- Email: [email protected]
Interferon-Gamma Release Assay-Based Screening for Pediatric Latent Tuberculosis Infection in an Urban Primary Care Network
ARTICLE IN PRESS THE JOURNAL OF PEDIATRICS • www.jpeds.com
ORIGINAL ARTICLES
Interferon-Gamma Release Assay-Based Screening for Pediatric Latent Tuberculosis Infection in an Urban Primary Care Network James Gaensbauer, MD, MScPH1,2,3,4, Bryn Gonzales, MD1, Robert Belknap, MD2,5, Michael L. Wilson, MD6,7, and Mary E. O'Connor, MD, MPH8,9,10 Objective To assess outcomes from a QuantiFERON-tuberculosis (TB) Gold (QFT)-based screening for pediatric latent TB infection (LTBI) in the Denver Health Community Health System (CHS), an urban primary-care network in the US. Study design We retrospectively analyzed all QFTs (n = 6685) performed on children aged 2-18 years between January 5, 2011, and August 18, 2014. Risk factors for positive testing in the CHS population were identified by logistic regression, and further assessed using a case-control comparison. Results from CHS were compared with higher-TB-risk populations (refugee and TB clinics) in our health system. Results Positive QFT occurred in 79 of 3745 (2.1%) CHS patients. Positive rates increased with age (0.3% in age 2-5 years to 4.9% in age 13-18 years). Indeterminate results were uncommon (0.8%) including in children <5 (1.3%). Risk factors for positive tests in the CHS population included non-Medicaid insured/uninsured and nonEnglish/Spanish preferred language. In the case-control analysis, birth/travel to/residence in a TB-endemic country was the only identified risk factor for positive testing (OR 5.2 [95% CI 1.04-25.5]). Rates of positive testing were lower in the CHS population than the refugee/TB clinic populations, including among children age 2-5. Discussion QFT-based LTBI screening was successfully introduced in our pediatric primary-care health system, and supported our programmatic goals of identifying LTBI cases while limiting unnecessary LTBI treatment courses. Increasing positive rates with age, and higher rates in the refugee/TB populations compared with CHS, add indirect evidence of adequate test sensitivity, even among young children, for whom data on interferon-gamma release assay performance are limited. (J Pediatr 2018;■■:■■-■■).
I
dentification and treatment of pediatric tuberculosis (TB) infection is an important national priority. Children are at greatest risk for progression to active disease during infancy and late adolescence, and young children have heightened morbidity and mortality because of disseminated disease.1 A subset of children with untreated latent TB infection (LTBI) will become symptomatic and infectious later in life, and, thus, treatment of children with LTBI is an essential component of long-term TB elimination programs in the US. Screening programs to identify LTBI in children play an important role in the prevention of childhood TB disease in the US, and much of this activity takes place within primary care clinics serving populations at higher risk of Mycobacterium tuberculosis infection.2 However, critical questions remain as to how to best define and identify candidates for LTBI screening in primary care based on demographic and travel-related risk factors, determine the optimal age for initial screening and timing of repeat testing, and optimize the use of available screening tools, including tuberculin skin tests (TSTs) and interferon-gamma release assays (IGRA).3 Denver Health is an urban safety-net hospital and a network of Community From the 1Department of Pediatrics, Denver Health and Hospital Authority; 2Denver Metro Tuberculosis Clinic, Health System (CHS) clinics that provide care for a large population of foreignDenver Public Health, Denver; 3Department of Pediatrics, born children and US-born children with foreign-born parents. In 2011, Section of Infectious Diseases, University of Colorado School of Medicine; 4Department of Epidemiology, QuantiFERON-TB Gold (QFT) (Qiagen, Germantown, Maryland) became the preColorado School of Public Health; 5Department of Infectious Diseases, University of Colorado School of ferred test for diagnosing LTBI in children ≥2 years old at all Denver Health amMedicine, Aurora; 6Department of Pathology and bulatory clinics. The programmatic goal of our screening program was to identify Laboratory Services, Denver Health and Hospital Authority, Denver; 7Department of Pathology, University of as many children with LTBI as possible while minimizing the burden to patients Colorado School of Medicine; 8Department of Pediatrics, University of Colorado School of Medicine, Aurora CO; and system resources from unnecessary investigation and treatment of children 9Children’s Hospital, Dartmouth-Hitchcock, Lebanon; and 10
Department of Pediatrics, Geisel School of Medicine at Dartmouth, Hanover, NH The authors declare no conflicts of interest.
ATS BCG CHS IGRA LTBI QFT TB TST
American Thoracic Society Bacille Calmette-Guérin Community Health System Interferon-gamma release assay Latent TB infection QuantiFERON-TB Gold Tuberculosis Tuberculin skin test
Portions of this study were presented as posters at the Union Against Tuberculosis (TB) and Lung Disease, North American Region/National TB Controllers Association Annual Conference, February 24-27, 2016, Denver, Colorado and the Union Against TB and Lung Disease North American Region Annual Conference, February 2225, 2017, Vancouver, British Colombia, Canada. 0022-3476/$ - see front matter. © 2018 Elsevier Inc. All rights reserved. https://doi.org10.1016/j.jpeds.2018.04.034
1 FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018
THE JOURNAL OF PEDIATRICS • www.jpeds.com with false positive testing. The rationale for an IGRA-based strategy was based on multiple potential advantages of QFT over TSTs, including increased specificity for LTBI in foreignborn, Bacille Calmette-Guérin (BCG)-vaccinated children and in children exposed to non-tuberculous mycobacteria and concern that false positive TSTs might be leading to unnecessary treatments and resource utilization.4,5 In addition, IGRAs offered the opportunity to perform screening at a single visit on any day of the week, and were able to be reported and retrieved more clearly in our medical record system.6,7 Finally, though need for phlebotomy for IGRA had been seen as a disadvantage over TSTs, particularly in young children, our clinics have expertise in pediatric phlebotomy, whereas we were concerned that the technical skills for TST placement and interpretation were waning. In the US, recent observations from an urban public healthbased TB control program demonstrated the advantages of an IGRA-based screening in a predominantly foreign-born, BCGvaccinated pediatric population.8 However, the performance of IGRAs for LTBI screening has not been clearly documented in US-born/non-BCG vaccinated children in primary care settings, in whom the principle risk factors for TB infection are contact with adults/relatives from TB-endemic countries or intermittent travel to these countries.9,10 In addition, the performance of IGRAs in younger children has yet to be definitively established, particularly in children less than 5 years of age, resulting in current recommendations that they be used with caution in this age group.2,7,11 We retrospectively reviewed the results of all children tested during the first 3.5 years after implementation of an IGRAbased LTBI screening program instead of TST testing. The primary objectives of the study were to assess performance of QFT testing from a programmatic perspective, and demonstrate the feasibility of IGRA-based TB screening in pediatric primary care. Secondary objectives included an analysis of children with positive testing to characterize risk factors for TB infection and inform ongoing development of efficient strategies for LTBI screening in our population. Finally, in the face of uncertainties related to IGRA performance, particularly in younger children, we sought to assess results in comparison with results in higher prevalence refugee and TB clinic patients in our health system.
Methods The study population comprised all pediatric patients aged 2-18 years screened for LTBI with QFT between January 5, 2011 and August 18, 2014 at Denver Health. The primary objective was to describe the performance of QFT in the CHS, comprised of 14 school-based health centers, 3 pediatric primary care clinics, and 6 family medicine clinics. In these sites, TB screening was performed as part of routine preventive care for children age 2 years and above who were born or traveled to a TB-endemic country, or whose parents were born in a TB-endemic country. Repeat testing in CHS was recommended for patients with new travel to a high-risk county or other new risk factor. In the CHS population, Mexico is the most common non-US country of
Volume ■■
•
■■ 2018
family origin (internal institutional data). To compare the results of IGRA-based screening in the CHS population with pediatric populations at higher risk of TB, we included patients from 2 Denver Health affiliated clinics: the Denver Health Refugee Clinic and the Denver Metro TB clinic. During the study period, patients in the Refugee Clinic had LTBI testing primarily as part of routine new immigrant health screening. The TB clinic is affiliated with the Denver Department of Public health and testing occurred primarily in the context of contact investigation or follow-up of external referrals for positive TSTs. Data Collection All QFTs drawn on pediatric patients during the study period were identified from a Denver Health hospital laboratory information system (Sunquest, Tucson, Arizona). Data extracted included test-order date, location, QFT result, and basic patient demographics including age, sex, insurance status, self/family-reported race/ethnicity, and primary spoken language. For all children with positive QFT’s in CHS, medical record review was performed to determine results of chest radiographs and LTBI treatment initiation. Chart review was not performed on refugee and TB clinic patients, many of whom have medical records outside the Denver Health system; available data was limited to QFT results and basic demographics. The project was approved as exempt review by the Colorado Multiple Institutional Review Board. Case-Control Methods To assess factors that might allow for a more targeted approach to LTBI testing in our population, all children in CHS with a positive QFT were matched 1:1 by age at time of testing, clinic site, and preferred language in the home (English, Spanish, or other) to children with a negative QFT. The majority of children were matched within 6 months of age, though a small number required a range of 12 months. Medical record review was performed on all cases and controls to determine the indication for LTBI testing. Indications were placed into 3 categories (limited by the granularity of data available in the medical record): born/lived or traveled to a TB-endemic country, exposure to jail/alcoholism/drug abuse/non-Englishpreferred household language, and no documented reason. This analysis was originally designed as an exploratory quality improvement project and no a priori power calculation was performed. A case:control ratio of 1:1 was selected based on available resources for chart review. Statistical Analyses All patient information was deidentified and analyzed using SPSS (IBM, Armonk, New York) v 23, using c2/Fisher exact test, Kruskal-Wallis, and logistic regression. Factors associated with positive QFT testing in the case-control analysis were assessed by multivariate logistic regression.
Results There were 6685 QFT tests performed on 6338 unique children in the 3 sites. We excluded 2 results from children who
2
Gaensbauer et al FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018
■■ 2018
ORIGINAL ARTICLES
Figure 1. Diagram of QuantiFERON-TB Gold tests and individual patients included in analysis, Denver Health, 2011-2014.
had been treated for LTBI in the past. Of the 6683 test results, 53 (0.8%) were indeterminate, 455 (6.8%) were positive, and 6175 (92.4%) were negative. Two or more tests were performed on 347 children, yielding 6336 unique children who were tested at least once (Figure 1). Of these, 34 children/36 tests had only an indeterminate result and were excluded from any analysis of patient characteristics based on result. Demographic information and QFT results by clinic type are presented in Table I. Results from children tested in CHS are described in Table II. Only 5 (0.3%) children aged 2-5 years had a positive QFT, increasing to 2.3% in children age 6-12 years and 4.9% over 12 years. Indeterminate results occurred in only 1.0% of CHS patients. In multivariate logistic regression analysis, risk factors for positive testing in CHS included increasing age (OR 3.0; 95% CI 2.1-4.2) per age group, uninsured 4.4 (95% CI 2.48.2) or non-Medicaid insured status (OR 4.0; 95% CI 1.79.3), and a preferred primary language in the home other than English or Spanish (OR 2.9; 95% CI 1.4-6.1). Case-Control Table III compares the children in CHS with a positive QFT matched to children with a negative QFT. The only variable associated with positive testing was a child who was born/ lived/traveled to a TB-endemic county with an aOR of 5.2 (95% CI 1.04-25.5) compared with those with no travel or family residence risk factors. The adjusted OR of children living in a non-English-preferred household language/exposure to drug abuse or alcoholism/jail was not statistically significantly different (OR 2.92; 95% CI 0.5-15.4).
Clinical Outcomes Of the 79 children with positive QFT, 73 had a chest radiograph completed (all normal) and 70 (89%) were initiated on LTBI treatment. Five patients were lost to follow-up after
Table I. Comparison of QFT results and demographic factors between testing sites, Denver Health, 2011-2014; n = 6302 patients* Characteristics
Number of patients QFT positive† Male sex Median age (y)‡ Age groups (y) 2-5* 6-12 13-18 Race/ethnicity Hispanic Native American Asian/Pacific Islander Black/African American White Unknown Preferred language in home English Spanish Other Unknown
Patient population CHS
Refugee clinic
TB clinic
3745 (59.4%) 79 (2.1%) 1836 (49.0%) 8
1286 (20.4%) 101 (7.9%) 649 (50.5%) 9
1271 (20.2%) 270 (21.2%) 647 (50.9%) 14
1497 (40.0%) 1292 (34.5%) 956 (25.5%)
364 (28.3%) 559 (43.5%) 363 (28.2%)
140 (11.1%) 362 (28.5%) 769 (60.5%)
3233 12 142 185 150 23
(86.3%) (0.3%) (3.8%) (4.9%) (4.0%) (0.6%)
2 0 217 121 52 894
(0.2%) (0%) (16.9%) (9.3%) (4.0 %) (69.5%)
207 1 177 108 81 697
(16.3%) (0.1%) (13.9%) (8.5%) (6.4%) (54.8%)
1240 2293 199 1
(33.4%) (61.1%) (5.3%) (0.1%)
46 6 811 423
(3.6%) (0.5%) (63.1%) (32.9%)
131 87 165 888
(10.3%) (6.8%) (13.0%) (69.8%)
*Patients with indeterminate result excluded. †P < .001, c2. ‡P < .001, Kruskal-Wallis.
Interferon-Gamma Release Assay-Based Screening for Pediatric Latent Tuberculosis Infection in an Urban Primary Care Network FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018
3
THE JOURNAL OF PEDIATRICS • www.jpeds.com
Volume ■■
Table II. Comparison of all QFT-positive and QFT-negative in patients age 2-18 years, Denver Health CHS 2011-2014; n = 3745 Characteristics Total Sex Female Male Age grouping 2-5 y 6-12 y 13-18 y Insurance Medicaid/pending State Children's Health Insurance Program Commercial Uninsured Discount § Racial/ethnic group Hispanic Black/African American Asian/Pacific Islander White Native American Language English Spanish Other
Statistical significance*
aOR†
1872 (51.1%) 1794 (48.9%)
NS
NS
5 (6.3%) 26 (32.9%) 48 (60.8%)
1492 (40.7%) 1266 (34.5%) 908 (24.8%)
P < .001
reference 3.0 (2.1-4.2)‡
39 4 7 15 14
(49.4%) (5.1%) (8.9%) (19.0%) (17.7%)
2876 299 77 123 290
(78.5%) (8.2%) (2.1%) (3.4%) (7.9%)
P < .001
reference NS 4.0 (1.7-9.3) 4.4 (2.4-8.2) NS
60 9 5 3 1
(75.9%) (11.4%) (6.3%) (3.8%) (1.3%)
3173 176 137 147 11
(86.6%) (4.8%) (3.7%) (4.0%) (0.3%)
P < .04
NS
1221 (33.3%) 2250 (61.4%) 190 (5.2%)
P < .05
Reference NS 2.9 (1.4-6.1)
QFT-positive
QFT-negative
79
3666
37 (46.8%) 42 (53.1%)
27 (34.1%) 43 (54.4%) 9 (11.4%)
NS, not significant. *c2 †Logistic regression with (95% CI). ‡aOR for each step up in age compared with 2- to 5-year-old cohort. §Community Health Center subsidized care for legally undocumented children.
screening, and 3 patients/families refused LTBI treatment. One patient was diagnosed with active TB; review of the medical record revealed that he was actually referred to a CHS provider by an ophthalmologist for evaluation of retinal disease suspicious for TB. As of February 27, 2018, none of the remaining 78 QFT-positive children have been identified with progression to active TB disease at the TBC (which
provides care or consultation to all pediatric TB cases in the state of Colorado). Repeat QFT Testing Repeat testing was performed on 346 children (Table IV). Only 1 child from CHS converted from negative to positive after living in Mexico for the summer. All other children who
Table III. Case-control comparison of QFT-positive and QFT-negative children, Denver Health CHS, 2011-2014; n = 79 each group Characteristics Sex Male Female Insurance type Uninsured Medicaid SCHIP Commercial Racial/ethnic group Hispanic Black/African Other White Risk None documented Non-English-preferred household/jail/ drug/alcohol exposure Born/traveled/lived TB-endemic country
QFT-positive
QFT-negative
P value*
aOR†
42 (58.3%) 37 (43.0%)
30 (41.7%) 49 (57%)
.06
NS
29 39 4 7
(36.7%) (49.4%) (5.0%) (9.0%)
12 54 8 5
(15.2%) (68.4%) (10.1%) (6.3%)
.01
NS
60 10 6 3
(75.9%) (12.7%) (7.6%) (3.8%)
66 6 4 3
(83.5%) (7.6%) (5.1%) (3.8%)
0.64
NS
2 (2.5%) 19 (24.0%)
8 (10.1%) 26 (32.9%)
.04
Reference 2.92 (.5-15.4)
58 (73.4%)
45 (57.0%)
5.2 (1.04-25.5)
2
*c †Multivariable logistic regression with aOR with 95% CI.
4
Gaensbauer et al FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018
■■ 2018
ORIGINAL ARTICLES
Table IV. Results of repeat QFT testing in CHS, refugee clinic, and TB clinic pediatric populations, Denver Health, 2011-2014 Initial results Indeterminate
Subsequent results
Number of patients
Negative Indeterminate
17 1
Positive
Positive
5
Negative
Positive
12
Positive
Negative
2
Negative
Negative
309
Comments Indeterminate value returned twice Confirmatory testing of first test in new setting 1 in CHS after prolonged travel to Mexico; remaining 11 in TB clinic patients Retests at 1 and 9 mo after initial result
converted from negative to positive QFT were in the TB clinic cohort and converted in the context of exposure to a case of active TB. Eighteen of the 36 children with an initial indeterminate result underwent retesting, of which 17 (94%) were negative and 1 (6%) remained indeterminate. Comparison of CHS Results with Refugee and TB Clinics The percentage of positive test results was highest in the TB clinic (21.2%), followed by the refugee (7.9%), and CHS clinics (2.1%) (P < .001, c2). Test positivity increased with age in each site (Figure 2). Among children age 2-5 years, higher rates of positive testing were observed in the TB (5.2%) and refugee clinics (4.9%) than among the CHS population (0.3%) (P < .0001, Fisher exact). Though low for all populations, indeterminate results were slightly more common in children aged 2-5 years (1.3%) compared with 0.6% in 6- to 12-year-old children and 0.4% in those ≥13 years of age (P < .001, Fisher exact) and slightly higher in CHS (1.0%) vs the refugee (0.4%) and TB clinics (0.5%) (P = .03, Fisher exact). Children tested in the CHS clinics tended to be younger than in the Refugee and TB clinics (P < .001, Kruskal-Wallis), with 40% tested in CHS being aged 2-5 years. Figure 2. Rates of positive and indeterminate QuantiFERONTB Gold results by age, Denver Health, 2011-2014. n = 6683 tests.
Discussion In screening more than 3700 children for LTBI using an IGRA in an urban primary care population, we observed an overall low-rate of QFT positivity, particularly in the 2- to 5-year age group, with rates increasing with patient age through adolescence. The primary risk factor associated with QFT positivity in this population was birth or residence in/travel to a TB-endemic country, a finding that supports the current recommendations from the American Academy of Pediatrics that screening children for LTBI or TB infection should focus on these children.10 Risk factors for children without travel including foreign-born parents and exposure to US-based highrisk populations (such as incarcerated persons) were not associated with increased odds of positive QFT. Higher rates
of positive testing were observed in the refugee clinic (7.9%) and TB clinic (21.2%). The observation that rates of positive testing increase with increasing putative prevalence of LTBI (based on higher rates of TB exposure in these groups) adds to our confidence that individuals in the primary care population with negative testing are truly uninfected. This correlation between rates of IGRA positivity and increasing exposure risk is consistent with other studies in low TB burden, otherwise healthy pediatric populations.12-14 Furthermore, inclusion of the refugee and TB clinic populations also allows a
Interferon-Gamma Release Assay-Based Screening for Pediatric Latent Tuberculosis Infection in an Urban Primary Care Network FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018
5
THE JOURNAL OF PEDIATRICS • www.jpeds.com sample size large enough to clearly demonstrate very low levels of indeterminate results across all age groups, which should be reassuring to primary care practitioners. Use of the QFT takes advantage of the increased specificity of the assay over TST to better define the epidemiology of LTBI in our population and devise more targeted screening strategies.4,5 Previously published results in the largest Denver Health CHS pediatric clinic demonstrated much higher rates of LTBI diagnosed by TST (12% of children tested in 20022003 and 6% in 2007-2008) than the 2.1% rate in 2011-2014 diagnosed by QFT.15,16 Although a significant proportion of this decline may reflect changing demographics in our system (a decrease in newly immigrated children) and the fact that more high-risk children are captured in screening over time, the decreasing rate likely also represents elimination of false positives, particularly among children who have been BCG vaccinated. Thus, use of QFT for screening has likely improved resource utilization related to fewer unnecessary chest radiographs and treatment courses for LTBI in our system.17 In the case control analysis, we identified children who were born, lived in, or traveled to TB- endemic countries as the most significant risk factor for LTBI in our population. This observation is consistent with current adult TB epidemiology in the US, in which the vast majority of cases occur in foreign-born individuals, though the majority of pediatric cases of active TB still occur among US-born individuals.18,19 It is also notable that the rate of positivity among Hispanic patients in our population is less than 2%. Nationally, Hispanic children still comprise the majority of pediatric TB cases, but the number in Hispanic children dropped approximately 85% between 1993 and 2015, and incidence rates of TB in Hispanic children have declined to below those of Asian and American Indian/ Alaska Native children.19 Thus, consideration of specific risk factors beyond Hispanic origin in the development of primary care LTBI screening strategies may improve resource utilization. We are unable to determine the significance of more specific foreign travel/habitation risk factors (how much time spent, what conditions, TB incidence in travel location) among USborn patients. The risk of TB infection will differ between a child who spends a week in Mexico City (TB incidence ~5/ 100 000) and another who spends a 3-month summer holiday with family outside Addis Ababa (TB incidence ~350/100 000), yet both would meet criteria for screening if “travel to a TBendemic country” is the standard criteria. Current guidance for screening in our institution specifies that only extended travel (>4 weeks) to a TB-endemic country warrants TB screening upon return. Future data incorporating details of specific travel location and time spent may further increase screening efficiency. The increase in QFT positivity with age in this primary care population is relevant to designing optimal screening strategies. Increasing rates of TB infection with age relate most directly to increased time of potential exposure.5 Though TB screening in children will have higher yields as patients age through adolescence, lower rates of positivity at young ages must be considered in the context of the increased rates of progression from LTBI to active disease in younger children.20 Thus,
Volume ■■ a uniform set of screening criteria across age groups may not yield the most efficient screening process. In current American Thoracic Society (ATS)/Infectious Disease Society of America/Centers for Disease Control and American Academy of Pediatrics guidelines, the TST remains the preferred test for children <5 years of age because of the lack of data on IGRA performance in this age group (though the ATS guideline is a “conditional recommendation” for children <5 years of age based on “very low-quality evidence”), while IGRAs are now preferred in the 2017 ATS/Infectious Disease Society of America/Centers for Disease Control guideline for children ≥5 years of age.7,10,11 Our data add to the understanding of the performance of IGRAs in young children and should contribute to future consideration of more widespread IGRA use in screening programs for younger children, particularly as additional supportive evidence of test performance becomes available. One challenge for QFTbased screening in younger children is the requirement for a venous blood sample, but in our clinics we were able to complete 1721 successful tests in children <5 years of age over the 3.5-year study period. Another barrier to more widespread utilization of QFT-based screening is the potential for higher rates of indeterminate QFT results in younger children. In hospitalized children, rates of indeterminate tests as high as 35% have been observed, with higher rates correlated with young age.21 Other studies report rates of indeterminate tests in children between 3% and 7%.8,22-24 In our study, the rate of indeterminate tests among all children age 2-5 years was 1.3%, and 0.5% in all ages combined, consistent with very low rates reported in LTBI testing in otherwise healthy children and adults.5,25 Indeterminate QFT testing may be the result of either intrinsic patient factors (eg, insufficient reaction in positive control tube) or technical factors in specimen processing, including a specific requirement for tube agitation immediately after phlebotomy. Our results suggest that this is not likely to be a limitation of QFT use in children age 2-5 years in a primary care population when utilizing an experienced laboratory to collect and process specimens.26 An additional area of concern using QFT for LTBI screening in younger children is the possibility of false negative tests resulting in some cases of LTBI being missed in IGRA-based screening programs.27 Data assessing the sensitivity of QFT for LTBI screening in otherwise healthy, presumably immune normal children from populations with low to moderate TB risk have been difficult to obtain given the lack of a goldstandard for LTBI diagnosis, and the combination of low overall rates of TB infection and low risk of progression in those that are infected. Furthermore, any assessment of QFT sensitivity for LTBI in young children must be compared with a similar assessment of TST performance, which suggest that both tests may have similarly imperfect sensitivity.28-30 Our data do not provide a quantifiable measure of sensitivity, and we cannot definitively state that the screening program has not missed cases that would have been identified by TST, but the absence of any incident cases as of February 2018 among our screened primary care population does add incrementally to our confidence that the assay has adequate sensitivity. More importantly,
6
Gaensbauer et al FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018
■■ 2018
ORIGINAL ARTICLES
the observed correlation between higher putative exposure risk and higher QFT positive rates in young children in the refugee and TB clinic populations demonstrates a gradient response, which has been used as indirect evidence of IGRA sensitivity in multiple studies of pediatric LTBI.31 The large number of patients in this project is a major strength, allowing us to identify risk factors for positive tests and demonstrate low rates of indeterminate results. However, there are important limitations to acknowledge in the use of a retrospective database to identify and characterize children who underwent TB screening. Because the larger database did not contain detailed information on TB risk factors including details of “travel/residence in TB-endemic countries,” we were unable describe the reasons for screening in the population not included in the case-control analysis. Even in the case-control population, these details were not well documented in the medical record, resulting in the need to define risk factors based on more general details. BCG status is also not typically captured in our database or medical records. Our primary care clinics have strong laboratory support with extensive pediatric phlebotomy experience and familiarity with the specific handling requirements for QFT; clinicians using external labs or those less familiar with IGRA performance may encounter higher rates of failed phlebotomy or indeterminate results. This project describes an IGRA-based screening program in a large safety-net pediatric primary care network for children and supports the use of QFT as a screening tool in children, with the inclusion of children age 2-5 years. Low rates of indeterminate results, and correlation between increasing putative exposure and rates of positive testing should add to confidence in the performance of QFT in this setting, though our findings do not provide a definitive demonstration of QFT sensitivity. Increased specificity of the assay for mycobacterium TB compared with TSTs, which is more clearly established in the literature, should allow for better resource utilization through fewer unnecessary radiographic evaluations and LTBI treatment courses, which is of major importance to safety net health systems such as Denver Health. Future studies to determine the optimal timing and frequency of repeat testing throughout childhood, identify better methods of capturing and incorporating details of travel time and location into screening procedures, and assess the benefit of any screening in children whose only risk factor is foreign family origin, are needed to help further refine screening guidelines for pediatric primary care settings. ■ Submitted for publication Dec 22, 2017; last revision received Apr 10, 2018; accepted Apr 17, 2018 Reprint requests: James Gaensbauer, MD, MScPH, Pediatrics, Denver Health and Hospital Authority, 777 Bannock St, Pavilion C, MC0590, Denver, CO 80204. E-mail: [email protected]
Data Statement Data will be made available on request.
References 1. Newton SM, Brent AJ, Anderson S, Whittaker E, Kampmann B. Paediatric tuberculosis. Lancet Infect Dis 2008;8:498-510. 2. American Academy of Pediatrics, Pediatric Tuberculosis Collaborative Group. Targeted tuberculin skin testing and treatment of latent tuberculosis infection in children and adolescents. Pediatrics 2004; 114(Suppl 4). 3. Bergamini BM, Losi M, Vaienti F, D’Amico R, Meccugni B, Meacci M, et al. Performance of commercial blood tests for the diagnosis of latent tuberculosis infection in children and adolescents. Pediatrics 2009;123:e41924. 4. Lighter J, Rigaud M, Eduardo R, Peng CH, Pollack H. Latent tuberculosis diagnosis in children by using the QuantiFERON-TB Gold In-Tube test. Pediatrics 2009;123:30-7. 5. Howley MM, Painter JA, Katz DJ, Graviss EA, Reves R, Beavers SF, et al. Evaluation of QuantiFERON-TB gold in-tube and tuberculin skin tests among immigrant children being screened for latent tuberculosis infection. Pediatr Infect Dis J 2015;34:35-9. 6. Starke JR. Interferon-gamma release assays for diagnosis of tuberculosis infection in children. Pediatr Infect Dis J 2006;25:9412. 7. Mazurek GH, Jereb J, Vernon A, LoBue P, Goldberg S, Castro K. Updated guidelines for using Interferon Gamma Release Assays to detect Mycobacterium tuberculosis infection—United States, 2010. MMWR Recomm Rep 2010;59:1-25. 8. Grinsdale JA, Islam S, Tran OC, Ho CS, Kawamura LM, Higashi JM. Interferon-gamma release assays and pediatric public health tuberculosis screening: the San Francisco Program Experience 2005 to 2008. J Pediatric Infect Dis Soc 2016;5:122-30. 9. Pang J, Teeter LD, Katz DJ, Davidow AL, Miranda W, Wall K, et al. Epidemiology of tuberculosis in young children in the United States. Pediatrics 2014;133:e494-504. 10. American Academy of Pediatrics, Committee on Infectious Diseases. Tuberculosis. In: Kimberlin DW, Brady MT, Jackson MA, Long SS, eds. Red Book: 2015 report of the Committee on Infectious Diseases. 30th ed. Elk Grove Village (IL): American Academy of Pediatrics; 2015. p. 80515. 11. Lewinsohn DM, Leonard MK, LoBue PA, Cohn DL, Daley CL, Desmond E, et al. Official American Thoracic Society/Infectious Diseases Society of America/Centers for Disease Control and Prevention Clinical Practice Guidelines: diagnosis of tuberculosis in adults and children. Clin Infect Dis 2017;64:111-5. 12. Ewer K, Deeks J, Alvarez L, Bryant G, Waller S, Andersen P, et al. Comparison of T-cell-based assay with tuberculin skin test for diagnosis of Mycobacterium tuberculosis infection in a school tuberculosis outbreak. Lancet 2003;361:1168-73. 13. Soysal A, Millington KA, Bakir M, Dosanjh D, Aslan Y, Deeks JJ, et al. Effect of BCG vaccination on risk of Mycobacterium tuberculosis infection in children with household tuberculosis contact: a prospective community-based study. Lancet 2005;366:1443-51. 14. Sali M, Buonsenso D, Goletti DD, Alfonso P, Zumbo A, Fadda G, et al. Accuracy of QuantiFERON-TB gold test for tuberculosis diagnosis in children. PLoS ONE 2015;10:e0138952. 15. Young J, O’Connor ME. Risk factors associated with latent tuberculosis infection in Mexican American children. Pediatrics 2005;115:e64753. 16. Young J, Edick T, Klee D, O’Connor ME. Successful treatment of pediatric latent tuberculosis infection in a community health center clinic. Pediatr Infect Dis J 2012;31:e147-51. 17. Nienhaus A, Schablon A, Costa JT, Diel R. Systematic review of cost and cost-effectiveness of different TB-screening strategies. BMC Health Serv Res 2011;11:247. 18. Centers for Disease Control and Prevention. Trends in tuberculosis, 2015. https://www.cdc.gov/tb/publications/factsheets/statistics/tbtrends.htm. Accessed October 23, 2017. 19. Centers for Disease Control and Prevention. Slide Set— Epidemiology of Pediatric Tuberculosis in the United States, 1993-2013.
Interferon-Gamma Release Assay-Based Screening for Pediatric Latent Tuberculosis Infection in an Urban Primary Care Network FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018
7
THE JOURNAL OF PEDIATRICS • www.jpeds.com
20. 21.
22.
23.
24.
https://www.cdc.gov/tb/publications/slidesets/pediatrictb/default.htm. Accessed October 23, 2017. Cruz AT, Starke JR. Clinical manifestations of tuberculosis in children. Paediatr Respir Rev 2007;8:107-17. Haustein T, Ridout DA, Hartley JC, Thaker U, Shingadia D, Klein NJ, et al. The likelihood of an indeterminate test result from a whole-blood interferon-gamma release assay for the diagnosis of Mycobacterium tuberculosis infection in children correlates with age and immune status. Pediatr Infect Dis J 2009;28:669-73. Thomas B, Pugalenthi A, Patel H, Woltmann G, Bankart J, Hoskyns W. Concordance between tuberculin skin test and interferon-gamma assay and interferon-gamma response to mitogen in pediatric tuberculosis contacts. Pediatr Pulmonol 2011;46:1225-32. Critselis E, Amanatidou V, Syridou G, Spyridis NP, Mavrikou M, Papadopoulos NG, et al. The effect of age on whole blood interferongamma release assay response among children investigated for latent tuberculosis infection. J Pediatr 2012;161:632-8. Blandinieres A, de Lauzanne A, Guerin-El Khourouj V, Gourgouillon N, See H, Pedron B, et al. QuantiFERON to diagnose infection by Mycobacterium tuberculosis: performance in infants and older children. J Infect 2013;67:391-8.
Volume ■■ 25. Pavic I, Topic RZ, Raos M, Aberle N, Dodig S. Interferon-gamma release assay for the diagnosis of latent tuberculosis in children younger than 5 years of age. Pediatr Infect Dis J 2011;30:866-70. 26. Bui DH, Cruz AT, Graviss EA. Indeterminate QuantiFERON-TB gold in-tube assay results in children: possible association with procedural specimen collection. Pediatr Infect Dis J 2014;33:220-2. 27. Mandalakas AM, Detjen AK, Hesseling AC, Benedetti A, Menzies D. Interferon-gamma release assays and childhood tuberculosis: systematic review and meta-analysis. Int J Tuberc Lung Dis 2011;15:101832. 28. Nakaoka H, Lawson L, Squire SB, Coulter B, Ravn P, Brock I, et al. Risk for tuberculosis among children. Emerg Infect Dis 2006;12:1383-8. 29. Tavast E, Salo E, Seppala I, Tuuminen T. IGRA tests perform similarly to TST but cause no adverse reactions: pediatric experience in Finland. BMC Res Notes 2009;2:9. 30. Connell TG, Ritz N, Paxton GA, Buttery JP, Curtis N, Ranganathan SC. A three-way comparison of tuberculin skin testing, QuantiFERON-TB gold and T-SPOT.TB in children. PLoS ONE 2008;3:e2624. 31. Trajman A, Steffen RE, Menzies D. Interferon-gamma release assays versus tuberculin skin testing for the diagnosis of latent tuberculosis infection: an overview of the evidence. Pulm Med 2013;2013:601737.
8
Gaensbauer et al FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018
ORIGINAL ARTICLES
Interferon-Gamma Release Assay-Based Screening for Pediatric Latent Tuberculosis Infection in an Urban Primary Care Network James Gaensbauer, MD, MScPH1,2,3,4, Bryn Gonzales, MD1, Robert Belknap, MD2,5, Michael L. Wilson, MD6,7, and Mary E. O'Connor, MD, MPH8,9,10 Objective To assess outcomes from a QuantiFERON-tuberculosis (TB) Gold (QFT)-based screening for pediatric latent TB infection (LTBI) in the Denver Health Community Health System (CHS), an urban primary-care network in the US. Study design We retrospectively analyzed all QFTs (n = 6685) performed on children aged 2-18 years between January 5, 2011, and August 18, 2014. Risk factors for positive testing in the CHS population were identified by logistic regression, and further assessed using a case-control comparison. Results from CHS were compared with higher-TB-risk populations (refugee and TB clinics) in our health system. Results Positive QFT occurred in 79 of 3745 (2.1%) CHS patients. Positive rates increased with age (0.3% in age 2-5 years to 4.9% in age 13-18 years). Indeterminate results were uncommon (0.8%) including in children <5 (1.3%). Risk factors for positive tests in the CHS population included non-Medicaid insured/uninsured and nonEnglish/Spanish preferred language. In the case-control analysis, birth/travel to/residence in a TB-endemic country was the only identified risk factor for positive testing (OR 5.2 [95% CI 1.04-25.5]). Rates of positive testing were lower in the CHS population than the refugee/TB clinic populations, including among children age 2-5. Discussion QFT-based LTBI screening was successfully introduced in our pediatric primary-care health system, and supported our programmatic goals of identifying LTBI cases while limiting unnecessary LTBI treatment courses. Increasing positive rates with age, and higher rates in the refugee/TB populations compared with CHS, add indirect evidence of adequate test sensitivity, even among young children, for whom data on interferon-gamma release assay performance are limited. (J Pediatr 2018;■■:■■-■■).
I
dentification and treatment of pediatric tuberculosis (TB) infection is an important national priority. Children are at greatest risk for progression to active disease during infancy and late adolescence, and young children have heightened morbidity and mortality because of disseminated disease.1 A subset of children with untreated latent TB infection (LTBI) will become symptomatic and infectious later in life, and, thus, treatment of children with LTBI is an essential component of long-term TB elimination programs in the US. Screening programs to identify LTBI in children play an important role in the prevention of childhood TB disease in the US, and much of this activity takes place within primary care clinics serving populations at higher risk of Mycobacterium tuberculosis infection.2 However, critical questions remain as to how to best define and identify candidates for LTBI screening in primary care based on demographic and travel-related risk factors, determine the optimal age for initial screening and timing of repeat testing, and optimize the use of available screening tools, including tuberculin skin tests (TSTs) and interferon-gamma release assays (IGRA).3 Denver Health is an urban safety-net hospital and a network of Community From the 1Department of Pediatrics, Denver Health and Hospital Authority; 2Denver Metro Tuberculosis Clinic, Health System (CHS) clinics that provide care for a large population of foreignDenver Public Health, Denver; 3Department of Pediatrics, born children and US-born children with foreign-born parents. In 2011, Section of Infectious Diseases, University of Colorado School of Medicine; 4Department of Epidemiology, QuantiFERON-TB Gold (QFT) (Qiagen, Germantown, Maryland) became the preColorado School of Public Health; 5Department of Infectious Diseases, University of Colorado School of ferred test for diagnosing LTBI in children ≥2 years old at all Denver Health amMedicine, Aurora; 6Department of Pathology and bulatory clinics. The programmatic goal of our screening program was to identify Laboratory Services, Denver Health and Hospital Authority, Denver; 7Department of Pathology, University of as many children with LTBI as possible while minimizing the burden to patients Colorado School of Medicine; 8Department of Pediatrics, University of Colorado School of Medicine, Aurora CO; and system resources from unnecessary investigation and treatment of children 9Children’s Hospital, Dartmouth-Hitchcock, Lebanon; and 10
Department of Pediatrics, Geisel School of Medicine at Dartmouth, Hanover, NH The authors declare no conflicts of interest.
ATS BCG CHS IGRA LTBI QFT TB TST
American Thoracic Society Bacille Calmette-Guérin Community Health System Interferon-gamma release assay Latent TB infection QuantiFERON-TB Gold Tuberculosis Tuberculin skin test
Portions of this study were presented as posters at the Union Against Tuberculosis (TB) and Lung Disease, North American Region/National TB Controllers Association Annual Conference, February 24-27, 2016, Denver, Colorado and the Union Against TB and Lung Disease North American Region Annual Conference, February 2225, 2017, Vancouver, British Colombia, Canada. 0022-3476/$ - see front matter. © 2018 Elsevier Inc. All rights reserved. https://doi.org10.1016/j.jpeds.2018.04.034
1 FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018
THE JOURNAL OF PEDIATRICS • www.jpeds.com with false positive testing. The rationale for an IGRA-based strategy was based on multiple potential advantages of QFT over TSTs, including increased specificity for LTBI in foreignborn, Bacille Calmette-Guérin (BCG)-vaccinated children and in children exposed to non-tuberculous mycobacteria and concern that false positive TSTs might be leading to unnecessary treatments and resource utilization.4,5 In addition, IGRAs offered the opportunity to perform screening at a single visit on any day of the week, and were able to be reported and retrieved more clearly in our medical record system.6,7 Finally, though need for phlebotomy for IGRA had been seen as a disadvantage over TSTs, particularly in young children, our clinics have expertise in pediatric phlebotomy, whereas we were concerned that the technical skills for TST placement and interpretation were waning. In the US, recent observations from an urban public healthbased TB control program demonstrated the advantages of an IGRA-based screening in a predominantly foreign-born, BCGvaccinated pediatric population.8 However, the performance of IGRAs for LTBI screening has not been clearly documented in US-born/non-BCG vaccinated children in primary care settings, in whom the principle risk factors for TB infection are contact with adults/relatives from TB-endemic countries or intermittent travel to these countries.9,10 In addition, the performance of IGRAs in younger children has yet to be definitively established, particularly in children less than 5 years of age, resulting in current recommendations that they be used with caution in this age group.2,7,11 We retrospectively reviewed the results of all children tested during the first 3.5 years after implementation of an IGRAbased LTBI screening program instead of TST testing. The primary objectives of the study were to assess performance of QFT testing from a programmatic perspective, and demonstrate the feasibility of IGRA-based TB screening in pediatric primary care. Secondary objectives included an analysis of children with positive testing to characterize risk factors for TB infection and inform ongoing development of efficient strategies for LTBI screening in our population. Finally, in the face of uncertainties related to IGRA performance, particularly in younger children, we sought to assess results in comparison with results in higher prevalence refugee and TB clinic patients in our health system.
Methods The study population comprised all pediatric patients aged 2-18 years screened for LTBI with QFT between January 5, 2011 and August 18, 2014 at Denver Health. The primary objective was to describe the performance of QFT in the CHS, comprised of 14 school-based health centers, 3 pediatric primary care clinics, and 6 family medicine clinics. In these sites, TB screening was performed as part of routine preventive care for children age 2 years and above who were born or traveled to a TB-endemic country, or whose parents were born in a TB-endemic country. Repeat testing in CHS was recommended for patients with new travel to a high-risk county or other new risk factor. In the CHS population, Mexico is the most common non-US country of
Volume ■■
•
■■ 2018
family origin (internal institutional data). To compare the results of IGRA-based screening in the CHS population with pediatric populations at higher risk of TB, we included patients from 2 Denver Health affiliated clinics: the Denver Health Refugee Clinic and the Denver Metro TB clinic. During the study period, patients in the Refugee Clinic had LTBI testing primarily as part of routine new immigrant health screening. The TB clinic is affiliated with the Denver Department of Public health and testing occurred primarily in the context of contact investigation or follow-up of external referrals for positive TSTs. Data Collection All QFTs drawn on pediatric patients during the study period were identified from a Denver Health hospital laboratory information system (Sunquest, Tucson, Arizona). Data extracted included test-order date, location, QFT result, and basic patient demographics including age, sex, insurance status, self/family-reported race/ethnicity, and primary spoken language. For all children with positive QFT’s in CHS, medical record review was performed to determine results of chest radiographs and LTBI treatment initiation. Chart review was not performed on refugee and TB clinic patients, many of whom have medical records outside the Denver Health system; available data was limited to QFT results and basic demographics. The project was approved as exempt review by the Colorado Multiple Institutional Review Board. Case-Control Methods To assess factors that might allow for a more targeted approach to LTBI testing in our population, all children in CHS with a positive QFT were matched 1:1 by age at time of testing, clinic site, and preferred language in the home (English, Spanish, or other) to children with a negative QFT. The majority of children were matched within 6 months of age, though a small number required a range of 12 months. Medical record review was performed on all cases and controls to determine the indication for LTBI testing. Indications were placed into 3 categories (limited by the granularity of data available in the medical record): born/lived or traveled to a TB-endemic country, exposure to jail/alcoholism/drug abuse/non-Englishpreferred household language, and no documented reason. This analysis was originally designed as an exploratory quality improvement project and no a priori power calculation was performed. A case:control ratio of 1:1 was selected based on available resources for chart review. Statistical Analyses All patient information was deidentified and analyzed using SPSS (IBM, Armonk, New York) v 23, using c2/Fisher exact test, Kruskal-Wallis, and logistic regression. Factors associated with positive QFT testing in the case-control analysis were assessed by multivariate logistic regression.
Results There were 6685 QFT tests performed on 6338 unique children in the 3 sites. We excluded 2 results from children who
2
Gaensbauer et al FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018
■■ 2018
ORIGINAL ARTICLES
Figure 1. Diagram of QuantiFERON-TB Gold tests and individual patients included in analysis, Denver Health, 2011-2014.
had been treated for LTBI in the past. Of the 6683 test results, 53 (0.8%) were indeterminate, 455 (6.8%) were positive, and 6175 (92.4%) were negative. Two or more tests were performed on 347 children, yielding 6336 unique children who were tested at least once (Figure 1). Of these, 34 children/36 tests had only an indeterminate result and were excluded from any analysis of patient characteristics based on result. Demographic information and QFT results by clinic type are presented in Table I. Results from children tested in CHS are described in Table II. Only 5 (0.3%) children aged 2-5 years had a positive QFT, increasing to 2.3% in children age 6-12 years and 4.9% over 12 years. Indeterminate results occurred in only 1.0% of CHS patients. In multivariate logistic regression analysis, risk factors for positive testing in CHS included increasing age (OR 3.0; 95% CI 2.1-4.2) per age group, uninsured 4.4 (95% CI 2.48.2) or non-Medicaid insured status (OR 4.0; 95% CI 1.79.3), and a preferred primary language in the home other than English or Spanish (OR 2.9; 95% CI 1.4-6.1). Case-Control Table III compares the children in CHS with a positive QFT matched to children with a negative QFT. The only variable associated with positive testing was a child who was born/ lived/traveled to a TB-endemic county with an aOR of 5.2 (95% CI 1.04-25.5) compared with those with no travel or family residence risk factors. The adjusted OR of children living in a non-English-preferred household language/exposure to drug abuse or alcoholism/jail was not statistically significantly different (OR 2.92; 95% CI 0.5-15.4).
Clinical Outcomes Of the 79 children with positive QFT, 73 had a chest radiograph completed (all normal) and 70 (89%) were initiated on LTBI treatment. Five patients were lost to follow-up after
Table I. Comparison of QFT results and demographic factors between testing sites, Denver Health, 2011-2014; n = 6302 patients* Characteristics
Number of patients QFT positive† Male sex Median age (y)‡ Age groups (y) 2-5* 6-12 13-18 Race/ethnicity Hispanic Native American Asian/Pacific Islander Black/African American White Unknown Preferred language in home English Spanish Other Unknown
Patient population CHS
Refugee clinic
TB clinic
3745 (59.4%) 79 (2.1%) 1836 (49.0%) 8
1286 (20.4%) 101 (7.9%) 649 (50.5%) 9
1271 (20.2%) 270 (21.2%) 647 (50.9%) 14
1497 (40.0%) 1292 (34.5%) 956 (25.5%)
364 (28.3%) 559 (43.5%) 363 (28.2%)
140 (11.1%) 362 (28.5%) 769 (60.5%)
3233 12 142 185 150 23
(86.3%) (0.3%) (3.8%) (4.9%) (4.0%) (0.6%)
2 0 217 121 52 894
(0.2%) (0%) (16.9%) (9.3%) (4.0 %) (69.5%)
207 1 177 108 81 697
(16.3%) (0.1%) (13.9%) (8.5%) (6.4%) (54.8%)
1240 2293 199 1
(33.4%) (61.1%) (5.3%) (0.1%)
46 6 811 423
(3.6%) (0.5%) (63.1%) (32.9%)
131 87 165 888
(10.3%) (6.8%) (13.0%) (69.8%)
*Patients with indeterminate result excluded. †P < .001, c2. ‡P < .001, Kruskal-Wallis.
Interferon-Gamma Release Assay-Based Screening for Pediatric Latent Tuberculosis Infection in an Urban Primary Care Network FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018
3
THE JOURNAL OF PEDIATRICS • www.jpeds.com
Volume ■■
Table II. Comparison of all QFT-positive and QFT-negative in patients age 2-18 years, Denver Health CHS 2011-2014; n = 3745 Characteristics Total Sex Female Male Age grouping 2-5 y 6-12 y 13-18 y Insurance Medicaid/pending State Children's Health Insurance Program Commercial Uninsured Discount § Racial/ethnic group Hispanic Black/African American Asian/Pacific Islander White Native American Language English Spanish Other
Statistical significance*
aOR†
1872 (51.1%) 1794 (48.9%)
NS
NS
5 (6.3%) 26 (32.9%) 48 (60.8%)
1492 (40.7%) 1266 (34.5%) 908 (24.8%)
P < .001
reference 3.0 (2.1-4.2)‡
39 4 7 15 14
(49.4%) (5.1%) (8.9%) (19.0%) (17.7%)
2876 299 77 123 290
(78.5%) (8.2%) (2.1%) (3.4%) (7.9%)
P < .001
reference NS 4.0 (1.7-9.3) 4.4 (2.4-8.2) NS
60 9 5 3 1
(75.9%) (11.4%) (6.3%) (3.8%) (1.3%)
3173 176 137 147 11
(86.6%) (4.8%) (3.7%) (4.0%) (0.3%)
P < .04
NS
1221 (33.3%) 2250 (61.4%) 190 (5.2%)
P < .05
Reference NS 2.9 (1.4-6.1)
QFT-positive
QFT-negative
79
3666
37 (46.8%) 42 (53.1%)
27 (34.1%) 43 (54.4%) 9 (11.4%)
NS, not significant. *c2 †Logistic regression with (95% CI). ‡aOR for each step up in age compared with 2- to 5-year-old cohort. §Community Health Center subsidized care for legally undocumented children.
screening, and 3 patients/families refused LTBI treatment. One patient was diagnosed with active TB; review of the medical record revealed that he was actually referred to a CHS provider by an ophthalmologist for evaluation of retinal disease suspicious for TB. As of February 27, 2018, none of the remaining 78 QFT-positive children have been identified with progression to active TB disease at the TBC (which
provides care or consultation to all pediatric TB cases in the state of Colorado). Repeat QFT Testing Repeat testing was performed on 346 children (Table IV). Only 1 child from CHS converted from negative to positive after living in Mexico for the summer. All other children who
Table III. Case-control comparison of QFT-positive and QFT-negative children, Denver Health CHS, 2011-2014; n = 79 each group Characteristics Sex Male Female Insurance type Uninsured Medicaid SCHIP Commercial Racial/ethnic group Hispanic Black/African Other White Risk None documented Non-English-preferred household/jail/ drug/alcohol exposure Born/traveled/lived TB-endemic country
QFT-positive
QFT-negative
P value*
aOR†
42 (58.3%) 37 (43.0%)
30 (41.7%) 49 (57%)
.06
NS
29 39 4 7
(36.7%) (49.4%) (5.0%) (9.0%)
12 54 8 5
(15.2%) (68.4%) (10.1%) (6.3%)
.01
NS
60 10 6 3
(75.9%) (12.7%) (7.6%) (3.8%)
66 6 4 3
(83.5%) (7.6%) (5.1%) (3.8%)
0.64
NS
2 (2.5%) 19 (24.0%)
8 (10.1%) 26 (32.9%)
.04
Reference 2.92 (.5-15.4)
58 (73.4%)
45 (57.0%)
5.2 (1.04-25.5)
2
*c †Multivariable logistic regression with aOR with 95% CI.
4
Gaensbauer et al FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018
■■ 2018
ORIGINAL ARTICLES
Table IV. Results of repeat QFT testing in CHS, refugee clinic, and TB clinic pediatric populations, Denver Health, 2011-2014 Initial results Indeterminate
Subsequent results
Number of patients
Negative Indeterminate
17 1
Positive
Positive
5
Negative
Positive
12
Positive
Negative
2
Negative
Negative
309
Comments Indeterminate value returned twice Confirmatory testing of first test in new setting 1 in CHS after prolonged travel to Mexico; remaining 11 in TB clinic patients Retests at 1 and 9 mo after initial result
converted from negative to positive QFT were in the TB clinic cohort and converted in the context of exposure to a case of active TB. Eighteen of the 36 children with an initial indeterminate result underwent retesting, of which 17 (94%) were negative and 1 (6%) remained indeterminate. Comparison of CHS Results with Refugee and TB Clinics The percentage of positive test results was highest in the TB clinic (21.2%), followed by the refugee (7.9%), and CHS clinics (2.1%) (P < .001, c2). Test positivity increased with age in each site (Figure 2). Among children age 2-5 years, higher rates of positive testing were observed in the TB (5.2%) and refugee clinics (4.9%) than among the CHS population (0.3%) (P < .0001, Fisher exact). Though low for all populations, indeterminate results were slightly more common in children aged 2-5 years (1.3%) compared with 0.6% in 6- to 12-year-old children and 0.4% in those ≥13 years of age (P < .001, Fisher exact) and slightly higher in CHS (1.0%) vs the refugee (0.4%) and TB clinics (0.5%) (P = .03, Fisher exact). Children tested in the CHS clinics tended to be younger than in the Refugee and TB clinics (P < .001, Kruskal-Wallis), with 40% tested in CHS being aged 2-5 years. Figure 2. Rates of positive and indeterminate QuantiFERONTB Gold results by age, Denver Health, 2011-2014. n = 6683 tests.
Discussion In screening more than 3700 children for LTBI using an IGRA in an urban primary care population, we observed an overall low-rate of QFT positivity, particularly in the 2- to 5-year age group, with rates increasing with patient age through adolescence. The primary risk factor associated with QFT positivity in this population was birth or residence in/travel to a TB-endemic country, a finding that supports the current recommendations from the American Academy of Pediatrics that screening children for LTBI or TB infection should focus on these children.10 Risk factors for children without travel including foreign-born parents and exposure to US-based highrisk populations (such as incarcerated persons) were not associated with increased odds of positive QFT. Higher rates
of positive testing were observed in the refugee clinic (7.9%) and TB clinic (21.2%). The observation that rates of positive testing increase with increasing putative prevalence of LTBI (based on higher rates of TB exposure in these groups) adds to our confidence that individuals in the primary care population with negative testing are truly uninfected. This correlation between rates of IGRA positivity and increasing exposure risk is consistent with other studies in low TB burden, otherwise healthy pediatric populations.12-14 Furthermore, inclusion of the refugee and TB clinic populations also allows a
Interferon-Gamma Release Assay-Based Screening for Pediatric Latent Tuberculosis Infection in an Urban Primary Care Network FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018
5
THE JOURNAL OF PEDIATRICS • www.jpeds.com sample size large enough to clearly demonstrate very low levels of indeterminate results across all age groups, which should be reassuring to primary care practitioners. Use of the QFT takes advantage of the increased specificity of the assay over TST to better define the epidemiology of LTBI in our population and devise more targeted screening strategies.4,5 Previously published results in the largest Denver Health CHS pediatric clinic demonstrated much higher rates of LTBI diagnosed by TST (12% of children tested in 20022003 and 6% in 2007-2008) than the 2.1% rate in 2011-2014 diagnosed by QFT.15,16 Although a significant proportion of this decline may reflect changing demographics in our system (a decrease in newly immigrated children) and the fact that more high-risk children are captured in screening over time, the decreasing rate likely also represents elimination of false positives, particularly among children who have been BCG vaccinated. Thus, use of QFT for screening has likely improved resource utilization related to fewer unnecessary chest radiographs and treatment courses for LTBI in our system.17 In the case control analysis, we identified children who were born, lived in, or traveled to TB- endemic countries as the most significant risk factor for LTBI in our population. This observation is consistent with current adult TB epidemiology in the US, in which the vast majority of cases occur in foreign-born individuals, though the majority of pediatric cases of active TB still occur among US-born individuals.18,19 It is also notable that the rate of positivity among Hispanic patients in our population is less than 2%. Nationally, Hispanic children still comprise the majority of pediatric TB cases, but the number in Hispanic children dropped approximately 85% between 1993 and 2015, and incidence rates of TB in Hispanic children have declined to below those of Asian and American Indian/ Alaska Native children.19 Thus, consideration of specific risk factors beyond Hispanic origin in the development of primary care LTBI screening strategies may improve resource utilization. We are unable to determine the significance of more specific foreign travel/habitation risk factors (how much time spent, what conditions, TB incidence in travel location) among USborn patients. The risk of TB infection will differ between a child who spends a week in Mexico City (TB incidence ~5/ 100 000) and another who spends a 3-month summer holiday with family outside Addis Ababa (TB incidence ~350/100 000), yet both would meet criteria for screening if “travel to a TBendemic country” is the standard criteria. Current guidance for screening in our institution specifies that only extended travel (>4 weeks) to a TB-endemic country warrants TB screening upon return. Future data incorporating details of specific travel location and time spent may further increase screening efficiency. The increase in QFT positivity with age in this primary care population is relevant to designing optimal screening strategies. Increasing rates of TB infection with age relate most directly to increased time of potential exposure.5 Though TB screening in children will have higher yields as patients age through adolescence, lower rates of positivity at young ages must be considered in the context of the increased rates of progression from LTBI to active disease in younger children.20 Thus,
Volume ■■ a uniform set of screening criteria across age groups may not yield the most efficient screening process. In current American Thoracic Society (ATS)/Infectious Disease Society of America/Centers for Disease Control and American Academy of Pediatrics guidelines, the TST remains the preferred test for children <5 years of age because of the lack of data on IGRA performance in this age group (though the ATS guideline is a “conditional recommendation” for children <5 years of age based on “very low-quality evidence”), while IGRAs are now preferred in the 2017 ATS/Infectious Disease Society of America/Centers for Disease Control guideline for children ≥5 years of age.7,10,11 Our data add to the understanding of the performance of IGRAs in young children and should contribute to future consideration of more widespread IGRA use in screening programs for younger children, particularly as additional supportive evidence of test performance becomes available. One challenge for QFTbased screening in younger children is the requirement for a venous blood sample, but in our clinics we were able to complete 1721 successful tests in children <5 years of age over the 3.5-year study period. Another barrier to more widespread utilization of QFT-based screening is the potential for higher rates of indeterminate QFT results in younger children. In hospitalized children, rates of indeterminate tests as high as 35% have been observed, with higher rates correlated with young age.21 Other studies report rates of indeterminate tests in children between 3% and 7%.8,22-24 In our study, the rate of indeterminate tests among all children age 2-5 years was 1.3%, and 0.5% in all ages combined, consistent with very low rates reported in LTBI testing in otherwise healthy children and adults.5,25 Indeterminate QFT testing may be the result of either intrinsic patient factors (eg, insufficient reaction in positive control tube) or technical factors in specimen processing, including a specific requirement for tube agitation immediately after phlebotomy. Our results suggest that this is not likely to be a limitation of QFT use in children age 2-5 years in a primary care population when utilizing an experienced laboratory to collect and process specimens.26 An additional area of concern using QFT for LTBI screening in younger children is the possibility of false negative tests resulting in some cases of LTBI being missed in IGRA-based screening programs.27 Data assessing the sensitivity of QFT for LTBI screening in otherwise healthy, presumably immune normal children from populations with low to moderate TB risk have been difficult to obtain given the lack of a goldstandard for LTBI diagnosis, and the combination of low overall rates of TB infection and low risk of progression in those that are infected. Furthermore, any assessment of QFT sensitivity for LTBI in young children must be compared with a similar assessment of TST performance, which suggest that both tests may have similarly imperfect sensitivity.28-30 Our data do not provide a quantifiable measure of sensitivity, and we cannot definitively state that the screening program has not missed cases that would have been identified by TST, but the absence of any incident cases as of February 2018 among our screened primary care population does add incrementally to our confidence that the assay has adequate sensitivity. More importantly,
6
Gaensbauer et al FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018
■■ 2018
ORIGINAL ARTICLES
the observed correlation between higher putative exposure risk and higher QFT positive rates in young children in the refugee and TB clinic populations demonstrates a gradient response, which has been used as indirect evidence of IGRA sensitivity in multiple studies of pediatric LTBI.31 The large number of patients in this project is a major strength, allowing us to identify risk factors for positive tests and demonstrate low rates of indeterminate results. However, there are important limitations to acknowledge in the use of a retrospective database to identify and characterize children who underwent TB screening. Because the larger database did not contain detailed information on TB risk factors including details of “travel/residence in TB-endemic countries,” we were unable describe the reasons for screening in the population not included in the case-control analysis. Even in the case-control population, these details were not well documented in the medical record, resulting in the need to define risk factors based on more general details. BCG status is also not typically captured in our database or medical records. Our primary care clinics have strong laboratory support with extensive pediatric phlebotomy experience and familiarity with the specific handling requirements for QFT; clinicians using external labs or those less familiar with IGRA performance may encounter higher rates of failed phlebotomy or indeterminate results. This project describes an IGRA-based screening program in a large safety-net pediatric primary care network for children and supports the use of QFT as a screening tool in children, with the inclusion of children age 2-5 years. Low rates of indeterminate results, and correlation between increasing putative exposure and rates of positive testing should add to confidence in the performance of QFT in this setting, though our findings do not provide a definitive demonstration of QFT sensitivity. Increased specificity of the assay for mycobacterium TB compared with TSTs, which is more clearly established in the literature, should allow for better resource utilization through fewer unnecessary radiographic evaluations and LTBI treatment courses, which is of major importance to safety net health systems such as Denver Health. Future studies to determine the optimal timing and frequency of repeat testing throughout childhood, identify better methods of capturing and incorporating details of travel time and location into screening procedures, and assess the benefit of any screening in children whose only risk factor is foreign family origin, are needed to help further refine screening guidelines for pediatric primary care settings. ■ Submitted for publication Dec 22, 2017; last revision received Apr 10, 2018; accepted Apr 17, 2018 Reprint requests: James Gaensbauer, MD, MScPH, Pediatrics, Denver Health and Hospital Authority, 777 Bannock St, Pavilion C, MC0590, Denver, CO 80204. E-mail: [email protected]
Data Statement Data will be made available on request.
References 1. Newton SM, Brent AJ, Anderson S, Whittaker E, Kampmann B. Paediatric tuberculosis. Lancet Infect Dis 2008;8:498-510. 2. American Academy of Pediatrics, Pediatric Tuberculosis Collaborative Group. Targeted tuberculin skin testing and treatment of latent tuberculosis infection in children and adolescents. Pediatrics 2004; 114(Suppl 4). 3. Bergamini BM, Losi M, Vaienti F, D’Amico R, Meccugni B, Meacci M, et al. Performance of commercial blood tests for the diagnosis of latent tuberculosis infection in children and adolescents. Pediatrics 2009;123:e41924. 4. Lighter J, Rigaud M, Eduardo R, Peng CH, Pollack H. Latent tuberculosis diagnosis in children by using the QuantiFERON-TB Gold In-Tube test. Pediatrics 2009;123:30-7. 5. Howley MM, Painter JA, Katz DJ, Graviss EA, Reves R, Beavers SF, et al. Evaluation of QuantiFERON-TB gold in-tube and tuberculin skin tests among immigrant children being screened for latent tuberculosis infection. Pediatr Infect Dis J 2015;34:35-9. 6. Starke JR. Interferon-gamma release assays for diagnosis of tuberculosis infection in children. Pediatr Infect Dis J 2006;25:9412. 7. Mazurek GH, Jereb J, Vernon A, LoBue P, Goldberg S, Castro K. Updated guidelines for using Interferon Gamma Release Assays to detect Mycobacterium tuberculosis infection—United States, 2010. MMWR Recomm Rep 2010;59:1-25. 8. Grinsdale JA, Islam S, Tran OC, Ho CS, Kawamura LM, Higashi JM. Interferon-gamma release assays and pediatric public health tuberculosis screening: the San Francisco Program Experience 2005 to 2008. J Pediatric Infect Dis Soc 2016;5:122-30. 9. Pang J, Teeter LD, Katz DJ, Davidow AL, Miranda W, Wall K, et al. Epidemiology of tuberculosis in young children in the United States. Pediatrics 2014;133:e494-504. 10. American Academy of Pediatrics, Committee on Infectious Diseases. Tuberculosis. In: Kimberlin DW, Brady MT, Jackson MA, Long SS, eds. Red Book: 2015 report of the Committee on Infectious Diseases. 30th ed. Elk Grove Village (IL): American Academy of Pediatrics; 2015. p. 80515. 11. Lewinsohn DM, Leonard MK, LoBue PA, Cohn DL, Daley CL, Desmond E, et al. Official American Thoracic Society/Infectious Diseases Society of America/Centers for Disease Control and Prevention Clinical Practice Guidelines: diagnosis of tuberculosis in adults and children. Clin Infect Dis 2017;64:111-5. 12. Ewer K, Deeks J, Alvarez L, Bryant G, Waller S, Andersen P, et al. Comparison of T-cell-based assay with tuberculin skin test for diagnosis of Mycobacterium tuberculosis infection in a school tuberculosis outbreak. Lancet 2003;361:1168-73. 13. Soysal A, Millington KA, Bakir M, Dosanjh D, Aslan Y, Deeks JJ, et al. Effect of BCG vaccination on risk of Mycobacterium tuberculosis infection in children with household tuberculosis contact: a prospective community-based study. Lancet 2005;366:1443-51. 14. Sali M, Buonsenso D, Goletti DD, Alfonso P, Zumbo A, Fadda G, et al. Accuracy of QuantiFERON-TB gold test for tuberculosis diagnosis in children. PLoS ONE 2015;10:e0138952. 15. Young J, O’Connor ME. Risk factors associated with latent tuberculosis infection in Mexican American children. Pediatrics 2005;115:e64753. 16. Young J, Edick T, Klee D, O’Connor ME. Successful treatment of pediatric latent tuberculosis infection in a community health center clinic. Pediatr Infect Dis J 2012;31:e147-51. 17. Nienhaus A, Schablon A, Costa JT, Diel R. Systematic review of cost and cost-effectiveness of different TB-screening strategies. BMC Health Serv Res 2011;11:247. 18. Centers for Disease Control and Prevention. Trends in tuberculosis, 2015. https://www.cdc.gov/tb/publications/factsheets/statistics/tbtrends.htm. Accessed October 23, 2017. 19. Centers for Disease Control and Prevention. Slide Set— Epidemiology of Pediatric Tuberculosis in the United States, 1993-2013.
Interferon-Gamma Release Assay-Based Screening for Pediatric Latent Tuberculosis Infection in an Urban Primary Care Network FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018
7
THE JOURNAL OF PEDIATRICS • www.jpeds.com
20. 21.
22.
23.
24.
https://www.cdc.gov/tb/publications/slidesets/pediatrictb/default.htm. Accessed October 23, 2017. Cruz AT, Starke JR. Clinical manifestations of tuberculosis in children. Paediatr Respir Rev 2007;8:107-17. Haustein T, Ridout DA, Hartley JC, Thaker U, Shingadia D, Klein NJ, et al. The likelihood of an indeterminate test result from a whole-blood interferon-gamma release assay for the diagnosis of Mycobacterium tuberculosis infection in children correlates with age and immune status. Pediatr Infect Dis J 2009;28:669-73. Thomas B, Pugalenthi A, Patel H, Woltmann G, Bankart J, Hoskyns W. Concordance between tuberculin skin test and interferon-gamma assay and interferon-gamma response to mitogen in pediatric tuberculosis contacts. Pediatr Pulmonol 2011;46:1225-32. Critselis E, Amanatidou V, Syridou G, Spyridis NP, Mavrikou M, Papadopoulos NG, et al. The effect of age on whole blood interferongamma release assay response among children investigated for latent tuberculosis infection. J Pediatr 2012;161:632-8. Blandinieres A, de Lauzanne A, Guerin-El Khourouj V, Gourgouillon N, See H, Pedron B, et al. QuantiFERON to diagnose infection by Mycobacterium tuberculosis: performance in infants and older children. J Infect 2013;67:391-8.
Volume ■■ 25. Pavic I, Topic RZ, Raos M, Aberle N, Dodig S. Interferon-gamma release assay for the diagnosis of latent tuberculosis in children younger than 5 years of age. Pediatr Infect Dis J 2011;30:866-70. 26. Bui DH, Cruz AT, Graviss EA. Indeterminate QuantiFERON-TB gold in-tube assay results in children: possible association with procedural specimen collection. Pediatr Infect Dis J 2014;33:220-2. 27. Mandalakas AM, Detjen AK, Hesseling AC, Benedetti A, Menzies D. Interferon-gamma release assays and childhood tuberculosis: systematic review and meta-analysis. Int J Tuberc Lung Dis 2011;15:101832. 28. Nakaoka H, Lawson L, Squire SB, Coulter B, Ravn P, Brock I, et al. Risk for tuberculosis among children. Emerg Infect Dis 2006;12:1383-8. 29. Tavast E, Salo E, Seppala I, Tuuminen T. IGRA tests perform similarly to TST but cause no adverse reactions: pediatric experience in Finland. BMC Res Notes 2009;2:9. 30. Connell TG, Ritz N, Paxton GA, Buttery JP, Curtis N, Ranganathan SC. A three-way comparison of tuberculin skin testing, QuantiFERON-TB gold and T-SPOT.TB in children. PLoS ONE 2008;3:e2624. 31. Trajman A, Steffen RE, Menzies D. Interferon-gamma release assays versus tuberculin skin testing for the diagnosis of latent tuberculosis infection: an overview of the evidence. Pulm Med 2013;2013:601737.
8
Gaensbauer et al FLA 5.5.0 DTD ■ YMPD9955_proof ■ June 1, 2018