Wide variations in the prevalence of drug-resistant HIV in newly infected and treatment-naive individuals have been reported, together with differences in the rates of change in prevalence over time.

Across Europe, several large recent studies have reported roughly 10% to 14% rates of drug resistance in newly acquired HIV:

• The CATCH study assessed resistance in over 1630 newly infected people in Europe between 1996 and 2002, finding primary resistance mutations in 10% of the group, NRTI resistance mutations in 7%, NNRTI mutations in 3%, and PI mutations in 2% ^[1].
• An analysis of 2208 newly infected Europeans, first presented in 2003, found drug resistance in 11% ^[2]. Similar rates were found in the SPREAD study of 1083 newly infected individuals from 17 European countries, with only 1% of the study group demonstrating dual-class resistance ^[3].
• In the Primo study of 303 patients with acute HIV infection, 14% had resistance mutations: 10% to NRTIs, 3% to NNRTIs and 4% to PIs ^[4].
• In the Odyssee cohort of 363 newly infected, treatment-naive patients, the overall prevalence was 6% (4% to NRTIs, 1% to NNRTIs and 1% to PIs) ^[5].

Other, smaller or more geographically limited studies have reported comparable findings ^{[6][7][8][9][10][11][12][13]}. Reports in certain areas and time periods have been considerably higher^[14][15].

In the United Kingdom, rates have varied with time, reaching a high around 2002 and then declining. From 1994 to 2000, 14% of newly infected individuals (seroconverters) had resistant virus, but 27% of people who contracted HIV in 2000 were infected with drug-resistant virus ^[16]. In 2002, 27% of gay men identified as recent seroconverters in England, Wales and Northern Ireland had resistance to at least one drug, compared to 20% in 2001^[17].

The UK Drug Resistance Database shows that the prevalence of genotypic resistance in antiretroviral-naive individuals in the United Kingdom has declined from 16% in 2002 to 12% in 2003 and 8% by the end of 2004. Most of the resistant cases were resistant to a single drug class (4.5% to NRTIs or NNRTIs, 2.1% to PIs); 17% to two drug classes, and 8% to three. Resistance in antiretroviral-experienced individuals is also showing a downward trend ^[18][19].

Estimates of drug-resistant new infections in the United States also vary, from reports of 8% to 9% prevalence in treatment-naive people ^[20][21][22] to nearly 25% in others ^[23][24]. Upward trends in overall resistance have been shown in many cities (3% between 1995 and 1998 versus 12% between 1999 and 2000) ^[25][26]. However, at least in San Francisco and New York, the prevalence of NRTI resistance has been declining: by 2001, to 6% in San Francisco and 3% in New York ^[27][28].

In Montreal, Canada, a decline in resistance amongst newly-infected individuals has been noted since 2000, and this is significantly correlated with the proportion of patients receiving antiretroviral treatment, a drop in the average viral load of chronically infected people and the availability of resistance testing ^[29].

A study carried out between June 2000 and March 2002 found no significant increase in transmitted drug resistance when compared with the periods 1995 to 1998 and 1998 to 2000. Indeed, NRTI and PI resistance in seroconverters declined significantly when 2000 to 2002 was compared with 1998 to 2000, with much of the resistance detected in the 2000 to 2002 period attributable to NNRTI resistance. Multidrug resistance declined significantly, from 7 to 1%.