cardiopulmonaryexercisetestingforpredicting ... · found for breast or brain cancers or lymphoma....

Journal of Anesthesia and Perioperative Medicine

This is an open-access article, published by Evidence Based Communications (EBC). This work is licensed un-der the Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution,and reproduction in any medium or format for any lawful purpose.To view a copy of this license, visit http://cre-ativecommons.org/licenses/by/4.0/.

From the 1Norfolk and Norwich Uni-versity Hospitals NHS FoundationTrust, Colney Lane, Norwich,UK; 2Norwich Medical School, Univer-sity of East Anglia, Norwich, UK.

Correspondence to Dr. Stephen Lamat [email protected].

Citation: Stephen Lam, Andrew Hart.Cardiopulmonary Exercise Testing forPredicting Early Outcomes after Ma-jor Cancer Resection: A SystematicReview. J Anesth Perioper Med 2018;5:136-48.

doi: 10.24015/JAPM.2018.0052

Background: Postoperative complications after major surgery are thought to be associatedwith reduced fitness. Surgical cancer patients are often malnourished, cachexic and subjectto neoadjuvant chemotherapy resulting in low preoperative fitness levels. This review ex-amined the associations between aerobic fitness, as determined objectively by preoperativecardiopulmonary exercise testing (CPEX), and short-term morbidity after cancer surgery.Methods: A literature search using databases of PubMed, Medline, Embase, CumulativeIndex to Nursing and Allied Health Literature (CINAHL) and the Cochrane Library forstudies that examined associations between preoperative CPEX variables and postopera-tive complications following surgery for the ten commonest cancers.Results: A total of 21 observational studies were identified with 4957 patients that under-went CPEX testing prior to lung, colorectal, liver, oesophagogastric, bladder and pancreasresections. The median sample size was 105 patients (range 64 - 1684). No studies werefound for breast or brain cancers or lymphoma. In lung cancer patients undergoing thoracot-omy, a VO2peak ≤ 15ml/kg/min was associated with an increased risk of respiratory complica-tions and death. None of the studies in other cancer types had adequate sample sizes to re-port on mortality. CPEX testing had mostly poor to average discriminatory accuracy to pre-dict postoperative morbidity in other cancer resection surgeries. Findings across studieswere inconsistent, and detection and selective reporting biases were likely to be significant.Conclusion: The utility of CPEX testing prior to cancer surgery is questionable and cur-rently should not be used as a discriminatory tool, except in patients undergoing lung can-cer resection by thoracotomy. Larger studies with more robust methodologies are currentlyrequired to determine the utility of CPEX.

ABSTRACT

Cardiopulmonary Exercise Testing for PredictingEarly Outcomes after Major Cancer Resection:

A Systematic ReviewStephen Lam1, 2, and Andrew Hart1, 2

Systematic Review

136

JAPM WWW.JAPMNET.COM May, 2018 Volume 5 Number 3

Surgical stress response following major sur-gery results in muscle wasting and systemic in-flammation,with a large increase in tissue oxy-

gen demand and consumption (1); increasing therisk of ischaemic events (2). Furthermore, postop-erative bed-rest and incisional pain inhibits normallung mechanics, promoting shallow breathing, at-electasis and infective lung consolidation (3).These physiological challenges are in part met by apatient’ s cardiopulmonary reserves, or their abili-ty to increase cardiac output and ventilation tomeet increased demand. Such reserves are greaterin physiologically ‘fitter’ patients. Cancer patientsrepresent a specific population, more likely tohave underlying malnutrition and cachexia thannon-cancer patients, with depleted fitness levels(4, 5). Neoadjuvant chemotherapy has also beenshown to independently reduce aerobic fitness (5).By measuring a patient’ s preoperative cardiopul-monary reserve, or functional capacity, wemay hy-pothetically be able to discriminate patients thatmay ormay not tolerate the physiological insult as-sociated with surgery. Therefore, measurement ofpreoperative fitness may serve as a preoperativerisk prediction tool for the development of compli-cationsprior tomajorcancer resection.

Cardiopulmonary exercise testing (CPEX) of-fers a quantitative and composite measure offunctional reserve, where a low score indicates apatient’ s inability to effectively transfer oxygento tissues on exertion. The test is carried out inlaboratory conditions usually on a static exercisebike, where pedal resistance, or workload, is sys-tematically increased until volitional termina-tion. Two important CPEX variables are cap-tured by analysis of gas exchange at the mouth,VO2peak (the maximal oxygen consumed at thepeak of exercise) and VO2 at estimated anaerobicthreshold (AT), a measure of sustainable aerobicactivity. Both parameters have shown promise inobservational studies in predicting both morbidi-ty and mortality after major elective surgery (6).In a seminal study from 1993, of 187 elderly pa-tients undergoing major abdominal surgery, apreoperative AT cut-off of 11ml / kg /min had asensitivity of 91% and specificity of 74% for pre-dicting postoperative mortality after major ab-dominal surgery (7). Since then, multiple, oftensmall, observational studies have been publishedin various cancer surgery specialties. However,

the results from these are inconsistent, possiblyexplained by biases inherent in the methodologysuch as performance and detection biases (due tounblinded clinical teams and outcome assessors,respectively) as well as selective reporting bias,which is likely to be substantial (8). Although 3reviews of CPEX testing and major surgerieshave been published (6, 9, 10), all included non-cancer patients and excluded common cancerssuch as lung and bladder. Furthermore, impor-tant sources of bias do not appear to be havebeen adequately addressed previously. The aimof this article was to assess the association be-tween CPEX testing and short-term postopera-tive morbidity and mortality after common ma-jor cancer resection surgery; with due consider-ation of potential biases both in terms of theirmagnitude and direction. If a convincing inverseassociation exists, it may not only support thepreoperative use of CPEX to determine operabil-ity and postoperative monitoring and manage-ment, but also identify fitness as a modifiablerisk factor for investigation in randomized con-trolled trials of surgical cancer patients.

METHODS

The preferred reporting items for systematic re-views (PRISMA) guidelines were used to stan-dardize the methods of conducting and report-ing this review. The ten commonest causes ofcancer deaths in the UK in 2014 were identified(lung, bowel, breast, prostate, pancreas, oesoph-agus, bladder, brain, liver, lymphoma) (11), anda literature search was conducted usingPubMed, Medline, Embase, CINAHL and theCochrane library (from commencement to 5thJuly 2017) for studies that examined associa-tions between preoperative CPEX and postoper-ative complications following cancer resectionsurgery. The search terms used were: CPEX,CPET, exercise testing, anaerobic threshold,VCO2, ventilatory inefficiency, oxygen consump-tion, VO2, preoperative exercise, aerobic exer-cise. For each cancer, additional specific searchterms were added (Supplementary Appendix 1).We included analytical studies (cohort, case-con-trol, randomized controlled trials) that investi-gated the association between preoperativeCPEX variables and short-term (up to 90 days)

Cardiopulmonary Exercise Testing for Predicting Early OutcomesStephen Lam et al.

137


JAPM WWW.JAPMNET.COM Volume 5 Number 3May, 2018

morbidity and mortality. Studies were excludedwith sample sizes <100 patients, unless therewere <2 studies in that surgical population inwhich case the total sample size needed to be atleast 60, which was considered the minimumsize to determine associations with a moderateevent rate. In order to select studies which exam-ined only cancer populations, studies with alarge proportion (≥ 25%) of non-cancer patientswere also excluded (unless cancer patients wereanalysed separately) as were investigations thatcombined multiple surgical patient populations(e.g. colorectal, urological and upper gastrointes-tinal cancer patients), unless these groups weresub-analysed.

RESULTS

A total of 21 hospital-based cohort studies (12prospective, 9 retrospective) were included inthis review with a total of 4957 patients that un-derwent CPEX testing prior to lung (12-17),colorectal (18-20), liver (21-23), oesophagogas-tric (24-26), bladder (27-29) and pancreas resec-tion (30-32) (Figure 1). No studies were found inbreast, brain or lymphoma cancers. No random-ized controlled trials of any cancer site were iden-tified. Data were extracted from each study (in-cluding study design, sample size, outcome mea-surement and effect sizes) and tabulated (Table 1).

Lung CancerFour hundred and three studies were identified,with 63 potentially relevant papers by title. A re-view of these abstracts identified 23 potentiallyrelevant papers. Six studies met the inclusion cri-teria and were included in this review (12-17).All were observational cohort studies (2 prospec-tive, 4 retrospective), with a total of 2814 pa-tients from hospitals within Europe and theUSA. The largest study was a retrospective analy-sis of 1684 patients who had lung cancer sur-gery identified from the European Society ofThoracic Surgeons (ESTS) database (12), whichis voluntarily contributed to by clinicians from235 sites across Europe. The authors reportedno association between VO2peak and all-causemorbidity (for either video-assisted thoracoscop-ic surgery (VATS) or thoracotomy resections),but a significant association with mortality in pa-

tients undergoing thoracotomy resection with aVO2peak < 15ml/kg/min (P = 0.008). Whilst thisstudy had several strengths including, a largesample size and use of propensity matching toreduce selection bias, it had significant limita-tions, which are acknowledged by the authors.The largest of these was the potential for mea-surement error (for complications) through theuse of a voluntary multi-institutional database,where the accuracy of the data entry has notbeen validated (12). Inaccurate recording of out-come would lead to an attenuation of any associ-ations between CPEX variables and outcome.

The second largest study was a multi-center (9centers) prospective observational investigationof 346 patients from the USA who underwentthoracotomy and lung cancer resection surgery(13). In contrast to the results of the previousstudy, the authors reported that VO2peak was sig-nificantly lower in the group with complications(15.2 ml / kg / min) compared to those without(16.7 ml /kg /min) although the mean differencewas small: -1.47 ml/kg/min (95% CI 0.55-2.4),P = 0.002. The authors also undertook a furthersub-analysis in patients with the outcomes of re-spiratory failure (n = 33) and death (n = 15).Both events were associated with a lower meanVO2peak (14.7 ml/kg/min), P = 0.041 compared tothose without complications (mean difference 2ml/kg/min), although the small number of eventsis noted. Whilst the large sample size, prospec-tive design and multi-center participation in-creases power, reduces bias and enables general-isability, respectively, there were limitations. Apre-defined criterion for what constituted eachcomplication was not established, nor were thepost-operative outcomes measured in a blindedfashion, both of which could contribute to detec-tion bias resulting in spurious over-estimation ofthe associations found. Furthermore, the authorsdefined postoperative morbidity as a compositeoutcome, which included complications lackinga clear plausible mechanistic relationship withaerobic fitness; such as red blood cell transfu-sions (n = 38). Such an event is more likely to beassociated with intra-operative blood loss (per-haps due to longer operating times in patientswith underlying lung disease), suggesting that re-sidual confounding may explain some of the as-sociations found.

Systematic Review Cardiopulmonary Exercise Testing for Predicting Early OutcomesStephen Lam et al.

138



Figure 1. The PRISMA Diagram.

The other 4 studies in lung cancer patientswere all hospital-based cohort studies, which re-ported inconsistent findings (14-17). Two of theinvestigations were by the same research group(15, 17), which reported an inverse associationbetween VO2peak and respiratory complications(P = 0.015) in one of their studies (15), but wereunable to subsequently replicate this finding(P = 0.5) in the other despite similar methodolo-gies and patient populations (17). Detection biasdue to non-blinded assessment may have contrib-uted to the variability in findings. The two otherstudies reported an inverse association betweenVO2peak and cardiopulmonary complications, OR0.05 (95% CI 0.01-0.58), P = 0.0216 and OR0.79, 95% CI 0.71-0.88, P ≤ 0.0001 (14). How-ever, the large variation in the effect size esti-mates and large confidence intervals reflects theimprecision of the findings. Again, outcome as-sessment was not blinded, which may have spuri-ously inflated the reported effect sizes.

Author’ s summary: Observational studieshave reported that VO2peak appears to be associat-ed with complications after lung cancer resectionsurgery. More specifically, the two largest studiesreported that a VO2peak <15 ml/kg/min was associ-ated with an increased risk of respiratory failure(13) and death following lung cancer resectionby thoracotomy (12, 13). However, only onestudy in lung cancer surgery examined outcomesafter VATS and found no association betweenVO2peak and morbidity or mortality (12). As VATSis increasingly becoming used for lung cancer re-section, the utility of CPEX testing needs to beupdated to determine if it is of value in less inva-sive surgeries for lung cancer resection.

Colorectal CancerThe literature search identified 431 studies, ofwhich 30 were considered relevant based ontheir title. These were reduced to 13 after read-ing the abstracts, of which 3 met the inclusion cri-teria and were included in this review (18-20).The main reason for excluding other investiga-tions was their inclusion of > 25% of non-cancerpatients. All 3 were observational studies fromthe same UK group. The largest was a retrospec-tive multi-center (6 sites) UK investigation of 703patients (20), most of which had a malignant dis-ease (87%). In contrast to the studies in lung can-

cer surgery, all-cause morbidity was measured us-ing a validated PostOperative Morbidity Survey(POMS) at postoperative day 5 (33, 34). The se-verity of complications was graded using the sys-tem devised by Clavien and Dindo (34). Thisgrading system is based upon the level of inter-vention required to treat a complication; fromnormal postoperative adjuncts such as supple-mentary oxygen, analgesia and anti-emetics(grade 1), to additional medicines above usualstandard care, including antibiotics (grade 2).Grade 3 is a complication requiring operative in-tervention and grade 4 complications require or-gan support in critical care. The authors reporteda significant difference in both median VO2peak

and AT in patients with and without all-causemorbidity of any Clavien-Dindo grade (P =0.031 and P = 0.002, respectively). Receiver Op-erator Characteristic (ROC) Curve analyses werealso undertaken to shows how sensitivity andspecificity varied with changing thresholds,

139



Lung

canc

er

Beg

um,2

015

(12)

Bru

nelli

,201

2(1

7)

Lick

er,2

011

(14)

Torc

hio,

2010

(16)

Bru

nelli

,200

9(1

5)

Loew

en,2

007

(13)

Col

orec

talc

ance

r

Wes

t,20

16

(20)

Wes

t,20

14

(18)

Ret

rosp

ectiv

e

case

-mat

ched

mul

ti-ce

ntre

coho

rtst

udy

Ret

rosp

ectiv

e

hosp

ital-b

ased

coho

rtst

udy

Ret

rosp

ectiv

e

hosp

ital-b

ased

coho

rtst

udy

Ret

rosp

ectiv

e

hosp

ital-b

ased

coho

rtst

udy

Pro

spec

tive

hosp

ital-b

ased

coho

rtst

udy

Pro

spec

tive

hosp

ital-b

ased

mul

ti-ce

ntre

coho

rtst

udy

Ret

rosp

ectiv

e

mul

ti-ce

ntre

coho

rtst

udy

Ret

rosp

ectiv

e

hosp

ital-b

ased

coho

rtst

udy

1684

225

210

145

204

346

703

95

In-h

ospi

tal

card

iopu

lmon

ary

com

plic

atio

ns(3

0%)

30-d

ay

card

iopu

lmon

ary

com

plic

atio

ns(2

3%)

In-h

ospi

tal

card

iopu

lmon

ary

com

plic

atio

ns(2

2%)

30-d

ay

card

iopu

lmon

ary

com

plic

atio

ns(1

5%)

In-h

ospi

tal

card

iopu

lmon

ary

com

plic

atio

ns(2

3%)

30-d

ay

card

iopu

lmon

ary

com

plic

atio

ns(4

0%)

All

caus

em

orbi

dity

atpo

st-

oper

ativ

eda

y5

(36.

7%)

All

caus

e

mor

bidi

tyat

post

oper

ativ

eda

y5

(48%

)

No

Yes Yes Yes Yes No

PO

MS

PO

MS

No

No

No

No

No

No

Cla

vien

Din

do

Cla

vien

Din

do

No

No

No

No

No

No

No

Yes

No

No

Yes

(OR

=0.

79,

95%

CI

0.71

-0.8

8,P

≤0.0

001)

Yes

(OR

=0.

05,

95%

CI

0.01

-0.5

8,P

=0.

017)

No

Yes

(No

RR

/OR

/HR

re-

port

ed)

Yes

(≤18

.2m

l/kg

/min

,

OR

2.15

,95

%C

I1.

01-

4.59

,P=

0.05

)

Yes

(≥18

.8m

l/kg

/min

,

OR

0.07

,95

%C

I0.

03-

0.19

,P≤0

.001

)

Not

mea

sure

d

Not

mea

sure

d

Not

mea

sure

d

Not

mea

sure

d

No

Not

mea

sure

d

Yes

(≤11

.1m

l/kg

/min

,

OR

7.56

,95

%C

I4.

44-

12.8

6,P

≤0.0

01)

Yes

(≥11

.2m

l/kg

/min

,

OR

0.07

,95

%C

I0.

03-

0.19

,P≤0

.001

)

Tab

le1.

Ch

arac

teri

stic

so

fS

tud

ies

Inve

stig

atin

gth

eA

sso

ciat

ion

bet

wee

nP

reo

per

ativ

eC

PE

Xan

dP

ost

op

erat

ive

Co

mp

licat

ion

sfo

llow

ing

Maj

or

Can

cer

Res

ec-

tio

nS

urg

ery.

Au

tho

r,

Yea

rS

tud

yD

esig

nS

amp

le

Siz

e(n

)

Po

sto

per

ativ

e

Co

mp

licat

ion

sM

easu

red

(rat

eo

fo

ccu

rren

ce)

Dia

gn

ost

ic

Cri

teri

afo

r

Co

mp

licat

ion

s?

Co

mp

licat

ion

Sev

erit

y

Cla

ssif

ied

?

Blin

ded

Co

mp

licat

ion

Ass

essm

ent?

Ass

oci

atio

nb

etw

een

VO

2pea

kan

dM

easu

red

Co

mp

licat

ion

s

Ass

oci

atio

nb

etw

een

AT

and

Mea

sure

d

Co

mp

licat

ion

s


140



Wes

t,20

14

(19)

Live

rca

ncer

Kas

ivis

vana

than

,

2015

(23)

Dun

ne,2

014

(22)

June

jo,2

012

(21)

Oes

opha

geal

canc

er

Moy

es,2

013

(25)

For

shaw

,200

8

(24)

Nag

amat

su,2

001

(26)

Pro

spec

tive

hosp

ital-b

ased

coho

rtst

udy

Pro

spec

tive

hosp

ital-b

ased

coho

rtst

udy

Ret

rosp

ectiv

e

hosp

ital-b

ased

coho

rtst

udy

Pro

spec

tive

hosp

ital-b

ased

coho

rtst

udy

Ret

rosp

ectiv

e

hosp

ital-b

ased

coho

rtst

udy

Ret

rosp

ectiv

e

hosp

ital-b

ased

coho

rtst

udy

Ret

rosp

ectiv

e

hosp

ital-b

ased

coho

rtst

udy

136

108

197

92 103

78 91

All

caus

e

mor

bidi

tyat

post

oper

ativ

eda

y5

(48%

)

3-da

yal

lcau

se

mor

bidl

y(7

0%)

All

caus

ein

-hos

pita

l

mor

bidi

ty(4

4%)

All

caus

ein

-hos

pita

l

mor

bidi

ty(5

1%)

All

caus

ein

-hos

pita

l

mor

bidi

ty(5

5%)

All

caus

ein

-hos

pita

l

mor

bidi

ty(n

otst

ated

)

In-h

ospi

tal

card

iopu

lmon

ary

com

plic

atio

ns(1

9%)

PO

MS

PO

MS

No

PO

MS

CT

CA

E

CT

CA

E

No

Cla

vien

Din

do

Cla

vien

Din

do

Cla

vien

Din

do

No

No

No

No

Yes Yes No

No

No

No

No

Yes

(No

RR

/OR

/HR

re-

port

ed)

Yes

(uni

nter

pret

able

from

data

prov

ided

)

No

No

No

Yes

(No

RR

/OR

/HR

re-

port

ed)

Yes

(No

RR

/OR

/HR

re-

port

ed)

Yes

(OR

0.77

,95

%C

I

0.66

-0.8

9,P

=0.

0005

)

Yes

(uni

nter

pret

able

from

data

prov

ided

)

No

No

Yes

(No

RR

/OR

/HR

repo

rted

)

No

No

Tab

le1.

Ch

arac

teri

stic

so

fS

tud

ies

Inve

stig

atin

gth

eA

sso

ciat

ion

bet

wee

nP

reo

per

ativ

eC

PE

Xan

dP

ost

op

erat

ive

Co

mp

licat

ion

sfo

llow

ing

Maj

or

Can

cer

Res

ec

-tio

nS

urg

ery

(con

tinue

d).

Au

tho

r,

Yea

rS

tud

yD

esig

nS

amp

le

Siz

e(n

)

Po

sto

per

ativ

e

Co

mp

licat

ion

sM

easu

red

(rat

eo

fo

ccu

rren

ce)

Dia

gn

ost

ic

Cri

teri

afo

r

Co

mp

licat

ion

s?

Co

mp

licat

ion

Sev

erit

y

Cla

ssif

ied

?

Blin

ded

Co

mp

licat

ion

Ass

essm

ent?

Ass

oci

atio

nb

etw

een

VO

2pea

kan

dM

easu

red

Co

mp

licat

ion

s

Ass

oci

atio

nb

etw

een

AT

and

Mea

sure

d

Co

mp

licat

ion

s

141



Bla

dder

canc

er

Lam

b,20

16(2

9)

Tolc

hard

,201

4

(28)

Pre

ntis

,201

3(2

7)

Pan

crea

ticca

ncer

Cha

ndra

bala

n,

2013

(31)

June

jo,2

013

(32)

Aus

ania

,201

2

(30)

PO

MS

=P

osto

pera

tive

Mor

bidi

tyS

urve

y(a

valid

ated

mea

sure

ofpo

stop

erat

ive

com

plic

atio

nou

tcom

es).

Cla

vien

-Din

dois

ast

anda

rdis

edth

erap

yor

ient

ated

grad

ing

syst

emfo

rth

ese

-

verit

yof

post

oper

ativ

eco

mpi

latio

nsin

surg

ical

prac

tice.

CT

CA

E=

Com

mon

Term

inol

ogy

forA

dver

seE

vent

s(a

desc

riptiv

ete

rmin

olog

yus

edin

Adv

erse

Eve

nt(A

E)r

epor

ting,

usua

llyin

clin

ical

drug

tria

ls).

ISG

PF

=In

tern

atio

nals

tudy

grou

pon

Pan

crea

ticF

istu

la(a

defin

ition

deriv

edby

anin

tern

atio

nalp

anel

ofpa

ncre

atic

surg

eons

).

Pro

spec

tive

hosp

ital-b

ased

coho

rtst

udy

Pro

spec

tive

hosp

ital-b

ased

coho

rtst

udy

Pro

spec

tive

hosp

ital-b

ased

coho

rtst

udy

Ret

rosp

ectiv

e

hosp

ital-b

ased

coho

rtst

udy

Pro

spec

tive

hosp

ital-b

ased

coho

rtst

udy

Ret

rosp

ectiv

e

hosp

ital-b

ased

coho

rtst

udy

82 105

69 100

64 124

30-d

ayC

lavi

en-D

indo

grad

e≥3,

allc

ause

mor

bidi

ty(1

2.6%

)

90-d

ayal

lcau

se

mor

bidi

ty(3

1%)

In-h

ospi

tal

allc

ause

mor

bidi

ty(5

6%)

Not

stat

ed

In-h

ospi

tal

allc

ause

mor

bidi

ty(6

4%)

In-h

ospi

tal

allc

ause

mor

bidi

ty(4

4%)

No

No

No

ISG

PF

ISG

PF

ISG

PF

&P

OM

S

Cla

vien

-Din

do

(gra

de≥3

only

)

Cla

vien

-Din

do

Cla

vien

-Din

do

Cla

vien

-Din

do

Fis

tula

byIS

-

GP

F

Fis

tula

byIS

-

GP

F

No

Yes Yes No

No

No

No

Yes

(No

RR

/OR

/HR

re-

port

ed)

No

Not

stat

ed

No

No

No

Yes

(AT

≥11.0

ml/

kg/

min

,O

R0.

72,

95%

CI

0.58

-0.9

1)

No

No

No

Yes

(AT

≤10.1

ml/

kg/

min

,O

R5.

79,

95%

CI

1.62

-20.

63)

Tab

le1.

Ch

arac

teri

stic

so

fS

tud

ies

Inve

stig

atin

gth

eA

sso

ciat

ion

bet

wee

nP

reo

per

ativ

eC

PE

Xan

dP

ost

op

erat

ive

Co

mp

licat

ion

sfo

llow

ing

Maj

or

Can

cer

Res

ec

-tio

nS

urg

ery

(con

tinue

d).

Au

tho

r,

Yea

rS

tud

yD

esig

nS

amp

le

Siz

e(n

)

Po

sto

per

ativ

e

Co

mp

licat

ion

sM

easu

red

(rat

eo

fo

ccu

rren

ce)

Dia

gn

ost

ic

Cri

teri

afo

r

Co

mp

licat

ion

s?

Co

mp

licat

ion

Sev

erit

y

Cla

ssif

ied

?

Blin

ded

Co

mp

licat

ion

Ass

essm

ent?

Ass

oci

atio

nb

etw

een

VO

2pea

kan

dM

easu

red

Co

mp

licat

ion

s

Ass

oci

atio

nb

etw

een

AT

and

Mea

sure

d

Co

mp

licat

ion

s


142



which was expressed as an area under the ROCcurve (AUC). The AUC takes into considerationthe accuracy of a diagnostic test (in terms of sen-sitivity and specificity) across a range of thresh-old values (35). In the context of CPEX testing,where the association may be inverse, the AUC isequal to the probability that if a pair of patients(one with a complication and one without) areselected at random, the patient with a complica-tion will have a lower CPEX value than the com-plication-free patient (35). An AUC of 1.0 indi-cates a perfect test and 0.5 a completely uninfor-mative test i. e. a result occurring by chance. AnAUC of <0.7 would be indicative of a poor pre-dictive test, 0.7-0.8 average accuracy and > 0.8good accuracy as a diagnostic test across a rangeof thresholds (35). In this study, AT had averagediscrimination (i. e. 0.70-0.80) with an AUC of0.79, 95% CI 0.76-0.83 with an optimal cut-point at 11.1 ml / kg / min (78% sensitivity and71% specificity). The AUC for VO2peak was 0.77,95% CI 0.71-0.82 with an optimal cut-point of18.2ml/kg/min (70% sensitivity and 72% speci-ficity). However, similar to the previous studiesin lung cancer, outcome assessors were not blind-ed to CPEX data, so detection bias could explainthe associations found. Indeed, there were signif-icant variations (P ≤ 0.001) in AUC values acrossrecruited hospital sites (supplementary material),where the largest recruiting center (n = 239) hadmore modest values for VO2peak (AUC 0.73) andAT (AUC 0.68) compared to the above-pooledvalues quoted in the study. Complications with-out a clear plausible relationship with aerobic fit-ness were associated with CPEX values includ-ing; postoperative pain and gastrointestinalsymptoms (such as ileus). Therefore, residualconfounding may explain some inverse associa-tions found. The same group previously pub-lished a prospective blinded observational studyin 136 patients undergoing colonic surgery, mostof whom (89%) had a malignant disease (19).With detection bias removed, the predictive per-formance of CPEX for day 5 morbidity was poor(i.e., AUC <70). For AT the AUC was 0.63, 95%CI 0.54-0.73, with a lower optimal cut-point at10.1 ml/kg/min (68% sensitivity and 58% speci-ficity). The AUC for VO2peak was 0.63, 95% CI0.53-0.73 with an optimal cut-point of 16.7 ml/kg / min (55% sensitivity and 69% specificity).

Furthermore, 14% of their sample who under-went CPEX testing and surgery were excluded asthey “lacked complete data”. It is unclear wheth-er this data was missing at random, and maytherefore represent a source of selection bias.

The final study by the same group investigat-ed 95 rectal cancer patients undergoing resec-tion surgery, 68 of whom received neoadjuvanttreatment (18). Morbidity and mortality weremeasured blinded, using the same methods re-ported in their other work (19, 20). Both VO2peak

and AT were associated with total morbidity atday postoperative day 5. For AT, the AUCshowed good accuracy at 0.87 (95% CI 0.79-0.95) with an optimal cut-point of 10.6 ml /kg/min (84% sensitivity and 92% specificity).VO2peak had an AUC of 0.85 (95% CI 0.77-0.93)and cut-point of 18.6 ml/kg/min (82% sensitivi-ty and 80% specificity). However, the small sam-ple size increases the risk of a chance finding.

Author’ s conclusion: Overall the associationbetween preoperative CPEX and postoperativeoutcome following colorectal cancer surgery isderived from observational studies by the sameresearch group. There does appear to be inverseassociations between VO2peak and AT and all-cause morbidity at day 5 post surgery. However,detection bias and residual confounding cannotbe excluded. Furthermore, 5-day POMS morbid-ity measured complications of a low severity(Clavien-Dindo ≤ 2) in the majority of patients,which makes the clinical usefulness of these find-ings questionable. POMS has also not been vali-dated as an index of overall morbidity (33). Thedecision of whether or not to undergo surgery isunlikely to be meaningfully informed by thiswork. Larger multicentre studies are required toaddress whether improved fitness prior tocolorectal surgery reduces the risk of major post-operative outcomes including death, with out-come assessment blinded to CPEX data.

Liver CancerOne hundred and eight studies were identifiedusing the search terms, of which 7 were consid-ered relevant based on their title. These were re-duced to 4 after reading the abstracts and 3 metthe inclusion criteria and were included in thisreview (21-23). All 3 were UK hospital-based co-hort investigations (2 prospective and 1 retro-

143



spective) of patients that underwent both majorand minor hepatectomies (21-23). The largest, aretrospective study of 197 patients, found no as-sociation between in-hospital morbidity andVO2peak or AT (22). The study did not measurecomplications using a validated outcome mea-sure or blind its assessors to CPEX values. How-ever, the result of such bias may inflate, ratherthan reduce, the effect size reported. The secondlargest study, a UK prospective cohort of 104 pa-tients, did report an association between bothVO2peak, AT and complications (23). This used avalidated outcome measure (POMS) with asses-sors blinded to the CPEX scores. However, theauthors chose to report complications at postop-erative day 3. This timeframe was not defendedin the study, and in the absence of a pre-definedprotocol, selective reporting bias cannot beruled out, which may have produced a false posi-tive result. Furthermore, the high complicationrate (70%) likely reflects routine, less severe,postoperative interventions (pain medication,urinary catheter, oxygen supplementation),which are likely to be clinically insignificant andrare beyond day 3 (33). When the authors grad-ed complication severity according to Clavien-Dindo they found no associations betweenCPEX variables and complications of grade 3(needing surgical intervention) or above. The fi-nal study was a UK prospective cohort investiga-tion of 92 patients, which reported no associa-tion between VO2peak or AT and 30-day morbidi-ty (as measured by POMS) (21). The authors diddocument that VE / VCO2 (a CPEX measure ofventilatory efficiency) was associated with com-plications, but its predictive value was poor;where a value of 34.5 provided a sensitivity of47% for complications.

Author’ s summary: The evidence base forCPEX and liver cancer surgery is derived fromsmall observational studies with poor qualityoutcome reporting. From the evidence to date,there is insufficient data demonstrating an associ-ation between CPEX and outcome following he-patic resection. A large well designed multicen-tre study is needed to assess the role of preopera-tive CPEX in liver cancer surgery.

Oesophageal CancerFour hundred and seventy-eight studies were

identified of which 11 were considered relevantbased on their title, reduced to 5 after readingthe abstracts. Of these, 3 met the inclusion crite-ria and were included in this review (24-26). All3 were retrospective hospital-based cohort stud-ies with small sample sizes from single institu-tions. The largest was a Japanese analysis of 91patients who underwent McKeown oesophagec-tomy for squamous cell carcinoma (26). Onlycardiopulmonary complications were measuredand occurred in 19% of patients. The meanVO2peak was found to be lower in those with, vs.those without cardiopulmonary complications(789 ml/min/m2 vs. 966 ml/min/m2, t-test, P ≤0.001). These values approximate to 20.9 ml/kg/min vs. 25.6 ml/kg/min [conversion using the av-erage height and weight of a Japanese male](36). No association was found between AT andcomplications (t-test, P = 0.12). The secondlargest study was a UK investigation of 78 pa-tients, predominantly with adenocarcinoma(74%), undergoing oesophagectomy (64% withneoadjuvant chemotherapy) (24). Cardiopulmo-nary outcomes occurred in 42% of patients (n=33) and non-cardiopulmonary in 24% (n=19).Similar to the Japanese study, a low mean VO2peak

was associated with cardiopulmonary complica-tions although the mean difference was small(19.2 ml/kg/min in those with complications vs.21.4 ml / kg / min in those without, t-test, P =0.04). AT was also not associated with complica-tions (13.2 ml / kg /min in those with complica-tions vs. 14.4 ml/kg/min in those without, t-test,P = 0.07). ROC curve analysis estimated thepredictive value of both VO2peak and AT to bepoor (i.e., < 70), AUC 0.63 (95% CI 0.50-0.76,P = 0.02) and 0.62 (95% CI 0.49-0.75, P =0.03), respectively. The same group subsequent-ly published a further study of 103 patients withboth oesophageal and gastric cancers that under-went CPEX testing prior to oesophagectomy(62%) and gastrectomy (25). The findings werethe reverse of their previous work, in that, thistime; a lower AT was associated with cardiopul-monary complications (9.9 ml / kg /min in thosewith complications vs. 11.2 ml/kg/min in thosewithout, P = 0.05), while VO2peak was not (16.6ml/kg/min in those with complications vs. 14.6ml / kg /min in those without, P = 0.07). ROCanalysis again reported both AT and VO2peak to


144



be poorly predictive of complications (AUC0.62 (95% CI 0.50-0.74, P = 0.06) and 0.60(95% CI 0.48-0.72, P = 0.08, respectively). Themost significant limitations of all three studies,apart from their small sample sizes and single in-stitution design are the potential for detection bi-as for complications due to unblinded outcomeassessments, particularly for complicationswhich can be subjectively diagnosed (e.g. pneu-monia). This could lead to an over-estimate ofthe association between CPEX variables and out-comes. Such methodological error may explainwhy the same group were unable to replicatetheir previous findings (24, 25).

Author’ s conclusion: Associations betweenCPEX variables and outcome after oesophagec-tomy are from small retrospective observationalstudies that did not use a validated measure ofpostoperative outcomes, or capture complica-tions with blinding to CPEX values. The ab-sence of blinding could result in an inflation ofthe association between CPEX values and post-operative outcomes. Further large studies wherecomplications are strictly defined and measuredby assessors blinded to CPEX values are needed.

Bladder CancerThirty-five studies were identified, with 9 poten-tially relevant papers identified by their title. Areview of these 9 abstracts identified 5 potential-ly relevant papers, but only 3 met the inclusioncriteria. All were prospective hospital-based co-hort studies, with a total of 256 patients fromhospitals in the UK (27-29). The largest was of105 patients who underwent preoperativeCPEX testing prior to either robot-assisted (n=38) or open (n=67) cystectomy (28). Complica-tions were measured at day 90 by blinded asses-sors and were associated with a significantly low-er median AT (10.6 vs 11.8, U-test, P = 0.007)and VO2peak (14.3 vs 15.4, U-test, P = 0.02) com-pared to those without complications. Addition-ally, VE/VECO2 was higher in the complicationgroup than in those without (33.3 vs 30.3, U-test P=0.007). Whilst these findings are convinc-ing in that there was consistency of associationsacross 3 CPEX values, the small study samplefrom a single institution presents a significantlimitation to a more generalized interpretationof the results. The second largest study was of

82 patients who underwent CPEX prior to intra-corporal robotic-assisted radical cystectomy(29). There were no associations between anyCPEX variable and outcome. However, thesmall sample size and small number of complica-tions (n=14) likely result in a lack of statisticalpower to detect a clinically meaningful associa-tions with CPEX. The smallest prospective co-hort study was of 69 patients undergoing radicalcystectomy. Again, no CPEX values were predic-tive of complications when the patients were di-vided into two groups, composed of those withand without a complication. However, sub-anal-ysis according to the presence of a Clavien-Din-do grade ≥ 3 complication (n = 13 vs n = 56)found an inverse association between AT andmajor complication risk, OR 0.74 (95% CI 0.57-0.97). Again, the results of this single institutionstudy with a small sample size are difficult to in-terpret, particularly when post-hoc analyses ac-cording to different complication severities wereundertaken, which increase the risk of a chancefinding in a small sub-group.

Author’s summary: The evidence of an associ-ation between CPEX and post-cystectomy out-come is from small single-institution studies,which report conflicting findings. Thereforelarger studies which limit sources of bias are re-quired to clarify whether an association exists.

Pancreatic CancerThirty-one studies were identified, with 8 poten-tially relevant papers by title. A review of these8 abstracts identified 3 potentially relevant pa-pers and all 3 met the inclusion criteria (30-32).These were UK hospital based-cohort studies (2retrospective, 1 prospective) with a total of 288patients. None of the studies used blinded out-come assessments. The largest was a retrospec-tive study of 124 patients who had CPEX test-ing prior to pancreaticoduodenectomy (30).Complications occurred in 44% of patients andwere defined using POMS (33) and the Interna-tional Study Group definition of Pancreatic Fis-tula (ISGPF) (37). There were no associationsbetween VO2peak and complications, includingpancreatic leak. AT was dichotomised using acut-point of 10.1 ml / kg / min (a value derivedfrom their previous work) (38), and included ina multivariable logistic regression model, which

145



estimated that AT <10.1ml / kg /min greatly in-creased the odds of a pancreatic leak, OR=5.79(95% CI 1.62-20.63). The imprecision of this es-timate likely reflects both the dichotomisation ofa continuous variable and the small pancreaticleak events in the total sample (n=29). The sec-ond largest study was a retrospective analysis of100 patients that underwent preoperative CPEXand major pancreatic surgery (98% pancreatico-duodenectomy) (31). Again, AT was dichoto-mized, rather than treated as a continuous vari-able. The point of dichotomization (10 ml / kg /min) was not justified in the report and unclearif chosen a priori. The results showed a greaterincidence of pancreatic leak (occurring in 25 pa-tients), when AT was < 10ml/kg/min (35.4% vs.16%, P = 0.028). However, statistical signifi-cance was lost when leaks were graded accord-ing to the ISGPF classification (P = 0.091). Se-lective reporting bias cannot be excluded, partic-ularly when a seemingly arbitrary threshold wasused to dichotomize a continuous variable. Fur-thermore, dichotomization during analysis re-sults in a loss of statistical power and increasesthe risk of a false positive result (39). The finalinvestigation was a prospective cohort study of64 patients undergoing pancreaticoduodenecto-my, which included a per-protocol statisticalanalysis (32). The authors reported no associa-tion between AT or VO2peak and all complica-tions, OR 1.07 (95% CI 0.83-1.39) and OR1.00 (95% CI 0.86-1.18), respectively.

Author’ s summary: The studies in pancreaticcancer have small patient numbers and potentialsources of detection and selective reporting biaseswhich makes interpretation of their findings diffi-cult. Similar to other cancer resection surgeries,large well-designed studies are needed to clarifywhether there is an association between fitness, asdetermined by CPEX testing, and outcome.

DISCUSSION

Overall, the evidence for associations betweenpreoperative CPEX values and postoperativeoutcome after cancer resection surgery is mostlyderived from small observational studies. Manywere underpowered to report on mortality risk.However, in the largest study of its kind, a lowVO2peak was associated with mortality after lung

cancer resection, but only in thoracotomy sur-gery (not in VATS) (12). For morbidity, preoper-ative CPEX testing has poor (19, 24, 25) to aver-age (20) discriminatory accuracy to predict post-operative outcomes after cancer resection sur-gery, so has limited utility as an isolated preoper-ative screening tool. Furthermore, investigationsoften used composite outcomes and includedlow Clavien-Dindo graded complications (19,20, 23), which reduces the clinical meaningful-ness of associations found (40). Much researchin CPEX testing has been a continuation of theseminal work of Older et al, in a paper pub-lished in 1993; reporting that AT may predictpostoperative cardiac-related death after majorsurgery (7). However, there seems to have beenincorrect interpretations of this original plausi-ble hypothesis. Cardiac-related death has a rela-tionship with aerobic fitness; mortality eventsare a reflection of how patients respond oncecomplications have occurred (41). Patients thatdie as a result of such complications are likely tolack the necessary cardiopulmonary and muscu-loskeletal reserves, which are required whenthere are ongoing physiological stresses and pro-longed ITU bed-rest (41). These outcomes arevery different to those measured in some of theCPEX studies in this review, which include surgi-cal wound infection on postoperative day 3 (23)or increased analgesia need due to postoperativepain (as measured by POMS) (20, 33). Ideally,observational studies investigating cardiac deathare required, but pragmatically may be difficultdue to the small number of such events. Howev-er, to justify the pre-operative use of CPEX incancer surgery, it should be shown that CPEXcan accurately identify patients at risk of signifi-cant postoperative complications.

This is the first review to assess the utility ofpreoperative CPEX to predict postoperativecomplications after major cancer resection sur-gery. And, whilst it is plausible that greater car-diopulmonary and skeletal muscle reserves (asmeasured by CPEX) could reduce postoperativecomplications following major cancer resectionsurgery, evidence from the reviewed observation-al studies suggests that the effect size, if presentat all, is likely to be small. This unintuitive find-ing may be explained by the size of the physio-logical insult associated with major cancer resec-


146



tion surgery. Taking oesophagectomy as an ex-ample; resection and reconstruction of the up-per gastrointestinal tract results in a complica-tion profile reflective of the operative field rath-er than fitness, where the cardiopulmonary sys-tem is directly affected. However, cardiorespira-tory and musculoskeletal reserves may be criticalto the ability of a patient to respond once a com-plication has occurred (41). This review high-

lights the need for large multi-center studies toassess the association between CPEX variablesand mortality in cancer resection surgery. Untilthis evidence is available, CPEX testing in isola-tion is unlikely to meaningfully inform cancersurgery practice.

The authors declare no other conflicts of competing interest for this work.

References

1.HemmingsHCJr,WlodyD,MahajanR,WebsterNR.The2014BJA/PGAspecial issue:aselectionofsixeduca-tionalreviews.BrJAnaesth2014;113Suppl2:ii1-2.2. Desborough JP. The stress response to trauma andsurgery. Br J Anaesth 2000;85(1):109-17.3. Wahba RW. Perioperative functional residual capac-ity. Can J Anaesth 1991;38(3):384-400.4. Harter J, Orlandi SP, Gonzalez MC. Nutritionaland functional factors as prognostic of surgical cancerpatients. Support Care Cancer 2017;25(8):2525-30.5. Jack S, West MA, Raw D, Marwood S, Ambler G,Cope TM, et al. The effect of neoadjuvant chemo-therapy on physical fitness and survival in patients un-dergoing oesophagogastric cancer surgery. Eur J SurgOncol 2014;40(10):1313-20.6. Smith TB, Stonell C, Purkayastha S, Paraskevas P.Cardiopulmonary exercise testing as a risk assessmentmethod in non cardio-pulmonary surgery: a systemat-ic review. Anaesthesia 2009;64(8):883-93.7. Older P, Smith R, Courtney P, Hone R. Preopera-tive evaluation of cardiac failure and ischemia in el-derly patients by cardiopulmonary exercise testing.Chest 1993;104(3):701-4.8. Chan AW, Altman DG. Identifying outcome report-ing bias in randomised trials on PubMed: review ofpublications and survey of authors. BMJ 2005; 330(7494):753.9. Hennis PJ, Meale PM, Grocott MP. Cardiopulmo-nary exercise testing for the evaluation of periopera-tive risk in non-cardiopulmonary surgery. PostgradMed J 2011;87(1030):550-7.10. Moran J, Wilson F, Guinan E, McCormick P,Hussey J, Moriarty J, et al. Role of cardiopulmonaryexercise testing as a risk-assessment method in pa-tients undergoing intra-abdominal surgery: a system-atic review. Brit J Anaesth 2016;116(2):177-91.11. CancerResearchUK.Cancer incidence for commoncancers. (Accessed June 1, 2017, at http://www.cancer-researchuk.org/health-professional/cancer-statistics/in-cidence/common-cancers-compared#collapseZero).12. Begum SS, Papagiannopoulos K, Falcoz PE, De-caluwe H, Salati M, Brunelli A. Outcome after video-assisted thoracoscopic surgery and open pulmonarylobectomy in patients with low VO2 max: a case-matched analysis from the ESTS databasedagger. EurJ Cardiothorac Surg 2016;49(4):1054-8.13. Loewen GM, Watson D, Kohman L, Herndon JE2nd, Shennib H, Kernstine K, et al. Preoperative exer-cise Vo2 measurement for lung resection candidates:results of Cancer and Leukemia Group B Protocol9238. J Thorac Oncol 2007;2(7):619-25.14. Licker M, Schnyder JM, Frey JG, Diaper J, Carti-er V, Inan C, et al. Impact of aerobic exercise capacityand procedure-related factors in lung cancer surgery.Eur Respir J 2011;37(5):1189-98.15. Brunelli A, Belardinelli R, Refai M, Salati M, Soc-ci L, Pompili C, et al. Peak oxygen consumption dur-ing cardiopulmonary exercise test improves risk strati-fication in candidates to major lung resection. Chest2009;135(5):1260-7.

16. Torchio R, Guglielmo M, Giardino R, ArdissoneF, Ciacco C, Gulotta C, et al. Exercise ventilatory in-efficiency and mortality in patients with chronic ob-structive pulmonary disease undergoing surgery fornon-small-cell lung cancer. Eur J Cardiothorac Surg2010;38(1):14-9.17. Brunelli A, Belardinelli R, Pompili C, Xiumé F, Re-fai M, Salati M, et al. Minute ventilation-to-carbon di-oxide output (VE/VCO2) slope is the strongest predic-tor of respiratory complications and death after pulmo-nary resection.AnnThorac Surg 2012;93(6):1802-6.18. West MA, Parry MG, Lythgoe D, Barben CP,Kemp GJ, Grocott MP, et al. Cardiopulmonary exer-cise testing for the prediction of morbidity risk afterrectal cancer surgery. Br J Surg 2014;101(9):1166-72.19. West MA, Lythgoe D, Barben CP, Noble L, KempGJ, Jack S, et al. Cardiopulmonary exercise variablesare associated with postoperative morbidity after ma-jor colonic surgery: a prospective blinded observa-tional study. Br J Anaesth 2014;112(4):665-71.20. West MA, Asher R, Browning M, Minto G, SwartM, Richardson K, et al. Validation of preoperativecardiopulmonary exercise testing-derived variables topredict in-hospital morbidity after major colorectalsurgery. Br J Surg 2016;103(6):744-52.21. Junejo MA, Mason JM, Sheen AJ, Moore J, Fos-ter P, Atkinson D, et al. Cardiopulmonary exercisetesting for preoperative risk assessment before hepat-ic resection. Br J Surg 2012;99(8):1097-104.22. Dunne DF, Jones RP, Lythgoe DT, Pilkington FJ,Palmer DH, Malik HZ, et al. Cardiopulmonary exer-cise testing before liver surgery. J Surg Oncol 2014;110(4):439-44.23. Kasivisvanathan R, Abbassi-Ghadi N, McLeodAD, Oliver A, Rao Baikady R, Jhanji S, et al. Cardio-pulmonary exercise testing for predicting postopera-tive morbidity in patients undergoing hepatic resec-tion surgery. HPB (Oxford) 2015;17(7):637-43.24. Forshaw MJ, Strauss DC, Davies AR, Wilson D,Lams B, Pearce A, et al. Is cardiopulmonary exercisetesting a useful test before esophagectomy? Ann Tho-rac Surg 2008;85(1):294-9.25. Moyes LH, McCaffer CJ, Carter RC, FullartonGM, Mackay CK, Forshaw MJ. Cardiopulmonary ex-ercise testing as a predictor of complications in oe-sophagogastric cancer surgery. Ann R Coll Surg Engl2013;95(2):125-30.26. Nagamatsu Y, Shima I, Yamana H, Fujita H, Shir-ouzu K, Ishitake T. Preoperative evaluation of cardio-pulmonary reserve with the use of expired gas analy-sis during exercise testing in patients with squamouscell carcinoma of the thoracic esophagus. J ThoracCardiovasc Surg 2001;121(6):1064-8.27. Prentis JM, Trenell MI, Vasdev N, French R,Dines G, Thorpe A, et al. Impaired cardiopulmonaryreserve in an elderly population is related to postop-erative morbidity and length of hospital stay afterradical cystectomy. BJU Int 2013;112(2):E13-9.28. Tolchard S, Angell J, Pyke M, Lewis S, Dodds N,Darweish A, et al. Cardiopulmonary reserve as deter-

mined by cardiopulmonary exercise testing correlateswith length of stay and predicts complications afterradical cystectomy. BJU Int 2015;115(4):554-61.29. Lamb BW, Tan WS, Eneje P, Bruce D, Jones A, Ah-mad I, et al. Benefits of robotic cystectomy with intra-corporeal diversion for patients with low cardiorespi-ratory fitness: A prospective cohort study. Urol Oncol2016;34(9):417.e17-23.30. Ausania F, Snowden CP, Prentis JM, Holmes LR,Jaques BC, White SA, et al. Effects of low cardiopul-monary reserve on pancreatic leak following pancre-aticoduodenectomy. Br J Surg 2012;99(9):1290-4.31. Chandrabalan VV, McMillan DC, Carter R, Kin-sella J, McKay CJ, Carter CR, et al. Pre-operative car-diopulmonary exercise testing predicts adverse post-operative events and non-progression to adjuvanttherapy after major pancreatic surgery. HPB (Oxford)2013;15(11):899-907.32. Junejo MA, Mason JM, Sheen AJ, Bryan A,Moore J, Foster P, et al. Cardiopulmonary exercisetesting for preoperative risk assessment before pan-creaticoduodenectomy for cancer. Ann Surg Oncol2014;21(6):1929-36.33. Grocott MP, Browne JP, Van der Meulen J, Matej-owsky C, Mutch M, Hamilton MA, et al. The Postop-erative Morbidity Survey was validated and used todescribe morbidity after major surgery. J Clin Epide-miol 2007;60(9):919-28.34. Dindo D, Demartines N, Clavien PA. Classifica-tion of surgical complications: a new proposal withevaluation in a cohort of 6336 patients and results ofa survey. Ann Surg 2004;240(2):205-13.35. The Cochrane Collaboration. Cochrane Hand-book for Systematic Reviews of Interventions (Ver-sion 5.1.0). (Accessed May 15, 2018, at http://hand-book-5-1.cochrane.org/).36. Ministry of Education, Culture, Sports, Scienceand Technology of Japan. Official statistics of the aver-age height and weight of Japanese male. (Accessed June1, 2017, at http://www.mext.go.jp/b_menu/toukei/001/022/2004/002.pdf).37. Bassi C, Dervenis C, Butturini G, Fingerhut A,Yeo C, Izbicki J, et al. Postoperative pancreatic fistu-la: an international study group (ISGPF) definition.Surgery 2005;138(1):8-13.38. Snowden CP, Prentis JM, Anderson HL, RobertsDR, Randles D, Renton M, et al. Submaximal cardio-pulmonary exercise testing predicts complicationsand hospital length of stay in patients undergoing ma-jor elective surgery. Ann Surg 2010;251(3):535-41.39. Altman DG, Royston P. The cost of dichotomis-ing continuous variables. BMJ 2006;332(7549):1080.40. Lim E, Beckles M, Warburton C, Baldwin D. Car-diopulmonary exercise testing for the selection of pa-tients undergoing surgery for lung cancer: friend orfoe? Thorax 2010;65(10):847-9.41. Hall A, Older P. Cardiopulmonary exercise test-ing accurately predicts risk of major surgery includ-ing esophageal resection: letter 1. Ann Thorac Surg2009;87(2):670-1.

147



Supplementary Appendix

This appendix has been provided by the authors to give readers additional information about their work.

Supplement to: Stephen Lam, Andrew Hart. Cardiopulmonary Exercise Testing for Predicting Early

Outcomes after Major Cancer Resection: A Systematic Review. J Anesth Perioper Med 2018;5:136-48. doi:

10.24015/JAPM.2018.0052

Search terms

Search terms for cardiopulmonary exercise testing:“cardiopulmonary exercise”, “CPEX”, “CPET”, “exercise testing”, “anaerobic threshold”,

“VCO2”, “ventilatory inefficiency”, “oxygen consumption”, “VO2”, “preoperative exercise”, “aero-bic exercise”.

Search terms for outcomes: “morbidity”, “mortality”, “outcome”, “complication”.

Additional search terms for lung cancer: “lung cancer surgery”, “lung resection”, “lobectomy”, “pneu-monectomy”.

Additional search terms for colorectal cancer: “colorectal”, “colon”, “colectomy”, “rectal”, “rectum”,“surgery”, “resection”.

Additional search terms for colorectal cancer: “hepatic”, “hepatectomy”, “liver surgery”.

Additional search terms for oesophageal cancer: “oesophagectomy”, “oesophagogastrectomy”.

Additional search terms for bladder cancer: “bladder cancer”, “cystectomy”.

Additional search terms for pancreatic cancer: “pancreatic surgery”, “pancreatic resection”, “pancreat-icoduodenectomy”.

An example for the whole search string for colorectal cancer: “(((cardiopulmonary exercise).ti,ab OR(cpex).ti,ab OR (cpet).ti,ab OR (exercise testing).ti,ab OR (anerobic threshold).ti,ab OR (VCO2).ti,abOR (ventilatory inefficiency).ti,ab OR (oxygen consumption).ti,ab OR (VO2).ti,ab OR (preoperativeexercise).ti,ab OR (aerobic exercise).ti,ab) AND (((colorectal).ti,ab OR (colon).ti,ab OR (colectomy).ti,ab OR (rectal).ti,ab OR (rectum).ti,ab) AND ((surgery).ti,ab OR (resection).ti,ab))) AND ((complica-tions).ti,ab OR (outcome).ti,ab OR (morbidity).ti,ab OR (mortality).ti,ab)”

ti=title, ab=abstract

Systematic Review

148

cardiopulmonaryexercisetestingforpredicting ... · found for breast or brain cancers or lymphoma....

Documents