Background:  Clinically oriented lists of HIV drug resistance mutations have limited utility for the epidemiology of transmitted resistance. Use of different lists has led to non-comparable prevalence estimates and inclusion of mutations that occur naturally without drug pressure lead to overestimates. We propose criteria for a mutation list designed for HIV drug resistance transmission epidemiology and surveillance. Mutations should be associated with HIV drug resistance (defined as inclusion in ³3 expert lists) and selected by ART, non-polymorphic (should not occur without drug pressure >0.5 to 1% in any HIV-1 subtype), parsimonious (should exclude rare mutations), and applicable to all subtypes.

Methods:  We examined 5 expert lists (IAS-USA, Los Alamos National Laboratory Sequence Database, lists based on algorithms from the Stanford HIV database interpretation program, Agence Nationale Reserche sur le Sida, and Rega Institute). We evaluated mutations appearing in ³3 lists for frequency of selection with ART and non-polymorphism using the Stanford HIV RT and Protease Sequence database. We excluded from the analysis any isolate from an untreated person with ³2 drug-resistance mutations known to be non-polymorphic, and from areas where primary HIV drug resistance is reported to be >5%. Our analysis was based on the following numbers of respective subtype B and non-B isolates:  3369 and 3653 from PI-naive persons, 3354 and 793 from PI-treated persons; 2077 and 2585 from RTI-naïve persons, 3205 and 755 from NRTI-treated persons, 1772 and 1042 from NNRTI-treated persons. Non-polymorphic mutations associated with ART among isolates from any subtype met the criteria.

Results:  These resistance mutations met the criteria:  34 at 16 PI positions (L10F/R,  l24I, D30N, V32I, M46I, I47V/A, G48V, I50V/L, F53L, I54V/M/L/T/S/A, A71I, G73S/T/C/A, V82A/F/T/S/M, I84V/A/C, N88D/S, L90M), 27 NRTI at 15 RT positions (M41L, K65R, D67N/G/del, T69D/N/ins, K70R, L74V, V75T/M/A/S, F77L, Y115F, F116Y, Q151M, M184V/I, L210W, T215Y/F/X, K219Q/E/R), and 19 NNRTI at 11 RT positions (L100I, K101E, K103N/S, V106A/M, V108I, Y181C/I, Y188L/H/C, G190A/S/E/Q, P225H, M230L, P236L). 

Conclusions:  A list of mutations based on standard criteria will facilitate comparable prevalence estimates and support global surveillance of transmitted resistance. For the definitive list, more non-B subtype sequences will be analyzed as they become available. The list will be publicly available and regularly updated using the stated criteria.