomes.
Our meta-analysis found only seven studies with controversial results on surgical interventions for sarcopenic patients with IBD. Although we did not detect a significant difference between the rate of surgical interventions for the sarcopenic patients versus the non-sarcopenic patients with IBD when unadjusted data were pooled, our analysis of covariate-adjusted data identified sarcopenia as an independent predictor of surgical interventions in patients with IBD. A potential explanation for this phenomenon could be the effect of bias (especially selection bias), which masked the true difference in the unadjusted analysis (Figs.
2 and
5), but was reduced, at least partly, with the introduction of multivariate models (Figs.
3 and
6). Regarding IBD subtype (CD or UC), no significant difference was observed in surgical interventions, although there was a trend towards a more frequent need for surgery in sarcopenic patients with UC.
Similarly, we could only detect significant differences between the groups with respect to postoperative complications when adjusted data were analyzed. This finding is consistent with previous meta-analyses on patients who underwent oncological abdominal surgery, where radiologically proven sarcopenia was associated with a significant increase in major postoperative complications as well as in 30-day mortality [
32,
33]. In our meta-analysis, most of the sarcopenic patients were men. Interestingly, a previous study identified sarcopenia as a predictor of a subsequent surgery only in women with IBD [
34]. Other clinical features such as age, disease duration, smoking, and prior therapies did not differ considerably between the groups. We also found lower BMI and serum albumin levels and higher CRP levels in the sarcopenic group. It should be noted that measuring these parameters can be helpful in the prediction of sarcopenia. Since cross-sectional imaging is frequently ordered preoperatively, the results of CT or MRI scans would provide a more accurate estimation of lean muscle mass.
Although the consequences of malnutrition, such as poor bone health, delayed puberty and growth failure, are more prevalent in pediatric patients with IBD, our meta-analysis highlighted that the assessment of body composition is also necessary in adult patients with IBD [
35]. The importance of evaluating the nutritional status of an IBD patient was emphasized in the recent European Crohn’s and Colitis Organisation (ECCO) guidelines [
36,
37]. Malnutrition and hypalbuminaemia are listed as the main risk factors of postoperative complications, including anastomotic leak, peritonitis, and intra-abdominal septic complications [
36]. Prior to surgery, not only the responsible adjustment of medical therapy, such as weaning off steroid treatment if possible, but also preoperative enteral or parenteral nutritional support may help to reduce the risk of surgical and postoperative complications [
38‐
42].
Several limitations of this meta-analysis must be considered. First, the vast majority of included studies were retrospective with a small sample size and a wide variety of surgical interventions, raising concerns about imprecision and indirectness. Second, none of the included studies used EWGSOP criteria to assess functional loss of muscle strength measurement. Third, only one study assessed the effect of nutritional therapy, in which preoperative enteral nutrition was a protective factor against major postoperative complications [
30]. Furthermore, minor differences were observed with respect to the covariates imputed in the logistic regression models, which may affect our results (Supplementary Table 2). As heterogeneity tests indicated homogeneous datasets (Figs.
3 and
6), we do not suspect rough distortion. Finally, there was considerable heterogeneity in cut-off points for the sarcopenia definition regarding ethnicity, carrying the potential of under- or overestimating the sarcopenia rate in the included studies.
The main strength of our work is its novelty, but we must mention the foremost systematic review by Ryan et al. [
14] investigating the prognostic role of sarcopenia in the outcomes of surgery for IBD. Our review revealed the need for longitudinal observational studies in this field. The highly transparent and reproducible methodology of this work was ensured by strictly adhering to the rules and recommendations of the PRISMA guidelines.
In conclusion, the findings of our analysis have implications for practice, particularly in the promotion of preoperative individualized risk prediction. In addition to simple anthropometric tests, anatomical and functional measurements should be performed. The SMI, measured on a CT scan, can be used as an objective assessment tool to identify sarcopenia in patients with IBD. To interpret the body composition and nutritional status of patients with IBD, a multidisciplinary approach is recommended. Education on nutritional issues is best provided by well-trained dietitians with a special interest in IBD. Since sarcopenia may be reversible with adequate nutritional support, targeted preoperative risk reduction strategies are recommended to optimize surgical outcomes. Further research through large prospective cohort studies is needed to confirm our findings and conclusions.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.