Download Screening older cancer patients: first evaluation of the G

original articles
combination also affect the same fraction fa of the cells. If the
sum of these two fractional terms is equal to 1, additivity is
indicated. If the CI value is <1, synergy is indicated, and if the
CI value is >1, antagonism is indicated.
The above relationships can be portrayed graphically by a
normalized isobologram. In these plots, the ratio of dose of
drug needed to affect a fraction x when used in combination
relative to when it is used alone is plotted (for different values
of x) for drug 1 (vertical axis) against drug 2 (horizontal axis,
Annals of Oncology
Figure 2C). If the combination data points fall on the curve, an
additive effect is indicated. If the combination data points fall
on the lower left or on the upper right, synergism or
antagonism is indicated, respectively.
Note that one difference between the two methods of
analysis is that the isobologram analysis of Steel and Peckham
[3] defines an envelope of additivity, but the median effect
method of Chou [2] defines only a single condition of
additivity.
Annals of Oncology 23: 2166–2172, 2012
doi:10.1093/annonc/mdr587
Published online 16 January 2012
Screening older cancer patients: first evaluation
of the G-8 geriatric screening tool
C. A. Bellera1,2*, M. Rainfray3,4, S. Mathoulin-Pélissier1,2,5, C. Mertens4,6, F. Delva1, M. Fonck6 &
P. L. Soubeyran6
1
Clinical Research and Clinical Epidemiology Unit, Institut Bergonié, Regional Comprehensive Cancer Centre, Bordeaux; 2INSERM CIC-EC7, ISPED, Bordeaux University,
Bordeaux; 3SFR Public Health, Bordeaux University, Bordeaux; 4Department of Clinical Gerontology, Bordeaux University Hospital, Bordeaux; 5INSERM U897 Epidemiology and Biostatistics, ISPED, Bordeaux University, Bordeaux; 6Department of Medical Oncology, Institut Bergonié, Regional Comprehensive Cancer Centre,
Bordeaux, France
Received 14 September 2011; revised 18 November 2011; accepted 21 November 2011
Background: Development of a geriatric screening tool is necessary to identify elderly cancer patients who would
benefit from comprehensive geriatric assessment (CGA). We develop and evaluate the G-8 screening tool against
various reference tests.
Patients and methods: Analyses were based on 364 cancer patients aged >70 years scheduled to receive first-line
chemotherapy included in a multicenter prospective study. The G-8 consists of seven items from the Mini Nutritional
Assessment (MNA) questionnaire and age. Our primary reference test is based on a set of seven CGA scales: Activities
Daily Living (ADL), Instrumental ADL, MNA, Mini–Mental State Exam, Geriatric Depression Scale, Cumulative Illness
Rating Scale-Geriatrics, and Timed Get Up and Go. We considered the presence of at least one questionnaire with an
impaired score as an abnormal reference exam. Additional reference exams are also discussed.
Results: The prevalence of being at risk varied from 60% to 94% according to the various definitions of the reference
test. When considering the primary reference test, a cut-off value of 14 for the G-8 tool provided a good sensitivity
estimate (85%) without deteriorating the specificity excessively (65%).
Conclusion: The G-8 shows good screening properties for identifying elderly cancer patients who could benefit from
CGA.
Key words: cancer, elderly, screening, sensitivity and specificity
introduction
The comprehensive geriatric assessment (CGA) is a
multidisciplinary evaluation of an older individual’s functional,
psychological and nutritional statuses, cognition, social
support, and comorbidity. This assessment can detect geriatric
*Correspondence to: Dr C. A. Bellera, Clinical Research and Clinical Epidemiology Unit,
Institut Bergonié, Regional Comprehensive Cancer Center, 229 Cours de l’Argonne,
33076 Bordeaux, France. Tel: +33-5-56-33-04-95; Fax: +33-5-56-33-04-85;
E-mail: [email protected]
problems and potentially improves survival, physical, and
cognitive state of patients as well as increases an elderly
person’s chances of staying at home longer [1–4]. With regard
to elderly cancer patients, CGA components have been shown
to be prognostic factors of survival (such as functional status
and quality of life) and even associated with changes in cancer
treatment (such as functional status and malnutrition) [5–7].
Although randomized controlled trials evaluating the impact
of CGA-based interventions in older cancer patients to
provide definite evidence on its efficacy to improve survival are
still rare, existing literature provides growing evidence that
© The Author 2012. Published by Oxford University Press on behalf of the European Society for Medical Oncology.
All rights reserved. For permissions, please email: [email protected]
original articles
Annals of Oncology
CGA can predict morbidity and mortality in older cancer
patients [8].
Since 2005, the application of the CGA for cancer patients
>70 has been recommended by the International Society of
Geriatric Oncology [9]. However, its systematic application to
this population does not seem essential as many patients can
be treated according to the standard treatment protocol
without major modification [10–12]. Similar findings have
been derived from the general population as improved
outcomes are observed in unfit patients, suggesting that healthy
older people should be excluded of the process [4].
The current challenge at hand is to identify individuals for
whom the CGA would be the most beneficial, justifying the
need for a validated screening instrument. Several screening
methods to identify frail patients are available [13–17]. They
usually have to be administered by clinical personnel [13–15,
17]. In addition, most have been validated in the clinical
geriatric context but not with cancer patients and thus do not
necessarily take into account the side-effects of cancer
treatment. The Vulnerable Elders Survey (VES-13), a selfadministered questionnaire, is one such tool that has been
developed based on a large population of 6000 patients aged
over 65 years [18]. It has since been evaluated in cancer
patients by several studies but sample sizes are limited with
heterogeneous cancer populations. Importantly, this tool has
been evaluated against various reference exams [19–23] leading
to different estimates of its screening performances, as
highlighted in an recent editorial [24].
Current work on geriatric screening tools raises two issues.
First, the development of a screening tool is necessary to
individualize older cancer patients who would benefit from a
CGA. This approach offers advantages both for the
rationalization of the use of medical resources for the health
care system and also to spare the patient from unnecessary
clinical and biological examinations. In addition, from a
methodological point of view, the reference exam to which the
screening tool will be compared needs to be defined with rigor
[25]. We propose to address these two issues.
Using data collected previously in a multicenter study
focusing on a cohort of elderly cancer patients, we carried out
a series of analyses to finalize the choice of a high-performing
screening tool. Specifically, for the development of this
screening test, we wish to ensure that none of the ‘at risk’
patients will be missed, thus prioritizing sensitivity over
specificity.
methods
patients
Our research was elaborated based on a regional multicenter prospective
study funded by a 2003 French clinical research program (NCT00210249
[26]). The primary objective of this study was to assess the role of the CGA
in the management of elderly patients treated for a cancer with
chemotherapy. Patients aged >70 years (no upper age limit) scheduled to
receive first-line chemotherapy for various types of cancer [colon, pancreas,
stomach, ovary, bladder, prostate, lung cancer, non-Hodgkin’s lymphoma
(NHL), or cancer of unknown primary origin] were eligible for inclusion.
Breast cancer patients were not included as very few elderly patients receive
first-line chemotherapy. Patients with known central nervous system
Volume 23 | No. 8 | August 2012
metastases were excluded. The protocol was approved by institutional
review boards and ethics committees and was conducted in accordance
with the Declaration of Helsinki, Good Clinical Practices, and local ethical
and legal requirements. All patients provided written informed consent for
their participation.
baseline assessment
The geriatric evaluation consultation was carried out after an initial
specialized oncology consultation. This evaluation assessed the patient’s
comorbidities, cognitive functions, mood disorders, nutritional and
functional status, disabilities as well as the main geriatric syndromes, and
the current medication of the elderly patient. The patient was seen by the
clinical research nurse for the following questionnaires and scales:
Instrumental (IADL) and Activities in Daily Living (ADL) questionnaires,
the Geriatric Depression Scale (GDS-15), and a quality of life selfquestionnaire (QLQ-C30). The geriatrician was responsible for completing
the Cumulative Illness Rating Scale for Geriatrics (CIRS-G), the Mini–
Mental Status Exam (MMSE), and the Mini Nutritional Assessment
(MNA) as well as for performing the ‘Timed Get Up and Go’ (GUG)
evaluation. As necessary, the oncologist scheduled interventions with
various health care practitioners (social worker, nutritionist,
physiotherapist, team of psychologists, and/or psychiatrist) as well as any
necessary supplementary examinations such as extra blood analyses.
Geriatric assessment data were blinded to the physician. These
questionnaires and scales have been validated including cut-off values for
defining abnormal test results. The domain covered by each of these
questionnaires, the range of their respective scores, as well as cut-off values
for defining abnormal scores are summarized in Table 1.
reference exam
In order to evaluate a screening tool, a clear definition of the reference
exam must first be provided. No recognized definition for such reference
tool is available in the literature [24]. We retained a set of seven validated
questionnaires (ADL, IADL, MMSE, GDS-15, MNA, CIRS-G, and GUG),
which are routinely used and often considered as components of the CGA
[8]. Similarly, a clear definition of what constitutes an abnormal reference
should be provided. We considered the presence of at least one
questionnaire with an impaired score as an abnormal reference exam. As
exploratory analyses, we also investigated five additional reference exams.
We considered using at least two abnormal scores based on the
aforementioned set of seven questionnaires. Next, we considered leaving
out either the ADL or the IADL and using at least one or two
questionnaires with an impaired score.
proposition of a screening tool
We chose to assess the G-8 screening tool consisting of eight items: a
selection of seven items from the MNA questionnaire as well as an
indication of age. As such, the G-8 tool shares common questions with the
MNA short form, although the latter was not specifically developed for
cancer patients [37]. Items selected from the MNA questionnaire included
questions relating to nutritional status [food intake (MNA item A), weight
loss (item B), body mass index (item F)], motor skills (item C),
psychological status (item E), number of medications (item H), and selfperception of health (item P). Age was considered in three categories (<80,
80–85, and >85) as it is typically included in frailty criteria [14, 18]. We
maintained the same scores as those used in the original MNA
questionnaire and used a score of 0 (>85) to 2 (<80) for age. Overall, the
G-8 score ranged from 0 (heavily impaired) to 17 (not at all impaired). The
G-8 questionnaire is provided in Table A1.
The choice of the items that constituted the G-8 was driven by earlier
findings. Following a preliminary analysis [26], we investigated factors
doi:10.1093/annonc/mdr587 | 
original articles
Annals of Oncology
Table 1. CGA questionnaires: reference values, cut-off values, and proportions of subjects with abnormal scores in a series of 364 subjects
Description of the questionnaire
Questionnaire Reference
ADL
IADL
MMSE
Katz 1963 [27]
Lawton 1969 [28]
Folstein 1975 [29]
GDS-15
Sheikh 1986 [30],
Yesavage 1988 [31]
Guigoz 1994 [32]
MNA
GUG
CIRS-G
Mathias 1986 [33],
Podsiadlo 1991 [34]
Linn 1968 [35], Miller
1991 [36]
Number of subjects with an abnormal score
Across all patients
From subset of 339 patients
(# abnormal scores
with all seven assessments
/# available) (%)
available (%)
Domain
Score range
Abnormal score if
Autonomy
Autonomy
Cognitive
functions
Mood
functions
Nutritional
status
Mobility
0–6
0–8
0–30
ADL ≤5
IADL ≤7
MMSE ≤23
66/353 (19%)
257/353 (73%)
70/351 (20%)
57 (17%)
244 (72%)
59 (17%)
0–15
GDS-15 ≥6
155/342 (45%)
153 (45%)
0–30
MNA ≤23.5
231/353 (65%)
218 (64%)
0–∞
≥20 seconds
88/353 (25%)
79 (23%)
99/364 (27%)
131 (39%)
Comorbidity
Up to 14
At least one
comorbidities
comorbidity
grade 3/4
ADL, Activities of Daily Living; CGA, comprehensive geriatric assessment; CIRS-G, Cumulative Illness Rating Scale for Geriatrics; GDS-15, Geriatric
Depression Scale 15; GUG, Get Up and Go; IADL, Instrumental Activities Of Daily Living; MMSE, Mini–Mental State Exam; MNA, Mini Nutritional
Assessment.
associated with early death risk (death within 6 months of treatment
initiation). Advanced disease stage, male gender, low MNA score, and
impaired GUG were significantly and independently associated with early
death risk. These preliminary results highlighted the value of the MNA
score when evaluating elderly cancer patients as compared with other test
scores specific to other domains. These results appear coherent as the
content of this questionnaire includes items not only related to the
nutritional status of the patient but also items oriented toward the
exploration of cognitive or mood troubles as well as disabilities or
comorbidities that are present in geriatric questionnaires of the CGA.
We thus expected items of the G-8 to be strongly correlated with scores
of geriatric questionnaires. Although in brief, items of the G-8 tend to
cover domains usually addressed by the CGA. As such, we anticipated
strong associations.
statistical analyses
Qualitative variables were described using counts and proportions.
Quantitative continuous variables were described using means and
standard deviations for normal data or medians and ranges otherwise.
Associations between qualitative variables were tested using the chi-square
test or the Fisher’s exact test depending on the sample sizes. Performance
of the G-8 tool was evaluated using sensitivity (Se), specificity (Sp), receiver
operating curve (ROC), and area under the ROC curve (AUC). Confidence
intervals (95% CI) are reported.
results
baseline patient characteristics and geriatric
evaluations
Between September 2002 and September 2005, 364 patients
were included across 12 regional centers (Table 2). Median age
was 77 years (70–99 years) with slightly more males (59%).
The main cancer types and locations were lymphomas (30%),
cancer of the colon (28%), stomach (10%), lung (10%),
pancreas (6%), prostate (6%), bladder (5%), and ovary (4%).
 | Bellera et al.
Table 2. Baseline characteristics of the patients
Characteristics
N (%)
Sex
Male
Female
Age (years)
<80
80–85 (inclusive)
>85
ECOG performance status
0–1
2–4
Cancer site
Non-Hodgkin’s lymphoma
Colon
Stomach
Lung
Pancreas
Prostate
Bladder
Ovary
Unknown primary
Disease stage—solid tumors
M0
M1
MX
Disease stage—non-Hodgkin’s lymphoma
aaIPI 0–1
aaIPI 2–3
214 (59%)
150 (41%)
247 (68%)
89 (24%)
28 (8%)
261 (76%)
83 (24%)
110 (30%)
101 (28%)
37 (10%)
37 (10%)
23 (6%)
20 (6%)
18 (5%)
14 (4%)
4 (1%)
62 (24%)
134 (53%)
58 (23%)
59 (54%)
51 (46%)
aaIPI, age-adjusted International Prognostic Index; ECOG, Eastern
Cooperative Oncology Group.
Most of the solid tumors were of advanced stage (53%).
Almost half of the NHL patients were of International
Prognostic Index (IPI) stage 2–3 (46%). Across all tumor types,
Volume 23 | No. 8 | August 2012
original articles
Annals of Oncology
Table 3. Proportion of patients with an abnormal Reference exam and performance of the G-8 screening tool according to six definitions of the reference
exam
Reference exam
Subjects with an abnormal
reference exam
N (total = 339) %
Seven questionnaires: ADL, IADL, MMSE, GDS-15, MNA, GUG, and CIRS-G
At least one impaired score
319 (94%)
At least two impaired scores
259 (76%)
Six questionnaires: IADL, MMSE, GDS-15, MNA, GUG, and CIRS-G
At least one impaired score
319 (94%)
At least two impaired scores
257 (76%)
Six questionnaires: ADL, MMSE, GDS-15, MNA, GUG, and CIRS-G
At least one impaired score
290 (86%)
At least two impaired scores
205 (60%)
Se (95% CI)
Sp (95% CI)
AUC (95% CI)
85% (81% to 89%)
92% (88% to 95%)
65% (41% to 85%)
0·48 (0·37–0·60)
87% (82% to 92%)
85% (80% to 89%)
85% (81% to 89%)
92% (88% to 95%)
65% (41% to 85%)
47% (36% to 58%)
87% (82% to 92%)
85% (80% to 89%)
89% (85% to 93%)
0·97 (93% to 99%)
60% (45% to 74%)
39% (31% to 48%)
90% (86% to 93%)
84% (80% to 88%)
For the performance of the G-8 questionnaire, sensitivity (Se), specificity (Sp), and area under the ROC curve (AUC) with 95% confidence intervals (95%
CIs) are presented assuming a cut-off value of 14 (≤14 versus >14).
ADL, Activities of Daily Living; CIRS-G, Cumulative Illness Rating Scale for Geriatrics; GDS-15, Geriatric Depression Scale 15; GUG, Get Up and Go;
IADL, Instrumental Activities of Daily Living; MNA, Mini Nutritional Assessment; MMSE, Mini–Mental State Examination.
the advanced stages (M1/MX or IPI 2–3) represented the
majority of inclusions (67%).
The seven evaluations were available for 339 patients (93%).
Complete assessment was significantly more often available for
subjects with localized disease (97% for localized versus 90%
for advanced disease), men (95% for men versus 90% for
women), and for patients with better performance status [98%
for Eastern Cooperative Oncology Group (ECOG) 0–1 versus
85% for ECOG 2–4]. Patients who completed the assessment
were also significantly younger (mean age 77·5 versus 82·2).
The questionnaire with the lowest completion rate was the
MMSE (92%) while the CIRS-G was available for all patients.
Among the patients who completed all seven evaluations, the
ADL questionnaire had the highest number of ‘normal’
nonimpaired scores (83%) compared with 28%, 36%, and 55%
for the IADL, MNA, and GDS-15, respectively. These
distributions of scores were similar when we considered the
global population (Table 1).
reference exams and rules
When the seven questionnaires were considered, 94% of the
population presented with at least one impaired score and 76%
presented with two or more impaired scores. Proportions of
subjects with an abnormal reference exam are presented in
Table 3 using other rules.
performance of the screening tool
As expected, items of the G-8 screening tool showed strong
associations with at least one of the geriatric evaluations
(Table 4). Approximately 82% of the population presented
an impaired G-8 score. When considering an impaired
reference standard as at least one impaired score, the
sensitivity and specificity estimates were close, whether the
reference standard included all seven questionnaires,
excluded the ADL, or excluded the IADL. Similar results
apply when considering an impaired reference standard as
two impaired scores, although the specificity fell drastically
Volume 23 | No. 8 | August 2012
(down to 45%) for relatively smaller gains in sensitivity. The
resulting ROC curves for the G-8 against the various
reference exams were very similar, leading to comparable
AUC estimates. We show the ROC curve representing the
G-8 screening tool against the reference exam consisting of
the seven questionnaires (Figure 1).
When retaining the seven questionnaires as part of the
reference exam, the ROC curves suggested that cut-off values
∼14 would provide good sensitivity estimates, of at least 80%,
without deteriorating the specificity excessively (not <60%).
discussion
Based on a multicenter cohort study, we carried out the first
evaluation of the G-8 screening tool for geriatric oncology. We
also discussed the definition of the reference exam it should be
compared with. The G-8 tool incorporating elements of the
MNA questionnaire and age is thus proposed with the selected
items covering multiple domains usually assessed by the
geriatrician when performing the CGA: disability, nutrition,
cognition, depression, and comorbidities. As in the MNA
questionnaire, items specific on mood or cognition disorders,
dependence, and comorbidities are highly subjective, and as
such cannot be considered as a valuable evaluation of these
domains, but rather global and/or subjective assessments.
Moreover, it should be acknowledged that the G-8 tool is not
aimed at replacing the expertise of geriatricians for the
diagnosis of frailty. Rather, it should be used as a screening
tool to identify patients in need for a further assessment and
appropriate care.
Selection of the optimal cut-off value of the screening tool
should rely on the desired properties of the tool under
development. Because of the duality between sensitivity and
specificity, both parameters cannot be simultaneously
maximized. When developing a screening test, we wish to
ensure that none of the at risk patients will be missed. Thus
the sensitivity should be prioritized over the specificity. Our
doi:10.1093/annonc/mdr587 | 
original articles
Annals of Oncology
Table 4. Exploratory tests for associations between individual items of the
G-8 tool and components of the CGA
Associations investigated
between MNA items and score
(normal/impaired) of common CGA
componentsa
P value for the test
statistics for the associationb
Association between MNA item ‘food intake’ and
ADL
0·06
IADL
<0·001
MNA
<0·001
Association between MNA item ‘weight loss’ and
ADL
0·02
IADL
0·39
MNA
<0·001
Association between MNA item ‘mobility’ and
ADL
<0·001
IADL
0·001
GUG
<0·001
Association between MNA item ‘neuropsychological problems’ and
MMSE
<0·001
GDS-15
<0·001
Association between MNA item ‘body mass index’ and
ADL
0·63
IADL
0·31
MNA
<0·001
Association between MNA item ‘more than three prescribed medications’
and
CIRS-G
0·003
Association between G-8 item ‘self-perception of health’ and
ADL
<0·001
IADL
0·001
GDS-15
<0·001
CIRS-G
0.79
P values for the tests of association are reported based on the 339 patients
with the complete set of questionnaires for the CGA.
a
CGA components are analyzed as binary variables (normal versus
impaired score).
b
Chi-square test or Fisher’s exact test, depending on the sample size.
ADL, Activities of Daily Living; CGA, comprehensive geriatric assessment;
CIRS-G, Cumulative Illness Rating Scale for Geriatrics; GDS-15, Geriatric
Depression Scale; GUG, Get Up and Go; IADL, Instrumental Activities of
Daily Living; MNA, Mini Nutritional Assessment; MMSE, Mini–Mental
State Examination.
analysis suggested that a threshold value at 14 (≤14 versus >14)
would provide good sensitivity estimates for the G-8 of at least
80%, without deteriorating the specificity under 60%.
As recently acknowledged by Molina-Garrido et al. [24],
currently, there is no one standard model of CGA accepted in
the scientific community. A standard model composed of
validated scales that would enable standardization of studies
and results is needed. Our selection of seven tests and
questionnaires that have been individually validated in geriatric
oncology is thus coherent and sensible.
Depending on our definition of being at risk, the prevalence
varied between 60% and 94%. Although these estimates could
appear high, it should be noted that our target population
included first-line chemotherapy elderly patients and as such a
large proportion of advanced disease (about two-thirds in our
 | Bellera et al.
Figure 1. Receiver operating curve (sensitivity versus 1-specificity) for the
G-8 screening tool against the reference exam consisting of seven
comprehensive geriatric assessment questionnaires (at least one abnormal
score versus none).
sample). We expect that a general population of elderly cancer
patients will involve a smaller proportion of advanced disease
leading to lower estimates of at risk subjects.
We defined being at risk as having at least one or two areas
of impairments when assessing the multiple dimensions of the
physical and psychological conditions of the older cancer
patient. We preferred the threshold of one abnormal
questionnaire since our objective is to be sensitive in the
detection of oncological frailty. Still, this choice may be
debatable since patients with only one abnormal questionnaire
do not necessarily need the intervention of a geriatrician in all
cases.
The simultaneous inclusion of the ADL and IADL can also
be discussed. Indeed, one could at first argue that both focus
on the patient’s autonomy and as such could be redundant.
However, the two scales do not address the same stage of
independence with IADL referring to skills required to
maintain usual activities in the community (ability to shop,
managing money, etc.), which are affected earlier than the
skills encompassed by the ADL that are required to maintain
very basic home activities (ability to bath, dress, etc.). It is
therefore expected that skills impaired on ADL will be
observed in the later stage of disability and should be identified
relatively easily. On the other hand, impaired IADL skills are
probably more subtle to identify during a standard
consultation. Identifying and observing abnormal IADL can be
considered as an early stage of disability. Furthermore, four
questions of the IADL questionnaire have been shown to be
highly correlated with cognitive deficiencies [38]. Given the
high prevalence of cognitive impairment in the elderly [39],
impairment on the IADL scale should thus be accounted for.
Finally, concerning the IADL, Lawton [28] originally suggested
that certain activities such as shopping, preparing meals, and
housework were gender-based and thus less representative for
males so should only be evaluated for female patients. While
this issue was also recently debated in the context of geriatric
screening tools [40], this distinction did not appear relevant in
Volume 23 | No. 8 | August 2012
original articles
Annals of Oncology
our dataset. Indeed, we applied both approaches when
performing our analysis but did not notice any gender-based
differences in terms of prevalence of subjects with an abnormal
IADL score (data not shown).
As of today, there is no clear definition of what should
constitute the reference exam when evaluating a screening tool
aimed at detecting old cancer patients in need for a CGA. The
existing literature focusing on the validation of the VES-13
questionnaire is an illustration of this issue since it has been
compared with very various reference exams [19–22].
Impairment on the CGA was defined either as meeting the cutoff scores for impairment on two of seven individual tests
within the CGA [19] (including ADL, IADL, Charlson index,
and other usual tests) or as at least two abnormal scores of
another set of seven questionnaires [21] or as having
impairment in two or more domains (ADL and IADL), or
being cognitively impaired (MMS ≤24) [20]. With regard to
the only large study available, the reference exam is not
systematically clearly defined [22]. The experience of the
evaluation of the VES-13 highlights the importance of properly
defining the reference exam.
This first evaluation has some limits. One could argue that
the social environment and biological variables were not
considered in the screening tool. While potentially informative,
some sociocultural variables were collected through a
questionnaire developed for the purpose of the ongoing study.
It was not validated previously, and we thus considered that it
should not be considered part of the reference exam. Although
some biological variables can be associated with
chemotherapy-related toxicity, their evaluation is often part of
the routine oncological assessment and as such cannot be
considered as a component of the CGA. Finally, the QLQ-C30
questionnaire provided valuable information but unlike other
questionnaires, it is not a usual component of the geriatric
evaluation. Given the design of the study, reproducibility of the
G-8 could not be evaluated and should be addressed in a
future study. Similarly, it would be of particular interest to
confirm these preliminary findings in a larger population
allowing us to address a broader range of treatments
(chemotherapy, surgery, and radiotherapy) or cancer types (e.g.
by including breast cancers) through subgroup analyses.
Finally, the VES-13 was not initially included in our study and
as such could not be assessed.
Although the lack of a common reference tool makes the
comparison of screening tools particularly complex, these
preliminary results are promising. When compared with a
reference exam defined as the set of seven geriatric evaluations
(ADL, IADL, MNA, MMSE, GDS-15, CIRS-G, and GUG), this
first evaluation suggests good screening classification properties
for the G-8 tool for use in geriatric oncology. We believe its
excellent sensitivity and reasonable specificity as well as the
ease of administration make the G-8 a promising screening
tool, which now requires validation in a larger-scale
prospective study.
acknowledgements
The authors thank Pippa McKelvie-Sebileau from Institut
Bergonié for medical writing assistance, the research nurses,
and geriatricians from all participating institutions: Centre
Volume 23 | No. 8 | August 2012
Hospitalier Universitaire (CHU) Bordeaux, Centre Hospitalier
Général (CHG) Agen, CHG Libourne, CHG Agen, CHG Mont
de Marsan, CHG Villeneuve sur Lot, Francheville Polyclinique
Périgueux, CHG Pau, CHG Bayonne, CHG Périgueux, and
CHG Le Bouscat. The authors also thank the reviewers for
their helpful comments.
funding
French Ministry of Health (Programme Hospitalier de
Recherche Clinique C2003CTT); Ligue Nationale Contre le
Cancer; and Sanofi-Aventis, Amgen, Chugaï, and Bristol-Myers
Squibb, pharmaceutical companies.
disclosure
The authors declare no conflict of interest.
references
1. Ellis G, Whitehead MA, Robinson D et al. Comprehensive geriatric assessment for
older adults admitted to hospital: meta-analysis of randomised controlled trials.
BMJ 2011; 343: d6553.
2. Huss A, Stuck AE, Rubenstein LZ et al. Multidimensional preventive home visit
programs for community-dwelling older adults: a systematic review and metaanalysis of randomized controlled trials. J Gerontol A Biol Sci Med Sci 2008; 63:
298–307.
3. Rubenstein LZ, Josephson KR, Wieland GD et al. Effectiveness of a geriatric
evaluation unit. A randomized clinical trial. N Engl J Med 1984; 311:
1664–1670.
4. Stuck AE, Siu AL, Wieland GD et al. Comprehensive geriatric assessment: a metaanalysis of controlled trials. Lancet 1993; 342: 1032–1036.
5. Maione P, Perrone F, Gallo C et al. Pretreatment quality of life and functional
status assessment significantly predict survival of elderly patients with advanced
non-small-cell lung cancer receiving chemotherapy: a prognostic analysis of the
multicenter Italian lung cancer in the elderly study. J Clin Oncol 2005; 23:
6865–6872.
6. Caillet P, Canoui-Poitrine F, Vouriot J et al. Comprehensive geriatric assessment in
the decision-making process in elderly patients with cancer: ELCAPA Study. J Clin
Oncol 2011; 29: 3636–3642.
7. Kanesvaran R, Li H, Koo KN et al. Analysis of prognostic factors of comprehensive
geriatric assessment and development of a clinical scoring system in elderly Asian
patients with cancer. J Clin Oncol 2011; 29: 3620–3627.
8. Extermann M, Hurria A Comprehensive geriatric assessment for older patients with
cancer. J Clin Oncol 2007; 25: 1824–1831.
9. Extermann M, Aapro M, Bernabei R et al. Use of comprehensive geriatric
assessment in older cancer patients: recommendations from the task force on
CGA of the International Society of Geriatric Oncology (SIOG). Crit Rev Oncol
Hematol 2005; 55: 241–252.
10. Balducci L, Yates J General guidelines for the management of older patients with
cancer. Oncology (Williston Park) 2000; 14: 221–227.
11. Repetto L, Biganzoli L, Koehne CH et al. EORTC Cancer in the Elderly Task Force
guidelines for the use of colony-stimulating factors in elderly patients with
cancer. Eur J Cancer 2003; 39: 2264–2272.
12. Smith TJ, Khatcheressian J, Lyman GH et al. 2006 update of recommendations
for the use of white blood cell growth factors: an evidence-based clinical practice
guideline. J Clin Oncol 2006; 24: 3187–3205.
13. Lachs MS, Feinstein AR, Cooney LM, Jr et al. A simple procedure for general
screening for functional disability in elderly patients. Ann Intern Med 1990; 112:
699–706.
14. Winograd CH, Gerety MB, Chung M et al. Screening for frailty: criteria and
predictors of outcomes. J Am Geriatr Soc 1991; 39: 778–784.
doi:10.1093/annonc/mdr587 | 
original articles
15. Gill TM, Williams CS, Tinetti ME Assessing risk for the onset of functional
dependence among older adults: the role of physical performance. J Am Geriatr
Soc 1995; 43: 603–609.
16. Brody KK, Johnson RE, Douglas RL Evaluation of a self-report screening
instrument to predict frailty outcomes in aging populations. Gerontologist 1997;
37: 182–191.
17. Rockwood K, Stadnyk K, MacKnight C et al. A brief clinical instrument to classify
frailty in elderly people. Lancet 1999; 353: 205–206.
18. Saliba D, Elliott M, Rubenstein LZ et al. The Vulnerable Elders Survey: a tool for
identifying vulnerable older people in the community. J Am Geriatr Soc 2001;
49: 1691–1699.
19. Mohile SG, Bylow K, Dale W et al. A pilot study of the vulnerable elders survey13 compared with the comprehensive geriatric assessment for identifying
disability in older patients with prostate cancer who receive androgen ablation.
Cancer 2007; 109: 802–810.
20. Kellen E, Bulens P, Deckx L et al. Identifying an accurate pre-screening tool in
geriatric oncology. Crit Rev Oncol Hematol 2010; 75: 243–248.
21. Molina-Garrido MJ, Guillen-Ponce C Comparison of two frailty screening tools in
older women with early breast cancer. Crit Rev Oncol Hematol 2011; 79:
51–64.
22. Luciani A, Ascione G, Bertuzzi C et al. Detecting disabilities in older patients with
cancer: comparison between comprehensive geriatric assessment and vulnerable
elders survey-13. J Clin Oncol 2010; 28: 2046–2050.
23. Falci C, Basso U, Fiduccia P et al. Is vulnerable elders survey 13 (VES-13) a
sensitive and specific screening tool for identifying vulnerable/frail elderly cancer
patients compared to full comprehensive geriatric assessment (CGA)?. Crit Rev
Oncol Hematol 2009; 72: 19–20.
24. Molina-Garrido MJ, Guillen-Ponce C Overvaluation of the vulnerable elders
survey-13 as a screening tool for vulnerability. J Clin Oncol 2011; 29(23):
3201–3202.
25. Bossuyt PM, Reitsma JB, Bruns DE et al. Towards complete and accurate
reporting of studies of diagnostic accuracy: the STARD Initiative. Ann Intern Med
2003; 138: 40–44.
26. Soubeyran P, Rainfray M, Mathoulin-Pelissier S et al. Prediction of early death
risk in the elderly with cancer: results of a prospective multicentric study of 364
patients under chemotherapy. J Clin Oncol 2007ASCO Annual Meeting
Proceedings Part 1 25: 9040.
27. Katz S, Ford AB, Moskowitz RW et al. Studies of illness in the aged. The index of
ADL: a standardized measure of biological and psychosocial function. JAMA
1963; 185: 914–919.
28. Lawton MP, Brody EM Assessment of older people: self-maintaining and
instrumental activities of daily living. Gerontologist 1969; 9: 179–186.
29. Folstein MF, Folstein SE, McHugh PR “Mini-mental state”. A practical method for
grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;
12: 189–198.
30. Sheikh JI, Yesavage JA Geriatric Depression Scale (GDS). Recent evidence and
development of a shorter version. Clin Gerontol 1986; 5: 165–171.
31. Yesavage JA Geriatric Depression Scale. Psychopharmacol Bull 1988; 24:
709–711.
32. Guigoz Y, Vellas B, Garry P Mini nutritional assessment: a practical assessment
tool for grading the nutritional state of elderly patients. Facts Res Gerontol 1994;
4 (Suppl 2): 15–59.
33. Mathias S, Nayak US, Isaacs B Balance in elderly patients: the “get-up and go”
test. Arch Phys Med Rehabil 1986; 67: 387–389.
34. Podsiadlo D, Richardson S The timed “Up & Go”: a test of basic functional
mobility for frail elderly persons. J Am Geriatr Soc 1991; 39: 142–148.
35. Linn BS, Linn MW, Gurel L Cumulative illness rating scale. J Am Geriatr Soc
1968; 16: 622–626.
36. Miller MD, Towers AA Manual of Guidelines for Scoring the Cumulative Illness
Rating Scale for Geriatrics (CIRS-G), Pittsburgh, PA; University of Pittsburgh
1991.
 | Bellera et al.
Annals of Oncology
37. Rubenstein LZ, Harker JO, Salva A et al. Screening for undernutrition in geriatric
practice: developing the short-form mini-nutritional assessment (MNA-SF).
J Gerontol A Biol Sci Med Sci 2001; 56: M366–M372.
38. Barberger-Gateau P, Commenges D, Gagnon M et al. Instrumental activities of
daily living as a screening tool for cognitive impairment and dementia in elderly
community dwellers. J Am Geriatr Soc 1992; 40: 1129–1134.
39. Mura T, Dartigues JF, Berr C How many dementia cases in France and Europe?
Alternative projections and scenarios 2010-2050. Eur J Neurol 2010; 17:
252–259.
40. Luciani A, Floriani I, Foa P Reply to C. Falci et al. J Clin Oncol 2010 Aug 16.
[Epub ahead of print].
appendix
Table A1. The G-8 questionnaire
Items
Possible responses (score)
A Has food intake declined over the past 3
months due to loss of appetite,
digestive problems, chewing, or
swallowing difficulties?
B Weight loss during the last 3 months?
C Mobility?
E Neuropsychological problems?
F BMI? (weight in kg)/(height in m2)
H Takes more than three prescription
drugs per day?
P In comparison with other people of the
same age, how does the patient
consider his/her health status?
Age
Total score
0 = severe decrease in food
intake
1 = moderate decrease in
food intake
2 = no decrease in food
intake
0 = weight loss >3 kg
1 = does not know
2 = weight loss between 1
and 3 kg
3 = no weight loss
0 = bed or chair bound
1 = able to get out of bed/
chair but does not go out
2 = goes out
0 = severe dementia or
depression
1 = mild dementia
2 = no psychological
problems
0 = BMI <19
1 = BMI 19 to <21
2 = BMI 21 to <23
3 = BMI ≥23
0 = yes
1 = no
0.0 = not as good
0.5 = does not know
1.0 = as good
2.0 = better
0: >85
1: 80–85
2: <80
0–17
BMI, body mass index.
Volume 23 | No. 8 | August 2012