Epidemiology & Social
Impact of HIV infection on invasive cervical cancer in Kenyan women
Gichangi, Peter B; Bwayo, Joba; Estambale, Bensona; De Vuyst, Hugob; Ojwang, Shadrack; Rogo, Khamac; Abwao, Henryd; Temmerman, Marleenb
From the Department of Obstetrics and Gynaecology, the aDepartment of Medical Microbiology, University of Nairobi, Kenya, the bInternational Centre for Reproductive Health, Ghent University, Belgium, cNairobi Oncology Centre, and the dRadiotherapy Unit, Kenyatta National Hospital, Kenya.
Correspondence to M. Temmerman, International Centre for Reproductive Health, University Hospital, De Pintelaan 185 P3, B-9000 Gent, Belgium.
Received: 4 October 2002; revised: 20 March 2003; accepted: 8 April 2003.
Objectives: To determine the association between invasive cervical cancer (ICC) and HIV infection in Kenyan women.
Study: design: Case-control, with ICC patients as cases, and women with uterine fibroids as controls.
Methods: Medical and socio-demographic data were collected from 367 ICC patients, and 226 women with fibroids. After informed consent, HIV testing was done.
Results: ICC patients were older than fibroid patients (48 versus 41 years; P < 0.001), with an HIV seroprevalence of 15% and 12% respectively (P > 0.05). However, cases younger than 35 years were 2.6-times more likely to be HIV positive than controls of similar age [35% versus 17%; odds ratio (OR), 2.6; P = 0.043]. ICC HIV-seropositive patients were, on average, 10 years younger than HIV-seronegative patients (40 versus 50 years; P < 0.001). Eighty per cent of HIV-seropositive and 77% of HIV-seronegative ICC patients were in FIGO stage IIb or above. However, the odds of having poorly differentiated tumours was three times higher for HIV-seropositive than for HIV-seronegative ICC patients (77% versus 52%; OR, 3.1; P = 0.038) after adjusting for histological cell type and clinical stage. Mean CD4 cell count was 833 × 106 cells/l in ICC and 1025 × 106 cells/l in fibroid patients (P = 0.001).
Conclusion: Young women with ICC were more often HIV infected than women with fibroids of the same age groups. HIV infection was associated with poor histological differentiation of the tumours. These findings suggest an accelerated clinical progression of premalignant cervical lesions to ICC in HIV-infected women.
Cervical cancer is now recognized as a sexually transmitted disease because of the causal role of a sexually transmitted aetiologic agent, human papilloma virus (HPV), and the association of cervical cancer with sexual behaviour . From 1973 to 1992, cervical cancer was the most common cause of sexually transmitted disease (STD)-related deaths among women in the USA with almost twice the number of deaths attributed to invasive cervical cancer (ICC) as to HIV infection . Risk factors for cervical cancer include early coitarche, multiple sex partners, smoking, history of STD and immunosuppression .
Although the association of HIV infection and premalignant cervical lesions is now well established [4–9], data on the role of HIV in ICC is scanty [7,10] or lacking [11–13]. The apparent association of HIV infection and ICC is of concern because in many areas with high incidence of cervical cancer, large numbers of women are also likely to be HIV infected, raising a theoretical possibility that HIV infection may increase the incidence of cervical cancer. Moreover, ICC and premalignant cervical lesions may increase transmission and acquisition of HIV infection. There are few studies from developing countries examining the relationship of HIV infection and ICC [14–17]. With limited resources for treatment of HIV infection, it is possible that HIV-infected women may die early before they develop ICC. However, if HIV infection causes rapid progression of premalignant cervical lesions [18–20], women with ICC may present earlier than usual [14,15] with possibly more advanced disease . It is a common clinical impression that HIV-infected women with cervical cancer present at earlier ages [21,22]. In Kenya, HIV infection and ICC are both very prevalent. This study was undertaken to determine the association between ICC and HIV infection in Kenyan women.
Materials and methods
A prospective case–control design was used. Patients were recruited from radiotherapy unit, obstetrics and gynaecology wards of Kenyatta National Hospital (KNH) from January 2000 to March 2002. KNH is a national referral and teaching hospital for the University of Nairobi. Patients accessing the hospital are self-referral from Nairobi city, and referred patients from the whole of Kenya. KNH is the only public facility offering radiotherapy treatment in Kenya. Cases were cervical cancer patients seeking radiotherapy treatment at KNH while controls were consecutive women with a clinical diagnosis of uterine fibroids confirmed by ultrasound. Over 98% of all eligible patients were recruited. All enrolled participants received HIV pre-test and post-test counselling. A blood sample was obtained for HIV testing. HIV screening was done using the enzyme-linked immunosorbent assay (ELISA) (Biochem Immuno Systems Kit, Montreal, Quebec, Canada) and positive samples were confirmed using double ELISA (Biotech Ltd, Cambridge, Ireland). Blood was obtained for CD4 cell count using flow cytometry (FAC scan, Becton-Dickson) from 137 consecutive ICC and 81 fibroid patients. The number of patients tested for CD4 cell count was limited because of limited funding. Study staff administered a structured questionnaire. Federation International Gynaecology Obstetrics (FIGO) 1999  clinical staging, histological subtypes and grading (scale of 1 to 3) were extracted from case records and histopathological reports. The pathologist and gynaecologists were not aware of the HIV serostatus of the patients.
Statistical analysis was carried out using SPSS version 10.0 (SPSS Inc. Chicago, Illinois, USA) statistical package. Fisher exact test and Yates corrected Chi-square testing was used to compare proportions. Differences between means were tested by t test. Odds ratio (OR) or adjusted OR (AOR) and their 95% confidence intervals (CI) were used to measure strength of associations. Multivariate logistic regression models included variables significant in univariate analysis or those that were thought to have biological influence on the dependent variable. Associations with two-sided P < 0.05 were considered statistically significant.
All study participants gave an informed consent. The KNH ethics and research committee and the University of Nairobi approved the study.
Three hundred and sixty-seven ICC patients, and 226 women with uterine fibroids were recruited. The seven ICC and four fibroid patients who declined to be enrolled were not different in terms of age or clinical stage from those enrolled in the study. ICC patients were older than fibroid patients (48 ± 12 versus 41 ± 9 years; P < 0.001). Forty-six of the 364 (12.6%) ICC and 59 of 226 (26.1%) fibroid patients were < 35 years old. Overall HIV seroprevalence was 15% in cases and 12% in controls (P > 0.05). This crude comparison is probably misleading due to confounding by age. After controlling for age, marital status, lifetime sex partners and past STD, the odds of HIV infection was twice as large in cases than in controls (AOR 2.0; 95%CI, 1.1–3.5; P = 0.016).
Table 1 shows selected characteristics of cases (ICC patients) and controls (fibroid patients). Out of 358 ICC patients, 101 (28.2%) had no formal education as compared with 18 of 223 (8.1%) fibroids patients (OR, 4.5; P < 0.001). For both ICC and fibroid patients, the median number of pregnancies was 4.0 with mean of 5.6 ± 3.3 pregnancies. A Mann–Whitney test comparing parity of ICC and fibroid patients found P < 0.001. Mean parity was 6.4 ± 3.3 for ICC and 4.2 ± 2.9 for fibroid patients (t test, P < 0.001). Two hundred and fifty-seven of 367 (68.9%) ICC patients had five or more children as compared with 35% (80/226) of the fibroid patients (P < 0.001). Mean number of lifetime sex partners was 2.32 ± 1.64 for ICC and 2.59 ± 1.94 for fibroid patients, P = 0.078. The median number of lifetime sex partners for the study population was two. There was no significant difference in the proportion of women with more than two lifetime sex partners or those reporting previous history of STD between cases and controls.
Seventy-four of 356 (20.8%) ICC patients had had a Papanicolaou smear in their lifetime as compared to 125 of 187 (66.8%) fibroid patients (P < 0.001). On multivariate analysis controlling for age, i.e., < 35 years, marital status, lack of education, parity, Papanicolaou smear testing, life time sex partners, history of STD and HIV, the odds of being HIV infected was three times as large for ICC patients 34 years or younger as compared with fibroid patients of similar age (AOR, 3.3). Controls were more likely to have had a Papanicolaou smear test in the past than cases (AOR, 5.8).
Among the patients with fibroids, HIV-seronegative and -seropositive women were similar in age, marital status, level of education, previous history of STD and parity (≤ 4 or > 4) (data not shown). HIV-seropositive women with ICC were, on average, about 10 years younger than ICC HIV-seronegative women (40 ± 10 versus 50 ± 12 years; P < 0.001).
Table 2 shows correlates of HIV infection in ICC patients. Sixteen out of 46 (34.8%) ICC HIV-seropositive patients were younger than 35 years old, as compared with 11.7% (37/316) of ICC HIV-seronegative patients (OR, 4.0; 95% CI, 2.0–8.1, P < 0.001). Two hundred and twenty-eight of 312 (73.1%) ICC HIV-seronegative patients as compared with 23 of 53 (43.4%) ICC HIV-seropositive patients have had more than four pregnancies, (P < 0.001). Sixteen of 51 (31.4%) HIV-seropositive as compared with 58 of 303 (19.1%) HIV-seronegative ICC patients had a Papanicolaou smear test in their lifetime (OR, 1.9; P = 0.047). There was no difference in marital status, level of education, history of STD and proportion of patients with more than two lifetime sex partners or mean number of sex partners between HIV-seropositive and -seronegative ICC patients. On multivariate analysis, including age (< 35 years), marital status, lack of education, parity, Papanicolaou smear testing, life time sex partners, history of STD, HIV seropositivity among ICC patients was significantly and independently associated with young age, < 35 years (AOR, 4.0) and lower parity, (four or fewer pregnancies) (AOR, 0.4).
Three hundred and forty-one out of 365 (93.4)% of the ICC patients were clinically staged. Overall, 10% were in FIGO clinical stage I, 43% stage II, 42% stage III and 6% stage IV. Twenty-six of 50 (52%) HIV-seropositive women, were in stage II, 42% in stage III, while for HIV-seronegative women, about equal proportions, [41.2% (120/291) and 41.6% (121/291)] were in stage II and III respectively. Twenty per cent (10/50) of the HIV-seropositive patients were in stage I–IIa as compared to 23% (67/283) of the HIV-seronegative patients (P > 0.05). FIGO clinical stage of ICC was not influenced by educational status or age of the patient (< 35 years) (Table 3). Three of 96 (3.1%) patients in stage IIb and above had CD4 cell count < 200 × 106cells/l as compared to 6% (2/32) of the patients in stage I–IIa, (OR, 2.1; P = 0.429).
Histological cell typing reports were available for 79.4% (290/365) of the ICC patients. Squamous cell histological subtype was the most common accounting for over 79% (230/290) of the patients with known histology. The degree of differentiation was indicated in 53.8% (156/290) histological biopsy reports. Twenty of 26 (76.9%) HIV-seropositive as compared with 67 of 130 (51.5%) HIV-seronegative ICC patients had poorly differentiated tumours even after controlling for histological cell type and clinical stage (AOR, 2.9; P = 0.038). Only 3.4% (2/58) of the poorly differentiated histological subtypes had CD4 cell count < 200 × 106cells/l as compared with 7% (2/30) of the patients with well differentiated tumours, (OR, 0.5; 95% CI, 0.1–3.7; P = 0.492).
ICC patients had significantly lower mean CD4 cell count than fibroid patients (883 ± 354 versus 1006 ± 454 × 106cells/l). Even after controlling for age, ICC patients had significantly lower CD4 cell count than fibroid patients. Overall, 3.6% (5/137) of the ICC patients had CD4 cell count < 200 × 106cells/l, 4.4% (6/137) 200–350 × 106cells/l, and 92% (126/137) > 350 × 106cells/l. Seventeen per cent (4/23) HIV-seropositive ICC patients had CD4 cell count < 200 × 106cells/l as compared with 0.9% (1/114) ICC HIV-seronegative patients. The number of HIV-seropositive cases was too small for further detailed stratified analysis. Five of the 81 (6%) fibroid patients had CD4 cell count < 200 × 106cells/l, 2.5% (2/81) 200–350 × 106cells/l, and 91.4% (74/81) > 350 × 106 cells/l (data not shown). Among HIV-seropositive patients, ICC and fibroid patients had similar CD4 cell counts (532 ± 320 × 106cells/l versus 588 ± 463 × 106cells/l; P = 0.673) while for HIV-seronegative patients, ICC patients had significantly lower CD4 cell count than fibroid patients (894 ± 331 × 106cells/l versus 1079 ± 415 × 106cells/l; P = 0.001).
Our data show that ICC HIV-seropositive patients are 10 years younger at presentation than ICC HIV-seronegative patients, even after controlling for education, number of sex partners and previous history of STD. This could suggest that HIV infection may shorten the progression from premalignant cervical lesions to ICC, resulting in earlier presentation. This confirms the common clinical impression that HIV-infected women with cervical cancer are younger as discussed in other studies [14,15,24]. Rapid progression from normal cervical cytology to squamous intraepithelial lesions (SIL) in HIV-infected women has been documented . Holcomb et al.  reported a case of rapid progression from carcinoma in situ to invasive cancer in an HIV-seropositive woman. There are other case reports of rapid progression of genital tract cancers in HIV-infected individuals [26,27]. On the other hand, the younger age at presentation of ICC HIV-positive patients could be related to altered health-seeking behaviour. There are currently no studies showing changes in health-seeking behaviour of cervical cancer patients with concurrent HIV infection. Another plausible explanation is that women who have earlier, more frequent sexual activity with more partners are at risk for both HPV and HIV, thus accounting for the younger age at presentation.
The overall HIV seroprevalence of 15% among women with ICC in Nairobi is lower than the 19% reported by Maiman et al.  among American women and 32% reported by Newton et al.  among cervical cancer patients in Uganda after adjusting for age and number of sex partners and 21% among South African women from Kwazulu Natal . In Gauteng province of South Africa, Lomalisa et al.  found HIV prevalence of 7.6% in cervical patients. These observed differences among various studies could be accounted for by differences in HIV seroprevalence in the general population, as shown by the two South African studies from sites with different background HIV seroprevalence, 32.5% in Kwazulu Natal  and 23.9% in Gauteng province , or by the degree of immunosuppression. Our finding of mean CD4 cell count of 588 × 106cells/l among HIV-infected women with ICC suggests that Kenyan cervical cancer patients are not as severely immunocompromised as South African women (mean CD4 cell count 316 × 106cells/l)  or cervical cancer patients in the USA (mean CD4 cell count 362 × 106cells/l) .
Our data does suggest a more rapid increase in HIV seroprevalence in ICC patients, than in the general population. The HIV seroprevalence in ICC patients has increased 10-fold from 1.5% in 1990  to 15% in this study, as compared to a threefold increase in HIV seroprevalence from 5% in 1990 to current level of 14–15% in 2002 among Kenyan women of reproductive age . Similar observations were reported from South African cervical cancer patients. Lomalisa et al.  observed doubling of HIV prevalence from 3.9% in 1995 to 7.2% in 2000, and Moodley et al.  reported quadrupling from 5 to 21%. ICC being an ulcerative condition may facilitate transmission and acquisition of HIV infection, which could account for the rapid increase in HIV prevalence in this population. Indeed, the odds of being HIV infected among ICC patients less than 35 years old was three times as large as that for women with uterine fibroids despite similar other risk factors for HIV transmission/acquisition in this study.
Outcome of ICC is dependent on FIGO clinical stage, age of the patient, tumour characteristics and treatment given. Five-year survival for stage I is 77–84%, stage II 54–67%, stage III 13–40% while for stage IV it is 5–13% . HIV infection was associated with moderate to poorly differentiated histological subtypes, which is a poor prognostic factor . Moodley et al.  reported that 37% of the HIV-seropositive patients had poorly differentiated tumours. HIV infection was not associated with advanced clinical stage at presentation in our study. Unlike in other studies [15,19,20], the lack of association of HIV infection and advanced stage of cervical cancer could be related to the relatively good immunostatus in our patients (mean CD4 cell count 532 × 106cells/l) or to lack of power since only about 37% (137/367) had CD4 cell count determination due to cost constraints. Advanced stages of HIV disease have been associated with severe immunosuppression [15,19,20].
Major limitations of the study include the was use of recall data, such as lifetime sex partners and past STD, which could not be validated, and the inclusion of hospital patients only. ICC patients accessing hospital facilities may be a biased population. The majority of ICC patients were from the neighbourhood of the hospital suggesting limited access for patients far from the hospital. Early death of HIV-infected women before they develop ICC, or before they come to hospital due to other HIV related opportunistic infections could have resulted in underestimating the HIV seroprevalence among ICC patients in our study. Power of the study to determine the association of immunosuppression with stage of cervical cancer was low due to financial limitations of CD4 cell count determination. Another limitation of the study was the use of fibroid patients as a control group. Indeed, fibroid patients were different from ICC in terms of younger age, less likely to be married, both risk factors for HIV infection, more educated and to have had a Papanicolaou smear in the past. Despite these differences, ICC and fibroid patients were similar in mean number of lifetime sex partners and previous history of STD. The foregoing suggests that fibroid patients may not be comparable to ICC patients, thus interpretation of these results has to be done with those differences in consideration.
In conclusion, our data provide evidence for a higher HIV prevalence in women with ICC as compared to women of the same age group with uterine fibroids. Although HIV infection per se did not confer an adverse effect on severity of ICC as assessed by FIGO clinical staging, it was associated with poor histological differentiation of the tumour, which is a poor prognostic factor for tumour spread and outcome of treatment and younger age at presentation. Further research is needed to examine the impact of the younger age at presentation of ICC HIV-seropositive patients and the poor histological differentiation as regards outcome of management of ICC.
The authors thank the director of Kenyatta National hospital for allowing access to patients, and the staff in radiotherapy and obstetrics and gynaecology wards. We acknowledge S. Vansteelandt, Department of Applied Mathematics and Informatics, Ghent University, Belgium for assistance in data analysis. Special thanks go R. Kilonzo, J. Mbithi, the study staff, and laboratory technicians in Department of Medical Microbiology, University of Nairobi.
Sponsorship: Supported by the VLIR (Flemish Interuniversity Council), Belgium.
1. Newcomb PA, Weiss NS, Daling JR. Incidence of vulvar carcinoma in relation to menstrual, reproductive, and medical factors. J Natl Cancer Inst
2. Ebrahim SH, Peterman TA, Zaidi AA, Kamb ML. Mortality related to sexually transmitted diseases in women, US, 1973 through 1992. Am J Public Health
3. Williams MA, Kenya PR, Mati JK, Thomas DB. Risk factors for invasive cervical cancer in Kenyan women. Int J Epidemiol
4. Maggwa BN, Hunter DJ, Mbugua S, Tikei P, Mati JK. The relationship between HIV infection and cervical intraepithelial neoplasia among women attending two family planning clinics in Nairobi, Kenya. AIDS
5. Miotti PG, Dallabetta GA, Daniel RW, Canner JK, Chiphagwi JD, Lioma GN, et al. Cervical abnormalities, Human Papillomavirus, and Human Immunodeficiency Virus Infections in women in Malawi. J Infect Dis
6. Ruche GL, Ramon R, Mensah-Ado I, Bergeron C, Diomande M, Sylla-Koko F, et al. squamous intraepithelial lesions of the cervix, invasive cervical carcinoma and immunosurpression induced by human immunodeficiency virus in Africa. Cancer
7. Serraino D, Carrieri P, Bidoli E, Dorruchi M, Ghetti E, Schiesari A, et al. Risk of invasive cervical cancer among women with, or at risk for, HIV infection. Int J Cancer
8. Temmerman M, Tyndall MW, Kidula N, Claeys P, Muchiri L, Quint W. Risk factors for human papillomavirus and cervical precancerous lesions, and the role of concurrent HIV-I infection. Int J Gynaecol Obstet
9. Wright TC, Koulos J, Schnoll F, Swanbeck J, Ellerbrock TV, Chiasson A, et al. Cervical intraepithelial neoplasia in women infected with the human immunodeficiency virus: outcome after loop electrosurgical excision. Gynecol Oncol
10. Fruchter RG, Maiman M, Sedlis A, Bartley L, Camilien L, Arrastia CD. Multiple recurrences of cervical intraepithelial neoplasia in women with the human immunodeficiency virus. Obstet Gynecol
11. Bassett MT, Chokunonga E, Mauchaza B, Levy L, Ferlay J, Parkin DD. Cancer in the African population of Harare, Zimbabwe, 1990–1992. Int J Cancer
12. Goedert JJ, Cote TR, Virgo P, Scoppa SM, Kingma DW, Gail MH, et al. Spectrum of AIDS-associated malignant disorders. Lancet
13. Rabkin CS, Biggar RJ, Baptiste MS, Abe T, Kohler BA, Nasca PC. Cancer incidence trends in women at high risk of human immunodeficiency virus (HIV) infection. Int J Cancer
14. Gichangi P, De Vuyst H, Estambale B, Rogo K, Bwayo J, Temmerman M. HIV and cervical cancer in Kenya. Int J Gynecol Obstet
15. Lomalisa P, Smith T, Guidozzi F. Human immunodeficiency virus infection and invasive cervical cancer in South Africa. Gynecol Oncol
16. Rogo KO. Human papillomavirus and human immunodeficiency virus infection in relation to cervical cancer: studies and observations of basic clinical and epidemiological aspects of cancer of the cervix with special references to Kenya, Africa.
Ph.D. thesis, University of Umea, Sweden; 1990.
17. Sitas F, Pacella-Norman R, Carrara H, Patel M, Ruff P, Sur R, J et al. The spectrum of HIV-1 related cancers in South Africa. Int J Cancer
18. Holcomb K, Maiman K, Dimaio T, Gates J. Rapid progression to invasive cervical cancer in a woman infected with HIV. Obstet Gynecol
19. Maiman M, Fruchter RG, Guy L, Cuthill S, Levine P, Serur E. Human immunodeficiency virus infection and invasive cervical carcinoma. Cancer
20. Six C, Heard I, Bergeron C, Orth G, Poveda JD, Zagury P, et al. Comparative prevalence, incidence and short-term prognosis of cervical squamous intraepithelial lesions amongst HIV-positive and HIV negative women. AIDS
21. Chokunonga E, Levy LM, Bassett MT, Borok MZ, Mauchaza BN, Chirenje MZ, Parkin DM. AIDs and cancer in Africa: the evolving epidemic in Zimbabwe. AIDS
22. Parkin DW, Wabinga H, Nambooze S, Wabwire-Mangen F. AIDS-related cancers in Africa: maturation of the epidemic in Uganda. AIDS
23. Pecorelli S, Benedet JL, Creasman WT, Shepherd JH. FIGO staging of gynecologic cancer. 1994-1997 FIGO Committee on Gynecologic Oncology. International Federation of Gynecology and Obstetrics. Int J Gynaecol Obstet
24. Moodley M, Moodley J, Kleinschmidt I. Invasive cervical cancer and human immunodeficiency virus (HIV) infection: a South African perspective. Int J Gynecol cancer
25. Ellerbrock TV, Chaison MA, Bush TJ, Sun XW, Swao D, Brundney K, et al. Incidence of cervical squamous intraepithelial lesions in HIV-infected women. JAMA
26. Giaquinto C, Del Mistro A, De Rossi A, Bertorelle R, Giacomet V, Ruga E, et al. Vulvar carcinoma in a 12-year-old girl with vertically acquired human immunodeficiency virus infection. Pediatrics
27. Lee Y-C, Holcomb K, Buhl A, Holden J, Abulafia O. Rapid progression of primary vaginal squamous cell carcinoma in a young HIV-infected woman. Gynecol Oncol
28. Newton R, Ziegler J, Beral V, Mbidde E, Carpenter L, Wabinga H, et al. A case-control study of human immunodeficiency virus infection and cancer in adults and children residing in Kampala, Uganda. Int J cancer
29. Ateka GK. HIV/AIDS. Global impact and human rights-A South African perspective. AIDS Scan
30. Rogo KO, Kavoo L. Human immunodeficiency virus seroprevalence among cervical cancer patients. Gyne Oncol
31. National AIDS/STDs Control Programme. AIDS in Kenya: Background, Projections, Impact and Interventions, 7th edn.
Nairobi, Kenya: Ministry of Health; 2002.
32. Shingleton HM, Bell MC, Fremgen A, Chmiel JS, Russell AH, Jones WB, Winchester DP, Clive RE. Is there really a difference in survival of women with squamous cell carcinoma, adnenocarcinoma and adenosquamous cell carcinoma of the cervix? Cancer
33. Chen RJ, Chang DY, Yen ML, Lee EF, Huang SC, Chow SN, et al
. Prognostic factors of primary adenocarcinoma of the uterine cervix. Gynecol Oncol
invasive cervical cancer; HIV; seroprevalence; CD4; Kenya
© 2003 Lippincott Williams & Wilkins, Inc.
Highlight selected keywords in the article text.