TYPHOID FEVER IN CHILDREN: DIAGNOSTIC AND THERAPEUTIC DIFFICULTIES

Abstract

As a result of increasing world travel, typhoid fever now presents more frequently in industrialized countries.1 There is little literature on typhoid fever in children in this setting.2 Most information on clinical presentation and therapy comes from countries where typhoid is endemic.3, 4 The presenting symptoms and signs of typhoid fever in children differ significantly from those in adults.2, 4 Studies from endemic areas show that the younger the children the more likely they are to present with a nonspecific febrile illness.4-6 Older children and adolescents have a presentation similar to that of adults.2 The relative rarity of typhoid fever in developed countries is likely to delay diagnosis.7 The purpose of this study was to highlight the diagnostic and therapeutic difficulties encountered in typhoid fever in children in an area where typhoid is not endemic. The aim was to determine whether certain clinical features at onset might hasten recognition and therapy and potentially reduce hospital stay. Methods. Cases of typhoid fever in children (age 0 to 14 years) were identified from two sources (laboratory records and discharge diagnoses) at two pediatric referral centers in Sydney (Royal Alexandra Hospital for Children and Westmead Hospital). At the Royal Alexandra Hospital for Children laboratory records were available from 1981 to 1995 and were the sole method of case ascertainment, whereas at Westmead Hospital laboratory records (1990 to 1995) and medical records (1981 to 1995) were used. A standard form was used for data extraction, which included details of ethnicity, recent travel or immigration, symptoms and signs at presentation, investigations, treatment and complications. Particular note was made of two outcome variables: time until defervescence and the duration of hospital admission. Results.Demographic data. Twenty-eight patients with Salmonella typhi isolated from blood or stool were identified. The age, sex and country of origin of these patients are summarized in Table 1.TABLE 1: Patient characteristics Initial diagnosis. In 18 patients (64%) the diagnosis of typhoid fever was not considered on admission. In 15 patients (53%) the diagnosis of typhoid fever was not entertained until blood cultures became positive. The time to identification of the organism in the blood culture was as long as 7 days, with a mean of 3.2 days, resulting in similar delays in instituting appropriate treatment. Many possible diagnoses were considered on admission. The most common were gastroenteritis (50%), pyrexia of unknown origin (44%) and malaria (39%). These diagnoses were often made by experienced pediatricians who had not considered typhoid fever. The most common presenting symptoms and signs are shown in Table 2. At least 1 abdominal symptom was present in 22 patients (78%). Typical symptoms in adults such as cough, headache and constipation were uncommon, tending to occur in older children. Common clinical signs of typhoid fever in adults8 such as relative bradycardia and rose spots were seldom documented.TABLE 2: Presenting symptoms and signs at hospital admission Laboratory investigations. Blood cultures were performed in all patients and grew S. typhi in 82%. Stool cultures were obtained from 25 patients and were positive in 72%. Widal titers were performed in 13 cases and were significantly elevated (O and H titers >1 in 160) in 3, although for most patients only acute serology was available. A full blood count was performed in 25 (90%) of patients. A white blood cell count of <5 × 109/liter was found in 3 patients; only 1 patient had a white blood cell count >15 × 109/liter. Antimicrobial therapy and hospital course. Before 1990 chloramphenicol (67%) was the most commonly used antimicrobial. Others used during this period, either alone or in combination with chloramphenicol, were amoxicillin and trimethoprim-sulfmethoxazole (TMP/SMX). After 1990 third generation cephalosporins were used in all cases; 80% of those patients also received ciprofloxacin when therapy was changed from intravenous to oral. Nine patients (32%) received only oral antibiotics (chloramphenicol, amoxicillin or TMP/SMX). Resistance to TMP/SMX or chloramphenicol was documented in five cases (18%), with multiple resistance (ampicillin, TMP/SMX and chloramphenicol) in one case. The mean hospital stay was 10 days, with a range up to 19 days. Intravenous fluids were required in 50% of patients for up to 7 days, with a mean of 2.3 days. The time to defervescence from institution of treatment was 2 to 11 days (mean, 5.5 days). Only two patients were discharged from the hospital while still febrile, both receiving ciprofloxacin. Complications developed in six cases (acalculous cholecystitis and arthritis in one, relapse in five). Discussion. Typhoid fever is uncommon in developed countries, with few reports on diagnosis and management in this setting.2 The symptoms and signs of typhoid fever in our patients were similar to those of other studies.3, 4 Given its uncommon presentation and nonspecific features, it was not surprising that the diagnosis of typhoid fever was not initially considered in two-thirds of cases. The change from chloramphenicol to third generation cephalosporins for primary therapy after 1990 reflects a world-wide trend caused by (plasmid-mediated) resistance to chloramphenicol.9 Resistance to chloramphenicol was seen in three cases, (1991, 1994) in this series, and to less frequently used drugs in three cases (sulfamethoxazole, ampicillin, trimethoprim). The average hospital stay in our series was 10 days, much longer than any adult series despite a lack of complications and the need for intravenous fluids for an average of only 2.3 days. Earlier discharge is facilitated by use of oral ciprofloxacin in adults, which is highly effective.8 Typhoid fever should be considered early in a child with fever and an appropriate travel history, especially if abdominal symptoms are present and malarial films are negative. Earlier diagnosis and institution of therapy with cefixime10 or ciprofloxacin (which can be used safely in children when indicated11) once oral intake has resumed will minimize hospital stay. Children need not be afebrile at discharge if they are clinically well. Michael O. Stormon, M.B.B.S. Peter B. McIntyre, Ph.D., F.R.A.C.P. Joanne Morris, M.B.B.S. Bruce Fasher, F.R.C.P., F.R.A.C.P. Department of Immunology and Infectious Diseases (MOS, PBM) Accident and Emergency Department (JM, BF) Royal Alexandra Hospital for Children; Parramatta, Australia