Am. J. Trop. Med. Hyg., 61(4), 1999, pp.

654657
Copyright q 1999 by The American Society of Tropical Medicine and Hygiene

RAPID SEROLOGIC DIAGNOSIS OF PEDIATRIC TYPHOID FEVER IN AN ENDEMIC

AREA: A PROSPECTIVE COMPARATIVE EVALUATION OF TWO DOT-ENZYME
IMMUNOASSAYS AND THE WIDAL TEST
ZULFIQAR AHMED BHUTTA AND NASEEM MANSURALI
Department of Paediatrics, The Aga Khan University Medical Center, Karachi, Pakistan

Abstract. We evaluated the diagnostic sensitivity and specificity of two dot-enzyme-linked immunoassays (Typhidott and Typhidot-Mt; Malaysian Biodiagnostic Research SDN BHD, Kuala Lumpur, Malaysia), assessing IgG
and IgM antibodies against the outer membrane protein (OMP) of Salmonella typhi, and the Widal test in comparison
with blood culture in a consecutive group of children with suspected typhoid fever. Of 97 children with suspected
typhoid fever, the disease was confirmed bacteriologically in 46 (47%), whereas 25 (26%) were considered to have
typhoid fever on clinical grounds. An alternative diagnosis was made in 26 (27%). The Typhidott and Typhidot-Mt
were superior to the Widal test in their diagnostic sensitivity and specificity, although values (sensitivity 5 8594%
and specificity 5 7789%) were significantly lower than in other regional reports. The lower specificity of the
Typhidott in our series may represent regional differences in the genomic structure and plasticity of the OMP of S.
typhi and merits further evaluation of these tests in diverse geographic locations.
response.13 A recent, commercially available, enzyme-linked
immunoassay (Typhidot-Mt; Malaysian Biodiagnostic Research SDN BHD) is reported to circumvent these blocking
antibodies by inactivating IgG antibodies, followed by an
immunoassay targeting specific IgM.14 Preliminary data using the Typhidott and Typhidot-Mt in combination have
shown sensitivity and specificity of 95% and 86%, respectively.15 Although the tests have shown promising results in
trials from Southeast Asia, given the genetic diversity and
plasticity of S. typhi strains,16,17 it is unknown if the test
would be of comparable sensitivity in other regions.
We prospectively evaluated the efficacy of the two dotEIA tests (Typhidott and Typhidot-Mt) in comparison with
the Widal test in a consecutive group of children with suspected typhoid fever in Karachi, Pakistan.

Typhoid fever is widely recognized as a major public

health problem in developing countries. It is estimated that
there are more than 13 million cases occurring annually in
Asia alone,1 of which a large proportion occur during childhood. In the wake of emerging multidrug-resistant strains of
bacteria causing typhoid fever, the disorder is known to be
associated with significant morbidity and mortality.2,3 It is
also recognized that a delay in diagnosis and institution of
appropriate therapy may significantly increase the risk of
adverse outcome and mortality.3 Although the isolation of
Salmonella typhi on blood culture remains the gold standard
for diagnosing typhoid fever, this may be problematic in endemic areas where adequate microbiologic facilities are limited. The widespread availability and use of antibiotics in
the community makes it frequently difficult to isolate the
organism on blood cultures and alternative methods such as
bone marrow cultures may be required.4 However, the latter
are invasive and difficult to obtain routinely in pediatric patients.
Despite improved methods of bacteriologic isolation, there
is a real need for rapid serologic diagnostic tests for typhoid
fever. The Widal test has been used for almost 100 years
old, is widely available in developing countries, and is still
regarded as a useful test in endemic areas.5 There is, however, considerable interest in newer methods of diagnosis of
typhoid fever such as latex agglutination,6 coagglutination,7
and the polymerase chain reaction.8 The dot-enzyme immunoassay (EIA) is a relatively newer serologic test based
upon the presence of specific IgG and IgM antibodies to a
specific 50-kD outer membrane protein (OMP) antigen on
S. typhi strains9 and has been commercially marketed as a
dot-EIA (Typhidott; Malaysian Biodiagnostic Research
SDN BHD, Kuala Lumpur, Malaysia). The test incorporates
nitrocellulose strips impregnated with the OMP antigen and
separately identifies IgM and IgG antibodies. Although the
test has shown promising results in preliminary studies from
Malaysia10,11 and The Philippines,12 the interpretation of the
IgG response in highly endemic areas remains problematic.
There is concern that in such endemic populations pre-existing IgG antibodies to S. typhi may increase rapidly following reinfection and potentially mask a concomitant IgM

MATERIALS AND METHODS

The study was performed at the Aga Khan University

Medical Center, a 450-bed teaching hospital in Karachi. All
children presenting to the ambulatory care services with suspected typhoid fever were evaluated according to a set protocol. Detailed clinical evaluation and profile were obtained
and a typhoid fever morbidity score was calculated as described previously.18 In all cases a complete blood count and
malarial blood film were obtained and liver function tests
were conducted. Five milliliters of blood were inoculated
into 2 blood culture bottles containing brain heart infusion
broth and thioglycollate broth, respectively. In cases who
had received antibiotic therapy for $ 72 hr, a bone marrow
culture was also obtained as described previously.18 The cultures were examined thereafter for bacterial growth at different stages, subcultured, and positive colonies were biochemically identified with analytical profile index 20E strips
(Analytab Products, Plainview, NY). The Widal test was performed by preparing serial tube dilutions of patient sera in
0.85% saline and mixing standard preparations of Salmonella O and H antigens (Wellcome Diagnostic, Dartford,
United Kingdom), followed by incubation at 508C for 24 hr.
H and O agglutination were characterized by large flocculent
or smaller granular aggregates, respectively. Results were

654

655

RAPID SEROLOGIC DIAGNOSIS OF TYPHOID FEVER IN AN ENDEMIC AREA

TABLE 1
Comparative characteristics of the study patients*
Culture-proven typhoid fever

Number
Age (years)
Weight (kg)
Gender (M:F)
Duration of illness (days)
Typhoid morbidity score
Alanine aminotransferase (IU/L)
C-reactive protein (mg/L)
Hemoglobin (g/L)
White blood cell count (109/L)

46
6.3 6 3.7
17.5 6 7.4
25:21
12.9 6 13.9
4.0 6 2.3
55.9 6 77.8
72.3 6 62.0
109.1 6 16.1
8.9 6 4.1

Clinical typhoid fever

Non-typhoid fever

25
7.9 6 4.3
21.9 6 11.6
13:12
12.7 6 6.4
3.7 6 1.7
61.6 6 142.1
57.5 6 48.8
108.2 6 19.0
11.6 6 12.3

26
5.3 6 4.1
18.1 6 11.1
12:14
9.1 6 10.4
3.0 6 1.9
29.6 6 31.4
65.0 6 63.6
102.3 6 25.2
15.8 6 15.1

* Unless otherwise stated, values are the mean 6 SD.

P , 0.05 compared with both culture-proven and clinical typhoid fever.

expressed as the inverse of the highest dilution expressing

agglutination.
The dot-EIA tests were performed using standard commercial kits (Typhidott and Typhidot-Mt (generous donations from Malaysian Biodiagnostics Research, Kuala Lumpur, Malaysia) containing nitrocellulose strips dotted with
0.3 mg of the 50-kD OMP. The Typhidottstrips were probed
with 1:100 dilution serum, washed with phosphate-buffered
saline, and 1 hr later the antigen-antibody complexes were
visualized by addition of horseradish peroxidaseconjugated
antiserum to human IgG and IgM and 4-chloro-1-naphthol.
Positive results were read in comparison with control sera
and represented a titer . 1:100. A positive IgG or IgM result, either alone or in combination, was regarded as a positive test result in our patients. The Typhidot-Mt differed in
the initial steps: the addition of an IgG-inactivation complex
and subsequent removal of the bound-IgG with anti-human
IgG labeled with horseradish peroxidase. The subsequent antibody binding and color change identified primarily IgM
antibodies against the OMP.
The protocol of the study was approved by the Human
Subjects Protection CommitteeAga Khan University Medical Center, and informed consent was obtained from parents/guardians of all children prior to sampling. In all cases
the sensitivity and specificity of the tests were calculated
along with corresponding positive and negative predictive
values. These parameters were calculated for the entire
group of typhoid fever patients (including both culture-proven and clinical typhoid cases) as well as the non-typhoidal
controls, as well as the subcohort of culture-proven typhoid
fever cases alone. Data were comparatively evaluated using
analysis of variance and the chi-square test, as appropriate.

RESULTS

A total of 97 children presented to the ambulatory services over a 6-month period with suspected typhoid fever. The
diagnosis was confirmed in 46 (47%) of 97 by the isolation
of S. typhi on blood and/or bone marrow cultures. In 25
cases (26%), although successive cultures were negative, a
clinical diagnosis of typhoid fever was made and these patients were treated as such with either first or second-line
antibiotics (oral amoxicillin or intravenous ceftriaxone). In
26 cases (27%), an alternative diagnosis was made after 48
72 hr, and treatment with the antibiotics was stopped. These
latter cases constituted negative controls and consisted of 8
cases with viral respiratory infection, 4 with pneumonia, 4
with urinary tract infections, 3 with malaria, 3 with bacterial
diarrhea, 3 with rheumatoid arthritis, and 1 with viral meningitis. All cases were followed up for 812 weeks after
presentation.
Table 1 compares the admission characteristics of the 3
groups of patients. Children with typhoid fever tended to be
older, and had significantly longer duration of illness at presentation (12.8 6 11.0 versus 9.1 6 4.0 days [mean 6 SD];
P , 0.05). They also had higher levels of alanine aminotransferase at admission (57.5 6 96.4 versus 29.6 6 31.4
IU/L; P , 0.05) and were significantly more ill with higher
values for the typhoid fever morbidity score (3.9 6 2.2 versus 3.0 6 1.9; P , 0.05).
Table 2 details the sensitivity, specificity, and positive and
negative predictive values for the Widal, Typhidott, Typhidot-Mt, and blood/bone marrow cultures for this cohort. The
combination of cultures and Typhidot-Mt offered the highest diagnostic sensitivity and specificity for diagnosing ty-

TABLE 2
Comparative evaluation of diagnostic tests for the entire cohort
No. positive among
clinical and cultureproven typhoid fever cases
(n 5 71)

Blood and/or bone

Widal test
Typhidott
Typhidot-Mt
Combined cultures
Combined cultures
Combined cultures

marrow cultures

and Widal test

and Typhidott
and Typhidot-Mt

46
39
50
52
56
57
59

(64%)
(54%)
(70%)
(73%)
(78%)
(80%)
(83%)

No. positive among

non-typhoid fever
cases
(n 5 26)

21
20
23
5
6
3

0
(80%)
(76%)
(88%)
(19%)
(23%)
(11%)

Sensitivity

Specificity

Positive
predictive
value

Negative
predictive
value

65%
55%
70%
73%
79%
80%
83%

100%
81%
77%
89%
81%
77%
89%

100%
89%
89%
95%
92%
91%
95%

51%
40%
49%
55%
58%
59%
66%

656

BHUTTA AND MANSURALI

TABLE 3
Comparative evaluation of diagnostic tests for culture-proven typhoid fever cases
No. positive among
culture-proven
typhoid fever cases
(n 5 46)

No. positive among

non-typhoid fever
cases
(n 5 26)

Sensitivity

Specificity

Positive
predictive
value

Negative
predictive
value

29 (63%)
43 (93%)
39 (84%)

5 (19%)
6 (23%)
3 (11%)

63%
94%
85%

81%
77%
89%

85%
88%
93%

55%
87%
77%

Widal test
Typhidott
Typhidot-Mt

phoid fever. When only the blood/bone marrow cultureproven cases (n 5 46) were analyzed, both the Typhidott
and Typhidot-Mt were significantly superior to the Widal
test in terms of the diagnostic predictive value (Table 3).
DISCUSSION

In contrast to findings from other parts of Asia,1922 our

data support the contention that the Widal test has poor diagnostic value in children with typhoid fever.23,24 Most of the
children presented in the second week of their illness and
we used a cut-off titer of 1:80. The use of a higher cut-off
titer would have further reduced the sensitivity of the test.
Although this has been questioned,25 antibiotic therapy has
been shown to alter the antibody response to S. typhi infection,26 and given that a large number of the patients had
previously received antibiotics, this factor may have altered
the antibody titers against O antigens. The Typhidott was
significantly more sensitive than the Widal test, although the
sensitivity and specificity were lower than those reported
from Malaysia and The Philippines.1012 Our findings of sensitivity and specificity were also lower than values . 90%
reported recently by Karamat and others27 from northern
Pakistan. These differences may be due to several factors
including the genomic diversity among S. typhi isolates in
the region28 and differences in antigenic epitopes. Other factors responsible for reported differences in areas of high endemicity are various stages of the illness and the rate of
increase of IgG antibodies to the OMPs, which may interfere
with identification of concomitant IgM antibodies. Most of
our patients presented in the second week of their illness,
whereas information on duration of illness is lacking in other
studies.10,11,28
The relative low sensitivity of the blood culture in diagnosing typhoid fever is understandable in the wake of widespread antibiotic use in Pakistan29 and the difficulties of obtaining large enough blood volumes for cultures from children. Although bone marrow cultures significantly increase
the yield from cultures,4,30 they are invasive and difficult to
obtain. It must be emphasized that although cultures are associated with a lag period of at least 48 hr for preliminary
confirmation of infection, with the recent emergence of drug
resistance among S. typhi, they remain an essential investigation. In many circumstances, especially among partially
treated cases presenting to health facilities, combining cultures with a rapid serologic test may reduce the diagnostic
difficulty in typhoid fever. Our data indicate that combining
the blood/bone marrow cultures with a Typhidot-Mt will
significantly improve the diagnostic yield of these investigations among children who have previously received antibiotics. We do not believe that our data support the use of

either the Widal or Typhidott tests as a substitute for cultures in typhoid fever.
The Typhidott offers an additional advantage among second-line serologic diagnostic tests for typhoid fever in that
the test strips do not require an ELISA reader for evaluation.
Also, only minimal operator training is required. Nevertheless, the 34-fold higher cost of the test in comparison with
the Widal test, as well as cold-storage requirements for test
reagents, are additional impediments in using this test in developing country. Although combining the Typhidott and
Typhidot-Mt tests may improve sensitivity, this is an expensive proposition. Given the recent call for an essential
diagnostics program in developing countries,31 it is important
that the Typhidott and Typhidot-Mt tests be evaluated on a
larger scale in different parts of the world with epidemiologically diverse strains of S. typhi.
Acknowledgments: We gratefully acknowledge the generous unrestricted donations of Typhidott and Typhidot-Mt kits by the Malaysian Biodiagnostic Research (Kuala Lumpur, Malaysia). We also
acknowledge the secretarial help of Ismail A. Rehmani.
Authors address: Zulfiqar Ahmed Bhutta and Naseem Mansurali,
Department of Paediatrics, The Aga Khan University Medical Center, PO Box 3500, Stadium Road, Karachi 74800, Pakistan.
REFERENCES

1. Ivanoff B, Levine MM, Lambert PH, 1994. Vaccination against

typhoid fever: present status. Bull World Health Organ 72:
957971.
2. Gupta A, 1994. Multidrug-resistant typhoid fever in children:
epidemiology and therapeutic approach. Pediatr Infect Dis J
13: 134140.
3. Bhutta ZA, 1996. Impact of age and drug resistant on mortality
in typhoid fever. Arch Dis Child 75: 214217.
4. Gilman RH, Terminel M, Levine MM, Hernandez-Menodoze P,
Hornick RB, 1975. Relative efficacy of blood, urine, rectal
swab, bone-marrow and rose spot cultures for recovery of
Salmonella typhi in typhoid fever. Lancet 1: 12111213.
5. Pang T, Puthucheary SD, 1983. Significance and value of the
Widal test in the diagnosis of typhoid fever in an endemic
area. J Clin Pathol 36: 471475.
6. Jesudason MV, Sridharan G, Mukundan S, John TJ, 1994. Vispecific latex agglutination for early and rapid detection of
Salmonella serotype typhi in blood cultures. Diagn Microbiol
Infect Dis 18: 7578.
7. Mukherjee C, Malik A, Khan HM, Malik A, 1993. Rapid diagnosis of typhoid fever by co-agglutination in an Indian hospital. J Med Microbiol 39: 7477.
8. Zhu Q, Lim CK, Chan YN, 1996. Detection of Salmonella typhi
by polymerase chain reaction. J Appl Bacteriol 80: 244251.
9. Ismail A, Hai OK, Kader ZA, 1991. Demonstration of an antigenic protein specific for Salmonella typhi. Biochem Biophys
Res Commun 181: 301305.
10. Choo KE, Oppenheimer SJ, Ismail AB, Ong KH, 1994. Rapid
serodiagnosis of typhoid fever by dot enzyme immunoassay
in an endemic area. Clin Infect Dis 19: 172176.

RAPID SEROLOGIC DIAGNOSIS OF TYPHOID FEVER IN AN ENDEMIC AREA

11. Jackson AA, Ismail A, Ibrahim TAT, Kader ZSA, Nawi NM,
1995. Retrospective review of dot enzyme immunoassay test
for typhoid fever in an endemic area. Southeast J Trop Med
Public Health 26: 625630.
12. Choo KE, Davis TM, Ismail A, Tuan Ibrahim TA, Ghazali WN,
1999. Rapid and reliable serological diagnosis of enteric fever: comparative sensitivity and specificity of Typhidot and
Typhidot-M tests in febrile Malaysian children. Acta Trop 72:
175183.
13. Immunochemical techniques, 1986. Wilson K, Goulding KH,
eds. A Biologists Guide to Principles and Techniques of
Practical Biochemistry. London: Arnold, 116152.
14. Choo KE, Davis TM, Ismail A, Ong KH, 1997. Longevity of
antibody responses to a Salmonella typhi-specific outer membrane protein: interpretation of a dot-enzyme immunosorbent
assay in an area of high typhoid fever endemicity. Am J Trop
Med Hyg 57: 656659.
15. Choo KE, Davis TME, Ismail A, Ibrahim TAT, Ghazali WNW,
1998. Comparative sensitivity and specificity of typhidot and
typhidot-M tests in the diagnosis of enteric fever in Malaysian
children. Med J Indonesia 7 (suppl 1): 284285.
16. Thong K-L, Puthucheary S, Yassin RM, Sudarmono P, Padmidewi M, Soewandojo E, Handojo I, Sarasombath S, Pang T,
1995. Analysis of Salmonella typhi isolates from southeast
Asia by pulsed-field gel electrophoresis. J Clin Microbiol 33:
19381941.
17. Shanahan PMA, Jesudason MV, Thomson CJ, Amyes SGB,
1998. Molecular analysis and identification of antibiotic resistance genes in clinical isolates of Salmonella typhi from
India. J Clin Microbiol 36: 15951600.
18. Bhutta ZA, Khan IA, Molla AM, 1994. Therapy of multidrugresistant typhoid fever with oral cefixime vs intravenous ceftriaxone. Pediatr Infect Dis J 13: 990994.
19. Bhutta ZA, Naqvi SH, Razzaq RA, Farooqui BJ, 1991. Multidrug-resistant typhoid fever in children: presentation and clinical features. Rev Infect Dis 13: 832836.
20. Chow C-B, Wang P-S, Cheung M-W, Yan W-W, Leung N-K,

21.

22.

23.
24.
25.
26.
27.
28.

29.

30.
31.

657

1987. Diagnostic value of the Widal test in childhood typhoid

fever. Pediatr Infect Dis J 6: 914917.
Choo KE, Razif AR, Oppenheimer SJ, Ariffin WA, Lau J, Abraham T, 1993. Usefulness of the Widal test in diagnosing childhood typhoid fever in endemic areas. J Paediatr Child Health
29: 3639.
Rasaily R, Dutta P, Saha MR, Mitra U, Bhattacharya SK, Manna
B, Mukherjee A, Chakravorty S, Pal SC, 1993. Value of a
single Widal test in the diagnosis of typhoid fever. Indian J
Med Res 97: 104107.
Saha SK, Ruhulamin M, Hanif M, Islam M, Khan WA, 1996.
Interpretation of the Widal test in the diagnosis of typhoid
fever in Bangladeshi children. Ann Trop Paediatr 16: 7578.
Onuigbo MAC, 1990. Diagnosis of typhoid fever in Nigeria:
misuse of the Widal test. Trans R Soc Trop Med Hyg 84:
129131.
Robertson RP, Abdel Wahab MF, 1970. Influence of chloramphenicol and ampicillin on antibody response in typhoid and
paratyphoid fever. Ann Intern Med 72: 219221.
Shukla S, Patel B, Chitnis DS, 1997. 100 years of Widal test
and its reappraisal in an endemic area. Indian J Med Res 105:
5357.
Karamat KA, Ahmed JU, Abbasi SA, 1998. Detection of serum
IgM against Salmonella typhi: a rapid diagnostic technique. J
Coll Physicians Surgeons Pak 8: 170173.
Pang T, Altwegg M, Martinetti G, Koh CL, Puthucheary S,
1992. Genetic variation among Malaysian isolates of Salmonella typhi as detected by ribosomal RNA gene restriction
patterns. Microbiol Immunol 36: 539543.
Sturm AW, van der Pol R, Smits AJ, van Hellemondt FM, Mouton SW, Jamil B, Minai AM, Sampers GHMA, 1997. Overthe-counter availability of antimicrobial agents, self-medication and patterns of resistance in Karachi, Pakistan. J Antimicrob Chemother 39: 543547.
Bhutta ZA, 1991. Bone marrow examination in prolonged fever
(letter). J Pediatr 119: 840.
Garner P, Kiani A, Supachutikul A, 1997. Diagnostics in developing countries: time for an essential diagnostics program.
Br Med J 315: 760761.