Women with HIV infection have an increased risk of cervical cytologic abnormalities1and cervical cancer2,3 compared to women without HIV infection. An American study containing almost 20,000 women with HIV infection found a standardised incidence ratio for cervical cancer (observed incidence of cervical cancer in women with HIV infection divided by expected incidence of cervical cancer based on population rates) of 2.9 (95%CI 1.9-4.2).2The current New Zealand and American recommendations are that women with HIV infection receive cervical screening when their HIV infection is diagnosed, 6 months later if the initial screen is normal and then yearly if the second screen is normal.4,5The Infectious Diseases and Sexual Health Services at Auckland City Hospital provide secondary level care for all women with HIV infection in the Auckland and Northland regions.We aimed to review our current cohort of women with HIV infection in the Auckland and Northland regions to document the number of women who had received a yearly cervical smear since their diagnosis of HIV infection and the number of women who were likely to have had undiagnosed HIV infection at the time of their first abnormal cervical smear.Method This audit was a retrospective review of the cervical smear history of all adult women (≥16 years) with HIV infection who were under active follow up by the Infectious Diseases and Sexual Health Services at Auckland City Hospital on 31 December 2007. We excluded women younger than 20 or older than 69 years of age on 31 December 2007 as routine cervical screening is not recommended for women of these ages by the National Cervical Screening Unit.4 Data including demographics, date of diagnosis of HIV infection, duration of follow-up, and whether the need for a yearly cervical smear was documented in one of the first two clinic letters sent to the woman's general practitioner (GP) were collected from clinical records at Auckland City Hospital. The cervical smear history, results and the person(s) performing the cervical smear(s) for each woman were obtained from the National Cervical Screening Unit. Cervical smears were defined as being taken yearly if the following criteria were met. The number of cervical smears performed divided by the number of years since the diagnosis of HIV infection was ≥0.8 and there were no gaps between cervical smears of ≥2 years, i.e. 4 cervical smears performed in a 5-year period without a gap of ≥2 years would meet our definition of a yearly cervical smear. For those women who were diagnosed with HIV infection before their arrival in New Zealand, we used the number of years they had resided in New Zealand instead of the number of years since the diagnosis of HIV infection. This audit received approval from the Northern X Regional Ethics Committee. The two-tailed Fisher's exact test was used to calculate univariate p values. The two-tailed Mann-Whitney test was used to calculate p values for age, time since diagnosis and duration of follow-up. Results At 31 December 2007, 128 adult women with HIV infection were under active follow-up by the Infectious Diseases or Sexual Health Services at Auckland City Hospital. Five women were excluded as they were younger than 20 or older than 69 years of age. The remaining 123 women form the basis of this audit. 113 of these women were under the care of the Infectious Diseases Service, 8 were under the care of the Sexual Health Service and 2 had been transferred from the care of the Sexual Health Service to the Infectious Diseases Service. The median age was 38 (range 22 to 59) years. The self-reported ethnicity was recorded as sub-Saharan African (n=69), Asian/South-East Asian (n=19), New Zealand European (n=18), European (n=9), Pacific Island Person (n=4), Māori (n=3) and other (n=1). Twenty-five women required an interpreter when they were seen in clinic. These women had been diagnosed with HIV infection for a median of 5 (range 0 to 22) years. They had been under the care of the Infectious Diseases or Sexual Health Services at Auckland City Hospital for a median of 4 (range 0 to 16) years. The median CD4 count at diagnosis (available for 112 women) was 310 (range 3 to 877) cells/mm3. The median CD4 count from the most recent test prior to 31 December 2007 (available for all women) was 448 (range 100 to 1341) cells/mm3. Eighty-three (67%) women were receiving antiretroviral treatment on 31 December 2007. These women had been on antiretroviral treatment for a median of 3 (range 0 to 18) years. Sixty-seven of the 83 (81%) women receiving antiretroviral treatment had an undetectable viral load on the most recent test prior to 31 December 2007. 120 (98%) women had had at least one cervical smear since 1991. Sixty-nine (56%) women met the definition for yearly cervical smears. Table 1 shows a number of parameters for women who did or did not meet the definition of a yearly cervical smear. The proportion of women who received a cervical smear for each yearly period from 2003 to 2007 is shown in Table 2. Table 1. Parameters for women who did or did not meet the definition of a yearly cervical smear Variables Yearly cervical smear P value Yes (n=69) No (n=54) Age (years) Median (range) 38 (22-59) 39 (24-56) 0.31 Ethnicity Sub-Saharan African 36 33 0.36 Asian/South-East Asian 12 7 0.62 New Zealand European 11 7 0.80 European 5 4 1 Pacific Island Person 2 2 1 Māori 2 1 1 Other 1 0 1 Required interpreter 15 10 0.82 Time since diagnosis (years) Median (range) 4 (0-18) 7 (0-22) 0.01 Duration of follow up (years) Median (range) 3 (0-15) 5 (0-16) 0.02 Those prescribed antiretroviral treatment 46 (67%) 37 (69%) 0.85 Those where one of the first two clinic letters discussed the need for a yearly cervical smear 15 10 0.82 Cervical smears performed by GP 13 20 0.04 HIV nurse specialist 13 11 1 Gynaecology Service 8 0 0.009 Sexual Health Service 3 0 0.26 Mangere Refugee Centre 0 1 0.44 Combination including Gynaecology Service 22 9 0.06 Other combination 8 7 1 Unknown 2 3 0.65 Those with at least one abnormal cervical smear 41 12 0.0001 The need for yearly cervical smears was documented in one of the first two clinic letters of 25 (20%) women. The cervical smears were performed by a GP (n=33), an HIV nurse specialist (n=24), the Gynaecology Service (n=8), the Sexual Health Service (n=3), the Mangere Refugee Centre (n=1), a combination including the Gynaecology Service (n=31), another combination (n=15) and an unidentified clinician (n=5). Fifty-three (43%) women had one or more abnormal cervical smears. The most abnormal cervical smear for each woman was abnormal squamous cells of undetermined significance (ASCUS) (n=4), cervical intraepithelial neoplasia (CIN) I (n=27), CIN II (n=6), CIN II/III (n=9) and CIN III (n=7). One patient with a cervical smear that showed CINII/III was found to have a poorly differentiated squamous cell carcinoma requiring chemo/radiotherapy. Table 2. The proportion of women who received a cervical smear for each yearly period from 2003 to 2007 Year Number of women diagnosed with HIV infection at the beginning of each year Number of women who received a cervical smear Proportion of women who received a cervical smear (%) P value (using 2003 as the comparator) 2003 2004 2005 2006 2007 60 70 82 98 113 30 45 58 61 73 50 64 71 62 65 - 0.11 0.01 0.14 0.07 Eleven (9%) women had one or more abnormal cervical smears before they were diagnosed with HIV infection. Their details are shown in Table 3. Not all these women would have had HIV infection at the time of their first abnormal cervical smear. Taking into account the women's CD4 count at the time of the diagnosis of their HIV infection it is very likely that seven women (patients 4 to 9 and 11) had undiagnosed HIV infection at the time of their first abnormal cervical smear. It is likely that two women (patients 1 and 2) had HIV infection at the time of their first abnormal cervical smear. It is unlikely that one woman (patient 3) had HIV infection at the time of her first abnormal cervical smear and one woman (patient 10) is known to have contracted HIV infection just after her first abnormal cervical smear. The median time between the first abnormal cervical smear and the diagnosis of HIV infection for the women who were very likely to have had undiagnosed HIV infection at the time of their first abnormal cervical smear was 24 (range 3 to 44) months. The self-reported ethnicity of these women was recorded as sub-Saharan African (n=5), New Zealand European (n=1) and South-East Asian (n=1). Table 3. Women who had one or more abnormal cervical smears before they were diagnosed with HIV infection. Patient Ethnicity Age at time of abnormal cervical smear (years) Date of abnormal cervical smear (abnormality) Date of HIV diagnosis CD4 count at time of HIV diagnosis (cells/mm3) Time between abnormal cervical smear and diagnosis of HIV infection (months) 1 Māori 19 1997 (CIN I) 2006 22 108 2 NZE

Women with HIV infection have an increased risk of cervical cytologic abnormalities1and cervical cancer2,3 compared to women without HIV infection. An American study containing almost 20,000 women with HIV infection found a standardised incidence ratio for cervical cancer (observed incidence of cervical cancer in women with HIV infection divided by expected incidence of cervical cancer based on population rates) of 2.9 (95%CI 1.9-4.2).2The current New Zealand and American recommendations are that women with HIV infection receive cervical screening when their HIV infection is diagnosed, 6 months later if the initial screen is normal and then yearly if the second screen is normal.4,5The Infectious Diseases and Sexual Health Services at Auckland City Hospital provide secondary level care for all women with HIV infection in the Auckland and Northland regions.We aimed to review our current cohort of women with HIV infection in the Auckland and Northland regions to document the number of women who had received a yearly cervical smear since their diagnosis of HIV infection and the number of women who were likely to have had undiagnosed HIV infection at the time of their first abnormal cervical smear.Method This audit was a retrospective review of the cervical smear history of all adult women (≥16 years) with HIV infection who were under active follow up by the Infectious Diseases and Sexual Health Services at Auckland City Hospital on 31 December 2007. We excluded women younger than 20 or older than 69 years of age on 31 December 2007 as routine cervical screening is not recommended for women of these ages by the National Cervical Screening Unit.4 Data including demographics, date of diagnosis of HIV infection, duration of follow-up, and whether the need for a yearly cervical smear was documented in one of the first two clinic letters sent to the woman's general practitioner (GP) were collected from clinical records at Auckland City Hospital. The cervical smear history, results and the person(s) performing the cervical smear(s) for each woman were obtained from the National Cervical Screening Unit. Cervical smears were defined as being taken yearly if the following criteria were met. The number of cervical smears performed divided by the number of years since the diagnosis of HIV infection was ≥0.8 and there were no gaps between cervical smears of ≥2 years, i.e. 4 cervical smears performed in a 5-year period without a gap of ≥2 years would meet our definition of a yearly cervical smear. For those women who were diagnosed with HIV infection before their arrival in New Zealand, we used the number of years they had resided in New Zealand instead of the number of years since the diagnosis of HIV infection. This audit received approval from the Northern X Regional Ethics Committee. The two-tailed Fisher's exact test was used to calculate univariate p values. The two-tailed Mann-Whitney test was used to calculate p values for age, time since diagnosis and duration of follow-up. Results At 31 December 2007, 128 adult women with HIV infection were under active follow-up by the Infectious Diseases or Sexual Health Services at Auckland City Hospital. Five women were excluded as they were younger than 20 or older than 69 years of age. The remaining 123 women form the basis of this audit. 113 of these women were under the care of the Infectious Diseases Service, 8 were under the care of the Sexual Health Service and 2 had been transferred from the care of the Sexual Health Service to the Infectious Diseases Service. The median age was 38 (range 22 to 59) years. The self-reported ethnicity was recorded as sub-Saharan African (n=69), Asian/South-East Asian (n=19), New Zealand European (n=18), European (n=9), Pacific Island Person (n=4), Māori (n=3) and other (n=1). Twenty-five women required an interpreter when they were seen in clinic. These women had been diagnosed with HIV infection for a median of 5 (range 0 to 22) years. They had been under the care of the Infectious Diseases or Sexual Health Services at Auckland City Hospital for a median of 4 (range 0 to 16) years. The median CD4 count at diagnosis (available for 112 women) was 310 (range 3 to 877) cells/mm3. The median CD4 count from the most recent test prior to 31 December 2007 (available for all women) was 448 (range 100 to 1341) cells/mm3. Eighty-three (67%) women were receiving antiretroviral treatment on 31 December 2007. These women had been on antiretroviral treatment for a median of 3 (range 0 to 18) years. Sixty-seven of the 83 (81%) women receiving antiretroviral treatment had an undetectable viral load on the most recent test prior to 31 December 2007. 120 (98%) women had had at least one cervical smear since 1991. Sixty-nine (56%) women met the definition for yearly cervical smears. Table 1 shows a number of parameters for women who did or did not meet the definition of a yearly cervical smear. The proportion of women who received a cervical smear for each yearly period from 2003 to 2007 is shown in Table 2. Table 1. Parameters for women who did or did not meet the definition of a yearly cervical smear Variables Yearly cervical smear P value Yes (n=69) No (n=54) Age (years) Median (range) 38 (22-59) 39 (24-56) 0.31 Ethnicity Sub-Saharan African 36 33 0.36 Asian/South-East Asian 12 7 0.62 New Zealand European 11 7 0.80 European 5 4 1 Pacific Island Person 2 2 1 Māori 2 1 1 Other 1 0 1 Required interpreter 15 10 0.82 Time since diagnosis (years) Median (range) 4 (0-18) 7 (0-22) 0.01 Duration of follow up (years) Median (range) 3 (0-15) 5 (0-16) 0.02 Those prescribed antiretroviral treatment 46 (67%) 37 (69%) 0.85 Those where one of the first two clinic letters discussed the need for a yearly cervical smear 15 10 0.82 Cervical smears performed by GP 13 20 0.04 HIV nurse specialist 13 11 1 Gynaecology Service 8 0 0.009 Sexual Health Service 3 0 0.26 Mangere Refugee Centre 0 1 0.44 Combination including Gynaecology Service 22 9 0.06 Other combination 8 7 1 Unknown 2 3 0.65 Those with at least one abnormal cervical smear 41 12 0.0001 The need for yearly cervical smears was documented in one of the first two clinic letters of 25 (20%) women. The cervical smears were performed by a GP (n=33), an HIV nurse specialist (n=24), the Gynaecology Service (n=8), the Sexual Health Service (n=3), the Mangere Refugee Centre (n=1), a combination including the Gynaecology Service (n=31), another combination (n=15) and an unidentified clinician (n=5). Fifty-three (43%) women had one or more abnormal cervical smears. The most abnormal cervical smear for each woman was abnormal squamous cells of undetermined significance (ASCUS) (n=4), cervical intraepithelial neoplasia (CIN) I (n=27), CIN II (n=6), CIN II/III (n=9) and CIN III (n=7). One patient with a cervical smear that showed CINII/III was found to have a poorly differentiated squamous cell carcinoma requiring chemo/radiotherapy. Table 2. The proportion of women who received a cervical smear for each yearly period from 2003 to 2007 Year Number of women diagnosed with HIV infection at the beginning of each year Number of women who received a cervical smear Proportion of women who received a cervical smear (%) P value (using 2003 as the comparator) 2003 2004 2005 2006 2007 60 70 82 98 113 30 45 58 61 73 50 64 71 62 65 - 0.11 0.01 0.14 0.07 Eleven (9%) women had one or more abnormal cervical smears before they were diagnosed with HIV infection. Their details are shown in Table 3. Not all these women would have had HIV infection at the time of their first abnormal cervical smear. Taking into account the women's CD4 count at the time of the diagnosis of their HIV infection it is very likely that seven women (patients 4 to 9 and 11) had undiagnosed HIV infection at the time of their first abnormal cervical smear. It is likely that two women (patients 1 and 2) had HIV infection at the time of their first abnormal cervical smear. It is unlikely that one woman (patient 3) had HIV infection at the time of her first abnormal cervical smear and one woman (patient 10) is known to have contracted HIV infection just after her first abnormal cervical smear. The median time between the first abnormal cervical smear and the diagnosis of HIV infection for the women who were very likely to have had undiagnosed HIV infection at the time of their first abnormal cervical smear was 24 (range 3 to 44) months. The self-reported ethnicity of these women was recorded as sub-Saharan African (n=5), New Zealand European (n=1) and South-East Asian (n=1). Table 3. Women who had one or more abnormal cervical smears before they were diagnosed with HIV infection. Patient Ethnicity Age at time of abnormal cervical smear (years) Date of abnormal cervical smear (abnormality) Date of HIV diagnosis CD4 count at time of HIV diagnosis (cells/mm3) Time between abnormal cervical smear and diagnosis of HIV infection (months) 1 Māori 19 1997 (CIN I) 2006 22 108 2 NZE