Med. J. Malaysia Vol. 46 No. 1 March 1991

APACHE II : Preliminary report on 100 Intensive Care Unit cases in University Hospital, Kuala Lumpur

*G.S. Teoh, MBBS (Malaya) Medical Officer *K.K.Mah, MBBS (Malaya) Medical Officer *Sufian Abd. Majid, MD (UKM) Medical Officer *Streram, MBBS (My sore) Medical Officer M.K. Yee**, MBBS (Malaya), FFARCS (I) Consultant Anaesthesiologist

*Department of Anaesthesia. University Hospital. Kuala Lumpur.

** Fatimah Hospital. /poh. Perak Formerly Lecturer. University Hospital. Kuala LUmf'Jur

Summary A good overall assessment of the severity of illnesses of patients admitted to a general intensive care unit (lCU) is not without problems. The APACHE (acute physiology -and chronic health evaluation) prognostic scoring system enables us to stratify acutely ill patients and compare efficiency of lCU therapy in different hospitals. This preliminary study carried out on 100 consecutive admissions to the ICU in University Hospital, Kuala Lumpur showed the spectrum of ICU admissions and the direct relationship between APACHE 11 score and mortality.

Key words: APACHE 11, outcome prediction, prognostic scoring system, preliminary usage in UHKL.

Introduction

The APACHE (acute physiology and chronic health evaluation) prognostic scoring system was fust developed in the United States in the George Washington University Medical.Centre by William Knaus et all. Their initial goal was to develop a better methodology of measuring case mix among lCU patients. It was noted that death rates varied among ICUs and it was not known whether the cause of the outcome variations was due to differences in therapeutic efficacy or case-selection criteria of admission to various ICUs.

The APACHE approach to severity measurement uses the conceptual model that a patient's prior risk of -death from an acute illness was a function of his major disease, his physiological reserve (which was influenced by his age and the presence of chronic diseases), along with a major contribution from the acute severity of disease as determined by derangements in acute physiologic balance. The basis for

72

APACHE's development was the hypothesis that the severity of acute disease can be measured by quantifying the degree of abnonnality of multiple physiologic variables.

The original APACHE system provides weightings for 34 potential physiologic measures, the sum of which yields an acute physiologic score (APS). In the APACHE IF, the number of physiologic measurements was reduced to 12. There is increased weightage for acute renal failure3, and Glasgow coma score4 •

The APACHE IT scoring system was commenced in the lCU in University Hospital, Kuala Lumpur (UHKL) on August 1, 1990 and this preliminary report is based upon the first 100 consecutive admissions.

Patients and Method Thefrrstl00consecutiveadmissionstothelCUinUHKLfrom 1 August ,1990 to September 23, 1990 were studied ill dus initial report. Admissions were from the operating theatres directly or from the recov~ r()()8;~ferrals from all wards (Medical, Surgical, Paediatrics); and from the Accident and Emergency ward. Patients admitted solely to lCU for the purpose of pain relief (epidural opiates or thoracic epi<hJrals) were excluded from this studY'i.f$ents who underwent cardiac ope_ons were admitted poSt:operatively to the Cardiac Intensive Care Unit and hence were excluded.

The APS is determined from the worst physiologic value during the initial 24 hours fo1lowing ICU admission. 'ftlewetst value may be the highest or lowest value from nonnal. An 12 physiologic variables must be scored to obtain the APS.

Age points were assigned accordingly. Chronic health points were assigned depending on the mode of ICU admissions; five points for non-operative or emergency posl:-opemtive patients and two points for elective post-operative patients.

The APACHE IT score is then calculated as the sum of the APS, age point and chronic health point (refer Appendix A2).

The death rate in this study is the observed hospital death rate i.e. in the JCU or in the wards, following discharge from ICU.

Results A total of 100 lCU patients were studied with ages ranging from 1 month to 82 years. The age distribution is illustrated in Table 1. The largest proportion of the patients (20%) fell into the 61-70 year age group.

Table 2 shows the diagnostic category into which the admissions were classified and the corresponding mortality. There was 100% mortality in those patients admitted with sepsis, haemorrhagic shock and post cardiac arrest (for cerebral resuscitation) and 42% mortality in trauma/emergency neurological patients.

The patients were divided into non-operative and operative categories. There were 43 non-opemtive and 57 operative patients. The distribution of APACHE IT scores in these admissions is illustrated in

73

-...l

.f>.

Appe

Ildh

A. TH

E A

PA

CH

E 1

1 S

EV

ER

ITY

OF

DIS

EA

SE

CLA

SS

IFIC

AT

ION

SY

STE

M

PH

YS

IOLO

GIC

VA

RiA

BL

E

fllG

H A

BNO

RMA

L RA

NG

E lO

W A

BNO

RMA

L RA

NClE

~

+4

. ) •

I .,

0 •

I .2

.3

•

4

VEMPE"ATUR~ -

,"cl

ol f

'Cl

0 )9,9

0 g'

0 0

0 0

0 ~~

9'

HI'

38

,'3

89

' J6

'·3

84

' 34

'3

5.9

' J2

'Jle

' 30

' 31

9'

ME

AN

ART

ERIA

L P

RE

SS

UR

E

-m

m H

O ,Ro

I J

O<i

,9

IIO

l(29

10

l{0'

3 J-i

9 ~9

H

EA

RT

I'IA

n

0 0

0 0

0 0

0 tl

t.nlr

tcul

lll f

eilp

Ontl

.~

~'80

laO

·,19

11

0·13

9 10

10'3

55

69

'O·~

09

R

UPI

A ...

VOI1Y

AA

Ye -

0 0

0 0

0 0

0 ln

on-v

onm

.ttd

0)1 •

• ni

il.l&

d!

~50

35·'

9 2

53

' 11

-2'

10 1

1 6·

9 $5

O

XY

GEN

ATI

ON

: A,.O

D. o

r p.

C.

(mm

HQI

0

0 0

0 $,

FI

O, ~ D

_S ,.c

e.d

A-,q

o,

1-2.~_ f-~4\l9,_ ~~9_

<2

00

/----'oPo.~~ 1-

;:<.--

--1

I-b:

flo.<ii.i~oT.f;.;ry P

.c-;-

--

-t-

----

UP

o,>

70

orO

,1,I-

lo

OP

O,<

55

/lFH

ER

IAlo

H

,H

16

?69

7

5C

f.5

9

7.3

l{.e

SE

RU

M S

OD

IUM

,"'M

911l

1 ):\i

o U

16

0· 17

9 IS

!:{

59

150<

i,.

130C

Jl'9

~ERU!A POTAS~IUM ~I1

'Mo'

/lI

,Cf

.29

0 3

59

,.

0 5.

5·5.

9 3-

3.'

SEFl

UM cR~AnNIII~ '''

'9110

0 m

ll 0

0 0

0 n)ou~le

po'n

l se

~l.

1111 ~

~~I.

,on

oll.

llu,

.\

~ 3

.5 2·

3.4

1.5-

1.9

0.6·

1.4

H~MI

.TOC

Rn ,

'""

So

50-;

~9

46~.

9 ;l

Q1~g

VJ

!IlITE

ilL

OQO

CO

UN

T !

lcU

l/mm

J)

Ro N9

s.9

0 0

Iln 1

.1)00

51

15 1

9.9

J·tu

GL

ASGO

W C

QMA

SCQ

RE ,G

CS}

: Se

ol~

• 15

mln

uo ~

elu.'

(iCS

I!Il

olo'

AC

UTe

PH

VSI

QlO

GV

SCO

RE ,A

PSI:

Sum

011

"1 1

2 In

dlvl

duol

var

iabl

e po

lnl.

:lis'V

'" H,

!<O,

CVl

nou"

mM

oIIL

I 0

0 IH

o, pr

.,."o

d. u

se If

no

ASOs

i :t

52

nH

9

" I\G~

~Oll

'll~

: . '

A

ni;

" pO

ln's

\0 '9

' 0$

'011

0".:

/lOft

yr.,

S 4

4 4S

·54

55·6

4 05

·74

:n~

roln

l,

o 2 3 ~ ~

&l C

HR

ON

IC H

eA

LT

H PQI

HT~

. n

the

pDlla

l'11

hil

i GI

hl,~

or)'

ra'

nv

e,,

, o

l9sn

sys

tem

In·

,u

!"el

sney

'o,

I. Im

mun

o·eo

mp,

oml •

• d 8

ulgn

pol

nla

0\1.

loll

ow

s'

"I. fO

f nO

f'lot

'er4

l1liv

e Of

em

erge

ncy

po

&io

per

allv

e po

llen,

s -

5 po

lnl.

'" b.

lo

r el

eeH

ve p

osl

op

erlll

tlve

pa

lie

nh

-2 po

l~IS

DE

FIN

ITIO

NS

Org

mh:

In

uu

llj.c

le,,_

cy o

r Im

mu

no

·co

mp

rom

ise

d s

lale

ro

uS1

h,v

e b

hn

evi

dent

ptlo

r to

Ihl

s ho

!pil!

/:d I

Idm

is,·

»:Io

n G

nd c

on

form

io

Ih

e fo

llo

win

g c

rlle

tls

'

UV

E.R

: B

ioP

sY"P

,ove

n C

lr'h

ofH

S a

nd

do

cu

me

"le

d p

Orl

sl

hype

rllQ

"slo

~. ~plnodes 0

1 p

ast

up

per

01

ble

ed

ing

1.l1

· (l

lbu

led

10

fPo

fhll

hyp

erte

ns.

ion

. O

f ~HiOf ~pj'bdes

01

tlep

allc

f~l

Iu'e

't'i

nCep

hljl

oPll

lhyl

cOma

0 0

32-0

0.9

22·3

1.9

CA

AO

IOV

ASC

UU

IR:

Now

YOl

k HO

B,I

IIss

oel.l

lo"

Cl ...

IV

.

RE

.SP

IRA

TO

RY

: C

hro

nic

11U

1,I,c

""'e

, o

bs

tru

cll

ve

, O

f v

lIS

cu

h,

dis

us

e f

uu

ltln

Q I

n u

ve

re e

lle

rclu

Hls

lrlc

· li

on

, I,e

., u

na

ble

10

clim

b 5

l1d

l$ o

r p

e,f

orm

ho

us

eh

old

d

ull

es

: 0'

do

cu

me

nte

d c

tuo

nlc

hv

po

.ia

, Il

YO

erca

pn

ia.

seco

nd

Ary

po

lyc

yll

lem

la,

SS

Veh

! p

ulm

on

ary

hvp

e,te

n.

slo

n (

>.4

0mm

HO

I. o

r re

splr

al,,)

1' d

ep

en

de

nc

y.

RE

NA

L:

Rec

eivi

ng

ch

lon

h.

Oio

lvsl

s IM

MU

NO

·CO

MP

RO

MtS

EO

th

e p

.tle

nt

ha

, ,e

ce

lve

d

~hel;py Ih

l! s

up

p'u

ses

resi

ila

nce

to

Infe

cti

on

. e.

g ..

Imm

un

o·s

up

pre

ss

lon

. c"

emo

lher

IllP

Y.

rad

ittl

lon

lo

ng

te

rm o

r 't

ee

nl

1"<'0

" d

ose

ale

lold

s.

or h

as

a d

ises

sE

Ih~1

is

su

llle

ten

llv

3d

... l/l

nced

10 5upp'es~

fes.

istm

nce

10

Inle

cOo

n.

e Q

. le

uke

mili

l. 'y

",p

ho

m&

!. A

.IDS

The

APA

CH

E I

! se

veri

ty o

f dis

ease

cla

ssili

catio

n sy

stem

.

U

/.1Hz

. <

H5

1.

25·7

.32

120l

?Z9

U

U

111·

\19

~ 1

10

0 0

25·2

.9

< 2.

5

0 I

< 0

.6

,

20~9

Ho

0

0 '·2

.9

< 1

0 0

0 IB

·21.

9 _ ~179

<U

APA

CH

E

II SC

OR

E

Sum

01

!!J

!!II

~

l!l A

PS

pO

ln" _

__

__

_

lID "

'0< p

Oln

,. _

__

__

__

_

fC]

Ch

ron

ic H

ea

lth

po

ints

101

.. 11 A

PA

CH

e: 1

1 _

__

__

__

_ _

Table 1 : Age distribution of patients

Age (years) Number

0-10 7 11-20 12 21-30 13 31-40 11 41-50 11 51-60 16 61-70 20 71-80 8

> 81 2

Table 2 : Diagnostic category and deaths

Diagnosis Patients

Neurologic Trauma I Emergency 12 Elective I Post -op 16

Resp.insufficiency Non-operative 21 Operative 10

Sepsis 3 Gastrointestinal 11 Cardiovascular 14 Drug overdose 1 Haemorrhagic shock 1 Metabolic disorder 2 Post-cardiac arrest 7 Multiple trauma 2

Deaths

5 1

o 1 3 3 1 o 1 o 7 1

Figure 1. Patients were admitted at all levels of severity with APACHE II scores ranging from one to 50. (The worst possible APACHE II score is 71).Of the patients, 51 % had an APACHE n score of between 10 -19. The non-operative patients scored higher and they also showed a wider scatter in their score. These non-operative patients were emergency medical admissions of varying severity.

The APACHE IT scores were correlated with hospital mortality in Figure 2. It showed that the mortality was higher with a higher APACHE II score. There were no deaths in the 0-4 APACHE n score group. In the higher ranges of APACHE IT score of 30 onwards, there was a 100% mortality except for APACHE n score of 45-49 for which there were no admissions within this group.

Table 3 shows the number of non-operative and operative patients and deaths in each category of APACHE II score. As the number of patients studied was small, there were none in the non-operative group with scores of 30-34 and 45-49, or in the operative group with a score of 35-54.

75

Pe

rce

nta

ge

of

ad

mis

sio

ns

351~--------------------------------------------------------~

30

25

1-.

20 I-

......

. ..

...... ~.

15

~

10 5 o

0-4

5-9

10-1

4 15

-19

20-2

4 2

5-2

93

0-3

43

5-3

94

0-4

44

5-4

95

0-5

4

AP

AC

HE

11

Sco

re

No

n-o

pe

rati

ve

_

Op

era

tive

Fig

1. D

istr

ibu

tio

n o

f A

PA

CH

E U

sco

res

in 1

00

IC

U a

dm

issi

on

s

-...!

-...!

100

80

60

40

20 o

De

ath

ra

te (

%)

0-4

5

-9

10-1

4 15

-19 20~24 2

5-2

93

0-3

43

5-3

94

0-4

44

5-4

95

0-5

4

AP

AC

HE

11

Sco

re

Fig

2,

AP

AC

HE

11

sco

res a

nd

ho

sp

ita

l m

ort

ality

in

100

IC

U a

dm

issio

ns

-.J

00

De

ath

ra

te (

%f

10

0

80

60

40

20 o

t?

"/A

0-4

5-9

10

-14

15-1

9 2

0-2

42

5-2

93

0-3

43

5-3

94

0-4

44

5-4

95

0-5

4

AP

AC

HE

11

Sco

re

_ N

on

-op

era

tive

_

Op

era

tive

Fig

3.

AP

AC

HE

11

sco

res

an

d h

osp

ita

l m

ort

ality

in

100

IC

U a

dm

issi

on

s.

Table 3 : APACHE n scores and hospital mortality in 100 ItU admissions

Non-operative Operative APACHE n Score Patient Death Patient Death

0- 4 1 0 8 0 5- 9 6 0 14 2

10-14 12 1 18 0 15 -19 12 5 9 2 20-24 4 2 5 3 25-29 2 0 2 1 30-34 0 0 1 1 35-39 3 3 0 0 40-44 2 2 0 0 45-49 0 0 0 0 50-54 1 1 0 0

Total 43 14 57 9

A plot of the death rate against the APACHE II score in non-operative and operative patients is illustrated in Figure 3.

A study of 5815 ICU admissions from 13 hospitals in the United States showed a direct relationship between APACHE II scores and observed hospital death rates as shown in Figure 42•

Discussion The analysis of the spectrum of our ICU admissions indicates that patients with respiratory insufficiency formed the largest proportion of admissions followed by neurological patients. The former included patients with asthma, chronic obstructive pulmonary disease, pulmonary oedema, infection and respiratory insuffic~ency after surgery. These patients had a good overall prognosis.

The septic patients had higher mortalities as also demonstrated in the study of ICU admissions in the United States2•

Post cardiac arrest patients generally had a poor outcome in our study. Whether this poor outcome can be attributed to severity of illness such that survival is improbable, or to inappropriate or inadequate ICU treatment needs to be looked into further.

It was noted in our series that longstaying patients may have a low APACHE II score on admission but eventually succumbed. It is perhaps better to do APACHE scoring at frequent intervals so that a better prediction of outcome can be obtained.

The major disadvantages about the APACHE IT prognostic scoring system is that the points allocated forth,e physiologic variables are based on adult values. As such we found that this system wa,s unf9!,tunately not suitable for neonatal cases. Physiologic variables which are abnormal for adults may be normal for neonates.

79

00

o

De

ath

ra

te (

%)

1001'--------------------------------------------------~--~

80

I--

60 1

----

40

f-.

.........

.

20

o L-I ---=~-

0-4

5-9

1

0-1

4

15

-19

2

0-2

4

25

-29

3

0-3

4

35

+

AP

AC

HE

11

Sco

re

_ N

on

-op

era

tive

_

Op

era

tive

Fig

4.

AP

AC

HE

"

sco

res a

nd

ho

sp

ita

l m

ort

ality

In

58

15

IC

U

ad

mis

sio

ns

fro

m 1

3 h

osp

ita

ls i

n th

e U

.S.

Another prognostic scoring system known as the Therapeutic Intervention Scoring System (TISS)S has been devised by Cullen DJ et al. The main disadvantage of TISS is that it cannot be used to compare results between different hospitals or even among different units in the same hospital. This is because the amount of therapeutic intervention required for each patient is very much dependant on the ability of the physician involved or the policy of the unit. The APACHE IT prognostic scoring system overcomes this problem by pre determining the physiologic variables that need to be scored. Thus standardisation of results can be achieved between different hospitals or even within the same unit in each hospital.

We have shown that the APACHE IT scoring system can be used in our ICU in Malaysia. We did not encounter much difficulty in implementing the system as most of the physiological variables were already monitored or could be made available to most ICUs.

It would be interesting to evaluate results among ICUs in Malaysia now that there is a fairly accurate and reliable system of comparing the different categories and severity of ICU patients and their final outcome. Self-assessment and proper audit in the ICU in UHKL would also be a possibility with the implementation of such a system.

Acknowledgement The authors would like to thank Professor A. E. Delilkan for his encouragement, the Anesthesia Department, UHKL and the nursing staff of the ICU, UHKL for their support and assistance in implementing the APACHE IT scoring in UHKL and the Medical Research Resource Centre, UHKL for helping out with the graphics.

References

1. Knaus WA, Zimmennan lE, Wagner DP, Draper EA, Lawrence DE. APACHE - acute physiology and chronic health evaluation: A physiologically based classification system. Critical Care Medicine 1981; 9: 591-597.

2. Knaus W A, Draper EA, Wagner DP,Zimmounman lE. APACHE ll. A severity of disease classification system. Critical Care Medicine 1985; 13: 818-829.

3. Sweet SI, Glenney CV, Fitzgibbons JP, Friedmann P, Teres D. Synergistic effect of acute renal failure and respiratory failure in the surgical intensive care unit. American Ioumal of Surgery 1981; 141 : 492-496.

81

4. Teres 0, Brown RB, Lemeshow S. Mortality of intensive care patients: The importance of coma. Critical Care Medicine 1982; 10 : 86 - 95.

5. Cullen 01, Civetta IM, Briggs BA. Therapeutic intervention scoring system: a method for quantita-tive comparison of patient care. Critical Care Medi-cine 1974; 2 : 57 - 63.