Abstract
BACKGROUND: Multidrug-resistant tuberculosis (MDR-TB) is an important global public health threat, but accurate estimates of MDR-TB burden among children are lacking. METHODS: We analyzed demographic, clinical, and laboratory data for newly diagnosed pediatric (age <15 years) TB cases reported to the US National TB Surveillance System during 1993-2014. MDR-TB was defined as culture-confirmed TB disease with resistance to at least isoniazid and rifampicin. To ascertain potential underestimation of pediatric MDR-TB, we surveyed high-burden states for clinically diagnosed cases treated for MDR-TB. RESULTS: Of 20789 pediatric TB cases, 5162 (24.8%) had bacteriologically confirmed TB. Among 4826 (93.5%) with drug susceptibility testing, 82 (1.7%) had MDR-TB. Most pediatric MDR-TB cases were female (n = 51 [62%]), median age was 5 years (interquartile range, 1-12 years), one-third were Hispanic (n = 28 [34%]), and two-thirds (n = 55 [67%]) were born in the United States. Most cases had additional resistance to ≥1 other first-line drug (n = 66 [81%]) and one-third had resistance to ≥1 second-line drug (24/73 tested). Of 77 who started treatment prior to 2013, 66 (86%) completed treatment and 4 (5%) died. Among the 4 high-TB-burden states/jurisdictions surveyed, there was 42%-55% underestimation of pediatric MDR-TB cases when using only culture-confirmed case definitions. CONCLUSIONS: Only one-quarter of pediatric TB cases had culture-confirmed TB, likely resulting in underestimation of true pediatric MDR-TB burden in the United States using strictly bacteriologic criteria. Better estimates of pediatric MDR-TB burden in the United States are needed and should include clinical diagnoses based on epidemiologic criteria.