Clinical Nutrition ESPEN 50 (2022) 148e154

Contents lists available at ScienceDirect

Clinical Nutrition ESPEN

journal homepage: http://www.clinicalnutritionespen.com

Original article

Evaluation of malnutrition in patients undergoing major abdominal

surgery using GLIM criteria and comparing CT and BIA for muscle
mass measurement
Maria Wobith a, Christian Herbst b, Markus Lurz c, Delia Haberzettl a, Martin Fischer a,
Arved Weimann a, *
a
Klinikum St. Georg gGmbH Leipzig, Department of General, Visceral, and Oncological Surgery, Clinical Nutrition Unit, Germany
b
Klinikum St. Georg gGmbH Leipzig, Department of Cardiology, Germany
c
Klinikum St. Georg gGmbH Leipzig, Department of Radiology, Germany

a r t i c l e i n f o s u m m a r y

Article history: Background & aims: GLIM criteria have become a worldwide standard for diagnosing malnutrition. They
Received 2 June 2022 emphasize the measurement of muscle mass but do not provide clear recommendations for the use of
Accepted 6 June 2022 different diagnostic tools and cut-offs. Measurements of body composition by using computerized to-
mography (CT) and bioelectrical impedance analysis (BIA) are both easily accessible in hospitalized
Keywords: patients. However, there is sparse data regarding the comparison for GLIM diagnosis of malnutrition and
GLIM criteria
its prognostic impact for postoperative outcome in patients undergoing major abdominal surgery for
Malnutrition
cancer.
Sarcopenia
Skeletal muscle mass
Methods: We retrospectively analysed 260 patients undergoing major abdominal surgery between
Major abdominal surgery January 2017 and December 2019. Patients were prospectively screened and assessed for malnutrition
with Nutritional Risk Score (NRS) and Subjective Global assessment (SGA). Body composition was ana-
lysed with CT scan and BIA within 30 days before surgery. GLIM criteria were retrospectively determined
referring to the Fat free Mass from BIA (FFMBIA) and Muscle Mass from axial CT scan at lumbar level 3
(MMCT). The prevalence of GLIM - malnutrition according to BIA and CT was evaluated. Multivariate
logistic regression analysis was used to determine association between malnutrition and outcome pa-
rameters. ROC-curves specified sensitivity and specificity of the different tools and areas under the curve
were calculated.
Results: From 260 patients in total, 179 patients (68.8%) had a confirmed malnutrition according to
MMCT, 178 patients (68.5%) were malnourished according to SGA (grade B or C), whereas 66 patients
(25.4%) were diagnosed with malnutrition using FFMBIA. The risk for developing a complication was
significant associated with both methods, FFMBIA (OR 2.116, 95% CI 1.185e3.778, p ¼ 0.01) and MMCT (OR
2.028, 95% CI 1..188e3.463, p ¼ 0.009). Sensitivity for the prediction of overall complications was: MMCT
76.3%, FFMBIA 31.9%, and SGA 73.3%; specificity: MMCT 40.0%, FFMBIA 81.6%, and SGA 36.8%.
Conclusion: When using GLIM criteria, the method for measuring muscle mass is pivotal resulting in
considerable differences in prevalence, sensitivity, and specificity. GLIM criteria are predictive for the risk
of developing complications in patients undergoing major abdominal surgery. With the pre-existing cut-
offs, BIA seems to diagnose patients at an more advanced stage of malnutrition and indicates an
advanced deterioration of nutritional status.
© 2022 European Society for Clinical Nutrition and Metabolism. Published by Elsevier Ltd. All rights
reserved.

1. Introduction
* Corresponding author. Department of General, Visceral, and Oncological Sur-
gery, Klinikum St. Georg gGmbH Leipzig, Delitzscher Str. 129, 04129 Leipzig, Despite new surgical techniques and improved patient care
Germany.
concepts, patients undergoing major abdominal surgery are at risk
E-mail addresses: maria.wobith@sanktgeorg.de (M. Wobith), delia.wirth@
sanktgeorg.de (D. Haberzettl), arved.weimann@sanktgeorg.de (A. Weimann).
to develop perioperative complications with a postoperative

https://doi.org/10.1016/j.clnesp.2022.06.004
2405-4577/© 2022 European Society for Clinical Nutrition and Metabolism. Published by Elsevier Ltd. All rights reserved.
M. Wobith, C. Herbst, M. Lurz et al. Clinical Nutrition ESPEN 50 (2022) 148e154

functional decline and a worsened short and long term outcome. 2. Methods
Thus, there is a focus on modifiable risk factors to improve outcome,
survival, and quality of life. Malnutrition is an important prognostic 260 patients undergoing major abdominal cancer surgery be-
factor in surgical cancer patients with a prevalence of up to 80%, tween January 2017 and December 2019 were analysed retro-
especially when the cancer is localised in the gastrointestinal tract spectively with prospectively collected data. The definition of major
[1]. Increasing prevalence of overweight and obesity in the popu- abdominal surgery included oesophageal, gastric, pancreatic, liver,
lation can mask malnutrition or sarcopenia in patients with cancer colon, or rectal cancer resection. All patients had been screened and
[2,3]. Therefore, a standardized screening for malnutrition is assessed according to Nutritional Risk Score (NRS) [13] and Sub-
necessary. Most screening and diagnostic tools, however, do not jective Global Assessment (SGA) [14] at the time of hospital
take into account that skeletal muscle mass, although related to admission. Abdominal CT scan and bioelectrical impedance analysis
cancer-related malnutrition, can be masked by increased fat mass. (BIA) had been performed within 30 days before surgery. The GLIM
Additionally, cancer patients tend to be older, which has also been criteria were evaluated retrospectively, which included etiologic
associated with sarcopenia. The Global Leadership Initiative on (reduced food intake or assimilation, disease burden/inflammatory
Malnutrition (GLIM) recently published their GLIM criteria to condition) and phenotypic criteria (non-volitional weight loss, low
standardize the clinical practice of malnutrition diagnosis. Accord- body mass index, reduced muscle mass) [15]. In our study, all pa-
ing to weight loss and BMI in surgical patients GLIM criteria without tients, irrespective of their NRS scores, were assessed with respect
measurement of muscle mass have shown malnutrition in 35.4% in to the GLIM criteria for malnutrition.
the Norwegian Registry of Gastrointestinal Surgery. Malnourished The thresholds used for the phenotypic criteria were, as rec-
patients had a significantly increased risk for severe postoperative ommended, for moderate malnutrition: weight loss of 5e10%
complications and mortality within 30 days (1.29, 95% CI: 1.13e1.47 within the past 6 months or 10e20% beyond 6 months, or a
and 2.15, 95% CI: 1.27e3.65, respectively) [4]. BMI < 20 kg/m2 in patients <70 years and a BMI < 22 kg/m2 in
To the best of our knowledge, data regarding the application of patients 70 years. Severe malnutrition was defined as: weight loss
GLIM criteria in surgical patients with gastrointestinal cancer is of >10% within the past 6 months or >20% beyond 6 months, or a
sparse. In absence of a gold standard for clinical routine, the aims of BMI < 18.5 kg/m2 in patients <70 years and a BMI < 20 kg/m2 in
this study were to compare CT and BIA with respect to 1) assess patients 70 years.
malnutrition in surgical patients for appropriate nutrition therapy For the measurement of muscle mass we relied on routine CT
and 2) to identify patients with an increased risk for a poor clinical which is typically acquired in cancer patients, as well as BIA which
outcome. has been part of our routine nutritional assessment for many years.
Due to a lack of consensus how to measure and define reduced The BIA results were used to calculate the Fat free Mass Index
muscle mass in clinical settings, the GLIM criteria refer to different (FFMBIA) and were evaluated according to the ESPEN definition of
methods and different cut-offs as part of the phenotypic criteria. malnutrition [16], i.e., a FFMI lower than 15 kg/m2 in women and
However, these cut-offs were derived from discrepant studies and a lower than 17 kg/m2 in men. Abdominal CT scan at lumbar level 3
validation has been sparsely investigated in clinical studies. The was used to compute sarcopenia measuring the skeletal muscle
prevalence of malnutrition in cancer patients varied from 72.2 to mass index (SMI) with the National Institutes of Health (NIH)
80% in previous studies depending on how muscle mass was ImageJ software [9]. The total abdominal muscle area including the
assessed [5]. Both, the Subjective Global Assessment (SGA) as well psoas, paraspinal, transversus abdominus, rectus abdominus,
as the GLIM criteria were associated with death at six months, quadratus lumborum, and internal and external oblique muscles
when muscle mass was assessed by using handgrip strength or the was demarcated using predetermined thresholds for the Houns-
Fat free Mass Index (FFMI) [5]. Nutritional decline defined as CT- field units on CT (HU -29 to 150). The SMI was calculated in skeletal
based sarcopenia, myosteatosis, and Patient-Generated (PG)-SGA muscle area/height (cm2/m2). According to previous studies, we
in overweight and obese oncological patients (BMI  25 kg/m2) defined a reduced muscle mass (MMCT) as a SMI < 52,4 cm2/m2 in
predicted reduced survival [6]. Data for outcome parameter in men and < 38.5 cm2/m2 in women [2,17].
surgical cancer patients, where malnutrition and sarcopenia are Malnutrition was classified with respect to all three introduced
important risk factors, is sparse. diagnostic tools: the SGA (i.e., Grade A ¼ no malnutrition, Grade
The GLIM recommends to use dual-energy absorptiometry, BIA, B ¼ moderate malnutrition, Grade C ¼ severe malnutrition), the
CT, or MRI for measuring the muscle mass. However, the locally NRS (i.e., 1e2 ¼ no malnutrition, 3e4 ¼ moderate malnutrition, 
available tools have to also be taken into account. Bioelectrical 5 ¼ severe malnutrition), and the GLIM criteria (as defined before).
impedance analysis (BIA) can be used to determine the body For BIA, the bioelectrical impedance analyzer Nutriguard-M
composition of each patient, especially in the course of the illness. (Data Input GmbH, 60487 Frankfurt, Germany) was used with
The resulting FFMI is a good surrogate parameter for muscle mass 800 mA current and 50 kHz signal. Measurement was performed
[7]. An abdominal CT scan in the portal vein phase is routinely done according to standard protocols in the fasting patient after
in patients with gastrointestinal cancer and can be used to calculate emptying the urine bladder in the morning [18].
skeletal muscle mass. Previous studies showed a good correlation The endpoints were defined as overall complication rate,
between skeletal muscle mass at lumbar level 3 and whole body complication grade classified by the ClavieneDindo scale, presence
muscle mass [8]. Several software solutions are available and partly of a severe complication defined as ClavieneDindo  3a, re-
free of charge, which allow computing body compartments based laparotomy, length of hospital stay, readmission within 31 days
on a CT scan [9]. However, body composition analysis remains to be after discharge, in-hospital mortality, and 6 months mortality.
time consuming which prevents routine application in clinical care.
Plenty of studies showed the impact of a CT-defined sarcopenia on 2.1. Statistical analysis
clinical outcome in surgical patients [10e12]. Little is known as to
how sensitivity and specificity compares across the different All continuous variables were described as median and range,
methods for measuring muscle mass. Furthermore, no cut-off while categorical variables were expressed as frequency and per-
values, neither for SMI nor for FFMI, are available to differentiate centage. The prevalence according to the methods was calculated.
between moderately and severely reduced skeletal muscle mass, as Correlations between every method and the outcome parameters
recommended in the GLIM criteria. overall complication rate, complication grade (ClavieneDindo),
149
M. Wobith, C. Herbst, M. Lurz et al. Clinical Nutrition ESPEN 50 (2022) 148e154

severe complications (ClavieneDindo  3a), re-laparotomy, length maximum of 85 days. In 156 patients (60%) data was available after
of hospital stay, readmission within 31 days, in-hospital mortality, six months. In those patients, 6 months mortality was 12.2%. The
and 6 months mortality were calculated. To explore univariate clinical outcome parameters and prevalence of malnourished and
associations in the distribution of categorical data, the c2 test or well-nourished patients according to the used tool are summarized
Fisher's exact test was used; for metric variables the Pearson's in Table 2.
coefficient was used when normal distribution and linearity was The diagnosis of a malnutrition according to the GLIM criteria
given, otherwise the Spearmans coefficient was used. To deter- was related to the method of muscle measurement. Whereas the
mine the diagnostic concordance between the diagnostic tools, minority of patients were malnourished according to GLIM criteria
Cohen's k statistics were calculated. Regarding the predictive using FFMBIA (n ¼ 66, 25.4%), more patients were malnourished
value for the occurrence of a complicated postoperative course the according to GLIM criteria using MMCT (n ¼ 179, 68.8%) and the SGA
different tools for the diagnosis of malnutrition were compared: (n ¼ 178, 68.5%), as shown in Fig. 1.
ROC curves were performed to calculate the sensitivity and In multivariate analysis adjusted for age, gender, cancer type,
specificity of MMCT and FFMBIA as well as the area under the curve cancer stage, ASA, malnutrition according to MMCT, malnutrition
for the prediction of overall complications. Multivariate logistic according to FFMBIA, and malnutrition according to SGA, a malnu-
regressions for the diagnostic tools were constructed to investi- trition according to GLIM criteria using MMCT remained an inde-
gate the association with the outcome parameters. A p- pendent risk factor for the overall complication rate (OR 2.330, 95%
value < 0.05 was considered statistically significant. Statistical CI 1.126e4.819, p ¼ 0.023).
analysis was performed using the Statistical Package for Social
Sciences version 25.
4. Malnutrition according to GLIM criteria using MMCT
The study was approved by the Ethical Review Board of the
Saxonian Medical Association (No. EK-BR-67/19-1).
The prevalence of malnutrition determined by abdominal CT
scan at lumbar level 3 was 68.8%. Mean MMCT in women was
3. Results 38,34 cm2/m2 (SD 7.518) and 48.58 cm2/m2 (SD 9.306) in men. The
50th percentile in our patient group was concordant to the pre-
In 260 patients undergoing major abdominal surgery, abdom- defined cut-off values for a reduced muscle mass.
inal CT scan, BIA, NRS, and SGA were available within 30 days before The risk for overall complication rate was significantly higher in
surgery. Patient characteristics are shown in Table 1. Major patients with malnutrition diagnosed by the GLIM criteria using
abdominal surgery included oesophageal resection (n ¼ 24, 9.2%), MMCT (OR 2.028, 95% CI 1.188e3.463, p ¼ 0,009). In the multivariate
gastrectomy (n ¼ 40, 15.4%), pancreatic resection (n ¼ 37, 14.2%), analysis a malnutrition diagnosed by CT scan remained an inde-
colon resection (n ¼ 92; 35.4%), rectal resection (n ¼ 59, 22.7%), and pendent risk factor for overall complication rate after adjustment
liver resection (n ¼ 8, 3.1%). The median age was 70.15 years ± 10.79 for age, gender, cancer type, cancer stage, and ASA (p ¼ 0.006), and
(ranged from 37 to 92 years), 113 patients were female (43.5%) and the age remained an independent risk factor for the overall
147 male (56.5%). complication rate (p ¼ 0.036). Cancer stage was an independent
The overall complication rate was 51.9%. Severe complications, risk factor for 6 months survival (p ¼ 0.022). Other clinical outcome
defined by ClavieneDindo  3a, were present in 21.5% of patients, parameters, such as length of hospital stay, re-laparotomy, and
hospital mortality was 6.2%. The median length of hospital stay was readmission to hospital within 31 days after discharge were not
17.36 days (SD 10.935) with a minimum stay of 6 days and a significantly associated with malnutrition diagnosed by CT scan.

Table 1
Patient characteristics.

Overall Malnutrition according Malnutrition according to Malnutrition according

n (%) to GLIM using MMCT n (%) GLIM using FFMBIA n (%) to the SGA n (%)

Sex
male 147 (56.5) 107 (72.8) 42 (28.6) 105 (71.4)
female 113 (43.5) 72 (63.7) 24 (21.2) 73 (64.6)
Age
< 70 years 112 (43.1) 73 (65.2) 26 (23.2) 80 (71.4)
 70 years 148 (56.9) 106 (71.6) 40 (27.0) 98 (66.2)
Location of the carcinoma
oesophageal 24 (9.2) 17 (70.8) 5 (20.8) 19 (79.2)
gastric 40 (15.4) 30 (75.0) 16 (40.0) 30 (75.0)
pancreatic 37 (14.2) 30 (81.1) 10 (27.0) 30 (81.1)
colon 92 (35.4) 63 (69.2) 22 (24.2) 56 (61.5)
rectal 59 (22.7) 36 (61.0) 12 (20.3) 36 (61.0)
liver 8 (3.1) 3 (37.5) 1 (12.5) 7 (87.5)
Cancer stage (UICC)
0 11 (4.2) 7 (63.6) 2 (18.2) 7 (63.6)
I 66 (25.4) 41 (62.1) 13 (19.7) 45 (68.2)
II 82 (31.5) 59 (72.0) 25 (30.5) 51 (62.2)
III 72 (27.7) 52 (72.2) 15 (20.8) 51 (70.8)
IV 29 (11.2) 20 (69.0) 11 (37.9) 24 (82.8)
ASA risk classification
1 18 (6.9) 14 (77.8) 4 (22.2) 11 (61.1)
2 76 (29.2) 40 (52.6) 9 (11.8) 45 (59.2)
3 161 (61.9) 122 (75.8) 49 (30.4) 117 (72.7)
4 5 (1.9) 3 (60.0) 4 (80.0) 5 (100.0)

ASA, American Society of Anaesthesiologists; BIA, bioelectrical impedance analysis; SGA, Subjective Global Assessment.

150
 Table 2

M. Wobith, C. Herbst, M. Lurz et al.

Clinical outcome parameters.

Overall Patients with Not malnourished p value Patients with Not malnourished p value Patients with Not malnourished p value
n (%) Malnutrition according to GLIM Malnutrition according to GUM Malnutrition using SGA n (%)
according using MMCT n (%) according using FFMBIA n (%) using SGA n (%)
to GLIM using to GLIM using
MMCT n (%) FFMBIA n (%)

n ¼ 260 n ¼ 179 (68.8) n ¼ 81 (31.2) n ¼ 66 (25.4) n ¼ 194 (74.6) n ¼ 178 (68.5) n ¼ 82 (31.5)

Complication 135 (51.9) 103 (57.5) 32 (39.5) 0.007 43 (65.1) 92 (47.4) 0.013 99 (55.6) 36 (43.9) 0.079
Severe complication 56 (21.5) 43 (24.0) 13 (16.0) 0.148 18 (27.3) 38 (19.6) 0.190 46 (25.8) 10 (12.2) 0.013
Complication grade Clavien-Dindo 0.063 0.062 0.007
I 17 (6.5) 13 (7.3) 4 (4.9) 4 (6.1) 13 (6.7) 10 (5.6) 7 (8.5)
II 62 (23.8) 47 (26.3) 15 (18.5) 21 (31.8) 41 (21.1) 43 (24.2) 19 (23.2)
IIIa 2(0.8) 2 (1.1) 0 0 2 (1.0) 2 (1.1) 0
IIIb 29 (11.2) 24 (13.4) 5 (6.2) 12 (18.2) 17 (8.8) 22 (12.4) 7(8.5)
IVa 7(2.7) 4 (2.2) 3 (3.7) 1 (1.5) 6 (3.1) 6 (3.4) 1(1.2)
IVb 2(0.8) 1 (0.6) 1 (1.2) 0 2 (1.0) 2 (1.1) 0
V 16 (6.2) 12 (6.7) 4 (4.9) 5 (7.6) 11 (5.7) 14 (7.9) 2 (2.4)
Length of hospital 17.36 ± 10.935 17.38 (SD 9.874) 17.31 (SD 12.973) 0.966 17.87 (SD 8.069) 17.18 (SD 11.757) 0.665 18.86 (SD 12.1) 14.15 (SD (6.954) 0.001
stay (days)
Re-Laparotomy 47 (18.1) n ¼ 244 36 (20.1) n ¼ 167 11 (13.6) n ¼ 77 0.205 18 (27.3) n ¼ 61 29 (14.9) n ¼ 183 0.025 39 (21.9) n ¼ 164 8(9.8) n ¼ 80 0.018
Readmission to hospital 35 (13.5) n ¼ 156 26 (15.6) n ¼ 104 9 (11.7) n ¼ 52 0.422 13 (21.3) n ¼ 40 22 (12.0) n ¼ 116 0.073 27 (16.5) n ¼ 114 8 (10.0) n ¼ 42 0.176
within 31 days
6 months mortality 19 (12.2) 13 (12.5) 6 (11.5) 0.863 5 (12.5) 14 (12.1) 0.943 16 (14.0) 3 (7.1) 0.243

Prevalence of clinical outcome parameters according to the different tools to diagnose a malnutrition. Severe complication ¼ ClavieneDindo  3a.
151

as the gold standard (area under the curve 0.577, p ¼ 0.031).

Fig. 2. ROC curve for the GLIM criteria according to MMCT using overall complications

factors (p ¼ 0.02, and p ¼ 0.024).

score and cancer stage remained to be independent prognostic
overall complication rate (p ¼ 0.026). For 6 months survival ASA-
cancer stage. Age remained to be an independent risk factor for
(p ¼ 0.016) after adjustment for age, gender, ASA, cancer type, and
mission within 31 days (p ¼ 0.03) and for overall complication rate
GLIM criteria using FFMBIA as an independent risk factor for read-

diagnosed by BIA.
rameters were not significantly associated with malnutrition
(OR 2.015, p ¼ 0.041, 95% CI 1.021 - 3.976). Other outcome pa-
patients were significantly more often in need of a re-laparotomy
FFMBIA (OR 2.116, p ¼ 0.01, 95% CI 1.185 - 3.778). Malnourished
the diagnosis of malnutrition according to GLIM criteria using

guidelines just met the 15th percentile in our patient group.

m2 (SD 2.36) in men. The recommended cut-off values of the ESPEN
The median FFMBIA in women was 17.4 kg/m2 (SD 1.82) and 19.7 kg/

5. Malnutrition according to GLIM criteria using FFMBIA

under the curve of 0.577 (p ¼ 0.031), as shown in Fig. 2.

calculated, which was 76.3% and a specificity of 40% with an area
according to MMCT for the prediction of overall complications was
The multivariate analysis revealed malnutrition diagnosed by

The overall complication rate was significantly associated with

Reduced skeletal muscle mass was present in 25.4% of patients.

With the help of a ROC curve the sensitivity of the GLIM criteria

Fig. 1. Prevalence of a malnutrition according to diagnostic tool.

Clinical Nutrition ESPEN 50 (2022) 148e154

M. Wobith, C. Herbst, M. Lurz et al. Clinical Nutrition ESPEN 50 (2022) 148e154

The sensitivity of the GLIM criteria according to FFMBIA was

31.9% and the specificity 81.6% with an area under the curve of
0.567 (p ¼ 0.061) as shown in Fig. 3.

6. Malnutrition diagnosed by SGA

Malnutrition assessed by the Subjective Global Assessment (SGA

B and C) had a prevalence of 68.5% in our patient group, which was
similar to the prevalence of malnutrition assessed by GLIM criteria
using MMCT.
Patients with malnutrition assessed by the SGA were at a
significantly higher risk for developing a severe complication (OR
2.528, p ¼ 0.012, 95% CI 1.528e5.309) and a need for a re-
laparotomy (OR 2.529, p ¼ 0.022, 95% CI 1.122e5.701). In the
multivariate analysis malnutrition diagnosed by the SGA remained
an independent risk factor for overall complications (p ¼ 0.032), a
severe complication (p ¼ 0.028), and the need of a re-laparotomy
(p ¼ 0.04). It was also independently associated with the length Fig. 4. ROC curve for the SGA for predicting overall complications (area under the
curve 0.551, p ¼ 0.158).
of hospital stay (0.004) and the grade of severity of complications
(p ¼ 0.022). ASA and cancer stage were independent risk factors for
6 months (p ¼ 0.042 and p ¼ 0.04).
The SGA showed a sensitivity of 73.3% and a specificity of 36.8% muscle mass, has also been associated with relevant outcome fac-
with an area under the curve of 0.551 (p ¼ 0.158), as shown in Fig. 4. tors [20e23].
In our patient group the prevalence of malnutrition was much
higher when using the GLIM criteria with MMCT or the SGA,
7. Discussion
compared to the GLIM criteria using FFMBIA. Whereas the numbers
following SGA and GLIM criteria using MMCT were comparable to
A standardized tool to diagnose malnutrition is of great
those found in previous studies [5], the prevalence of malnutrition
importance, especially in the preoperative course in abdominal
using FFMBIA was a much lower. Applying the recommended cut-
cancer patients, as they are at high risk to be malnourished with a
offs on the FFMI as provided by the ESPEN guidelines [16], we
worsening of their condition perioperatively and an increased risk
found a much lower rate of patients with reduced muscle mass
for complications. The criteria for malnutrition of the Global
compared to applying the previously defined cut-offs for the SMI
Leadership Initiative for Malnutrition include the measurement of
[2,17]. The cut-off values recommended by the ESPEN guidelines
muscle mass, which is not considered in the ESPEN guidelines, NRS,
are very strict and seem to present patients in a worse nutritional
SGA, or other diagnostic tools for malnutrition. The recommenda-
condition than a sarcopenia defined by previous CT cut-off values.
tions in the GLIM criteria about how to measure the muscle mass
This became evident by using different percentiles in our patient
include dual-energy absorptiometry, BIA, CT, or MRI. Regarding
group. Whereas the 50th percentile of the MMCT just met the cut-
local facilities, we chose routine CT which is always available in
off values for a CT-defined sarcopenia, the ESPEN recommended
gastrointestinal cancer patients and BIA which has been part of our
cut-off values of the FFMI where just beneath the 15th percentile in
nutritional assessment for many years.
our patient group. This clearly explains the prevalence differences.
Several studies showed significant correlations between a CT-
The poor correlation of MMCT and FFMBIA may be explained by the
defined sarcopenia and outcome factors in surgical patients
pre-defined cut-off values, but also by the fact that the FFMI is just a
[10e12,19]. By using bioelectrical impedance analysis as a non-
surrogate parameter for the muscle mass and that both represent
invasive bedside tool, body composition can be easily measured.
different compartments. It could be discussed whether the phase
The derived Fat free mass index (FFMI), a surrogate parameter for
angle may be also useful for defining malnutrition, as it has been
shown to be a good predictor for surgical complications in previous
studies [24].
Nevertheless, we found significant correlations between
malnutrition and overall complication rate independently of the
utilized diagnostic tool. In the multivariate regression analysis,
malnutrition according to the GLIM criteria using MMCT remained
to be an independent risk factor for overall complications. Previous
studies identified the GLIM criteria also to be predictive for the
perioperative complication rate [25e28]. Sarcopenia assessed by
BIA and physical performance measurements (handgrip strength
and 4-m walking test) was also predictive for overall complications,
major complications, and delayed hospital discharge in patients
undergoing surgery for esophageal cancer [29].
Whereas in oncological patients, who did not undergo surgery, a
correlation between death at six months and malnutrition has been
found in studies which diagnosed malnutrition by SGA, GLIM
criteria using FFMI and GLIM criteria using handgrip strength [5], in
our study only cancer stage and ASA-score remained independent
Fig. 3. ROC curve for the GLIM criteria according to FFMBIA for the prediction of overall prognostic factors for the 6 months mortality. Patients who un-
complications (area under the curve 0.567, p ¼ 0.061). dergo surgery for cancer treatment, in most cases do not suffer
152
M. Wobith, C. Herbst, M. Lurz et al. Clinical Nutrition ESPEN 50 (2022) 148e154

from an advanced cancer disease, in which a malnutrition seems to prognostic factor, independent of body mass index. J Clin Oncol 2013;31(12):
1539e47.
be more severe because of ongoing cancer cachexia, disease pro-
[4] Skeie E, Tangvik RJ, Nymo LS, Harthug S, Lassen K, Viste A. Weight loss and
gression, and ongoing chemotherapy. BMI criteria in GLIM's definition of malnutrition is associated with post-
The sensitivity of the GLIM criteria using MMCT and the SGA for operative complications following abdominal resections - results from a Na-
predicting overall complication rate was much higher than the tional Quality Registry. Clin Nutr 2020;3 9(5):1593e9.
[5] Contreras-Bolívar V, Sa
nchez-Torralvo FJ, Ruiz-Vico M, Gonz
alez-Almendros I,
sensitivity of GLIM criteria using FFMBIA. This suggests that by Barrios M, Padín S, et al. GLIM criteria using hand grip strength adequately
utilizing a CT scan, patients may be detected at an earlier state of predict six-month mortality in cancer inpatients. Nutrition 2019;11(9):2043.
malnutrition than using the FFMI from BIA. Plauth et al. (2019) [6] Martin L, Gioulbasanis I, Senesse P, Baracos VE. Cancer-associated malnutri-
tion and CT-defined sarcopenia and myosteatosis are endemic in overweight
reported a sensitivity of 93.3% for using the SMI to predict hospital and obese patients. JPEN - J Parenter Enter Nutr 2020;44(2):227e38.
mortality in hospitalized patients, but so far, there are no studies [7] Janssen I, Heymsfield SB, Baumgartner RN, Ross R. Estimation of skeletal
which calculated the sensitivity for predicting the complication muscle mass by bioelectrical impedance analysis. J Appl Physiol 2000;89(2):
465e71.
rate in surgical patients [30]. [8] Waduud MA, Adusumilli P, Drozd M, Bailey MA, Cuthbert G, Hammond C,
Limitations in our study are the retrospective design, we did not et al. Volumetric versus single slice measurements of core abdominal muscle
assess physical performance, i.e. handgrip strength, and that we did for sarcopenia. Br J Radiol 2019;92(1097):20180434.
[9] Gomez-Perez SL, Haus JM, Sheean P, Patel B, Mar W, Chaudhry V, et al.
not differentiate between obese and non-obese patients when us- Measuring abdominal circumference and skeletal muscle from a single cross-
ing the SMI. Our strengths are our homogeneous population of sectional computed tomography image: a step-by-step guide for clinicians
patients undergoing major abdominal surgery for cancer and the using National Institutes of Health ImageJ. JPEN - J Parenter Enter Nutr
2016;40(3):308e18.
prospectively collected data with BIA in all of our patients.
[10] Tamandl D, Paireder M, Asari R, Baltzer PA, Schoppmann SF, Ba-Ssalamah A.
Markers of sarcopenia quantified by computed tomography predict adverse
8. Conclusion long-term outcome in patients with resected oesophageal or gastro-
oesophageal junction cancer. Eur Radiol 2016;26(5):1359e67.
[11] Xiao J, Caan BJ, Cespedes Feliciano EM, Meyerhardt JA, Kroenke CH,
The GLIM criteria represent a useful tool for the standardized Baracos VE, et al. The association of medical and demographic characteristics
diagnosis of malnutrition. Independently of the method for muscle with sarcopenia and low muscle radiodensity in patients with nonmetastatic
colorectal cancer. Am J Clin Nutr 2019;109(3):626e34.
mass measurement, they are predictive for the clinical outcome. [12] Martin L, Hopkins J, Malietzis G, Jenkins JT, Sawyer MB, Brisebois R, et al.
Identifying surgical patients at metabolic risk for complications Assessment of computed tomography (CT)-Defined muscle and adipose tissue
allows to start preconditioning and prehabilitation early. Regarding features in relation to short-term outcomes after elective surgery for colo-
rectal cancer: a multicenter approach. Ann Surg Oncol 2018;25(9):2669e80.
the measurement of muscle mass or FFM, CT and BIA are reliable for [13] Kondrup J, Allison SP, Elia M, Vellas B, Plauth M. Educational and clinical
the identification of patients with a high risk of complications and practice committee, European society of parenteral and enteral nutrition
prolonged hospital stay. While CT with high sensitivity may over- (ESPEN). ESPEN guidelines for nutrition screening 2002. Clin Nutr 2003;22:
415e21.
estimate malnutrition, BIA with high specificity will detect patients
[14] Detsky AS, McLaughlin JR, Baker JP, Johnston N, Whittaker S, Mendelson RA,
at a more severe state of malnutrition. et al. What is subjective global assessment of nutritional status? JPEN - J
Parenter Enter Nutr 1987;11:8e13.
[15] Jensen GL, Cederholm T, Correia MTD, Gonzalez MC, Fukushima R,
Author contributions Higashiguchi T, et al. GLIM Core Leadership Committee; GLIM Working Group.
GLIM criteria for the diagnosis of malnutrition. a consensus report from the
Wobith: formal analysis, methodology, writing original draft; global clinical nutrition community. JPEN - J Parenter Enter Nutr 2019;43:
32e40.
Herbst: Data evaluation; Lurz: data analysis; Haberzettl: Data [16] Cederholm T, Bosaeus I, Barazzoni R, Bauer J, Van Gossum A, Klek S, et al.
curation, investigation; Fischer: writing e review & editing; Wei- Diagnostic criteria for malnutrition - an ESPEN consensus statement. Clin Nutr
mann: Methodology, Supervision, Validation, writing e review & 2015;34(3):335e40.
[17] Montano-Loza AJ, Meza-Junco J, Prado CMM, Lieffers JR, Baracos VE, Bain VG,
editing. et al. Muscle wasting is associated with mortality in patients with cirrhosis.
Clin Gastroenterol Hepatol 2012;10(2):166e73.
[18] Fischer M, Obera €nder N, Kaufmann J, Kirsche V, Prodehl G, Sch€ afer AO, et al.
Declaration of competing interest
Changes in intra- and extramyocellular lipids in morbidly obese patients after
non-surgical weight loss-a pilot study using magnetic resonance spectros-
Arved Weimann received grants from B Braun, Seca, and Mucos, copy. Clin Nutr ESPEN 2018;28:121e6.
as well as speakers honorarias from Baxter, Germany, Aesculap, [19] Yassaie SS, Keane C, French SJH, Al-Herz FAJ, Young MK, Gordon AC. Decreased
total psoas muscle area after neoadjuvant therapy is a predictor of increased
Fresenius Kabi, and Falk Foundation, Germany. mortality in patients undergoing oesophageal cancer resection. ANZ J Surg
Maria Wobith received grants from B Braun, and speakers 2019;89(5):515e9.
honoraria from Fresenius Kabi. [20] Thibault R, Makhlouf AM, Mulliez A, Gonzalez CM, Kekstas G, Kozjek NR, et al.
Fat-free mass at admission predicts 28-day mortality in intensive care unit
Delia Haberzettl, Christian Herbst Martin Fischer, and Markus patients: the international prospective observational study Phase Angle
Lurz declare to have no conflicts of interest. Project. Intensive Care Med 2016;42:1445e53.
[21] Pichard C, Kyle UG, Morabia A, Perrier A, Vermeulen B, Unger P. Nutritional
assessment: lean body mass depletion at hospital admission is associated
Acknowledgements with an increased length of stay. Am J Clin Nutr 2004;79(4):613e8.
[22] Tamura T, Sakurai K, Nambara M, Miki Y, Toyokawa T, Kubo N, et al. Adverse
We appreciated a lot the help of the IMISE (institute for medical effects of preoperative sarcopenia on postoperative complications of patients
with gastric cancer. Anticancer Res 2019;39(2):987e92.
informatics, statistics and epidemiology) of the University of Leip- [23] Tsaousi G, Kokkota S, Papakostas P, Stavrou G, Doumaki E3, Kotzampassi K.
zig for the help in statistical matters. Body composition analysis for discrimination of prolonged hospital stay in
colorectal cancer surgery patients. Eur J Cancer Care 2017;26(6). https://
doi.org/10.1111/ecc.12491.
References [24] H€arter J, Orlandi SP, Gonzalez MC. Nutritional and functional factors as
prognostic of surgical cancer patients. Support Care Cancer 2017;25:2525e30.
[1] Arends J, Bachmann P, Baracos V, Barthelemy N, Bertz H, Bozzetti F, et al. [25] Kakavas S, Karayiannis D, Bouloubasi Z, Poulia KA, Kompogiorgas S,
ESPEN guidelines on nutrition in cancer patients. Clin Nutr 2017;36(1):11e48. Konstantinou D, et al. Global leadership initiative on malnutrition criteria
[2] Prado CM, Lieffers JR, McCargar LJ, Reiman T, Sawyer MB, Martin L, et al. predict pulmonary complications and 90-day mortality after major abdominal
Prevalence and clinical implications of sarcopenic obesity in patients with surgery in cancer patients. Nutrition 2020;12(12):1e11.
solid tumours of the respiratory and gastrointestinal tracts: a population- [26] Sanchez-Rodriguez D, Locquet M, Reginster JY, Cavalier E, Bruye re O,
based study. Lancet Oncol 2008;9(7):629e35. Beaudart C. Mortality in malnourished older adults diagnosed by ESPEN and
[3] Martin L, Birdsell L, MacDonald N, Reiman T, Clandinin MT, McCargar LJ, et al. GLIM criteria in the SarcoPhAge study. J Cachexia Sarcopenia Muscle
Cancer cachexia in the age of obesity: skeletal muscle depletion is a powerful 2020;11(5):1200e11.

153
M. Wobith, C. Herbst, M. Lurz et al. Clinical Nutrition ESPEN 50 (2022) 148e154

[27] Skeie E, Tangvik RJ, Nymo LS, Harthug S, Lassen K, Viste A. Weight loss and [29] Wang P-Y, Chen X-K, Liu Q, Yu Y-K, Xu L, Liu X-B, et al. Highlighting sarco-
BMI criteria in GLIM's definition of malnutrition is associated with post- penia management for promoting surgical outcomes in esophageal cancers:
operative complications following abdominal resections e results from a evidence from a prospective cohort study. Int J Surg 2020;83:206e15.
National Quality Registry. Clin Nutr 2020;39(5):1593e9. [30] Viertel M, Bock C, Reich M, Lo€ ser S, Plauth M. Performance of CT-based low
[28] Maurício SF, Xiao J, Prado CM, Gonzalez MC, Correia MITD. Different nutri- skeletal muscle index, low mean muscle attenuation, and bioelectric imped-
tional assessment tools as predictors of postoperative complications in pa- ance derived low phase angle in the detection of an increased risk of nutrition
tients undergoing colorectal cancer resection. Clin Nutr 2018;37(5):1505e11. related mortality. Clin Nutr 2019;38(5):2375e80.

154