Invest New Drugs (2014) 32:382–387
DOI 10.1007/s10637-013-0053-6
SHORT REPORT
Low skeletal muscle is associated with toxicity in patients
included in phase I trials
Sophie Cousin & A. Hollebecque & S. Koscielny & O. Mir &
A. Varga & V. E. Baracos & J. C. Soria & S. Antoun
Received: 7 October 2013 / Accepted: 26 November 2013 / Published online: 17 December 2013
# Springer Science+Business Media New York 2013
Summary Background Low muscle mass has been associat-
ed with chemotherapy toxicity. We conducted this prospective
study to evaluate the effects of body composition on the
occurrence of toxicity in phase I trials. Patients and Methods
Patients were consecutively enrolled irrespective of the type
of tumor or the type of drug. The Skeletal Muscle Index
(SMIndex) and visceral and subcutaneous adipose tissue were
assessed with computed tomography imaging by measuring
cross-sectional areas of the tissues (cm2/m2). Dose-limiting
toxicity (DLT) corresponded to toxicities occurring during the
1st cycle that necessitated dose reduction, postponement or
interruption of drug administration and severe toxicity events
(STE) corresponded to DLT or permanent treatment with-
drawal due to toxicity. Results 93 patients were evaluated.
Ten percent of patients experienced DLT and had a lower
SMIndex: 40.8±4.6 vs. 48.1±9.6 cm2/m2 (p =0.01). STE
occurred in 14 % of the patients. The only factor associated
with STE was a low SMIndex: 42.4±5.8 vs. 48.4±9.7 cm2/m2
(p =0.02). STE were observed in 25.5 % of the patients when
the SMIndex was below the median value compared to 6.5 %
of patients with a high SMIndex (p =0.02). Conclusion
S. Cousin : A. Hollebecque : O. Mir : A. Varga : J. C. Soria
Phase I Department, Gustave Roussy, Villejuif, France
S. Koscielny
Department of Biostatistics and Epidemiology, Gustave Roussy,
Villejuif, France
V. E. Baracos
Department of Oncology, Cross Cancer Institute, Edmonton, Canada
S. Antoun
Department of Ambulatory Care, Gustave Roussy, Villejuif, France
S. Cousin (*)
Department of Medical Oncology, Centre Oscar Lambret,
3, rue Frédéric Combemale, 59000 Lille, France
e-mail: sophie_cousin@hotmail.fr
Muscle mass is a critical predictor of severe toxicity events
in phase I patients, suggesting that sarcopenia may be consid-
ered in assessing patients for eligibility of phase-1 studies.
Keywords Skeletal muscle . Sarcopenia . Drug interruption .
Severe toxicity . Phase I studies
Introduction
Phase I clinical trials are dose- and toxicity-finding studies. As
potential phase I patients are in a vulnerable position with an
advanced and progressive malignancy, proper selection of
these patients remains challenging. Conventional eligibility
criteria in phase I trials therefore include a good performance
status and adequate organ function, with an expected life
expectancy exceeding 3 months. These criteria are prognostic
factors for survival [1–3]. However, these criteria are suboptimal
for predicting toxicity among phase I patients [4]. A better
determination of predictive toxicity criteria could improve the
selection of phase I trial candidates.
Results suggesting a link between a low body mass index
(BMI) and drug toxicities are few and far between [5, 6].
Cancer patients have heterogeneous body composition (BC)
characteristics which could potentially engender variations in
drug concentration and metabolism; thus BC parameters
should be considered. BC assessment is currently possible
through the analysis of Computed Tomography (CT) images,
which has been demonstrated to be accurate and reproducible
[7, 8]. Sarcopenia refers to a state where muscle mass is
abnormally low, and was found to be linked to poor overall
survival (OS) [9, 10]. Recently, sarcopenia was shown to be
associated with toxicity among cancer patients treated with
cytotoxic drugs [11, 12] or targeted therapy [13–16].
We hypothesized that enrolling patients with a high risk of
experiencing toxicity in phase I trials could bias the evaluation
Invest New Drugs (2014) 32:382–387
of drugs. We therefore conducted a prospective study among
consecutively treated phase I patients in order to evaluate
the relationship between the BMI, BC and the incidence
of toxicity events.
Patients and methods
Patients
This prospective observational study included all patients with
advanced cancer or hematological malignancies who were
consecutively treated in 34 phase I trials at Gustave Roussy
from January to July 2011. All patients had given their in-
formed consent for their inclusion in the original phase I study.
The study was approved by the local Research Ethics
Committee according to good clinical practices and applicable
laws. Further analysis of clinical toxicity data and the assess-
ment of BC based on CT images were also approved by the
local Ethics Committee. No patient had organ dysfunction at
trial inclusion.
Toxicities were defined using V3.0 or V4.0 of the Common
Terminology Criteria for Adverse Events. DLT was defined
for each phase I study by the protocol and corresponded to
intolerable toxicities requiring the postponement of treatment,
a drug dose reduction or definitive interruption of drug
administration when they occurred during the 1st treatment
cycle. Drug interrupting toxicity (DIT) was defined as any
severe toxicity which necessitated definitive interruption of
the drug at anytime during the phase I treatment. We defined
severe toxicity events (STE) as either DLT or DIT.
Anthropometry and body composition
The BMI was calculated (i.e. weight [kg]/height2 [m2]). BC
was evaluated by screening CT images acquired within
30 days before inclusion. Images were analyzed with Slice-
O-Matic software V4.3 (Tomovision, Montreal, Canada). The
following BC variables were measured: cross-sectional area
(cm2) of SM, visceral adipose tissue (VAT), and subcutaneous
adipose tissue (SAT). These values are indexed for height and
are expressed in cm2/m2 [8–10]. An independent researcher
was trained to correctly identify and quantify both psoas and
paraspinal muscles as well as the transversus abdominis ,
abdominal external and internal oblique muscles and the
rectus abdominus. There was no risk of mistaking ascites,
peritoneal carcinomatosis or digestive structures for muscles.
The third lumbar vertebra (L3) was used as a landmark to
measure a cross-sectional area of SM and adipose tissue,
because it has been proven to be linearly related to whole
body measurements [8, 9, 17].
Since the distribution of VAT, SAT and SM differs consid-
erably according to gender, the median of the observed
383
distribution for each parameter by patient gender was
(arbitrarily) chosen as the cutoff. The cut point between a
low and high skeletal muscle index (SMIndex) was equal to
the median for patients of the same gender: 54.1 cm2/m2 for
males and 40.8 cm2/m2 for females.
Statistical analysis
Descriptive statistics were used: number of cases, percentages,
mean and standard deviation, median and extreme values. The
surface areas of BC parameters in patients with and without
toxicity were compared with Wilcoxon tests.
BC parameters and the BMI were entered as continuous
covariates in univariate logistic regression models. A multi-
variate logistic regression analysis including all the variables
with a p -value<0.20 in the univariate models was finally
performed using a backward selection strategy to eliminate
non-significant variables. The threshold p-value for variable
elimination was 0.05.
Analyses were done using the SAS® software version 9.2,
Cary, North Carolina, United States. All p-values were two-
tailed and the level of significance was p <0.05.
Results
Patient characteristics
The study population consisted of 93 patients (50 men). The
median age was 57 (range, 21–77). The median follow-up was
9 months (range, 5.5–12). Patient baseline characteristics are
reported in Table 1. BC features are reported in Table 2. There
were significant differences between women and men
for the following parameters: weight (p =0.0001), SMIndex
(p =0.0001, Table 2). At trial inclusion, 44 % were overweight
or obese (BMI>25 kg/m2) and in 29 % of these patients, a low
SMIndex was observed.
Body composition and toxicity
A DLT occurred in 9 patients (10 %). The only factor associ-
ated with a DLT was a low SMIndex (cm2/m2): 40.8±4.6
compared to 48.1±9.6 in patients without a DLT (p =0.01).
A DIT occurred in 13 patients (14 %). Male patients with a
DLT had a lower SMIndex (40.6±5.3) and a lower BMI (21.1±
1.4 kg/m2) compared to those without a DLT: SMIndex (53.1±
8.4; p =0.008) and BMI (24.9±3.5 kg/m2 P =0.02). The
SMIndex was significantly lower in men who experienced a
DIT (45.6±5.9 vs. 53.5±9.0 cm2/m2 p =0.04). No association
was found in women for DLT or DIT.
Patients who experienced a STE (n =15) (16 %), had a
lower SMIndex (42.4±5.8) compared to those without a STE
(48.4 ± 9.7 p = 0.02) (Table 3). VAT and SAT were not
384 Invest New Drugs (2014) 32:382–387

Table 1 Patient baseline clinical

and biological characteristics Age (years) median [range] 57 [21–77]
Gender: # patients (%) Male/Female 50 (53.8 %)/43 (46.2 %)
ECOG Performance Status # patients (%)
0–1 32 (34 %); 59 (64 %)
≥2 2 (2 %)
Tumor type: # patients (%)
Colorectal n =13 (14 %) Breast n =12 (13 %) Non Small Cell Lung n =11 (12 %)
Melanoma n =7 (7.5 %) Lymphoma n =7 (7.5 %) Mesothelioma n =6 (6.5 %)
Pancreas n =6 (6.5 %) Cervix n =(5.4 %) Other n =26 (28 %)
Number of chemotherapy regimens prior to phase I drug: # patients (%)
0 1 2 3 4 ≥5
11 (12 %) 16 (17 %) 17 (18 %) 14 (15 %) 11 (12 %) 12 (13 %)
Baseline biological parameters
Average (SD) Median [range]
Albumin (g/l) 35.5 (4.4) 36 [18–44]
Transthyretin (g/l) 0.22 (0.084) 0.2 [0.03–0.52]
LDH (IU/L) 260 (153) 199 [112–863]
CRP (mg/l) 33.1 (62.9) 12.6 [1–456]
Fibrinogen (g/l) 5.1 (1.5) 5.1 [3–12]

significantly associated with toxicity. The association between
the BMI and STE was close to significance (p =0.06).
Overall, 25.5 % of the patients with a low SMIndex expe-
rienced STE compared to 6.5 % of patients with a high
SMIndex (p =0.02). The difference between a high and low
SMIndex was significant only for men, with toxicities in 4 %
and 29 % respectively (p =0.02) (Fig. 1).
Patients with a BMI <25 kg/m2 (normal or low referring to
WHO categories) experienced more STE (22.5 %) than those
with a BMI ≥25 kg/m2 (7.5 %) (p =0.09). The difference was
significant exclusively for men: no STE occurred in men with
a BMI ≥25 kg/m2 whereas 27.5 % with a BMI<25 kg/m2
experienced STE (p =0.02) (Fig. 2).
Univariate analyses testing the impact of the BMI and
SMIndex as continuous factors confirmed this finding. A
multivariate analysis was conducted to predict toxicity with
the BMI and SMIndex. The BMI was not retained as an
independent factor predictive of toxicity (p =0.24). Only
the SMIndex remained significantly associated with STE
(p =0.013).
Finally, none of the following parameters were found to be
associated with a DIT or DLT: loss of weight over 6 months
>5 % or >10 %, Albumin <35 g/l, Lacticodeshydrogenase
>250 IU/L, transthyretin <21 g/L, C-Reactive protein
>6 mg/L, or any other BC parameters (SAT < median,
VAT < median).

Table 2 Body composition

baseline parameters according to Male Female Total
gender (n =50) (n =43) (n =93)

Age, years: mean (SD) 54 (12.9) 55.7 (11) 54.8 (12.1)

Weight: mean (SD) 75.5 (11.9)* 65 (10.4) 71 (12.4)
BMI (kg/m2): mean (SD) 24.6 (3.5) 24.6 (4.1) 24.6 (3.7)
Underweight (BMI<18.5), n (%) 1 (2 %) 1 (2 %) 2 (2 %)
Normal weight (18.5≤BMI≤24.9), n (%) 27 (54 %) 21 (49 %) 48 (52 %)
Overweight (25≤BMI≤29.9), n (%) 16 (32 %) 15 (35 %) 31 (33 %)
Obese (30≤BMI), n (%) 6 (12 %) 4 (9 %) 10 (11 %)
Skeletal muscle area (cm2) mean (SD) 157.4 (28.3)* 109.7 (18.7) 136.2 (34.1)
Skeletal muscle index (cm2/m2) mean (SD) 52.1 (8.9)* 41.8 (6.7) 47.5 (9.5)
VAT L3 area (cm2) mean (SD) 101.7 (81) 64.4 (46.9) 85.1 (70.2)
BMI body mass index, VAT VAT L3 index (cm2/m2) mean (SD) 34.1 (27.5) 24.8 (18.6) 30 (24.2)
visceral adipose tissue, SAT SAT L3 area (cm2) mean (SD) 129.1 (58.5) 185 (100.2) 154.2 (84.2)
subcutaneous adipose tissue
SAT L3 index (cm2/m2) mean (SD) 43.2 (20.3) 70.9 (38.9) 55.7 (33)
*Student test P =0.0001
Invest New Drugs (2014) 32:382–387 385

Table 3 Body composition characteristics according to dose-limiting

toxicity (DLT) or dose-interrupting toxicity (DIT) (mean (SD))

Patients with Patients without Wilcoxon

dose-limiting or dose-limiting or test
-interrupting toxicity -interrupting toxicity

Male
BMI 22.2 (1.8) 25.1 (3.6) 0.03
SMIndex 44.4 (6.5) 53.6 (8.5) 0.01
VAT 27.8 (16.4) 35.3 (29.1) 0.67
SAT 41.5 (12.4) 43.6 (21.6) 0.89
Female
BMI 24.2 (4.7) 24.7 (4.0) 0.68
SMIndex 40.0 (4.2) 42.0 (7.0) 0.41
VAT 25.5 (15.0) 24.6 (19.2) 0.84
SAT 54.7 (27.7) 74.3 (40.4) 0.27
Total
Fig. 2 Prevalence of severe toxicity events according to Body Mass
BMI 23.2 (3.5) 24.9 (3.8) 0.06 Index category. Overall patients: Patients with a BMI <25 kg/m2 (normal
SMIndex 42.4 (5.8) 48.4 (9.7) 0.02 or low referring to WHO categories) experienced more severe toxic
VAT 26.7 (15.2) 24.7 (25.5) 0.86 events (22.5 %) than those with a BMI ≥25 kg/m2 (7.5 %) (p =0.09).
For men: The difference was significant exclusively for men: no STE
SAT 47.7 (21.3) 49.4 (34.8) 0.43
occurred in men with a BMI ≥25 kg/m2 whereas 27.5 % of those with a
BMI<25 kg/m2 experienced severe toxic events (p =0.02)
BMI body mass index (kg/m2 ), SMIndex skeletal muscle index
(cm2 /m2 ), VAT visceral adipose tissue index (cm2 /m2 ), SAT subcutaneous
adipose tissue index (cm2 /m2 )
parameters may be useful for predicting toxicity among phase
Discussion I patients.
The use of CT imaging to assess BC is readily feasible and
We found that a lower SMIndex was associated with the reproducible without the need for additional invasive proce-
occurrence of STE and DLT among phase I patients. To our dures [7, 8]. This method has a minimal additional cost.
knowledge, this is one of the first studies to suggest that BC Currently an experienced technician takes 15 min to analyze
a CT image. Although there is a linear relationship between
the surface areas measured at L3 and the whole BC [17], all
statistical analyses were done with the surface areas measured,
because this linear relationship did not take into account the
different BMI categories and the differences in adipose tissue
distribution according to gender. The cut-off value defining
sarcopenia was based on the median value of this series, which
is a limitation of this study. Defining and validating the cut-off
values for a low SMindex is important since in most of the
previous studies, the cut-off values were defined using sur-
vival studies of North American, and especially of Canadian
cancer patients [9, 11, 12].
Very few data are available regarding the association be-
tween BMI and toxicity. However, the study by Telli et al. [6]
raised concerns about flat dosing of targeted therapies. Of 45
patients treated with sunitinib, 7 experienced cardiac toxicity;
the mean BMI (kg/m2) of these patients (23.9) was consider-
ably lower than that of patients who did not experience cardiac
Fig. 1 Prevalence of severe toxicity events according to skeletal muscle toxicity (mean BMI=27.1; p =0.03). In our study there was
category. Overall patients: Patients with a low skeletal muscle index only a trend towards an association between the BMI and the
experienced more severe toxic events 25.5 % compared to 6.5 % of
patients with a high SMIndex (p =0.02). For men: The difference between
occurrence of STE.
a high and low SMIndex was significant with toxicities in 4 % and 29 % Recently, BC studies highlighted the prognostic role of the
respectively (p =0.02) SMIndex. Its assessment is important since loss of muscle
386
mass may be masked in overweight or obese patients [10, 11]:
29 % of obese or overweight patients in the present study had
a SMIndex below the median value. A low SMIndex has been
associated with the toxicity caused by cytotoxic chemotherapy
[11, 12]. Recent data highlighted the impact of BC on the
occurrence of DLT in patients treated with targeted therapies
for metastatic renal cell carcinoma [13, 15] or for hepatocel-
lular carcinoma [14]. The present study confirms the previous
findings and emphasizes the importance of muscle mass,
independently of weight or the BMI, as a key parameter for
evaluating the risk of toxicity and thus for better selection of
phase I patients.
The relationship between toxicity and low SMIndex values
is increasingly being described but the underlying mecha-
nisms fuel debate. The pharmacokinetic (Pk) hypothesis de-
serves discussion. Gusella et al. [18] highlighted the role of
BC on the Pk parameters of 5-fluorouracil. Other studies
suggested that the Pk parameters involved are not related to
the criteria commonly used (weight, height, BSA) but to BC
parameters [19, 20]. Later, Prado et al. [12] demonstrated that
the doses of 5-fluorouracil, when divided by the kilograms of
lean body mass, were higher in the subgroup of patients who
experienced a DLT compared to those without a DLT
(p =0.034). Mir et al. [14] showed that in patients with
hepatocellular carcinoma, sarcopenic patients experienced sig-
nificantly more DLT than non-sarcopenic patients (82 % ver-
sus 31 %, p =0.005) and that the area under the concentration
time curve of sorafenib was significantly higher in sarcopenic
patients (p =0.013). These results and others observed in med-
ullary thyroid carcinoma [16] suggest that a hidden loss of
muscle mass can be related to a possible drug overdose and an
increased rate of toxicity.
Nevertheless other mechanisms of increased toxicity
in patients with low skeletal muscle mass can be hypoth-
esized. Sarcopenia may lead to a greater proportion of
circulating unbound drug and consequently to increased
toxicity. An inflammatory process was also suggested
since SM loss in patients with cancer can often be
associated with inflammation [21]. However, the relationship
between inflammation and chemotherapy toxicity is currently
not well established.
In conclusion, our findings suggest that the occurrence
of STE in phase I trials might be associated with a low
SM mass. Unlike previous studies, which evaluated a
specific cancer population treated with one drug, our
data concern the association between a low SM mass
and treatment toxicity in a heterogeneous population,
regardless of the primary tumor or whatever the drugs
used. Due to this heterogeneous population and the small
number of patients caution should be exercised in general-
izing the use of a low SMIndex. Further studies with a
larger number of patients and validated criteria for sarcopenia
are warranted.
Invest New Drugs (2014) 32:382–387
Acknowledgments The authors express their gratitude to Lorna Saint
Ange for editing.
Funding This research did not receive any specific grant from any
funding agency in the public, commercial or not-for-profit sector.
Disclosure The authors have declared no conflicts of interest.
References
1. Italiano A, Massard C, Bahleda R et al (2008) Treatment outcome
and survival in participants of phase I oncology trials carried out from
2003 to 2006 at Institut Gustave Roussy. Ann Oncol 19:787–792
2. Wheler J, Tsimberidou AM, Hong D et al (2012) Survival of 1181
patients in a phase I clinic: the MD anderson clinical center for
targeted therapy experience. Clin Cancer Res 18:2922–2929
3. Olmos D, A’hern RP, Marsoni S et al (2012) Patient selection for
oncology phase I trials: a multi-institutional study of prognostic
factors. J Clin Oncol 30:996–1004
4. LoConte NK, Smith M, Alberti D et al (2010) Amongst eligible
patients, age and comorbidity do not predict for dose-limiting
toxicity from phase I chemotherapy. Cancer Chemother
Pharmacol 65:775–780
5. Meyerhardt JA, Tepper JE, Niedzwiecki D et al (2004) Impact of
body mass index on outcomes and treatment-related toxicity in
patients with stage II and III rectal cancer: findings from Intergroup
Trial 0114. J Clin Oncol 22:648–657
6. Telli L, Witteles RM, Fisher GA, Srinivas S (2008) Cardiotoxicity
associated with the cancer therapeutic agent sunitinib malate. Ann
Oncol 19:1613–1618
7. Mitsiopoulos N, Baumgartner RN, Heymsfield SB et al (1998)
Cadaver validation of skeletal muscle measurement by magnetic
resonance imaging and computerized tomography. J Appl Physiol
85:115–122
8. Mourtzakis M, Prado CMM, Lieffers JR et al (2008) A practical and
precise approach to quantification of body composition in cancer
patients using computed tomography images acquired during routine
care. Appl Physiol Nutr Metab 33:997–1006
9. Prado CMM, Lieffers JR, McCargar LJ et al (2008) Prevalence and
clinical implications of sarcopenic obesity in patients with solid
tumours of the respiratory and gastrointestinal tracts: a population-
based study. Lancet Oncol 9:629–635
10. Tan BHL, Birdsell LA, Martin L et al (2009) Sarcopenia in an
overweight or obese patient is an adverse prognostic factor in pan-
creatic cancer. Clin Cancer Res 15:6973–6979
11. Prado CMM, Baracos VE, McCargar LJ et al (2009) Sarcopenia as a
determinant of chemotherapy toxicity and time to tumor progression
in metastatic breast cancer patients receiving capecitabine treatment.
Clin Cancer Res 15:2920–2926
12. Prado CMM, Baracos VE, McCargar LJ et al (2007) Body
composition as an independent determinant of 5-fluorouracil-based
chemotherapy toxicity. Clin Cancer Res 13:3264–3268
13. Antoun S, Baracos VE, Birdsell L et al (2010) Low body mass index
and sarcopenia associated with dose-limiting toxicity of sorafenib in
patients with renal cell carcinoma. Ann Oncol 21:1594–1598
14. Mir O, Coriat R, Blanchet B et al (2012) Sarcopenia predicts early
dose-limiting toxicities and pharmacokinetics of sorafenib in patients
with hepatocellular carcinoma. Plos One 7:e37563
15. Huillard O, Mir O, Peyromaure M et al (2013) Sarcopenia and body
mass index predict sunitinib-induced early dose-limiting toxicities in
renal cancer patients. Br J Cancer 108:1034–1041
16. Massicotte MH, Borget I, Broutin S et al (2013) Body composition
variation and impact of low skeletal muscle mass in patients with
Invest New Drugs (2014) 32:382–387
advanced medullary thyroid carcinoma treated with vandetanib:
Results from a placebo-controlled study. J Clin Endocrinol Metab
98:2401–2408
17. Shen W, Punyanitya M, Wang Z et al (2004) Total body skeletal
muscle and adipose tissue volumes: estimation from a single abdominal
cross-sectional image. J Appl Physiol 97:2333–2338
18. Gusella M, Toso S, Ferrazzi E et al (2002) Relationships between
body composition parameters and fluorouracil pharmacokinetics. Br
J Clin Pharmacol 54:131–139
387
19. Nawaratne S, Brien JE, Seeman E et al (1998) Relationships among
liver and kidney volumes, lean body mass and drug clearance. Br J
Clin Pharmacol 46:447–452
20. Morgan DJ, Bray KM (1994) Lean body mass as a predictor of drug
dosage. Implications for drug therapy. Clin Pharmacokinet 26:
292–307
21. Khamoui AV, Kim JS (2012) Candidate mechanisms underlying
effects of contractile activity on muscle morphology and energetics
in cancer cachexia. Eur J Cancer Care 21:143–157