The Role of Youths in Within-Household Tuberculosis Transmission: A Household Contact Cohort Study

Abstract

Background

Youth aged 15–24 years are significantly impacted by tuberculosis globally. Their expanding social networks heighten exposure risks, potentially amplifying tuberculosis transmission, but their role in within-household transmission remains poorly understood.

Methods

A household contact cohort study in Lima, Peru (2009–2012) enrolled individuals (>15 years) with incident tuberculosis (index patients) and their household contacts (HHCs), following them for 12 months. We evaluated whether the age of index patients and HHCs modified the risk of tuberculosis infection among HHCs. Study outcomes included: (1) prevalent tuberculosis infection among child contacts (<15 years) at enrollment and (2) incident tuberculosis infection among HHCs. Whole-genome sequencing of Mycobacterium tuberculosis (Mtb) isolates from index-HHC pairs assessed whether index patient age modified the likelihood of within-household transmission, defined as ≤12 single-nucleotide polymorphism differences.

Results

Child contacts of youth index patients had a lower risk of prevalent tuberculosis infection than those with adult index patients (aRR = 0.77; 95% CI: .67–.87). Among index-secondary patient pairs, 62% (26/42) of pairs with a youth index patient were genetically linked, compared to 72% (48/67) of pairs with an adult index patient (P = .3). Child and youth contacts had lower incident tuberculosis infection risk at 12-month follow-up compared to adult contacts (child vs adult: HR = 0.33; 95% CI: .28–.38; youth vs adult: HR = 0.61; 95% CI: .52–.71).

Conclusions

Youth play a limited role in household tuberculosis transmission. Compared to adults, youth transmitted less TB as index patients and were less likely to be infected as HHCs. Research should explore youth social interactions and community-based transmission to inform targeted prevention strategies.

Compared to adults, youth were less likely to transmit tuberculosis as index patients and be infected as household contacts. These findings suggest that youth play a smaller role in household-based TB transmission than other age groups.

Tuberculosis (TB) significantly impacts youth, with an estimated 1.6 million individuals aged 15–24 falling ill in 2023, representing 15% of global TB incidence [1]. Among them, 535 000 are older adolescents (15–19 years) and 1 049 000 are young adults (20–24 years) [2]. In 2019, the World Health Organization reported that TB claimed the lives of 150 000 youths, including 60 000 older adolescents and 90 000 young adults [3, 4]. The prevalence of TB infection rises throughout adolescence and into young adulthood [5], with a corresponding increase in the progression from TB infection to active disease [6].

Youth face unique challenges in accessing healthcare, which can delay diagnosis and treatment. Once diagnosed, youth, and particularly young adults, exhibit worse medication adherence and are more likely to be lost to follow-up compared to other age groups [7, 8]. Their larger social networks [9] and propensity to transmit TB to peers under age 25 years [10] further complicate transmission dynamics. Despite this, current TB services frequently fail to address the unique needs of youth, and few studies have evaluated interventions aimed at improving outcomes for this group [11, 12]. Understanding how TB risk differs among youths compared to other age groups is crucial for developing effective elimination strategies. Traditional contact investigations focus mainly on identifying and screening household contacts (HHCs) of individuals diagnosed with TB. Because age has not been the main exposure of interest in prior analyses, the majority of the extensive literature on TB transmission revolves around adults and younger children. Less is available for youth, though previous studies acknowledge the presence of highly transmissible forms of TB in this group [13], and no association between HHCs and risk of TB infection among youth 15–22 years old has previously been observed [14]. Little is known about how the age of the index patient affects TB infection risk among HHCs, particularly for youths. This gap hinders the development of targeted interventions.

To address this, our study investigates how the age of individuals diagnosed with TB influences transmission risk to their household contacts. Because TB infection in children is more likely to represent recent transmission—given their limited lifetime exposure and higher test specificity in this age group—we focus primarily on TB infection risk among child household contacts (≤15 years). We first assess the prevalence of TB infection at enrollment as a proxy for baseline transmission. Next, we evaluate the incidence of infection over a 12-month follow-up period to capture ongoing transmission risk. Finally, we analyze the genetic linkage of Mycobacterium tuberculosis (Mtb) strains between index patients and their household contacts to corroborate presumed transmission events. By integrating these approaches, our study aims to clarify how index patient age, particularly among youth (15–24 years), influences TB transmission within households. These findings can help inform targeted, age-specific screening and

METHODS

Setting and Study Population

The cohort study was conducted in Lima, Peru, between September 2009 and August 2012. The detailed study design and methods have been previously reported [15]. In brief, adults (age ≥15 years) suspected of incident pulmonary TB were identified from 106 health centers serving a population of approximately three million. Patients suspected of having TB underwent smear microscopy and culture; those positive on either test were classified as index patients. Within one month of diagnosing the index patient, a study nurse visited their household to invite exposed HHCs to participate in a 12-month follow-up study to monitor TB outcomes, including incident TB infection and disease. HHCs without a previous positive tuberculin skin test (TST) or TB diagnosis underwent a baseline TST; those with induration ≥10 mm were considered TB-infected at enrollment. Repeat TSTs were conducted at 6 and 12 months for initially TST-negative HHCs. HHCs who were not TB-infected at enrollment but later tested positive with the TST or developed TB during follow-up were classified as having incident TB infection. HHCs diagnosed with TB within 14 days of enrollment were classified as co-prevalent TB patients, while those diagnosed after 14 days were classified as secondary patients. Social-demographic data collected at baseline included age, gender, tobacco and alcohol use, socioeconomic status (SES), and incarceration history. Given that it was common for smokers in this setting to smoke only one cigaret per day, we categorized participants into three groups: nonsmokers (no cigarette use), light smokers (1 cigarette per day), and moderate to heavy smokers (>1 cigarette per day). Similarly, participants were classified as nondrinkers (no alcohol consumption), light drinkers (<40 g or <3 alcoholic drinks per day), and heavy drinkers (≥40 g, ≥3 alcoholic drinks per day). We developed a continuous variable to assess household SES by using data on housing quality, water supply, and sanitation in a principal component analysis. The resulting SES scores were then divided into tertiles, corresponding to “low,” “middle,” and “high” SES categories [16]. Among HHCs <20 years, nutrition status was classified as underweight if the body mass index (BMI) z-score < −2 or overweight if z-score >2 according to the World Health Organization age and gender-specific BMI z-score tables [17]. Among HHC ≥20 years, nutritional status was: underweight (BMI <18.5), normal (BMI 18.5 to <25), and overweight (BMI ≥25) [18].

Age Categorization of Index Patients and HHCs

Index patients were categorized into two age groups: Youth (15–24 years) and Adults (25–60 years). Household contacts (HHCs) were categorized into three groups: Children (0–14 years), Youth (15–24 years), and Adults (25–60 years). Participants older than 60 years were excluded due to potential generationally related mixing patterns within households [19].

TB Outcomes Among HHCs

We investigated whether TB outcomes among HHCs varied by the age of the index patient or the HHC. Outcomes comprised (1) the prevalence of TB infection at enrollment, (2) the incidence of TB infection during 12 months of follow-up, and (3) the likelihood of within-household transmission among index-HHC patient pairs using Whole-Genome Sequencing (WGS) data of Mtb.

Data Analysis

To account for prior infections or community-acquired TB among adult HHCs, risk analyses were restricted to child contacts. We employed a modified Poisson generalized estimating equation to assess TB infection risk in child HHCs at enrollment and a Cox frailty proportional hazards model to evaluate incident TB infection risk. Empirical standard error estimates were used for inference and to construct Wald 95% confidence intervals (CI).

We conducted univariate analyses for known TB infection risk factors among HHCs, followed by multivariate adjustments. These factors included index patient characteristics (eg, gender, smoking, alcohol use, sputum smear status, cavitary disease, duration of coughing symptoms prior to diagnosis) and HHC characteristics (eg, gender, BCG scars, socioeconomic status). We did not adjust for diabetes mellitus (DM) or HIV status, as the prevalence of DM is low in the youth population, and the prevalence of HIV is low in our study population.

To investigate the mechanisms underlying the differential TB risks among HHCs, we evaluated the associations between known risk factors and TB outcomes, restricting the analysis to HHCs exposed to youth index patients. This approach enabled us to assess whether well-established risk factors for TB transmission applied to youth index patients.

Because only approximately 10% of HHCs would be excluded due to missing data for each study setting, we conducted a complete case analysis approach.

Genetically Linked Mycobacterium Tuberculosis Isolates of Index-secondary Patient Pairs

Mtb isolates from culture-positive index patients and HHC secondary patients underwent WGS using Illumina HiSeq in paired-end mode, with a read length of 100–150 base pairs and a coverage of at least 30-fold. We mapped the raw WGS reads to the Mtb H37Rv reference genome using the BWA-MEM algorithm [20, 21]. Subsequently, we identified single-nucleotide polymorphisms (SNPs), insertions, and deletions using an alignment-based approach [21, 22]. Mtb isolates with <95% genome mapping coverage were excluded from the analysis. A nucleotide call of an isolate was considered missing if the valid depth of coverage at that site was less than 10% of the mean coverage of the isolate, if the mean read mapping quality at that site did not reach 10, or if the alternative allele occupied less than 85% of the valid depth of coverage at the site. Given the high error rate of alignment quality for repetitive genes when using short-read sequencing, we excluded variants in the proline-glutamate (PE)/proline-proline-glutamate (PPE) family and in regions with a high error rate in previous reports from the analyses [23 , 24]. Additionally, we excluded Mtb isolates with evidence of mixed infection from the analyses [25].

For each index-HHC secondary patient pair, we used the SNP difference to determine the genetic distance between them. An index-HHC secondary patient pair was considered genetically linked if the genetic distance between their Mtb strains was less than or equal to 12 base pairs. We used a chi-squared test to examine whether the proportion of genetically-linked index-HHC secondary patient pairs varied by index patients’ age groups. All analyses were conducted using R 4.2.2.

Research Ethics

The Institutional Review Board of the Harvard School of Public Health (Boston, MA, USA) and the Research Ethics Committee of the National Institute of Health of Peru (Lima, Peru) approved the parent household contact cohort study. Study participants or their guardians provided voluntary, written informed consent prior to study enrollment. All procedures were performed in accordance with relevant local guidelines and regulations.

RESULTS

Among the 4500 index patients, 4044 (90%) had microbiologically confirmed TB disease, defined as having a positive sputum smear test, culture test, or both. Of these, 3109 index patients had at least one household contact (HHC), resulting in a total of 12 767 HHCs. After excluding participants >60 years old, 10 745 HHCs of 2771 index patients were used for the data analyses of this study.

Prevalence of TB Infection Among Child Household Contacts

Among the 10 745 HHCs, 3886 (36%) were younger than 15 years old. Of these, baseline TB infection status was available for 3711 (96%), and 1012 (27%) were found to have TB infection at baseline enrollment (Table 1). At enrollment, 22% (350/1557) of the child contacts with a youth index patient were TB-infected, compared to 31% (662/2154) of those with an adult index patient (univariate prevalence risk ratios [PRR] = 0.75; 95% CI: .66–.85). After multivariate adjustment, the results remained largely unchanged (aPRR = 0.77; (95% CI: .67–.87) (Table 2).

Table 1.

Baseline Characteristics of Child Contacts Stratified by TB Infection Status at Enrollment

	Total		Non-TB-infected		TB-Infected
	N	%	N	%	N	%
Index age
Adult	2154	58	1492	55	662	65
Youth	1557	42	1207	45	350	35
Index sex
Female	1727	47	1196	44	531	53
Male	1984	54	1503	56	481	48
Index smoking status
Non-smoker	3538	97	2584	97	954	96
Light-smoker	55	2	40	2	15	2
Heavy-smoker	56	2	33	1	23	2
Index drinking status
Non-drinker	2001	56	1457	56	544	56
Light-drinker	1184	33	880	34	304	32
Heavy-drinker	377	11	259	10	118	12
Index smear status
Negative	1042	28	822	31	220	22
+	1025	28	723	27	302	30
++	645	17	442	16	203	20
+++	986	27	704	26	282	28
Index coughing duration
0–13 d	717	20	581	22	136	14
14–28 d	1233	34	907	34	326	33
29–55 d	907	25	639	24	268	27
≥56 d	802	22	534	20	268	27
Index cavity
No	2680	73	2025	76	655	66
Yes	981	27	640	24	341	34
Child contact age
0–5	1447	39	1159	43	288	29
6–10	1198	32	852	32	346	34
11–14	1066	29	688	26	378	37
Child contact sex
Female	1851	50	1336	50	515	51
Male	1860	50	1363	51	497	49
Child contact number of BCG scars
0	711	19%	505	19	206	20
1	2922	79	2144	79	778	77
>1	78	2	50	2	28	3
Child contact nutritional status
Normal	2937	80	2127	80	810	81
Underweight	103	3	83	3	20	2
Overweight	632	17	459	17	173	17
Household SES
High	1404	39	993	38	411	42
Medium	1601	44	1190	45	411	42
Low	607	17	456	17	151	16

Table 2.

Associations Between the Characteristics of Index Patients and Child Contacts and the Risk of TB Infection in Child Contacts at Enrollment

	Univariate (N = 3711)		Multivariate (N = 3277)
	PRR (95% CIs)	P-value	PRR (95% CIs)	P-value
Index age
Adult	Ref		Ref
Youth	0.75 (.66–.85)	<.001	0.77 (.67–.87)	<.001
Index sex
Female	Ref		Ref
Male	0.80 (.71–.89)	<.001	0.82 (.72–.93)	.003
Index smoking status
Non-smoker	Ref		Ref
Light-smoker	0.99 (.55–1.76)	.965	0.95 (.48–1.88)	.885
Heavy-smoker	1.46 (1.01–2.11)	.044	1.59 (1.15–2.21)	.005
Index drinking status
Non-drinker	Ref		Ref
Light-drinker	0.91 (.79–1.04)	.153	0.90 (.78–1.04)	.157
Heavy-drinker	1.16 (.97–1.39)	.110	1.11 (.91–1.36)	.300
Index smear status
Negative	Ref		Ref
+	1.35 (1.15–1.60)	<.001	1.20 (1.01–1.42)	.039
++	1.43 (1.19–1.71)	<.001	1.34 (1.1–1.61)	.002
+++	1.32 (1.12–1.56)	.001	1.17 (.98–1.39)	.078
Index coughing duration
0–13 d	Ref		Ref
14–28 d	1.38 (1.14–1.67)	.001	1.22 (1.00–1.48)	.050
29–55 d	1.57 (1.29–1.90)	<.001	1.39 (1.13–1.69)	.001
≥56 d	1.71 (1.41–2.08)	<.001	1.38 (1.12–1.70)	.002
Index cavity
No	Ref		Ref
Yes	1.40 (1.24–1.57)	<.001	1.32 (1.16–1.49)	<.001
Child contact age
0–5	Ref		Ref
6–10	1.46 (1.29–1.66)	<.001	1.49 (1.30–1.71)	<.001
11–14	1.72 (1.51–1.95)	<.001	1.78 (1.54–2.03)	<.001
Child contact sex
Female	Ref		Ref
Male	0.97 (.88–1.07)	.585	0.96 (.87–1.07)	.492
Child contact number of BCG scars
0	Ref		Ref
1	0.92 (.81–1.04)	.165	0.93 (.82–1.06)	.300
>1	1.20 (.88–1.62)	.246	1.17 (.85–1.61)	.334
Child contact nutritional status
Normal	Ref		Ref
Underweight	0.72 (.49–1.05)	.091	0.85 (.58–1.25)	.418
Overweight	1.02 (.90–1.15)	.810	0.99 (.87–1.14)	.939
Household SES
High	Ref		Ref
Medium	0.87 (.76–.99)	.030	0.89 (.78–1.02)	.096
Low	0.84 (.71–1.00)	.053	0.83 (.69–.98)	.026

Incidence of TB Infection Among Household Contacts

Among the 10 745 HHCs, 5850 (54%) were uninfected at the time of enrollment. Of these, 4553 (78%) either underwent at least one repeat TST or developed incident TB infection during the 12-months of follow-up (Table 3). The age group of index patients was not associated with incident TB infection among HHCs (youth index patient vs adult index patient: hazard ratio (HR) = 1.00; 95% CI: .88–1.13; aHR = 0.94; 95% CI: .82–1.08). The incidence of TB infection was lower among both child contacts and youth contacts compared to adult contacts (child contacts vs adult contacts: HR = 0.29; 95% CI: .25–.33; youth contacts vs adult contacts: HR = 0.56; (95% CI: .49–.64). These associations remained after adjustment for potential confounders (child contacts vs adult contacts: aHR = 0.34; 95% CI: .29–.40; youth contacts vs adult contacts: aHR = 0.64; 95% CI:.54–.75) (Table 4).

Table 3.

Characteristics of Household Contacts Who Were not TB-infected at Enrollment Stratified by Incident TB Infection at 12-month Follow-up

	Total		Without Incident TB Infection		With Incident TB Infection
	N	%	N	%	N	%
Index age
Adult	2327	51	1549	51	778	51
Youth	2226	49	1482	49	744	49
Index sex
Female	1907	42	1268	42	639	42
Male	2646	58	1763	58	883	58
Index smoking status
Non-smoker	4353	97	2889	97	1464	98
Light-smoker	53	1	37	1	16	1
Heavy-smoker	63	1	44	2%	19	1
Index drinking status
Non-drinker	2476	56	1675	57	801	54
Light-drinker	1515	34	987	34	528	36
Heavy-drinker	414	9%	262	9%	152	10%
Index smear status
Negative	1270	28%	893	30%	377	25%
+	1282	28%	830	28%	452	30%
++	783	17	508	17%	275	18
+++	1202	27	790	26%	412	27
Index coughing duration
0–13 d	946	21	649	22	297	20
14–28 d	1490	33	1012	34	478	32
29–55 d	1134	25	720	24	414	28
≥56 d	912	20	604	20	308	21
Index Cavity
No	3344	74	2299	77	1045	69
Yes	1169	26	703	23	466	31
Contact age
Adult (≥25)	1423	31	680	22	743	49
Youth (15–24)	974	21	646	21	328	22
Children (<15)	2156	47	1705	56	451	30
Contact sex (n = 4773)
Female	2529	56%	1671	55%	858	56%
Male	2024	45%	1360	45%	664	44%
Contact number of BCG scars
0	648	14%	515	17%	133	9%
1	3158	69%	2161	71%	997	66%
>1	747	16%	355	12%	392	26%
Contact nutritional status
Normal	3049	65	2155	69	894	56
Underweight	86	2	67	2	19	1
Overweight	1587	34	891	29	696	43
Household SES
High	1619	36	1080	36	539	36
Medium	1986	45	1368	46	618	42
Low	859	19	526	18	333	22

Table 4.

Associations Between the Characteristics of Index Patients and HHCs and the Risk of Incident TB Infection Among HHCs During a 12-month Follow-up

	Univariate (N = 4553)		Multivariate (N = 4088)
	HR (95% CIs)	P-value	HR (95% CIs)	P-value
Index age
Adult	Ref		Ref
Youth	1.00 (.88–1.13)	>.999	0.94 (.82–1.08)	.395
Index sex
Female	Ref		Ref
Male	0.98 (.86–1.11)	.748	0.91 (.79–1.05)	.186
Index smoking status
Non-smoker	Ref		Ref
Light-smoker	0.81 (.44–1.50)	.508	0.84 (.44–1.60)	.600
Heavy-smoker	0.80 (.46–1.40)	.44	0.80 (.44–1.48)	.482
Index drinking status
Non-drinker	Ref		Ref
Light-drinker	1.07 (.94–1.23)	.301	1.12 (.96–1.30)	.146
Heavy-drinker	1.18 (.95–1.47)	.132	1.29 (1.02–1.63)	.036
Index smear status
Negative	Ref		Ref
+	1.28 (1.08–1.52)	.004	1.24 (1.03–1.49)	.020
++	1.29 (1.07–1.56)	.009	1.40 (1.14–1.72)	.001
+++	1.24 (1.04–1.47)	.015	1.23 (1.02–1.48)	.029
Index Coughing duration
0–13 d	Ref		Ref
14–28 d	1.03 (.86–1.23)	.756	1.05 (.87–1.27)	.639
29–55 d	1.25 (1.04–1.50)	.017	1.21 (.99–1.48)	.057
≥56 d	1.10 (.91–1.34)	.334	1.03 (.83–1.27)	.777
Index cavity
No	Ref		Ref
Yes	1.40 (1.22–1.60)	<.001	1.30 (1.13–1.51)	<.001
Contact age
Adult (≥25)	Ref		Ref
Youth (15–24)	0.56 (.49–.64)	<.001	0.64 (.54–.75)	<.001
Children (<15)	0.29 (.25–.33)	<.001	0.34 (.29–.4)	<.001
Contact sex
Female	Ref		Ref
Male	0.93 (.84–1.04)	.202	1.10 (.98–1.23)	.123
Contact number of BCG scars
0	Ref		Ref
1	1.72 (1.42–2.08)	<.001	1.55 (1.27–1.9)	<.001
>1	3.5 (2.84–4.31)	<.001	1.88 (1.48–2.38)	<.001
Contact nutritional status
Normal	Ref		Ref
Underweight	0.69 (.43–2.08)	<.001	0.67 (.39–1.14)	.142
Overweight	1.76 (2.84–4.31)	<.001	1.08 (.95–1.24)	.246
Household SES
High	Ref		Ref
Medium	0.94 (.82–1.09)	.412	0.93 (.80–1.09)	.357
Low	1.26 (1.07–1.49)	.006	1.14 (.96–1.37)	.136

In analyses restricted to HHCs exposed to youth index patients, most well-known risk factors for TB transmission remained strongly associated with both the risk of TB infection in child contacts and the risk of incident TB in HHCs, except for the duration of coughing prior to disease diagnosis (Supplementary Tables 1 and 2). Further, analyses did not identify associations among the youth age groups for either outcome when broken down into 15–19 and 20–24 years (Supplementary Tables 3 and 4).

Genetically-linked Mtb Isolates of Index-secondary Patient Pairs

Among the 12 676 HHCs exposed to a microbiologically confirmed TB index patient, 692 (5%) HHCs were diagnosed with TB at either enrollment or during the follow-up. Among these, 246 (36%) had a positive culture, and 225 (91%) of their isolates were sequenced. Among the 190 index-HHC patient pairs in which both index and HHC patients had a positive culture, 15 (8%) pairs had at least one isolate with a mixed infection and were excluded, leaving 175 (92%) pairs with high-quality WGS data for both index and HHC patients. Of these, 109 (62%) pairs were index-HHC secondary patient pairs. The proportion of genetically linked pairs was lower for pairs with a youth index patient (26/42, 62%) compared to those with an adult index patient (48/67, 72%), although this difference did not reach statistical significance (P = .3, Fisher's exact test) (Supplementary Table 5).

DISCUSSION

Our study found that, within the household setting, youth were less likely to transmit TB as index patients and less likely to be infected as HHCs compared to adults. This was supported by a lower risk of TB infection in child contacts of youth index patients and a smaller proportion of genetically-linked index-HHC pairs. Likewise, youth HHCs were less likely to acquire TB from their index patients than adult HHCs. Overall, these findings suggest that—in comparison to adults—youth play a smaller role in TB transmission within their households.

Several potential mechanisms may explain our finding that youth index patients transmitted TB to household contacts less frequently than adult index patients. First, youth index patients might simply not transmit TB; however, this is not the case in our study, nor in others. Second, youth index patients may differ from adults in terms of factors that influence TB transmission—such as disease severity or symptom duration—but this is unlikely to explain our findings, as effect sizes remained nearly unchanged after adjusting for these known risk factors. Third, young adults tend to have highly assortative social contact patterns, primarily interacting with peers rather than household members of different ages. Also set in Western Europe, which may have limited extrapolation to this context, Mossong et al found that this pattern results in fewer interactions with older or younger household members and more frequent contact with similarly aged community members [26]. Similarly, a study in Peru by Grijalva et al, which surveyed 558 members from 114 households, reported that youth had the lowest proportion of within-household social interactions among all age groups [27]. Together, these findings suggest that young adults may have fewer household-associated contacts and spend more time in community settings. While our study did not collect data on within-household interaction patterns, we found that cough duration—a marker of the potential duration of contagiousness—did not appear to influence TB infection risk among child contacts of youth index patients. In contrast, other transmission-related factors such as smear positivity and cavitary disease primarily reflect disease severity and bacillary burden. The lack of association between cough duration and transmission in youth may suggest that even when symptoms persist, reduced time spent at home limits opportunities for transmission. In this context, total cough duration may overestimate actual exposure time within the household for younger individuals.

No previous studies have specifically examined the relative within-household TB transmission risk between youth contacts and contacts of other age groups. However, two studies have highlighted age-related differences in community-acquired TB infection risk among HHCs. In the first study, using a modeling approach, McIntosh et al analyzed HHC data from Brazil and Uganda and found that the probability of community-acquired TB transmission among HHCs increased with age. However, this study did not further categorize individuals older than 15 years, making it unclear which specific age group contributed most to this trend [28]. In the second study, Middelkoop et al investigated TB infection risk among 628 children and 472 adolescents in a high-burden South African community. They found that exposure to a TB patient in the home doubled the odds of TB infection in children, while the increase was only 1.4-fold for adolescents [14].

Currently, contact investigation is a key component of TB control efforts [29]. Its effectiveness relies on the assumption that Mtb transmission to household contacts primarily occurs within the household rather than from the wider community. Previous studies have reported that adolescents and young adults may have an increased TB exposure in schools and public transit [13]. While household-based contact investigation is known to be highly effective for preventing TB disease among people who have been exposed at home to TB, it is not an intervention that is designed to address the particular role of youths in TB transmission, which may be relatively more important outside their homes. This is an area requiring new research.

Our study has several limitations. First, in a high TB-burden setting such as Lima, HHCs may have been infected outside the household. It is unclear how this might have affected the results, as the true risk of community-acquired infection likely varies by age group due to differences in social mixing patterns. Second, our study's location in urban Lima, Peru, may limit its generalizability to other settings, particularly those with different social mixing patterns influenced by factors such as age structure or cultural differences. Third, we did not collect data on within-household social interaction patterns, which limits our ability to further explore the mechanisms underlying our findings. Further, it is always possible that unmeasured confounders or effect modifiers may have impacted the results.

CONCLUSIONS

In conclusion, our findings suggest that youths contribute less to within-household TB transmission compared to adults, underscoring the importance of considering age-specific social contact patterns in transmission dynamics. These results highlight the need to complement household-focused TB strategies with community-based interventions that reflect the broader social behaviors of youth. Expanding TB screening and surveillance beyond the household setting may improve case detection and prevention in this age group. In parallel, novel tools—such as TB vaccines targeting youth—offer promising opportunities to prevent infection and disease. Future research should further explore youth social networks and community-based transmission pathways to inform the development of effective, age-tailored TB control strategies.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.