In the USA and Canada there were an estimated one million people living with HIV at the end of 2004, 945,000 in the USA and 55,000 in Canada.

In the USA an estimated 0.6% of the population is living with HIV. 31,886 people were newly diagnosed with HIV in 2003 in the 33 states that used names-based reporting. The other 18 states (and the District of Columbia) currently use code-based anonymised reports which the Centers for Disease Control regard as unreliable. If they were included the estimated number of people newly infected during 2003 was 50,000; in that year 14,000 deaths occurred due to AIDS. The availability of antiretroviral treatment has brought a decline in death rates.

However there is continuing transmission of HIV, and it is increasingly concentrated among ethnic minority populations in the USA. African-Americans, for example, are 12% of the population of the USA but constituted 54% of new HIV diagnoses and 25% of all AIDS cases in 2002.

Previously the epidemic was primarily concentrated among gay men with large minorities of intravenous drug users and heterosexuals. Of the 877,274 people ever diagnosed with AIDS in the USA 48% were men who acquired HIV through sex with men and another 7% were gay men who injected drugs, so could have caught HIV through either route. Another 27% cited injecting drug use as their main risk factor and 15% appear to have caught HIV heterosexually. (It is important to note that for AIDS surveillance purposes the USA now uses a different definition of AIDS from the one employed in Europe. Since 1993 anyone with a CD4 count below 200 is considered to have AIDS. In Europe a diagnosis of AIDS can only occur after the diagnosis of specified opportunistic infections).

Last year the proportion who caught HIV through gay sex had stayed about the same (46%, plus 4% of gay men who injected drugs) but heterosexuals now accounted for 34% of new diagnoses and injecting drug users 15%. Most of those heterosexuals are African-Americans: African American women accounted for 72% of all new HIV diagnoses in women, meaning they are 21 times as likely to get HIV as their white counterparts. In the 34 areas of the United States with confidential HIV reporting, the bulk of HIV infections among 13–19-year-olds reported in July 2000–June 2001 were among females (56%), a disproportionate percentage of them African-American. Most young women had acquired the virus through heterosexual intercourse.

Furthermore the gay epidemic had become far more a black gay epidemic; in one study HIV prevalence in black gay men in five urban centres was nearly 50%, five to ten times what it was in white men. In the 33 areas with names-based reporting 46% of new cases of HIV in 2003 were in whites, 35% in blacks and 17% in Hispanics (with the other two per cent is Asians, Pacific Islanders and Native Americans), but the 18 areas excluded from these figures include urban areas with high black populations.

The ethnicity of people with AIDS has changed so that in 1985 60% of PWAs where white, 25% black and about 15% Hispanic. By 1994 AIDS cases in blacks were equalling those in whites, and in 2002 about 30% of AIDS cases were in whites, over 50% in blacks and about 20% in Hispanics. ~In 2002 the death rate due to AIDS was twice as high in blacks as in whites.

In a sense the USA is seeing a whole new epidemic of HIV superimposed on top of the old one; not an imported epidemic from high-prevalence countries as in the UK and other European countries, but an epidemic among the ‘third world in the first’ of impoverished black US citizens, many with multiple HIV risk factors and with limited access to adequate healthcare.

HIV reporting in the United States is not universal (though AIDS reporting is); 18 states have anonymised testing systems which prevent reporting, and in none of these states does HIV reporting extend back as far as 1985. Data on HIV infection are thus highly incomplete. However most states are now moving towards names-based reporting.

1996 was the first year in which the US reported a fall in the incidence of AIDS since the epidemic first appeared. The fall is attributed to improvements in treatment. The number of deaths declined 12% from 24,900 in the first six months of 1995 to 22,000 in the first half of 1996. During the first half of 1997 AIDS diagnoses fell by a further 15% compared with the first half of 1996.

40,000 Americans died of AIDS in 1993, the highest number in any year of the epidemic. However, there are significant variations between racial groups and between genders. Whilst deaths declined by 21% amongst whites in 1996, they declined only 2% amongst African/ Americans, and deaths have increased by 3% amongst women. In part these variations can be explained by the differing rates at which HIV spread amongst these population groups during the 1980s, but they are also a reflection of access to, and uptake of, treatment.

Death rates in the major cities declined further than the national average. New York, San Francisco, Los Angeles, and Miami all reported a decline of 20% in the death rate in the first half of 1996 compared to 1995.

According to a 2002 CDC report, AIDS-related illnesses remained the leading cause of death for African-American men aged 25–44 and the third-leading cause of death for Hispanic men in the same age group.

HIV prevalence levels are exceptionally high among African-American men who have sex with men in the USA—up to 48% among 23–29-year-olds, according to one six-city survey. About 64% of the women diagnosed with HIV in 2001 in the United States were African-American. A significant number of these women acquired the virus from men who also have sex with men.

Surveys, regularly conducted among several thousand gay men in San Francisco, have identified a continuing increase in the proportion of men who are reporting multiple partners and unprotected sex over the past six months. The San Francisco survey reveals that risk behaviour has increased most rapidly among men under 25. The survey identified an increase in the percentage of men reporting unprotected sex and multiple partners from 22% in 1994 to 32% in 1997. The prevalence of HIV among men presenting for voluntary counselling and testing almost tripled from 1.3% in 1997 to 3.7% in 1999.

Since then unprotected sex levels have stabilised, but at a much higher rate, with around a third of HIV-negative men reporting unprotected sex and about 50% of HIV-positive men. Despite this, HIV incidence has actually fallen in some areas such as San Francisco (halving from 2.4% to 1.2% a year) despite continuing high levels of STIs. This appears to be because gay men, when they have unprotected sex, are increasingly finding ways of doing it solely with men of their own HIV status, the amount of serodiscordant (different-status) unprotected sex has decreased since 2001 from 30% to 20% in HIV positive men and from 20% to 5% in HIV-negative men.

Nonetheless, unprotected sex will still expose men to STIs, and rectal gonorrhoea among gay men had been declining until 1993, but since that time has been steadily increasing, in line with the reported increase in unprotected anal sex. In New York, it is estimated that 9 out of 10 cases of HIV acquired through sex between men and women are related to sex with a drug user.

In Canada, meanwhile, aboriginal persons accounted for 9% of new HIV infections in 1999, although they constituted less than 3% of the general population.

In Canada there are an estimated 55,000 people living with HIV/AIDS, or 0.3% of the adult population.