Enteric fever, despite its presence for a long time, very little progress has been made in the diagnosis of the disease. Diagnosis is still made either clinically or by Widal test and very rarely by blood culture. Blood culture is not only the gold standard for diagnosis but also suggests the sensitivity pattern of the organism.

Till 1990’s enteric fever was managed by cheap effective first line drugs. With the emergence of multi drug resistant typhoid the treatment has become complex with several antibiotics being used. This has further added to the problem of resistance with some strains showing resistance to both fluoroquinolones and cephalosporins. Indian Academy of Pediatrics has set up a National Task Force under IAP Action Plan 2006 to form a uniform guideline on management of enteric fever, particularly antibiotics to be used, which can be followed by all members of IAP in general. The list of the members who took part in the workshop is given in Annexure I.

The correct and rapid diagnosis of enteric fever is of paramount importance for instituting appropriate therapy and also for avoiding unnecessary therapy.

For practical purposes the CBC in enteric fever is unremarkable. The hemoglobin is normal in the initial stages but drops with progressing illness. Severe anemia is unusual and should make one suspect intestinal hemorrhage or hemolysis or an alternative diagnosis like malaria. The WBC count is normal in most cases and leukocytosis makes the diagnosis less probable. Leukopenia perceived to be an important feature of typhoid fever and has been reported in only 20-25% cases(1). The differential count is usually unremarkable except for eosinopenia. Eosinopenia often absolute may be present in 70-80% cases(2,3). Presence of absolute eosinopenia offers a clue to diagnosis but does not differentiate enteric fever from other acute bacterial or viral infections. Conversely, a normal eosinophil count does make typhoid fever a less likely possibility. Platelet counts are normal to begin with and fall in some cases by the second week of illness. Overall prevalence of thrombocytopenia is around 10-15%(4).

Blood cultures are the gold standard diagnostic method for diagnosis of enteric fever(5). The sensitivity of blood culture is highest in the first week of the illness and reduces with advancing illness(6). Overall sensitivity is around 50% but drops considerably with prior antibiotic therapy(6). Failure to isolate the organism may be caused by several factors which includes inadequate laboratory media, the volume of blood taken for culture, the presence of antibiotics and the time of collection. For blood culture it is essential to inoculate media at the time of drawing blood.

Salmonella can be easily cultured in most microbiologic laboratories with use of routine culture media (Hartley’s media, Blood agar and MacConkey agar). Automated blood culture systems such as BACTEC certainly enhance the recovery rate. Sufficient amount of blood should be collected for culture as the median bacterial count in the peripheral blood is only 0.3 CFU/ml (inter quartile range 0.1 to 10; range, 0.1 to 399)(7). At least 10 mL of blood in adults and 5 mL in children should be collected. Dilution should be appropriate in order to adequately neutralize the bactericidal effect of serum and a ratio of 1:5 to 1:10 of blood to broth is recommended. Clot cultures, wherein the inhibitory effect of serum is obviated have not been found to be of superior sensitivity as compared to blood cultures in several clinical studies(8-10). In the laboratory, blood culture bottles should be incubated at 37°C and checked for turbidity, gas formation and other evidence of growth after 1, 2, 3 and 7 days. For days 1, 2 and 3 only bottles showing signs of positive growth are cultured on agar plates. On day 7 all bottles should be sub-cultured before being discarded as negative.

There are considerable advantages of routine blood cultures in investigation of suspected enteric fever. Not only are they 100% specific, but they also provide information on the antimicrobial sensitivity of the isolate. This is vital in today’s scenario of multidrug resistance. Moreover, alternative methods for diagnosis particularly serology are rather unsatisfactory as will be discussed later. Blood cultures may actually turn out to be cheaper and very cost effective in the long run, as positive cultures unequivocally establish the diagnosis of enteric fever and all other investigations for PUO can be safely deferred.

Salmonella typhi is an intracellular pathogen in the reticuloendothelial cells of the body including the bone marrow(5). Studies have revealed that the median bacteremia in the bone marrow is 9 CFU/mL (IQR 1 to 85; range 0.1 to 15,805) compared to 0.3 CFU/mL (IQR 0.1 to 10; range 0.1 to 399) in blood. This bone marrow: peripheral blood ratio which is around 4.8 (IQR 1 to 27.5) in the first week of the illness increases to 158 (IQR 60 to 397) during the third week owing to disappearance of bacteria from the peripheral blood(7). The overall sensitivity of bone marrow cultures ranges from 80-95% and is good even in late disease and despite prior antibiotic therapy(5,11-13).

The invasive nature of bone marrow aspiration deters from its use as a first line investigation for diagnosis of typhoid fever. It is however a very useful and valid investigation in evaluation of PUO wherein the marrow should be inoculated in the culture bottle at the bedside.

Stool specimen should be collected in a sterile wide mouthed container. Specimens should preferably be processed within 2 hours after collection. If there is a delay the specimen should be stored in a refrigerator at 4°C or in a cool box with freezer packs. The sensitivity of stool culture depends on the amount of feces cultured, and the positivity rate increases with the duration of the illness. Rectal swabs should be avoided as these are less successful. Stool cultures are positive in 30% of patients with acute enteric fever(5). For the detection of carriers, several samples should be examined because of irregular shedding of salmonella. Urine cultures are not recommended for diagnosis in view of poor sensitivity(5,14). Other methods such as duodenal string and skin snip culture of rose spots have been reported to be more efficacious than blood cultures but are mainly of academic importance(14-16).

The crucial issue here pertains to fluoroquinolone susceptibility testing. Fluoroquinolones were introduced in 1989 and during the past decade there has been a progressive increase in the MICs of ciprofloxacin in Salmonella typhi and paratyphi(5). Since the current MIC’s are still below the National Committee for Clinical Laboratory Standards (NCCLS) susceptibility breakpoint, laboratory reports will continue to report Salmonella typhi/paratyphi as ciprofloxacin/ofloxacin sensitive(17). However, use of fluoroquinolones in this scenario is associated with a high incidence of clinical failure(5,17). It has also been demonstrated that resistance to nalidixic acid is a surrogate marker for high cipro-floxacin MIC’s, predicts fluoroquinolone failure and can hence be used to guide antibiotic therapy (i.e, if culture results show resistance to nalidixic acid irrespective of the results of ciprofloxacin/ofloxacin sensitivity, quinolones should not be used or if used high doses should be given)(18). Since MIC testing is not within the scope of most laboratories, nalidixic acid susceptibility testing is mandatory to help guide choice of antibiotics.

This test first described by F Widal in 1896, detects agglutinating antibodies against the O and H antigens of Salmonella typhi and H antigens of paratyphi A and B(6,19). The "O" antigen is the somatic antigen of Salmonella typhi and is shared by Salmonella paratyphi A, paratyphi B, other Salmonella species and other members of the Enterobacteriaceae family(20). Antibodies against the O antigen are predominantly IgM, rise early in the illness and disappear early(20). The H antigens are flagellar antigens of Salmonella typhi, paratyphi A and paratyphi B. Antibodies to H antigens are both IgM and IgG, rise late in the illness and persist for a longer time(19,20). Usually, O antibodies appear on day 6-8 and H antibodies on days 10-12 after the onset of disease. The test is usually performed on an acute serum (at first contact with the patient). A convalescent serum should preferably also be collected so that paired titration’s can be performed.

Conventionally, a positive Widal test result implies demonstration of rising titers in paired blood samples 10-14 days apart(19). Unfortunately, this criterion is purely of academic interest. Decisions about antibiotic therapy cannot wait for results from two samples. Moreover, antibiotics may dampen the immune response and prevent a rise in titers even in truly infected individuals. Therapeutic decisions have to be generally based on results of a single acute sample. In endemic areas, baseline anti O and anti H antibodies are present in the population owing to repeated subclinical infections with Salmonella typhi/paratyphi, infections with other Entero-bacteriaceae and other tropical diseases such as dengue and malaria(19-21). These antibody titers vary with age, socio economic strata, urban or rural areas and prior immunization with the TAB vaccine. Establishing appropriate cut offs for distinguishing acute from past infections is thus important for the population where the test is applied. In one study from Central India, anti O and anti H titer of more than 1:80 was seen in 14% and 8% respectively of a study sample of 1200 healthy blood donors(22).

While interpreting the results of the WIDAL test, both H and O antibodies have to be taken into account. There is controversy about the predictive value of O and H antibodies for diagnosis of enteric fever. Certain authorities claim that O antibodies have superior specificity and positive predictive value (PPV) because these antibodies decline early after an acute infection(23). Other studies report a poorer positive predictive value of O antibodies probably due to rise of these antibodies in other salmonella species, gram-negative infections, in unrelated infection and following TAB vaccination(21). For practical purpose and for optimal result this test should be done after 5-7 days of fever by tube method and level of both H and O antibodies of 1 in 160 dilution (four fold rise) should be taken as cut off value for diagnosis. H antibodies once positive can remain positive for a long time.

The Widal test as a diagnostic modality has suboptimal sensitivity and specificity(19-21). It can be negative in up to 30% of culture proven cases of typhoid fever. Sub optimal sensitivity results from negativity in early infection, prior antibiotic therapy and failure to mount an immune response by certain individuals(19). Poor specificity, an even greater problem and is a consequence of pre-existing baseline antibodies in endemic areas, cross reactivity with other Gram-negative infections and non-typhoidal salmonella, anamnestic reactions in unrelated infections and prior TAB or oral typhoid vaccination. The purity and standardization of antigens used for the WIDAL test is a major problem and often results in poor specificity and poor reproducibility of test results(19). The slide Widal test should also be discouraged owing to high rate of false positives(20).

Not withstanding these problems, the WIDAL test may be the only test available in certain resource poor set ups for diagnosis of enteric. In Vietnam, using a cutoff of >1/200 for the O agglutinin or >1/100 for H agglutinin test performed on acute-phase serum the Widal test could correctly diagnose 74% of blood culture positive typhoid fever, however 14% results would be false positive and 10% false negative(21). Hence, it is important to realize the limitations of the Widal test and interpret the results carefully in light of endemic titers so that both over diagnosis and under diagnosis of typhoid fever and the resulting consequences are avoided(24).

In view of the limitations of the Widal test and need for a cheap and rapid diagnostic method, several attempts to develop alternative serologic tests have been made. These include rapid dipstick assays, dot enzyme immuno-assays and agglutination inhibition tests(25-27).

Enzyme Immunoassay (EIA) test or Typhidotâ test: A dot enzyme immunoassay that detects IgG and IgM antibodies against a 50 KD outer membrane protein distinct from the somatic (O), flagellar (H) or capsular (Vi) antigen of Salmonella typhi is commercially available as Typhidotâ(27). The sensitivity and specificity of this test has been reported to vary from 70-100% and 43-90% respectively(28-33). This dot EIA test offers simplicity, speed, early diagnosis and high negative and positive predictive values. The detection of IgM reveals acute typhoid in the early phase of infection, while the detection of both IgG and IgM suggests acute typhoid in the middle phase of infection. In areas of high endemicity where the rate of typhoid transmission is high the detection of specific IgG increase. Since IgG can persist for more that 2 years after typhoid infection(34) the detection of specific IgG can not differentiate between acute and convalescent cases. Furthermore, false positive results attributable to previous infection may occur. On the other hand IgG positivity may also occur in the event of current reinfection. In cases of reinfection there is a secondary immune response with a significant boosting of IgG over IgM, such that the later can not be detected and its effect masked. A possible strategy for solving this problem is to enable the detection of IgM by ensuring that it is unmasked(35). The original Typhidotâ test was modified by inactivating the total IgG in the serum samples. Studies with modified test, Typhidotâ M, have shown that inactivation of IgG removes competitive binding and allows the access of the antigen to the specific IgM when it is present.

The Typhidotâ M that detects only IgM antibodies of Salmonella typhi has been reported to be slightly more specific in a couple of studies(26,33).

IDL Tubexâ test: The Tubexâ test is easy to perform and takes approximately 2 minutes time(36). The test is based on detecting antibodies to a single antigen in S. typhi only. The 09 antigen used in this test is very specific found in only sero group D salmonellae. A positive result always suggest a salmonellae infection but not which group D salmonella is responsible. Infection by other serotypes like S. paratyphi A give negative result. This test detects IgM antibodies but not IgG which is further helpful in the diagnosis of current infections.

IgM dipstick test(26): The test is based on the binding of S. typhi specific IgM antibodies to S. typhi lipopolysaccharide (LPS) antigen and the staining of the bound antibodies by an antihuman IgM antibody conjugated to colloidal dye particles. This test will be useful in places where culture facilities are not available as it can be performed without formal training and in the absence of specialized equipments. One should keep in mind that specific antibodies appear a week after the onset of symptoms so the sensitivity of this test increases with time.

Antigen detection tests: Enzyme immuno-assay’s, counterimmune electrophoresis and co-agglutination tests to detect serum or urinary somatic/flagellar/Vi antigens of Salmonella typhi have been evaluated(37-40). Sensitivity of Vi antigen has been found to be superior to somatic and flagellar antigen and has been reported as ranging from 50-100% in different studies(37-40). Similarly, specificity estimates have been reported to vary from 25-90%(37-40). The suboptimal and variable sensitivity and specificity estimates, inability to detect Salmonella paratyphi infection and Vi antigen negative strains of S typhi are serious limitations of the Vi antigen detection tests.

The limitations of cultures and serologic tests advocate for development of alternative diagnostic strategies. PCR as a diagnostic modality for typhoid fever was first evaluated in 1993 when Song, et al. successfully amplified the flagellin gene of S. typhi in all cases of culture proven typhoid fever and from none of the healthy controls(41). Moreover, some patients with culture negative typhoid fever were PCR positive suggesting that PCR diagnosis of typhoid may have superior sensitivity than cultures. Over the next 10 years a handful of studies have reported PCR methods targeting the flagellin gene, somatic gene, Vi antigen gene, 5S-23S spacer region of the ribosomal RNA gene, invA gene and hilA gene of Salmonella typhi for diagnosis of typhoid fever(42-50). These studies have reported excellent sensitivity and specificity when compared to positive (blood culture proven) and healthy controls. The turn- around time for diagnosis has been less than 24 hours.

These reports should be viewed within the context of certain limitations. Clinical utility of PCR tests has been inadequately evaluated. Performance of the test in individuals with febrile illnesses other than typhoid, in those with past history of typhoid, carriers of S typhi, and those vaccinated with typhoid vaccine is not known. Patients with a clinical diagnosis of typhoid fever who are culture negative but PCR positive may in fact be false positives. Comparison of PCR to bone marrow cultures as a gold standard may be a superior way of evaluating the sensitivity and specificity of these tests, but has not been done. The tests claim to detect as few as 10 organisms, but it should be remembered that in typhoid the median bacteremia is 0.3 CFU/ml of blood(7). Using small volumes of blood for DNA extraction may significantly lower the sensitivity of these tests. The cost and requirement for sophisticated instruments is also a potential drawback of molecular methods.

The complete blood count is the logical first investigation. Presence of a normal or low leukocyte count with eosinopenia points to possible enteric fever. It also helps in evaluation of alternative diagnoses such as malaria, dengue and other bacteremias. Blood culture remains the most effective investigation for diagnosis of enteric till date. They should be sent early in the course of the illness and prior to starting antibiotic therapy. Susceptibility testing for nalidixic acid should be routinely done for all isolates to aid choice of antibiotics. Bone marrow culture is a highly sensitive diagnostic test even in late stages of the illness and with prior antibiotic therapy. It should be performed in all patients with prolonged pyrexia if routine investigations have failed to establish a diagnosis. The Widal test has several limitations and should be requested for in the second week of the illness and its results interpreted with caution.