Current concepts in the diagnosis and treatment of typhoid fever

Introduction

Although advances in public health and hygiene have led to the virtual disappearance of enteric fever (more commonly termed typhoid fever) from much of the developed world, the disease remains endemic in many developing countries. Typhoid fever is caused by Salmonella enterica serovar Typhi (S typhi), a Gram negative bacterium. A similar but often less severe disease is caused by S paratyphi A and, less commonly, by S paratyphi B (Schotmulleri) and S paratyphi C (Hirschfeldii). The common mode of infection is by ingestion of an infecting dose of the organism, usually through contaminated water or food. Although the source of infection may vary, person to person transmission through poor hygiene and sewage contamination of water supply are the most important.

Have the epidemiology and burden estimates of typhoid changed?

Few established surveillance systems for typhoid exist in the developing world, especially in community settings, so the true burden is difficult to estimate. This is shown by recent revisions in the global estimates of the true burden of typhoid. In contrast to previous estimates, which were 60% higher,1 investigators from the US Centers for Disease Control and Prevention estimate that there are 21.6 million typhoid cases annually, with the annual incidence varying from 100 to 1000 cases per 100 000 population.2 The global mortality estimates from typhoid have also been revised downwards from 600 000 to 200 000, largely on the basis of regional extrapolations.2 Recent population based studies from South Asia suggest that the incidence is highest in children aged less than 5 years, with higher rates of complications and hospitalisation, and may indicate risk of early exposure to relatively large infecting doses of the organisms in these populations.3^–5 These findings contrast with previous studies from Latin America^w1 and Africa,^w2 which suggested that S typhi infection caused a mild disease in infancy and childhood.

There may be other factors that affect the changing epidemiology of typhoid. Although the overall ratio of disease caused by S typhi to that caused by S paratyphi is about 10 to 1, the proportion of S paratyphi infections is increasing in some parts of the world (Dong Mei Tan, personal communication 2005).6 Also, in contrast to the Asian situation, the HIV and AIDS epidemic in Africa has been associated with a concomitant increase in community acquired bacteraemia due to non-typhoidal salmonellae such as S typhimurium,7 8 an illness that may be clinically indistinguishable from typhoid. The exact reasons for these differences in the epidemiology and spectrum of salmonella infections between Asia and Africa remain unclear.

Another worrying development has been the emergence of drug resistant typhoid. After sporadic outbreaks of chloramphenicol resistant typhoid between 1970 and 1985, many strains of S typhi developed plasmid mediated multidrug resistance to the three primary antimicrobials used (ampicillin, chloramphenicol, and co-trimoxazole).9 This was countered by the advent of oral quinolones, but chromosomally acquired quinolone resistance in S typhi and S paratyphi^w3 has been recently described in various parts of Asia, possibly related to the widespread and indiscriminate use of quinolones.10 11

Can typhoid be diagnosed clinically where it matters?

Typhoid fever is among the most common febrile illnesses encountered by practitioners in developing countries. The advent of antibiotic treatment has led to a change in the presentation of typhoid, and the classic mode of presentation with a slow and “stepladder” rise in fever and toxicity is rarely seen. However, rising antimicrobial resistance has been associated with increased severity of illness and related complications.

Many other factors influence the severity and overall clinical outcome of the infection. They include the duration of illness before the start of appropriate treatment, the choice of antimicrobial, the patient's age and exposure or vaccination history, the virulence of the bacterial strain, the quantity of inoculum ingested, and several host factors affecting immune status. Recent data from South Asia indicate that the presentation of typhoid may be more dramatic in children younger than 5 years, with higher rates of complications and hospitalisation.3^–5 Diarrhoea, toxicity, and complications such as disseminated intravascular coagulation are also more common in infancy, with higher mortality. Table 1 shows some of the common clinical features and complications of typhoid in children and adults based on our experience in Karachi of hospitalised children and those diagnosed and treated in a community setting,5 12 indicating the significantly higher morbidity and complications among children presenting to hospital.

The presentation of typhoid fever may be altered by coexisting morbidities and early administration of antibiotics. In areas where malaria is endemic and where schistosomiasis is common the presentation of typhoid may be atypical.13 14 Multidrug resistant typhoid and paratyphoid infections are more severe with higher rates of toxicity, complications, and mortality than infections with sensitive strains.12 This may be related to the increased virulence of multidrug resistant S typhi as well as a higher number of circulating bacteria.15 Although clinical diagnosis of typhoid may be difficult, there are indications that simple algorithms can be developed for diagnosis and patient triage in endemic areas.16 Such algorithms would have implications for diagnostic and treatment protocols in endemic areas: in particular, diagnosis and triage of typhoid among febrile children must be included among the protocols for integrated management of childhood illnesses in South Asia, which currently largely focus on malaria as a cause of fever without localising signs.

The challenge of appropriate diagnostics in typhoid

Although the mainstay of diagnosing typhoid fever is a positive blood culture, the test is positive in only 40-60% of cases,17 usually early in the course of the disease. Stool and urine cultures become positive after the first week of infection, but their sensitivity is much lower. In much of the developing world, widespread antibiotic availability and prescribing is another reason for the low sensitivity of blood cultures. Although bone marrow cultures are more sensitive, they are difficult to obtain, relatively invasive, and of little use in public health settings.

Other haematological investigations are non-specific. Blood leucocyte counts are often low in relation to the fever and toxicity, but the range is wide; in younger children leucocytosis is a common association and may reach 20 000-25 000/mm³.12 ^w4 Thrombocytopenia may be a marker of severe illness and accompany disseminated intravascular coagulation. Liver function test results may be deranged, but significant hepatic dysfunction is rare.

The classic Widal test measures antibodies against O and H antigens of S typhi and is more than 100 years old.^w5 Although robust and simple to perform, this test lacks sensitivity and specificity, and reliance on it alone in areas where typhoid is endemic may lead to overdiagnosis.^w6 Newer diagnostic tests have been developed—such as the Typhidot^{w7 w8} or Tubex,^{w9 w10} which directly detect IgM antibodies against a host of specific S typhi antigens—but these have not proved to be sufficiently robust in large scale evaluations in community settings. A nested polymerase chain reaction using H1-d primers has been used to amplify specific genes of S typhi in the blood of patients and is a promising means of making a rapid diagnosis.^w11 Table 2 compares the performance of the various tests for typhoid.^w12-w14

Despite these new developments, the diagnosis of typhoid in much of the developing world is made on clinical criteria. This poses problems, since typhoid fever may mimic many common febrile illnesses without localising signs. In children with multisystem features, the early stages of enteric fever may be confused with conditions such as acute gastroenteritis, bronchitis, and bronchopneumonia. Subsequently, the differential diagnosis includes malaria; sepsis with other bacterial pathogens; infections caused by intra-cellular organisms such as tuberculosis, brucellosis, tularaemia, leptospirosis, and rickettsial diseases; and viral infections such as dengue fever, acute hepatitis, and infectious mononucleosis. There is thus an urgent need to develop a multipurpose “fever stick” that may allow the rapid and specific diagnosis of common febrile illnesses, especially malaria, dengue fever, and typhoid.^w15

How has drug resistance affected treatment?

Early diagnosis of typhoid fever and prompt institution of appropriate antibiotic treatment are essential for optimal management, especially in children. Although most cases can be managed at home with oral antibiotics and regular follow-up, patients with severe illness, persistent vomiting, severe diarrhoea, and abdominal distension require hospitalisation and parenteral antibiotic treatment. In addition to antibiotics, supportive treatment and maintenance of appropriate nutrition and hydration are crucial (box 1).

Appropriate antibiotic treatment (the right drug, dose, and duration) is critical to curing typhoid with minimal complications.18 Standard treatment with chloramphenicol or amoxicillin is associated with a relapse rate of 5-15% or 4-8% respectively, whereas the newer quinolones and third generation cephalosporins are associated with higher cure rates.17 The emergence of multidrug resistant typhoid in the 1990s led to widespread use of fluoroquinolones as the treatment of choice for suspected typhoid, especially in South Asia and South East Asia where the disease was endemic.19 In recent years, however, the emergence of resistance to quinolones has placed tremendous pressure on public health systems in developing countries as treatment options are limited.20 21

Table 3 shows the World Health Organization's recommendations for treating uncomplicated and severe cases of typhoid fever.17 Studies of short course antibiotic treatment for multidrug resistant typhoid have shown that fluoroquinolones can achieve satisfactory cure rates,^{w16 w17} but parenteral ceftriaxone was associated with higher rates of relapse.^w18 A recent Cochrane review of antimicrobial treatment of typhoid fever concludes that there is little evidence to support administration of fluoroquinolones to all cases of typhoid and that satisfactory cure rates can be achieved in drug sensitive cases with first line agents such as chloramphenicol.22 Although some open studies have suggested that cure rates may be better with oral fluoroquinolones compared with chloramphenicol,23 these case series also include multidrug resistant cases. Given the signs of rapidly increasing resistance of S typhi to fluoroquinolones, it is imperative that the widespread use of these antibiotics for fever and their availability over the counter are restricted, although it may already be too late.24 However, treatment regimens must restrict as much as possible the use of further second and third line antibiotics for treating typhoid in primary care settings.25

The prognosis for a patient with enteric fever depends on the rapidity of diagnosis and treatment with an appropriate antibiotic. Other factors include the patient's age, general state of health, and nutrition; the causative Salmonella serotype; and the appearance of complications. Infants and children with underlying malnutrition and those infected with multidrug resistant isolates are at higher risk of adverse outcomes. Although additional treatment with dexamethasone (3 mg/kg for the initial dose, followed by 1 mg/kg every 6 hours for 48 hours) has been recommended among severely ill patients with shock, obtundation, stupor, or coma,^w19 this must be done only under strictly controlled conditions and supervision, and signs of abdominal complications may be masked.

Despite appropriate treatment, some 2-4% of infected children relapse after initial clinical response to treatment.17 Individuals who excrete S typhi for more than three months after infection are regarded as chronic carriers. However, the risk of becoming a carrier is low in children and increases with age, but in general it occurs in less than 2% of all infected children.17

In summary, many challenges remain for the effective control and management of typhoid in endemic countries. Although these include establishing rapid clinical diagnosis and confirmation, the fact that both S typhi and S paratyphi are rapidly becoming resistant to commonly used antibiotics is of great concern. Addressing this issue would require a host of measures, including adequate investments in safe water and sanitation services, community education, control over antimicrobial prescribing and over the counter sales, and large scale vaccination strategies. Box 2 details some of the preventive strategies and advice for travellers to areas where typhoid is endemic.