To the Editors:
The global HIV-1 pandemic does not recognize national borders. The contribution of migrants to the HIV epidemic in Western Europe has long been recognized,1 and recently some local health jurisdictions in the United States have reported that the proportion of new HIV diagnoses occurring among foreign-born people is increasing over time.2,3 Understanding the epidemiology of HIV in this population, including when and where people acquired HIV, is instrumental in developing effective strategies to prevent HIV among international migrants.
Valverde et al4 recently reported in this journal (Valverde 76(5):445-452, December 15, 2017.) that most HIV-infected foreign-born US residents with a genetic link to another infection were linked to a US-born individual. This finding was based on the identification of genetic similarity clusters in a large data set of HIV-1 pol sequences from 77,686 individuals in the National HIV Surveillance System (curated by the US Centers for Disease Control and Prevention). From this result, Valverde et al concluded that most HIV-infected foreign-born individuals in the putative US transmission network were infected after immigrating to the US.
We commend the authors for addressing this important issue with molecular epidemiological methods, which can be informative for this type of public health question. However, we believe that the main conclusion may not be fully supported by the data.
In their study, Valverde et al first identified putative transmission linkages as any 2 or more sequences that were closer than a 1.5% Tamura and Nei5 distance threshold (0.015 substitutions/site; implemented through HIV-TRACE6). Next, they performed a series of descriptive and statistical analyses to assess associations between epidemiological characteristics (including region of birth) and linkage status. Of the 77,686 individuals in the data set, 12,064 were foreign-born individuals (16%)—thus 84% were US born. A total of 28% of the foreign-born individuals were genetically linked to at least one other individual in the data set. Of those, 28% of foreign-born individuals, 62% were linked to a US-born individual. This last proportion is the source of the main conclusion that most HIV-infected foreign-born individuals in the transmission network revealed by linkage analysis were infected after immigrating to the United States.
We see 2 main problems with the analyses.
First, the analysis does not establish an expected proportion of (linked) foreign-born individuals that link to a US-born individual, ie, although it is technically true that most (62%) linked foreign-born individuals have a link to a US-born individual, it is unclear whether this is more or less than expected by chance alone. A null expectation might be established as follows: in a randomly mixing population, we can expect that any single individual in the current data set, if linked, would link at 84% probability to a US-born individual and at 16% probability to a foreign-born individual, based on the respective frequencies in the database (for simplicity, we ignore potential assortative mixing.) The observed rate of linkage between foreign-born and US-born individuals is 62%. Thus, the proportion of foreign-born individuals linked to US-born individuals was actually smaller than expected by chance; this would reverse the authors' conclusion. What might be anticipated is an exploration of why linked HIV-infected foreign-born individuals are not linked more often to US-born individuals.
Second, the analysis does not consider the possible direction of transmission in each putative transmission pair. If we assume that 50% of the genetic linkages between a foreign-born and a US-born individual represent a series of one or more transmissions that originate from the US-born individual, then 31% (62% × 0.5) of linked foreign-born individuals are potentially infection sources rather than recipients. Absent epidemiological evidence that may alter this directionality assumption (ie, that would point to either foreign-born or US-born individuals as more likely to be the source in an HIV transmission pair), one might conclude that a minority of HIV-infected foreign-born individuals in the United States were infected after immigrating.
We concede that we arrive at this alternate conclusion through simplifying assumptions about mixing patterns and transmission dynamics, and without incorporating recent advances in phylodynamics,7 phylogeography,8 or source attribution.9 Yet the discrepancy between our and Valverde et al's estimates highlights the importance of establishing null expectations when inferring epidemiological parameters from genetic similarity data.
Finally, we would like to highlight the 72% of HIV-infected foreign-born individuals diagnosed and sampled in the United States who were not linked to any other individual in the data set. A heuristic assumption might be that the sources of these individuals' infections are foreign-born individuals currently living outside the United States and therefore not sampled nor included in the data set. This suggests that most HIV-infected foreign-born individuals in the United States were not infected in the United States. In the context of Valverde et al's conclusion, this is a critical point that should not be overlooked.
1. Hernando V, Alva´rez-del Arco D, Alejos B, et al. HIV infection in migrant populations in the European union and European economic area in 2007–2012: an epidemic on the move. J Acquir Immune Defic Syndr. 2015;70:204–211.
2. Wiewel EW, Torian LV, Nasrallah HN, et al. HIV diagnosis and utilisation of HIV-related medical care among foreign-born persons in New York City, 2001–2009. Sex Transm Infect. 2013;89:380–382.
3. Kerani R, Bennett AB, Golden M, et al. Foreign-born individuals with HIV in King County, WA: a glimpse of the future of HIV? AIDS Behav. 2017:1–8. doi:10.1007/s10461-017-1914-3.
4. Valverde EE, Oster AM, Xu S, et al. HIV transmission dynamics among foreign-born persons in the United States. J Acquir Immune Defic Syndr. 2017;76:445–452.
5. Tamura K, Nei M. Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol Biol Evol. 1993;10:512–526.
6. Wertheim JO, Leigh Brown AJ, Hepler NL, et al. The global transmission network of HIV-1. J Infect Dis. 2014;209:304–313.
7. Hughes GJ, Fearnhill E, Dunn D, et al. Molecular phylodynamics of the heterosexual HIV epidemic in the United Kingdom. PLoS Pathog. 2009;5:e1000590.
8. Paraskevis D, Pybus O, Magiorkinis G, et al. Tracing the HIV-1 subtype B mobility in Europe: a phylogeographic approach. Retrovirology. 2009;6:49.
9. Volz EM, Frost SD. Inferring the source of transmission with phylogenetic data. PLoS Comput Biol. 2013;9:e1003397.