

S92
A B S T R A C T S
Methods:
Patients were randomized to either usual care
or geriatric intervention. Patients randomized to the geriatric
intervention received a multidisciplinary assessment which
included the G8 and VES-13 screening tools. Modified FFC
were calculated based on self-reported low physical activity,
self-reported exhaustion, unintentional weight loss
4.5kg in
the last 6 months, low grip strength (
16kg for females and
26kg for males) and gait speed
1m/s on a 4m walk test. G8,
VES-13 and modified FFC were considered abnormal for a
score
14,
3 and
3, respectively.
Patients received a CGA which evaluated 8 domains:
comorbidities, polypharmacy, IADLs, mobility and falls,
nutrition, cognition, mood and social isolation. CGA-based
geriatric interventions were defined as modifications to onco-
logic treatments, management of comorbidities or referral
to an allied health specialist (e.g. physiotherapy, social work,
dietician, etc).
SPSS v23 was used to calculate sensitivity, specificity,
positive predictive value (PPV) and negative predictive value
(NPV) and 95% confidence intervals (CI) for both an abnormal
CGA and a CGA-based medical intervention.
Results:
Thirty patients were randomized to the geriatric
intervention arm. Two patients were not able to complete
a CGA. For the 28 patients analyzed, 22 (79%) had at least 2
abnormal domains on the CGA. CGA-based interventions
occurred in 23 (82%) patients: 18 had modification to
management of comorbidities and 15 were referred to an
allied health specialist. It was not feasible to modify oncologic
management during this trial based on the CGA as all patients
were accrued after starting treatment. The sensitivity,
specificity, PPV and NPV for each screening tool is shown in
Table 1.
Table 1 (abstract P095)
2 abnormal
1 CGA-based
CGA domains
intervention
G8
VES-13
FFC
G8
VES-13
FFC
Sensitivity
0.727
0.455
0.409
0.609
0.304
0.348
Specificity
0.833
1.00
1.00
0.400
0.400
0.800
PPV
0.941
1.00
1.00
0.824
0.700
0.889
NPV
0.454
0.333
0.316
0.189
0.111
0.211
Conclusion:
As previously described, the G8 is more
sensitive than the VES-13 or FFC while VES-13 and FFC
are more specific for detection of patients who may have
an abnormal CGA. If instead of considering these tools as
predicting an abnormal CGA but rather for predicting whether
an intervention will occur based on the CGA, the G8 continues
to be the most sensitive tool although the sensitivity declines
substantially for each. The FFC has the highest specificity
and PPV for predicting a CGA-based intervention. The NPV
remains poor for all 3 measures.
Disclosure of interest:
None declared
Keywords:
Fried frailty criteria, G8, geriatric assessment,
screening tools, VES-13
P096
EFFECTIVENESS OF THE FLEMISH VERSION OF TRIAGE RISK
SCREENING TOOL IN DETECTING FRAILTY IN ELDERLY
PATIENTS UNDERGOING EMERGENCY SURGERY: A PILOT
STUDY
D. Zattoni
1,
*, M. Cervellera
1
, P. Calogero
2
, C. Galetti
2
, A. Garutti
2
,
M. L. Bacchi Reggiani
3
, V. Tonini
1
1
Emergency Surgery,
2
Geriatrics,
3
Cardiology, Policlinico S. Orsola-
Malpighi, Bologna, Italy
Introduction:
Individual frailty is the most important risk
factor for postoperative complications. So far, no frailty-
screening tool has been made available in the emergency
surgical setting where a rapid assessment is highly needed.
Objectives:
The aim of this study is to verify if the Flemish
version of theTriage Risk ScreeningTool (fTRST) is adequate to
predict postoperative outcomes in a group of elderly patients
undergoing emergency abdominal surgery.
Methods:
Consecutive patients, 70-year and older, requir-
ing emergency abdominal surgery were prospectively enrolled
from December 2015 to March 2016.The fTRST was performed
on admission. Demographic, surgical and perioperative data
were collected in a dedicated database. Thirty-day post-
operative mortality, complications and functional outcomes
were recorded.
Results:
79 patients (41 female) were enrolled.The majority
of the population was 80 years and older (54.4%). ‘Major’
abdominal surgery was performed in 42 cases (53%), while
26 (33%) and 12 (14%) patients underwent ‘intermediate’ or
‘minor’ procedures.
Thepostoperative30-daymortalitywas20.2%(16/79) but the
majority of them (56%) occurred during the first postoperative
week. 14 were recorded after major surgery (87.5%). Logistic
regression analysis showed a significant correlation between
30-day mortality and fTRST
2, ASA
4, Charlson Age adjusted
Comorbidity Index (CACI)
7 and ‘major’ surgery. Patients’ age
was not statistically relevant. (Table 1)
Postoperative morbidity rate was 59.5%. Non-fatal major
complications (Clavien-Dindo 3-4) occurred in 5% of cases,
while 68% patients in this group experienced two or more
complications. No significant relationship was shown
between morbidity rate and fTRST or any other risk factor.
Of the 63 patients who survived surgery, 17% (11/63)
developed severe functional loss (loss of walking capacity/
Table 1 (abstract P096) – Logistic regression of 30 days mortality
Sensibility (95% CI) Specificity (95% CI)
PPV (95% CI)
NPV (95% CI)
OR (95% CI)
AUC
R2
p
fTRST
2
93.7% (70%
100%)
41.3% (29%
54%) 28.85% (17%
43%) 96.3% (81%
100%)
10.5 (1.31-84.8)
0.675
0.11
0.027*
ASA
4
69% (41%
89%)
83% (71%
91%)
50% (28%
72%)
91% (81%
97%)
10.4 (3.01-35.98)
0.756
0.1913 <0.001*
CACI
7
75% (48%
93%)
71% (59%
82%)
40% (23%
59%)
92% (80%
98%)
7.5 (2.13-26.35)
0.81
0.1448 0.002*
Major surgery 88% (62%
98%)
88% (62%
98%)
33% (20%
50%)
95% (82%
99%)
8.75 (1.83- 41.75)
0.715
0.1330 0.007*