Dynamic Incidence of Typhoid Fever over a 10-Year Period (2010-2019) in Kibera, an Urban Informal Settlement in Nairobi, Kenya

Abstract

Typhoid fever burden can vary over time. Long-term data can inform prevention strategies; however, such data are lacking in many African settings. We reexamined typhoid fever incidence and antimicrobial resistance (AMR) over a 10-year period in Kibera, a densely populated urban informal settlement where a high burden has been previously described. We used data from the Population Based Infectious Diseases Surveillance platform to estimate crude and adjusted incidence rates and prevalence of AMR in nearly 26,000 individuals of all ages. Demographic and healthcare-seeking information was collected through household visits. Blood cultures were processed for patients with acute fever or lower respiratory infection. Between 2010 and 2019, 16,437 participants were eligible for blood culture and 11,848 (72.1%) had a culture performed. Among 11,417 noncontaminated cultures (96.4%), 237 grew Salmonella enterica serovar Typhi (2.1%). Overall crude and adjusted incidences were 95 and 188 cases per 100,000 person-years of observation (pyo), respectively. Annual crude incidence varied from 144 to 233 between 2010 and 2012 and from 9 to 55 between 2013 and 2018 and reached 130 per 100,000 pyo in 2019. Children 5-9 years old had the highest overall incidence (crude, 208; adjusted, 359 per 100,000 pyo). Among isolates tested, 156 of 217 were multidrug resistant (resistant to chloramphenicol, ampicillin, and trimethoprim/sulfamethoxazole [71.9%]) and 6 of 223 were resistant to ciprofloxacin (2.7%). Typhoid fever incidence resurged in 2019 after a prolonged period of low rates, with the highest incidence among children. Typhoid fever control measures, including vaccines, could reduce morbidity in this setting.

Figure 1.

Flow diagram showing overall clinic visits, blood cultures performed, pathogens isolated, and sequenced Salmonella enterica serovar Typhi in Kibera, Nairobi, Kenya (2010–2019). ∞Number of those eligible for blood culture is adjusted for 1,022 blood cultures with missing clinical records. A distribution of clinical syndromes similar to that in participants with clinical data was assumed for these records in the different years. *Contaminants defined as Staphylococcus species coagulase negative, Staphylococcus epidermidis, Staphylococcus xylosus, gram-positive rods, Micrococcus species, Serratia odorifera, gram-positive bacillus, and Erwinia species. **NTS refers to non-typhoidal Salmonella. †Considered repeat if S. Typhi was isolated twice in < 14 days in the same individual. ***Denominator is blood cultures that grew a pathogen. S. Typhi isolation was 2.1% overall among non-contaminated blood cultures.

Figure 2.

Proportion of Salmonella enterica serovar Typhi isolated from non-contaminated blood cultures by year at Tabitha Clinic, Kibera, Nairobi, Kenya, 2010–2019.

Figure 3.

Distribution of crude and adjusted incidence rates of typhoid fever by age in Kibera, Nairobi, Kenya, 2010–2019. Bars reflect 95% CI for the incidence estimate adjusted for both sample collection and healthcare seeking. To adjust for sample collection, overall crude incidence for each age category was divided by the proportion of all acute febrile illness and acute lower respiratory illness cases among active PBIDS participants in the respective age category seen in the clinic who had a blood culture performed. To adjust for sample collection and healthcare seeking, the sample collection–adjusted estimate was divided by the proportion of all medically attended acute febrile illnesses and acute lower respiratory illnesses in the respective age category reported in household surveys that had sought care at the Tabitha Medical Clinic surveillance facility, based on illnesses reported during 14 days preceding a household interview. PBIDS = Population-Based Infectious Disease Surveillance; PYO = person-years of observation.

Figure 4.

Crude and adjusted incidence trends of typhoid fever by year in Kibera, Nairobi, Kenya, 2010–2019. Shaded area reflects 95% confidence bands for the incidence estimates. Crude incidence was calculated by dividing the annual number of typhoid fever cases by person-years of observation among active PBIDS participants in that year. To adjust for sample collection, crude incidence was divided by the proportion of all acute febrile illness and acute lower respiratory illness cases among active PBIDS participants seen in the clinic who had a blood culture performed in that year. To adjust for sample collection and healthcare seeking, the sample collection–adjusted estimate was divided by the proportion of all medically attended acute febrile illnesses and acute lower respiratory illnesses reported in household surveys that had sought care at the Tabitha Medical Clinic surveillance facility in that year, based on illnesses reported during 14 days preceding a household interview. PBIDS = Population-Based Infectious Disease Surveillance; PYO = person-years of observation.

Figure 5.

Crude and adjusted age group–stratified typhoid fever incidence by year in Kibera, Nairobi, Kenya, 2010–2019. Crude incidence was calculated by dividing the annual number of typhoid fever cases in each age category by person-years of observation among active PBIDS participants in the age category in that year. To adjust for sample collection, crude incidence was divided by the proportion of all acute febrile illness and acute lower respiratory illness cases among active PBIDS participants in the respective age category and year seen in the clinic who had a blood culture performed. To adjust for sample collection and healthcare seeking, the sample collection–adjusted estimate was divided by the proportion of all medically attended acute febrile illnesses and acute lower respiratory illnesses in the respective age category and year reported in household surveys that had sought care at the Tabitha Medical Clinic surveillance facility, based on illnesses reported during 14 days preceding a household interview. PBIDS = Population-Based Infectious Disease Surveillance; PYO = person-years of observation.

Figure 6.

Multidrug resistance, ciprofloxacin resistance, and intermediate ciprofloxacin susceptibility over time among Salmonella enterica serovar Typhi blood culture isolates in Kibera, Nairobi, Kenya, 2010–2019. (A and B) Resistance proportion was calculated by dividing the number of S. Typhi isolates resistant to a specific antibiotic by the number of S. Typhi isolates tested for resistance to the antibiotic. Isolates were not tested for resistance where antibiotic disks were missing or when archived isolates could not be revived for later testing.