Clusters of HIV transmission can serve as alerts for public health action. These clusters can comprise (i) geographic clusters of diagnosis in time and space or (ii) clusters of genetically linked cases. We assessed the impact of geography across the Los Angeles County (LAC) HIV genetic transmission network and its relevance to directing HIV prevention and treatment.
Deidentified surveillance data reported for 8150 persons residing in LAC (2010 through 2016) were used to construct a transmission network using HIV-TRACE. Persons were linked if their pairwise genetic distance was ≤0.015 substitutions/site. Residential information at diagnosis was examined at ZIP code, health district (HD), and service planning area (SPA). We used 3 approaches to characterize this relationship: (i) geographic assortativity, the tendency for people to link to others within the same geography (ii) concordant time-space pairs, the proportion of genetically linked pairs from the same geography and diagnosis year, and (iii) Jaccard coefficient (JC), the intersection divided by the union of geography and cluster sets. Significance was determined using 1000 random network permutations. A generalized linear model was used to identify characteristics associated with the JC.
The 4150 (50.6%) HIV cases that were clustered were assortative by ZIP, HD, and SPA (0.02, 0.09, 0.15; p<0.001). Geography was less assortative than race/ethnicity and transmission risk. 58% of genetically linked cases were diagnosed in the same year, and 44% were diagnosed in the same HD; however, only 19% were diagnosed in the same year and HD. This time-space concordance among genetically-linked pairs was also low across ZIP and SPA (p<0.001). In the JC analysis, cis-men (b=0.20; p<0.001) and those younger at diagnosis (b=0.189; p=0.01) had more overlap between clusters and geography; we observed an inverse association for trans-women (b=-0.51; p<0.001) and African-Americans (b=-0.18; p<0.001).
We found significant, but weak associations between the HIV transmission network and residence at diagnosis. Within an urban setting with endemic HIV, genetic clustering may serve as a better indicator than time-space clustering to understand transmission patterns and direct public health action.