In the United Kingdom (UK), men who have sex with men (MSM) and heterosexual men and women accounted for most persons newly diagnosed as having HIV in 2011.1 Of 6250 persons diagnosed as having HIV in the UK in that year, an estimated 3010 were MSM and 2950 were heterosexual men and women.1 By way of comparison, only 140 were people who inject drugs, and 150 were people who acquired HIV through other routes (e.g., mother to child transmission).1
Compared with those acquiring HIV through other routes, heterosexual men and women in the UK are more likely to remain undiagnosed (2011: 27%) and to be diagnosed beyond the point at which treatment should have started (2011: 60%).2 These factors increase the risk of onward transmission3 and mortality.4 It is critical, therefore, that HIV testing and care services are both accessible and acceptable to people at risk for heterosexually acquired HIV in the UK.
In this article, we describe trends in HIV diagnoses among heterosexual adults (≥15 years) in England, Wales, and Northern Ireland (E,W&NI) between 1992 and 2011 using national surveillance data. To provide context for these trends, the number of HIV diagnoses among MSM are presented for the period 1992 to 2011. As measures of prompt HIV testing and quality of care, we also examine late HIV diagnoses, first AIDS diagnosis, prompt integration into HIV care, uptake of antiretroviral therapy (ART), and short-term mortality, for the decade 2002 to 2011. In addition, we examine social deprivation, as measured by the Index of Multiple Deprivation (IMD) score5 associated with participant postcode.
Information on heterosexual adults (≥15 years) first as having HIV in E,W&NI between 1992 and 2011 (referred to as newly diagnosed), or seen for HIV care in 2011, was obtained from the national HIV and AIDS Reporting System (HARS) held at Public Health England.6 The number of MSM diagnosed between 1992 and 2011 was also obtained from HARS.
There are 3 national surveillance components to HARS. The HIV and AIDS New Diagnoses Database7 collects clinical and risk factor information on all adults newly diagnosed as having HIV in E,W&NI. The Survey of Prevalent HIV Infections Diagnosed collects clinical and risk factor information on all adults seen for HIV care in NHS sites in E,W&NI on an annual basis.8 Information on CD4 cell counts is supplemented by laboratory reports.9
Strict attention to confidentiality is maintained at every stage of data collection, analysis, and storage. No names are collected or held in HARS. A soundex code (a 4-character coding of the surname),10 together with sex and date of birth, are reported for each person, providing a unique identifier. To reduce the potential of underreporting, reporting sites that either have not reported or have reported fewer patients than might be expected based on previous reports are contacted. Reports received by the 3 HARS components are linked using a person’s unique identifier to create a single record for an HIV-diagnosed adult and identify underreporting to a single component.
The first (but not subsequent) diagnosis of an AIDS defining illness and deaths from any cause among HIV-diagnosed adults are reported directly to HARS by clinicians. Death reports are also obtained through data linkage to the death register of the Office for National Statistics (ONS).11 The Office for National Statistics provides a data file containing all deaths with a known HIV and/or AIDS-related cause in all age-groups, as well as a data file of all causes of deaths (coded using ICD10 codes) reported in persons 65 years and younger at the time of death.
Overall trends in new HIV diagnoses are presented for the 20-year period, 1992 to 2011. In 2002, a new clinic HIV report form was introduced to obtain more detailed information and to improve data completion. Consequently, the analysis of late HIV diagnosis, AIDS, prompt integration into care, uptake of ART, and deaths is restricted to the 10 year period, 2002 to 2011. Analysis of heterosexual adults seen for HIV care focuses on 2011, the last year for which data are available. All data are as reported to end June 2012.
Ethnicity was classified as white, black African, black Caribbean, or other (includes black other, Indian, Pakistani and Bangladeshi, other/mixed, and other Asian). A CD4 cell count within 91 days of diagnosis was defined as a CD4 cell count at diagnosis. Probable country of infection was assigned in 2 stages. Among adults born abroad, for whom a CD4 cell count at diagnosis and year of arrival in the UK were available, assignment was based on a published algorithm.12 In summary, the algorithm assigns an adult born abroad as having probably acquired HIV while living in the UK if their estimated year of infection is later than their year of arrival.12 Among all other adults, assignment was based on clinic HIV reports where clinicians or health advisors indicate whether the patient was presumed to have been infected in the UK or abroad.
Areas of residence and care were grouped into London, rest-of-England (roE), and Wales and Northern Ireland (W&NI). Based on their postcode of residence, adults residing in England were assigned an IMD score (part of the English Indices of Deprivation 2010)5 and ranked between 1 (most deprived) and 32,482 (least deprived). Postcodes ranked in the bottom third were defined as being in a deprived area.
A late HIV diagnosis was defined as having a CD4 cell count less than 350 cells/mm3 at diagnosis (only those who had a CD4 cell count measured within 91 days of diagnosis could be classified); prompt integration into care was defined as having a CD4 cell count measured within 28 days of HIV diagnosis; short-term mortality was defined as death from any cause within 1 year of HIV diagnosis; uptake of ART was defined as being in receipt of treatment within 1 year of HIV diagnosis. The definition for AIDS was having a first AIDS defining illness reported to HARS at the time of, or after an HIV diagnosis. Variables found to be significantly associated with these outcomes in univariate analysis were included in multivariate logistic regression models.
In univariate and multivariate analysis, a late HIV diagnosis was not included as an explanatory variable for AIDS as both are markers of severe immunosuppression; a late HIV diagnosis was not included as an explanatory variable for prompt integration into care because both are defined by having a CD4 cell count measured soon after HIV diagnosis. Prompt integration into care was not included as an explanatory variable for AIDS because these 2 variables were highly correlated (P < 0.01). Because of interaction (P < 0.01), country of birth and probable country of infection were combined for multivariate analysis.
Descriptive analyses were conducted using Mann-Whitney tests (when comparing across 2 groups) or Kruskal-Wallis tests (when comparing across ≥3 groups). Rates were compared using Pearson χ2 test. Cochran-Armitage χ2 trend tests were used for time trend analyses. Testing values and all confidence intervals (CI) were at the 99% significant level. Proportions were calculated based on available data. STATA 12.0 (Stata Corp, College Station, TX) was used for analyses.
Between 1992 and 2011, 50,791 heterosexual adults were newly diagnosed as having HIV in E,W&NI (women: 61% [31,229]; men: 39% [19,562]). Among these adults, year of diagnosis, sex, and age at diagnosis were available for all and ethnicity for 99% (50,282).
The annual number of heterosexual adults as having HIV increased rapidly from 731 in 1992 to a peak of 4676 in 2004, before declining to 2631 in 2011 (Fig. 1A). In contrast, the number of MSM as having HIV increased steadily year on year from 1729 in 1992 to 2618 in 2011 (Fig. 1A). Despite the decline in new diagnoses, since 2001, heterosexual adults accounted for 49% (2631/5423) of all adults newly diagnosed as having HIV in E,W&NI.
In all years, the largest number of HIV diagnoses among heterosexual adults was among women (Fig. 1B) and those of black African ethnicity (Fig. 2A). In recent years, HIV diagnoses have declined among both black African women and men (Fig. 2A). Despite this decline, black African adults accounted for 59% (1518/2570) of heterosexuals newly diagnosed as having HIV in 2011. Over the study period, median age at HIV diagnosis increased among men from 33 to 41 years, and among women from 28 to 36 years (both P < 0.01), and increased among all sex and ethnicity groups (Fig. 2B).
Over the 20-year period, 16% (8158) of heterosexual adults diagnosed as having HIV had an AIDS diagnosis and 7.2% (3,679) died. The number of heterosexual adults with an AIDS diagnosis decreased from 311 in 1992 to 185 in 2011. The average annual number of heterosexual adults who died declined from 209 between 1992 and 2005 to 127 between 2006 and 2011. Both the decline in AIDS diagnoses and deaths were largely explained by steep declines among black African adults (Fig. 2C and D).
During the period 2002 to 2011 (the period for which there was additional clinical and risk information), 38,228 heterosexual adults were newly diagnosed as having HIV. Among these adults, information on subsequent HIV care was available for 95% (36,467), country of birth for 83% (31,668), and CD4 cell count at diagnosis for 81% (31,072). Probable country of infection was available for 92% (35,113) of adults (17,233 assigned according to a published algorithm12 and 17,880 according to clinic HIV reports).
Of heterosexual adults newly diagnosed as having HIV between 2002 and 2011, 72% (27,384/37,984) were black African, of whom 99% (22,524/22,860) were born abroad. The decline in diagnoses among black African adults (Fig. 2A) was largely explained by an overall decline among adults born abroad, in particular among those born in Eastern Africa (Table 1). At the country level, the largest decline was observed among adults originating from Zimbabwe (from a peak of 946 in 2003 to 241 in 2011 among women; from a peak of 466 in 2003 to 134 in 2011 among men).
The percentage of heterosexual adults probably acquiring HIV in the UK increased from 15% in 2002 to 44% in 2011 (Table 1). Restricting the analysis to black African adults, this increase was from 9.1% (276/3019) in 2002 to 37% (444/1202) in 2011 (P < 0.01).
Late Diagnosis of HIV and AIDS
Of heterosexual adults as having HIV during the decade 2002 to 2011, 64% were diagnosed late, declining from 66% in 2002 to 61% in 2011 (P < 0.01; Table 2). In multivariate analysis, a late HIV diagnosis was associated with being male, of older age at diagnosis, of nonwhite ethnicity, and being born and probably acquiring HIV abroad (all, P < 0.01; Table 3). Late HIV diagnosis was not associated with year of diagnosis in multivariate analysis.
During this decade, 12% of heterosexual adults as having HIV were reported to have AIDS; this percentage decreased between 2002 and 2011 (P < 0.01; Table 2). Most of these adults were diagnosed late with HIV (93%; 3781/4048) and were diagnosed as having AIDS within 91 days of their HIV diagnosis (87%; 4109/4722).
Among the 4722 adults with AIDS, almost half (46%) had either pneumocystis pneumonia (1111) or pulmonary tuberculosis (1083) reported as the primary AIDS defining illness. In multivariate analysis, AIDS was associated with earlier year of HIV diagnosis, being male, being of older age at HIV diagnosis, and being born and probably acquiring HIV abroad (all, P < 0.01) (Table 3).
Prompt Integration Into Care and ART Within 1 Year of HIV Diagnosis
Sixty-eight percent of heterosexual adults diagnosed as having HIV between 2002 and 2011 were integrated into care within 28 days, and half were in receipt of ART within 1 year of their HIV diagnosis (Table 2). Both percentages increased over the period (both, P < 0.01; Table 2). In multivariate analysis, prompt integration into care was associated with more recent year of HIV diagnosis, older age at diagnosis, and being of white ethnicity (all, P < 0.01; Table 3). Uptake of ART was associated with more recent year of HIV diagnosis, being female, being of older age at diagnosis, and being diagnosed late with HIV (all, P < 0.01; Table 3).
Of heterosexual adults as having HIV between 2002 and 2011, 1713 (4.5%) subsequently died within the study period. Short-term mortality (died within a year of HIV diagnosis) accounted for almost two-thirds of these deaths (64%; 1102). Among the adults who died, 87% (1101/1272) were diagnosed late with HIV, increasing to 95% (740/783) among those with short-term mortality.
In multivariate analysis, short-term mortality was strongly associated with late HIV diagnosis and not receiving ART within 1 year of HIV diagnosis. There was also an increased likelihood of short-term mortality among men and adults of older age at HIV diagnosis (P < 0.01; Table 3).
Applying the short-term mortality rate of 1.8% among those in receipt of ART within 1 year of diagnosis (Table 2) to all heterosexual adults would reduce short-term mortality from 1102 deaths to 688 deaths. Applying the rate of 0.4% among those diagnosed as having a CD4 cell count of at least 350 cells/mm3 (Table 2) would reduce the figure even further from 1102 to 153 deaths.
Heterosexual Adults Seen for HIV Care in 2011
In 2011, 34,822 heterosexual adults were seen for HIV care in E,W&NI, accounting for 52% (34,822/67,538) of all adults seen for care that year. Sex, age, and site of care were available for all, with ethnicity and area of residence available for more than 99% (34,680 and 34,658, respectively). An IMD score was assigned to 97% (32,989/33,841) of adults residing in England.
Black African women accounted for 44% (15,367/34,680) of heterosexual adults seen for care in 2011, whereas black African men accounted for a further 21% (7151). Median age of those receiving HIV care in 2011 was higher among men (44 years) compared with women (40 years; P < 0.01); overall, 1 in 5 (20%; 6,963) were 50 years or older.
Forty-one percent (14,275) of heterosexual adults seen for care in 2011 resided in London, 56% (19,566) in the roE, and 2% (817) in W&NI. There was concordance between area of care and area of residence for 95% (32,873/34,658) of adults. Among the 1785 adults seen for HIV care outside their area of residence, 85% (1,509) resided in the roE and were seeking care in London. Overall, 65% (21,361/32,989) of adults lived in a deprived area. This percentage was higher in London (73%; 10,445/14,252) than in the roE (58%; 10,916/18,730) and higher among black Caribbeans (76%; 1105/1453), black Africans (70%; 15,034/21,514), or adults of “other” ethnicity (61%; 1937/3200) compared with those of white ethnicity (48%; 3191/6688; all, P < 0.01).
An increase in new HIV diagnoses among heterosexual adults in E,W&NI between 1992 and 2004 was followed by a steady decline to 2011. Despite this recent decline, the number of HIV diagnoses among heterosexual adults in 2011 was similar to that among MSM (2950 and 3010, respectively).
In all years, black African adults, most of whom were born abroad and acquired HIV abroad, accounted for most of HIV diagnoses among heterosexuals in E,W&NI. However, the percentage of black African adults probably acquiring HIV in the UK increased significantly from less than 1 in 10 in 2002 to more than 1 in 3 in 2011. Rather than reflecting changes in HIV transmission within the UK, the overall rise and subsequent decline in HIV diagnoses among heterosexuals were most likely due to changes in the number of people with HIV arriving in the UK from countries in sub-Saharan Africa.
Close historical, economic, and sociopolitical ties, often strengthened through the Commonwealth of Nations,13 have facilitated the movement of people from sub-Saharan African countries to the UK. This is true for Zimbabwe, a country with a high prevalence of HIV14 and a country from where many heterosexual adults diagnosed as having HIV in E,W&NI originated. Migration from Zimbabwe to the UK has declined since the introduction in November 2002 of stricter requirements for its citizens to travel to the UK15 and its withdrawal from the Commonwealth of Nations in December 2003.16
Among heterosexual adults newly diagnosed as having HIV between 2002 and 2011, the percentage with AIDS declined, whereas the percentage promptly integrated into care or starting ART within 1 year of HIV diagnosis increased. These trends provide evidence of improvements in clinical care over a 10-year period for patients once they are diagnosed as having HIV. Over the study period, we also reported median age at HIV diagnosis to have increased among all sex and ethnicity groups. Further analyses are warranted to explain this increase. Although there were differences between ethnic groups in prompt integration into care, we observed no significant disparities in uptake of ART by ethnicity.
In E,W&NI, heterosexual adults remain the group most likely to be diagnosed late with HIV (i.e., beyond the point at which ART should have started),2 and we report particularly elevated rates of late diagnosis among black African adults and/or those probably infected abroad. Heterosexual adults and/or migrants have also been reported as being most at risk for a late HIV diagnosis in Belgium and France.17,18 Reducing the percentage of heterosexual adults diagnosed late would contribute to a further decline in AIDS diagnoses and deaths. More than 9 in 10 adults with AIDS were diagnosed late, and adults diagnosed late were almost 10 times more likely to die within 1 year of diagnosis than those diagnosed as having a higher CD4 cell count.
Although overall deaths rates have significantly declined among heterosexual adults with diagnosed HIV since the introduction of ART,19 short-term mortality has remained stable over the past decade in our analysis. Late diagnosis was the strongest predictor of short-term mortality. Although the rate of late diagnosis among heterosexual adults decreased slightly over the study period, this change was not statistically significant. The rate of late diagnosis among heterosexual adults in the UK remains unacceptably high. If all heterosexual adults had been diagnosed promptly with a CD4 cell count of at least 350 cells/mm3 over the last decade, we estimate that nearly 1000 deaths that occurred within 1 year of diagnosis would have been prevented.
In 2002, it was reported that many Africans accessing HIV services in London lived in poor socioeconomic circumstances.20 It seems that little has changed in 10 years. We report that most heterosexual adults seen for HIV care in 2011 were living in a deprived area. Black African and black Caribbean adults were more likely to be living in a deprived area than white adults.
Our study has some limitations. Our definition of heterosexual is based on probable route of infection as reported by clinics, and there is potential for misclassification. For example, some adults who acquire HIV via other routes may report heterosexual contact instead. There is also potential for misclassification of probable country of infection due to, for example, persons acquiring HIV while traveling abroad from the UK. Limitations in assigning country of infection have been discussed elsewhere.12
It is unlikely that underreporting has a major impact on our results. Reports of new HIV diagnoses are received and followed up from multiple sources (e.g., primary care, sexual health clinics, and diagnostic laboratories) and are cross checked with persons attending HIV care and reported to the Survey of Prevalent HIV Infections Diagnosed (SOPHID). Reporting to SOPHID is linked to the local commissioning of HIV services and underreporting could result in the underfunding of services in subsequent years. Previous analyses have highlighted comprehensive reporting to national HIV surveillance.21 In this article, data completion exceeded 90% for all variables, except country of birth (83%) and CD4 cell count at diagnosis (81%). A missing CD4 count may be due to either no count having been taken or a count having been taken but not reported. If it is the latter, we may have underestimated prompt integration into care.
To identify potential differences between groups, we applied a strict definition of prompt integration based on having a CD4 cell count measured within 28 days of HIV diagnosis. Lengthening this period would increase the percentage of adults promptly integrated. Among all adults newly diagnosed as having HIV in 2011, 97% had a CD4 cell count measured within 3 months of diagnosis.2
In this article, we have highlighted the impact of migration on the epidemiology of heterosexually acquired HIV in the UK and provided evidence of improvements in postdiagnosis clinical care. We have also shown continuing high rates of late HIV diagnosis and increased transmission within the UK among heterosexual adults, particularly black African adults. These findings underscore the need for targeted HIV prevention strategies and the routine offer of HIV testing to black Africans living in the UK, in a range of health care and community settings as recommended by national guidelines.22,23