Under Detection Leaves Typhoid Incidence Potentially Underestimated

A new study examining conjugate vaccine trial data has concluded that typhoid incidence estimates should be adjusted for the proportion of cases which go undetected due to lack of blood cultures.

In Nepal, the Typhoid Vaccine Acceleration Consortium has been studying a typhoid conjugate vaccine and recently announced promising interim results. Trial participants will be followed for several years, but blood tests indicate good protection thus far.

Data from the clinical trial of the typhoid conjugate vaccine may also help provide a better picture of the true burden of typhoid in endemic populations.

A new study published in PLOS Neglected Tropical Diseases has examined data from the aforementioned trial and concluded that typhoid incidence estimates should be adjusted for the proportion of cases which go undetected due to lack of blood cultures. The caveat is that patients who do not receive a blood test are less likely to have the disease.

The clinical trial randomized 20,019 children between 9 months and <16 years of age in Kathmandu, Nepal, to receive either the typhoid conjugate vaccine or Group A meningococcal conjugate vaccine between November 2017 and April 2018.

As part of surveillance efforts integrated into the clinical trial, participants who experience fever are encouraged to go to a tertiary-care hospital or a community clinic involved in the study. Those with a self-reported fever duration of 2 or more days or a temperature above 100.04 degrees Fahrenheit are eligible for blood culture.

Clinical staff were trained to administer blood tests to all children meeting screening criteria, but in practice some patients did not provide consent or were not offered a blood culture due to provider assessment. Investigators also examined factors associated with a positive culture among those who were tested.

Blood culture data were analyzed using chi-square tests to assess associations with blood culture collection and growth of Salmonella typhi and paratyphi. For participants without blood culture data, the investigators used multiple imputations to create 100 datasets to model which participants were likely to have been blood culture positive.

Within the first year of study, 1903 children with eligible fevers visited surveillance clinics 2392 times. Among these children, blood was drawn for culture during 1615 visits (68%). Blood samples were positive for S typhi or S paratyphi 35 times (2.2%).

Median age at fever presentation was 4 years. The participants had a median of 2 days of fever. A temperature of at least 100.04 degrees Fahrenheit was detected in 52% of patients with fever.

Despite the screening criteria, investigators found there were significantly more blood cultures performed when there was clinical suspicion of typhoid fever, when a child was older, or during a longer duration of fever.

The strongest predictor of a positive culture was clinical suspicion of typhoid fever, yet the authors indicate that “a third of blood culture-positive fever cases were children with alternative diagnoses or for whom typhoid was not suspected as their cause of illness, even though participants were enrolled in a typhoid vaccine trial.”

However, clinical suspicion that a participant had an upper respiratory tract infection or other alternative diagnosis significantly reduced the probability of blood being collected for a culture (P < .0001 and P = .006, respectively.)

Blood was more likely to be sampled if a urinary tract infection was the initial diagnosis (P = .006).

The 35 blood culture positive results represented 2.17% of fever cases. In the 777 participants with fever who did not have a blood culture, a median of 9 out 777 (1.16%) were imputed. This calculation accounts simultaneously for both the lower risk of typhoid and the likelihood of missed infections in those who did not have a blood culture.

By accounting for competing sources of bias, the investigators were able to better estimate incidence of typhoid fever among the study population. They report a high degree of accuracy in clinical predictions of typhoid alongside a substantial proportion of culture positive cases with an alternative diagnosis.

The study authors acknowledged clinical opinion, as well as the resource intensive nature of typhoid surveillance, will always impact whether blood cultures are performed or not. But the investigators concluded that their study ultimately indicates that “crude typhoid incidence estimates should be adjusted for the proportion of cases that go undetected due to missing blood cultures while adjusting for the lower likelihood of culture-positivity in the group with missing data.”