JOURNAL OF CLINICAL MICROBIOLOGY, June 1978,p.539-545 0095-1137/78/0007-0539$02.00/0
Copyright i 1978 American Society for Microbiology
Vol.7, No. 6 Printed in U.S.A.
Evaluation of the API 20E System for Identification of
Nonfermentative Gram-Negative Bacteria
MEHDI SHAYEGANI,* PEGGY S. MAUPIN,ANDDORIS M. McGLYNN
Divisionof Laboratoriesand Research,New York StateDepartmentofHealth, Albany, New York12201
Received for publication 13 December 1977
The API 20E system forEnterobacteriaceae, recently broadened toinclude identification of nonfermentative gram-negative bacteria, was evaluated and
compared with the conventional method for complete identification of221
non-fermenters, which werewelldistributed into48speciesorbiotypes and included
organismsnotlisted in the API20E data base.The results of16testscommonto bothsystemswerein closeagreement. TheAPI 20Esystemcorrectly identified
71 (43%) of the 165 organisms included in the API 20E data base. However,
almost 90% of Acinetobacter calcoaceticus, three species of Pseudomonas, and Bordetella bronchisepticawerecorrectly identifiedtospecies.
The API 20E system is a plastic strip with
microtubes containing dehydrated substrates, originally designedfor theidentification of En-terobacteriaceae. Later, API introduced the
Profile Recognition Systemfor numerical
iden-tification and then, using a computer-assisted
program,developedtheAnalytical Profile Index, supplemented by the API Computer System Service. In1976, theAnalyticalProfile Index for API 20E wasexpandedtoinclude other
fermen-tative andnonfermentative gram-negative
bac-teria. Five separate tests, not included in the strip, were then added to complete the system
for identificationofthenonfermenters.
Various investigators (1, 5, 10, 12) evaluated
the API 20E system for the identification of
Enterobacteriaceaeandreportedahighlevelof agreement with conventional methods in both
biochemical reactions and identifications. The
Analytical Profile Index (or Register) has also beenmathematically evaluatedasexcellent (7).
Two evaluations have been made of the com-plete API 20E system,including the five sepa-ratetests-fermentation ofglucose (OFF),
oxi-dationofglucose (OFO), motility, oxidase, and
MacConkey-for identificationof
nonfermenta-tive
gram-negative
bacteria(3, 6). Bothreports found the system useful for identification ofclinical isolates of Pseudomonas and Acineto-bacter. We have now evaluated the system, whichincludes the API 20E and thefive sepa-rate tests, for the complete identification (to
speciesorbiotypes)of 221isolates of nonfermen-tative gram-negative bacteria asrequired by a referencelaboratory.
MATERIALS AND METHODS
Bacteria.Of the 221 isolates used inthisstudy,166 werefrom aculturecollectionmaintainedby the
Di-visionof LaboratoriesandResearch, NewYorkState Department of Health. All isolates originated from clinical specimens that had been submitted to our laboratory for identification or confirmation. Forty-sevenof thecultureswerereceivedthrough the
cour-tesyofAnalytab Products Inc.,Plainview, New York. Eightwerekindly provided by G. L.Gilardi,Hospital for Joint Diseases and Medical Center, New York City. Inourlaboratoryallcultureswereeither lyoph-ilizedormaintainedonbloodagarslants.
Theorganisms usedintheevaluationwere nonfer-mentativegram-negativerods,welldistributedamong 48species and including56 isolatesnotlisted in the API 20E data base.Nofermentative organismswere used.
Conventional media andprocedures. The
me-diawereprepared by the Division's media sectionas describedpreviously (9). The inoculated mediawere incubated forupto5daysat35to37°C.Theorganisms wereidentified byusinggenerallyaccepted criteria (2, 4, 11,13).
API 20E system for nonfermenters. The API 20E strip (this strip is the same one used for the identification ofEnterobacteriaceae) contains20 mi-crotubeswithsubstrates for thefollowing23tests: 0-nitrophenyl-fi-D-galactosidase (ONPG); arginine di-hydrolase;lysineandornithinedecarboxylase; citrate utilization; hydrogen sulfide; urease; tryptophan
de-aminase; indole; Voges-Proskauer (acetoin); gelatin
liquefaction; fermentation of the carbohydrates glu-cose, mannitol, inositol, sorbitol, rhamnose, sucrose, melibiose, amygdalin, and arabinose; nitrate reduction and nitrogen gas production, tested in the glucose microtube; and catalaseproduction,in any other car-bohydratemicrotube. The catalasetest wasnotused inthis study. Acomplete description ofthe strip is giveninotherreports(1,5, 10,12).
Additional media arerequiredfor the five separate
tests not on thestrip.Mediaforthree of these tests
are available from the manufacturer in snap-open ampules:API Mfor themotilitytestand APIOFfor boththe glucose oxidation and glucosefermentation
tests.Also available is the API oxidase testkit
con-539
on February 7, 2020 by guest
http://jcm.asm.org/
540 SHAYEGANI, MAUPIN, AND McGLYNN taining oxidasereagent(a 1% solution of tetramethyl-p-phenylenediamine dihydrochloride) and plastic chambers with filterpaper.Conventional MacConkey agarwasused for the fifth separatetest.Inthisstudy, methodology for theuseof the systemand interpre-tationof resultswereaccordingtothe manufacturer's recommendations.
Twenty-two of the initial tests werechecked and
recorded after 24hofincubation andagainat48h.At thispoint,reagentswereadded for thefiveremaining tests(tryptophan deaminase, indole, Voges-Proskauer, nitrate reduction, and nitrogengas production), and
those results were recorded. The Analytical Profile
Index and (if necessary) the computer service were
consulted for identification of the isolates. When in-dicated,supplementaltestsrecommendedbythe
man-ufacturer for identification weredone using
conven-tional media. When sufficient reactionswereclearat 24h,identificationwasdeterminedonthosereadings.
RESULTS
Biochemical reactions of the 221 isolates in the API 20E system (excluding fermentative carbohydratetests notusedintheconventional method) werein close agreement with
conven-tional method results (Table 1). The lowest agreementwasinthe citrateutilizationand
mo-tilitytests.
The level of identification expected tobe
at-tainedby theAPI system forall221isolateswas
compared with the actual identification achieved (Table 2). Ofthe organisms included in the API data base (165 isolates), 43% were
correctly identifiedtothelevelexpected bythe
manufacturer, and only 29% were completely
identified to species and biotype. The system
misidentified 22.4%of theorganisms includedin
theAPI20Echartsand32.2%of thosenotlisted.
Unidentified by the system were 24.8% of the
organismsincluded and 26.8% of thosenotlisted. Ananalysis of the reactions and identification obtained with the APIsystemforeach of the48
conventionally identified species used in this evaluation is giveninTable 3. Of the 221
orga-nisms tested, 119 (53.8%) required fromone to
TABLE 1. Comparisonof16biochemicalreactions of221isolates, using the API 20E andconventional
(C)tests'
No.in agree- No. differing ment
Tests
API+, API-, API+, API-,
C+ C- C- C+
ONPG 6 195 14 6
Argininedihydrolase 16 195 2 8
Lysinedecarboxylase 3 208 0 10
Ornithine decarboxyl- 2 213 0 6
ase
Citrateutilization 74 66 23 58
Hydrogen sulfide 2 219 0 0
Urease 17 168 0 36
Tryptophan deami- 0 195 1 25
naseb
Indole 3 210 0 8
Gelatinliquefaction 22 163 30 6
Oxidase 182 28 8 3
Nitrate to nitrate 44 148 13 16
(NO3toNO2)
Nitrate to gas(NO3to 31 162 9 19
gas)
Motility 93 68 25 35
Glucose, oxidative 38 150 1 32
(OFO)
Glucose, fermentative 0 217 0 4 (OFF)
+,Positivereaction;-,negativereaction.
Phenylalaninedeaminase used inconventionalmethod.
TABLE 2. Extentof identification of221isolatesby the API system
API identification [no. (%)] Complete
identification
API'sexpected level ofidentifi- No.of isolates (asrequired
cation Toexpected Correct only Misidentified Unidentified by our
refer-level togenus ence lab)[no.
(%)]
Genus' 64 23(35.9) _b 12(18.8) 29(45.3)
Species' 87 47(54.0) 8(9.2) 22(25.3) 10(11.5) 47(54.0)
Biotype' 14 1(7.1) 8(57.2) 3(21.4) 2(14.3) 1(7.1)
Total 165 71 (43.0) 16(9.7) 37(22.4) 41 (24.8) 48(29.1)
No expected identification 56 23(41.1) 18(32.2) 15(26.8) 0
(not included in API sys-tem)
Total 221 71 (32.1) 39(17.6) 55(24.9) 56(25.3) 48(21.7)
a Includesorganismsdesignated"genus-like"whenidentifiedasthatgenus, e.g.,Alcaligenes-like group IVe identifiedasAlcaligenessp.
b For the64isolates in thiscategory, "togenusonly"was theexpectedlevel of identification.
CUnnamedorganismsincluded inspeciesandbiotype wheneverapplicable.
J. CLIN. MICROBIOL.
on February 7, 2020 by guest
http://jcm.asm.org/
In cis boc
;vX.xDDo
be|
4a ,= I.0 fe 9
.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~g
e3
':
t2 2
= gco
}
it3 338 ;
3
s
Et t tiT= n nn =
4-|
32_,t-_____
___
ci .0 +
C' ;,-
oC?N>ea,
3Q| = -~~~~t0 _ N - t
-t~~~~~~~~~~~~tC|a4_N6.4.1 4- Go et4- 4
ko ED O O Oee
cQFit ,3 } j &i
3
<
3
31
3
ee
]~~~~~~~~~~~~~~~~~C
541
on February 7, 2020 by guest
http://jcm.asm.org/
cc
cj
0
0
az
SzZ
zZZ X0 + 0+ ; +
E--' - '
00-0
+ + w
C a)Z
I+ Cos+ +
3 000000
0 22 ~
+ 0 c
2021+0
H H..E
OZOO0 202z
O-I 0
X O CD
+ +
+^
i o++
CO C O0e 0 u
:S >~ZZz
"0 "0 U) U) 1-r r r -lbs>"A;A|xs X X.%} X r
a 4> U)U)U)"0"Q.0UM "0 "0 o "0 "0
#0 #o 0 0 a)a)
s
a)t)).a0
a))as=ss<s
a) a)a a)a
a
0 0t: a
n s s
a)a)aEt
a) Ea)a)0a) a) aa)a
"0 D"0 g toQ"0QRgR>"0" to "0Q"0"0m"0Q "0 "00 0_ 2 0 0C00 0222 N 2 .o0 -4 _ __ _
- - Coo -4 0 O
- LO Z
cmc1
-cqIl0 -0
1-a)a) a)a) a)a4) a) a
00 00 0 0 0 0
0 30 5-5-8 *5-5 3 030 3 3.5.5 .5.5
00C
ZZc,e
ZZ000c
OZZ Z Z542
2
0
U)
Q
Ca a)
0
0
'-.;
.e)
Z0a)
"0
Q
r.
C)
0
0)
fl)
0
CQ
0-4
0 0
Q Q
6 t
a)
a2)
10 0L
>6 cn
0
O Qo
"0
0;a)0 la
a2) 10
0
0
r
z
a4
"00
It
0
t0 a.) a)
4)
*-k
on February 7, 2020 by guest
http://jcm.asm.org/
+I+
f.~L ++0C
10 lala Da 'a U
* 2i .I E*EO
co CZ x w 0 0 V
CdQ CCC.° Q
<
«
- : <<<: U
C>
x
0
+ cc
co
O C
omiz
-++ +
4) 4) (1
cr
E E
r4) (P C) S.; C- <:
ci: 4
0
z
+ 6T rrzE.r,, tD
o oo o
E
0
C?
L0C000~~ ~ _ 0 0 0 x
> ~C_.~ - e - c N
- _ 0 O0> - 0
ci orN
cl cl
co --CN CN
CC -Z - C-~
-_ - -_ _
-4 _ C4 _4 _4Lo
- Cmm CN
cli
C9
_ _
_- CD
0s 0
V. 0 004 )0 0) 0)
0A
cco 0) 0) (1 0) 0)r. 4) 0)4
Z Z Q ci: (L cii (.4 c.c
V; C C
cc 0 0)
0. cc 0
co cccc
0. 0.
D
0 0
U
0 0 t4444
~~Cd cc~~54
_-4
cc . o
C.).0 C. _ 0
0 0.
4 44 0
0
zo
0
++f+ +
0H F0
Wo o o
0S2 >
O &
.4
Q
-o
co=
0
.O
0
z
*0
-w %.4
04
H
0 co 0
_'
on February 7, 2020 by guest
http://jcm.asm.org/
++
- .1E" , COX OcOO
rz O)Z2Z
~ r~c. 0
0q C,)
z -z
cN c.s csi
-c
-0 0 0 . 0
z z m z
0 0
544 C9cq
o o
04)
10
0o~00
o
Z X xT 0 0 0
++
X X
0 0
0
o -.
z
0
.
.. -.
0 .
o
0 0
o
0 .
0
0.
C.) 0>.
.
*0 00
0 0
. bC0
o
.
z .0.
o .
0
0>)
o :..0
0 . b. 00 0
.
oo
5
.0 0.o)
00
.
o .o
.
& .
0
0o
z :00 co
ci
C -.) 0
0 0
0
10
co
6 Y
Co
-o
5._
5-o
Zo
.0
00E
00t8 0.00
*0 Ct
2 0 o 0
0
6.
0P4
.10C.)
Ct
C.
._
0
la
-o
0. 10
0. OD 0
O >
Uf
on February 7, 2020 by guest
http://jcm.asm.org/
EVALUATION OF THE API 20E FOR NONFERMENTERS 545
fivesupplementaltests.Eighty-one isolates
gen-eratedprofile numbersnotfound in the Analyt-ical Profile Index and hadtobereferredtoAPI's computerservice, yet 63 ofthese were in cate-gories included in theAPI 20E system. Ofthe
63,only26 werecorrectly identifiedto theAPI
system's expectedlevelofidentification (15
re-quiring supplemental
tests),
while23 weremisi-dentified (12requiring supplemental tests)and 14 remained unidentified (13 requiring
supple-mental
tests).
DISCUSSION
Weused 221isolates whichbelongto 48
spe-ciesandbiotypesofnonfermentative
gram-neg-ative rods, 15 of which are not included in the
API20Esystemcharts. The results of the initial
API 20E biochemical tests (excludingthe
car-bohydrate fermentation tests) showed close
agreement with the conventional method
re-sults. This compares with the close agreement
forthesame commontests asreportedbySmith etal. (10) inastudyusing only Enterobacteri-aceae.
Despite this close agreement in biochemical reactions, only110(50%) of the221isolateswere
correctly identifiedtogenus,andonly48 (22%) to species andbiotype. However, the API 20E systemdoesnotclaimto
completely
identify all nonfermenter isolates used in this study to ge-nus, species, and biotype. When the expected level of identificationwasbased on theAnalyti-calProfile Index and reaction chart, 43%ofthe
isolates included in the system were correctly
identified.
All isolates ofPseudomonas aeruginosa, P. cepacia, and Bordetella bronchiseptica and
most of the two
subspecies
ofAcinetobacter
calcoaceticus and P. maltophilia isolates werecorrectly
identified to species bythe API 20E system. Three of theseorganisms
(P. aerugi-nosa and Acinetobacter species) are among those more commonly found in clinical speci-mens.Thesefindingssuggestthat the API 20E may
be usefulfor theidentificationofthe nonfermen-tative gram-negative bacteria more commonly
encountered in the clinical laboratory, but, as
with the Oxi/Ferm tube system (8) the API
systemisnotsuitable foruseby reference
labo-ratories.
LITERATURE CITED
1. Brooks, K. A., M.Jeno, and T. M. Sodeman. 1974. A clinical evaluation of the API microtube system for identification of Enterobacteriaceae. Am. J. Med. Technol. 40:55-61.
2. Buchanan, R. E., and N. E. Gibbons(ed.).1974. Ber-gey's manual ofdeterminativebacteriology,8th ed. The Williams and WilkinsCo.,Baltimore, Md.
3. Dowda,H. 1977. Evaluation oftwo rapidmethods for identification ofcommonlyencounterednonfermenting
oroxidase-positivegram-negative rods. J. Clin. Micro-biol. 6:605-609.
4. Hugh, R.,and G. L Gilardi. 1974. Pseudomonas, p. 250-269. In E. H.Lennette,E. H.Spaulding,and J. P. Truant (ed.), Manual of clinicalmicrobiology,2nd ed. AmericanSocietyforMicrobiology,Washington,D.C. 5. Nord, C. E., A. A.Lindberg,and A. Dahlback.1974.
Evaluation of fivetestkitsAPI,AuxoTab,Enterotube,
Patho Tech andR/Bforidentification of Enterobac-teriaceae. Med.Microbiol. Immunol. 159:211-220. 6.Nord, C. E., B. Wretlind, and A. Dahlback. 1977.
Evaluation of two test kids-API and Oxi Ferm Tube-for identification of oxidative-fermentative gram-negative rods. Med. Microbiol. Immunol. 163:93-97.
7. Robertson,E.A.,and J. D.MacLowry.1974. Mathe-maticalanalysis of the API enteric20profile register
usinga computerdiagnostic model. Appl.Microbiol.
28:691-695.
8. Shayegani, M.,A. M.Lee,and D. M.McGlynn.1978.
Evaluation of theOxi/Fermtube system for identifica-tionof nonfermentativegram-negativebacilli. J. Clin. Microbiol.7:533-538.
9. Shayegani, M., A. M.Lee,andL. M.Parsons.1977.A
scheme for identification ofnonfermentative gram-neg-ative bacteria. Health Lab.Sci. 14:83-94.
10. Smith,P.B., K. M.Tomfohrde,D.L.Rhoden,and A. Balows. 1972.API system:amultitube micromethod
for identification ofEnterobacteriaceae.Appl. Micro-biol.24:449-452.
11. Tatum, H.W., W. H.Ewing,and R. E.Weaver.1974.
Miscellaneousgram-negativebacteria,p.270-294. In E. H.Lennette, E. H.Spaulding,andJ. P. Truant (ed.),
Manual of clinicalmicrobiology,2nded. American
So-cietyforMicrobiology,Washington,D.C.
12. Washington,J.A.,P.K. W.Yu,and W. J. Martin.
1971.Evaluation of accuracy of multitest method
sys-temfor identification of Enterobacteriaceae.Appl. Mi-crobiol.22:267-269.
13. Wilson, G.S.,and A. Miles.1975.Topley and Wilson's principles ofbacteriology, virology,andimmunity, 6th ed. The Williams andWilkinsCo.,Baltimore,Md.
VOL. 7, 1978
