The diagnosis of typhoid fever is usually based on clinical criteria and in the majority of cases it will be made without formal proof. Perforation of the terminal ileum is quasi pathognomonic for typhoid fever, but it is of course better to make the diagnosis before this complication arises. In clinical practice there are actually few disorders which cause perforations in the terminal ileum: typhoid fever, tuberculosis and Crohn’s disease. In many developing countries, two diseases often act as default diagnoses: malaria and typhoid fever. This illustrates the difficulties and uncertainties with which clinicians are confronted, together with the fact that both diseases are relatively frequent and are treatable (low threshold for diagnosis). Further, too much importance is attached to a Widal test and the interpretation of a thick smear is often not reliable in a local laboratory (the problem is not the thick smear itself but the reading of it).

Cultures (bone marrow, blood, faeces, urine, duodenal aspirate or string test) will often be positive, but are often not feasible in practice. The chance of obtaining a positive culture is higher if repeated cultures are taken. Positive cultures still form the gold standard for diagnosis.

Serological tests for antibodies to O and H antigens can be carried out (Widal). A Widal test is only positive in 50 % at the beginning of hospitalisation, and may be positive due to salmonellosis suffered previously (S. enteritidis non-Typhi) or due to an earlier vaccination. Routinely requesting this test under third world conditions makes no sense. The test can be used to detect a rising titre (seroconversion). Antibodies to the O antigen rise swiftly, and return to negative or to low titres in a couple of months (in particular type IgM antibodies). Anti-H antibodies rise rather more slowly but will stay positive for longer (in particular type IgG antibodies). If the presence of advanced typhoid fever is suspected on clinical grounds and if malaria is ruled out and a single Widal test is carried out (preferably using O antigen), then a high titre of these antibodies is a relatively strong argument that the patient does indeed have typhoid fever. Nothing can be decided from a negative result.

A complete blood count and differential often shows normal or reduced white blood cells. The eosinophils will be low or zero. In intestinal perforation there is leukocytosis, and in intestinal bleeding there is significant anaemia. A chest X-ray is often normal, in spite of the frequent presence of respiratory symptoms.