Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

The epidemiology of conjunctival squamous cell carcinoma in Uganda

The epidemiology of conjunctival squamous cell carcinoma in Uganda British Journal of Cancer (2002) 87, 301 – 308 ã 2002 Cancer Research UK All rights reserved 0007 – 0920/02 $25.00 www.bjcancer.com The epidemiology of conjunctival squamous cell carcinoma in Uganda ,1 2 2 3 1 1 2 4 R Newton* , J Ziegler , C Ateenyi-Agaba , L Bousarghin , D Casabonne , V Beral , E Mbidde , L Carpenter , 1 5 2 2 6 7 7 3 G Reeves , DM Parkin , H Wabinga , S Mbulaiteye , H Jaffe , D Bourboulia , C Boshoff , A Touze ´ , 3 8 P Coursaget and the Uganda Kaposi’s Sarcoma Study Group 1 2 Cancer Research UK, Epidemiology Unit, Gibson Building, Radcliffe Infirmary, Oxford OX2 6HE, UK; Uganda Cancer Institute and Makerere University Medical School, Kampala, Uganda; Laboratoire de Virologie Mole ´culaire, INSERM EMIU 00-10 and USC INRA, Faculte ´ de Pharmacie, 37200 Tours, 4 5 France; MRC Programme on AIDS, Uganda Virus Research Institute, PO Box 49, Entebbe, Uganda; International Agency for Research on Cancer, 150 6 7 Cours Albert-Thomas, Lyon, France; Centers for Disease Control and Prevention, 1600 Clifton Road, Atlanta, Georgia, GA 30333, USA; Wolfson Institute of Medical Research, University College London, 46 Cleveland Street, London, UK As part of a larger investigation of cancer in Uganda, we conducted a case – control study of conjunctival squamous cell carcinoma in adults presenting at hospitals in Kampala. Participants were interviewed about social and lifestyle factors and had blood tested for antibodies to HIV, KSHV and HPV-16, -18 and -45. The odds of each factor among 60 people with conjunctival cancer was compared to that among 1214 controls with other cancer sites or types, using odds ratios, estimated with unconditional logistic regression. Conjunctival cancer was associated with HIV infection (OR 10.1, 95% confidence intervals [CI] 5.2 – 19.4; P50.001), and was less common in those with a higher personal income (OR 0.4, 95% CI 0.3 – 1.2; P50.001). The risk of conjunctival cancer increased with increasing time spent in cultivation and therefore in direct sunlight 2 2 (w trend=3.9, P=0.05), but decreased with decreasing age at leaving home (w trend=3.9, P=0.05), perhaps reflecting less exposure to sunlight consequent to working in towns, although both results were of borderline statistical significance. To reduce confounding, sexual and reproductive variables were examined among HIV seropositive individuals only. Cases were more likely than controls to report that they had given or received gifts for sex (OR 3.5, 95% CI 1.2 – 10.4; P=0.03), but this may have been a chance finding as no other sexual or reproductive variable was associated with conjunctival cancer, including the number of self-reported lifetime sexual partners (P=0.4). The seroprevalence of antibodies against HPV-18 and -45 was too low to make reliable conclusions. The presence of anti-HPV-16 antibodies was not significantly associated with squamous cell carcinoma of the conjunctiva (OR 1.5, 95% CI 0.5 – 4.3; P=0.5) and nor were anti-KSHV antibodies (OR 0.9, 95% CI 0.4 – 2.1; P=0.8). The 10-fold increased risk of conjunctival cancer in HIV infected individuals is similar to results from other studies. The role of other oncogenic viral infections is unclear. British Journal of Cancer (2002) 87, 301 – 308. doi:10.1038/sj.bjc.6600451 www.bjcancer.com ã 2002 Cancer Research UK Keywords: conjunctival carcinoma; HIV; HPV; KSHV; HHV-8; Uganda Squamous cell carcinoma of the conjunctiva is thought to be an carcinoma of the conjunctiva, because the tumour was relatively extreme form of a spectrum of clinical conditions, collectively frequent there, even before the onset of the HIV epidemic known as ‘ocular surface squamous neoplasias’, which range in (Templeton, 1973; Wabinga et al, 2000). In this report, we examine severity from mild dysplasia, to carcinoma in situ and, ultimately, the association of conjunctival tumours with over 50 possible risk to invasive carcinoma. Symptoms can range from none, to severe factors, including evidence of infection with HIV-1, HPV-16, -18 pain and visual loss. Lesions generally arise on exposed areas of and -45 and Kaposi’s sarcoma-associated herpesvirus (KSHV; the eye, particularly on the nasal side, and treatment involves local human herpesvirus type 8 [HHV-8]), using data from a case- excision, or in more severe cases, orbital clearance. Metastases are control study of cancer in adults (Ziegler et al, 1997; Newton et rare and the prognosis is usually favourable. al, 2001). Although relatively rare everywhere, conjunctival carcinoma is more frequent in parts of sub-Saharan Africa. Uganda offers a good MATERIALS AND METHODS setting in which to investigate the epidemiology of squamous cell Study subjects *Correspondence: R Newton; E-mail: rob_newton@cancer.org.uk The subjects included in these analyses were selected from a large The Uganda Kaposi’s Sarcoma Study Group includes named authors and V cross-sectional study of risk factors for cancer in Uganda, which Sembajwe, M Kalinaki, R Byansi, C Rwatooro, S Nambooze, B Tushimiere, included, in total, 2091 individuals with malignancies other than N Byabazaire (deceased), E Bitamazire, E Katabira, J Mugerwa (deceased), Kaposi’s sarcoma. Between August 1994 and February 1998, any D Tindyebura, J Whitworth, B Richardson, R Weiss and K de Cock adults (aged 15 years and over) with a provisional new diagnosis Received 8 February 2002; revised 19 April 2002; accepted 22 April 2002 of incident cancer were eligible for recruitment into the study from Epidemiology Epidemiology Conjunctival cancer in Uganda R Newton et al all the wards and out-patient clinics of the four main hospitals in culaire in Tours, France, for HPV testing and to University College Kampala, Uganda: Mulago (including the Uganda Cancer Institute), London, UK, for KSHV testing. Nsambya, Mengo and Rubaga. Further details of the methods can be found elsewhere (Ziegler et al, 1997; Newton et al, 2001). The HPV-16, -18 and -45 All assays were performed by a single subjects included in this report are restricted to 60 cases of conjunc- investigator (LB), who was unaware of each patient’s personal char- tival cancer and 1214 controls. Thirty-two people with eye cancers acteristics, and diagnosis. HPV VLPs were produced in Sf21 insect that were not specifically diagnosed clinically as being squamous cell cells using recombinant baculoviruses encoding the L1 gene of carcinomas, or had a pathological diagnosis of malignancy of uncer- HPV-16, -18, and -45, according to previously described procedures tain morphology were excluded. The controls comprised people (Touze ´ et al; 1998; Combita et al, 2002a,b). These subtypes were with other incident cancers, excluding those with cancer sites or chosen because they are known to be prevalent in tumour speci- types that are known to be associated with infection with HIV, mens from women with cancer of the uterine cervix in Uganda HPV or KSHV, or exposure to solar ultraviolet radiation – that (Bosch et al, 1995). Briefly, the HPV L1 genes were first amplified is, Kaposi’s sarcoma, non-Hodgkin’s lymphoma, Hodgkin’s disease, from an HPV DNA-positive biopsy using primers containing BglII cancers of the uterine cervix, anus and penis, and skin cancers sites. The amplified product was then inserted into pFastBacI to (IARC, 1992, 1995; Beral and Newton, 1998). The control group generate HPV L1 VLPs. VLPs were produced according to a proce- included men and women with cancers of the oral cavity (57), oeso- dure used for HPV-16 VLPs (Touze et al, 1998; El Mehdaoui et al, phagus (150), stomach (74), liver (103), breast (178), ovary (67), 2000). Sf21 cells, maintained in Grace’s insect medium supplemen- prostate (56), and other cancer sites or types (405). In addition, ted with 10% foetal calf serum (FCS) were infected with the 124 patients with a provisional diagnosis of cancer, but who subse- different recombinant baculoviruses at a m.o.i. of 10 and were incu- quently turned out to have benign tumours, were also included in bated for 72 h at 278C. Cells were harvested by centrifugation, re- the control group. suspended in PBS containing 0.5% NP40 and allowed to stand at room temperature for 30 min. Cell lysates were then centrifuged at 14 0006g for 15 min at 48C. The nuclear fractions were further Recruitment and questionnaire re-suspended in ice cold PBS and sonicated by three 15 s bursts at Interviewers (who were also trained HIV counsellors) approached 60% maximal power (Vibra Cell, Bioblock Scientific, Strasbourg, ward or clinic staff to gain permission to interview potential France). Fractions were then loaded on the top of a preformed CsCl recruits. If permission was granted, the patient was approached gradient and centrifuged at equilibrium in a Beckman SW28 rotor by one of four interviewers and invited to participate in the study. (20 h, 27 000 r.p.m., 48C). Gradient fractions were analysed for A HIV test was requested and appropriate counselling was given. density by refractometry and were tested for the presence of L1 Patients who had been HIV tested within a month of recruitment protein by ELISA. Immunoreactive fractions were finally pooled and had a medical certificate indicating the test result, were not re- and pelleted by ultracentrifugation in a Beckman SW 28 rotor tested. The patient was interviewed about social and demographic (3 h, 28 000 r.p.m., 48C). VLPs were resuspended in PBS (pH 7.4) factors, and sexual and reproductive history. Where possible, the and protein content was evaluated using the microBCA kit (Pierce, patient was interviewed by a counsellor of the same sex, and in Touzart et Matignon, France). Each preparation was tested for the their native language. Blood was drawn, primarily for the HIV test, presence of VLPs by electron microscopy. For this purpose, VLP but also for storage of serum and leukocytes. Interviewers reported preparations were applied to 400-mesh carbon-coated grids, nega- the HIV test results back to the patient, together with post-test tively stained with 1.5% uranyl acetate and then examined at a counselling. The study was approved by the Committee on Human nominal magnification of 50 000 with a Jeol 1010 electron micro- Research (VA Medical Centre and University of California, San scope. VLPs were then diluted in PBS and used in the following Francisco, CA, USA) and by the Uganda National Council for ELISA tests (Touze ´ et al, 1998; Combita et al, 2002a). Science and Technology. Flat-bottomed wells of 96-well microplates (Maxisorp, Nunc, Life Technologies, Eragny, France) were coated overnight at 48C with 200 – 400 ng of VLPs (test well) or 200 ng of BSA (control well) in Laboratory diagnoses PBS, pH 7.4. Each serum sample was tested twice against each of Histology Diagnoses of cancer were established by histology or the seven VLP types and BSA at the same time on the same plate. other laboratory investigation, where possible. Diagnoses made After washing with PBS, 0.1% Tween 20 and 200 ml of PBS contain- on clinical grounds alone, were reviewed by the investigators (C ing 1% newborn bovine serum (NBS, Sigma, St Quentin Favallier, A-A, JZ, EM and SM). Among the cases, 36 (60%) underwent France) were added (2 h at 378C). The blocking solution was histological review (all were invasive tumours) and the remainder replaced by 100 ml of sera diluted 1 : 20 in 5xPBS-10% NBS and were diagnosed clinically. Among controls, 63% of tumour diag- 2% Tween 20, and plates were incubated at 458C for 60 min. After noses were verified by laboratory investigation (histology, four washes, bound antibodies were detected with a goat anti-human cytology, blood chemistry or ultrasound examination). IgG immunoglobulin (diluted 1 : 5000) conjugated to horseradish peroxidase (Sigma). Following incubation at 458C for 1 h and four washes, 100 ml of a substrate solution containing ortho-phenylene- HIV-1 HIV-1 serostatus was determined using a single ELISA, the Cambridge Bioscience Recombigen enzyme linked immunosor- diamine and H O was added. After 30 min incubation, the 2 2 bent assay (Cambridge, MA, USA). Periodic laboratory quality reaction was stopped by addition of 100 mlof 4N H SO , and optical 2 4 control assays using blinded standards from the United States densities (OD) were read at 492 nm with an automated plate reader Centers for Disease Control (Atlanta, GA, USA) revealed test sensi- (BioRad, model 550). For each serum sample the background reac- tivity and specificity of 99%. In the initial months of the study, tivity found in the BSA coated wells was subtracted from the OD 40 – 50% of adults declined to have venepuncture, mostly because found in each of the HPV – VLP coated wells. Negative values were they were too ill. Therefore, a GACELISA saliva test for HIV was adjusted to zero. The cut-off values for positivity to HPV-16, -18 introduced as an option (with sensitivity and specificity similar and -45 were set up at 0.2 (OD test well minus OD control well) to the blood test), and the refusal rate dropped to 10%. In total, and those with a value of 0.4 or greater were considered to have a HIV test results were available on 57 cases and 826 controls. high titre of anti-HPV antibodies. In total, 457 samples were available However, for one case and seven controls the result was indetermi- for testing for antibodies against HPV-16 (39 cases and 418 controls). nate. Any remaining aliquots of sera were stored at 7808C and In addition, there were sufficient sera for 453 of those to be tested for were later shipped on dry ice to the Laboratoire de Virologie Mole ´ - antibodies against HPV-18 and -45 (39 cases and 414 controls). British Journal of Cancer (2002) 87(3), 301 – 308 ã 2002 Cancer Research UK Conjunctival cancer in Uganda R Newton et al Table 1 Distribution of region of birth, region of residence, tribe, nation- KSHV All assays were performed by a single investigator (DB), ality, HIV-1 sero-status, income, age left home and time spent cultivating who was unaware of each patient’s personal characteristics, and diag- among cases with conjunctival carcinoma and controls with other can- nosis. A body-cavity-related B-cell lymphoma cell, BCP-1, which is cers, in Uganda positive for KSHV and negative for Epstein – Barr virus, was used for an indirect immunofluorescence assay to detect IgG antibodies Number of Odds ratio and 95% against KSHV antigen (Gao et al, 1996; Boshoff et al, 1998). Latently cases/controls confidence intervals infected BCP-1 cells were fixed in 4% paraformaldehyde and were made permeable with 0.2% Triton X-100. Cells were re-suspended Region of birth Kampala 4/54 1.0 in phosphate-buffered saline and fixed on glass slides. The samples Rest of Uganda 59/982 1.2 were diluted 1 : 100 in phosphate-buffered saline with 3% foetal calf w heterogeneity (1df)=0.1; P=0.7 serum. The diluted serum was added to fixed BCP-1 cells and incu- Region of residence bated at 228C for 45 min. After the slides were washed in Kampala 25/272 1.0 phosphate-buffered saline with 3% foetal calf serum, rabbit antihu- Rest of Uganda 35/922 0.6 man IgG labelled with fluorescein isothiocyanate (Dako, High w heterogeneity (1df)=2.3; P=0.1 Wycombe, UK), diluted 1 : 40 in phosphate-buffered saline with Tribe 3% foetal calf serum, was added and the slides were incubated at Baganda 27/631 1.0 Other 33/560 1.4 228C for 20 min. The slides were then washed in phosphate-buffered w heterogeneity (1df)=1.3; P=0.3 saline without foetal calf serum and screened by ultraviolet micro- Nationality scopy for the nuclear stippling pattern characteristic of antibodies Ugandan 56/1132 1.0 against the latent nuclear antigen of KSHV encoded by orf73 (Gao Other 4/59 1.4 et al, 1996). In total, 416 samples were available for testing for anti- w heterogeneity (1df)=0.3; P=0.6 bodies against KSHV (32 cases and 384 controls). HIV-1 sero-status Negative 17/700 1.0 Positive 39/119 10.1 Statistical methods w heterogeneity (1df)=52.9; P50.001 Personal income (SH) Data were computerised by trained clerks using EPI-INFO-5 soft- 520,000 30/408 1.0 ware, and statistical analyses were conducted using STATA 7.0 20,000+ 29/552 0.4 (Dean et al, 1990; STATA Corp, 2001). The odds ratio (OR) for each w heterogeneity (1df)=9.5; P50.001 variable in cases versus controls, was estimated with unconditional Household income (SH) logistic regression. All odds ratios were adjusted for age (15 – 24, 530,000 33/485 1.0 25 – 34, 35 – 44 and 45+), sex, HIV-1 sero-status and personal income 30,000+ 22/419 0.6 (less than 20 000 Uganda Shillings per year or 20 000+ Shillings). w heterogeneity (1df)=1.9; P=0.2 Additional adjustment was made for region of residence (Kampala Age left home (years) or elsewhere), because most cases had been identified from out- 21+ (or never left) 22/330 1.0 patient clinics, which tend to draw patients from the local area, 15 – 20 26/375 0.7 1 – 14 11/292 0.4 whereas controls had been hospital in-patients, who are from a wider w trend (1df)=3.9; P=0.05 area (Ziegler et al, 1997). These factors were selected a priori, with the Time spent cultivating exception of personal income, which was included following preli- (Hours per week) minary analyses. HIV infection is sexually transmitted and is 0 – 9 9/241 1.0 strongly associated with many of the sexual and reproductive vari- 10 – 19 23/401 1.9 ables studied (data not shown). In order to avoid confounding, 20+ 26/348 2.4 these factors were examined among HIV seropositive individuals w trend (1df)=3.9; P=0.05 only (there were too few HIV seronegative people to justify analysis). Odds ratios adjusted for age group, sex, current region of residence, HIV-1 status All P values are two-sided. Note that numbers of cases and controls in and personal income. df=degrees of freedom. the tables do not always add to the total, because of missing values. RESULTS Among those with conjunctival cancer, 43% (26 out of 60) were men Table 2 shows the results for anti-HPV and KSHV antibodies. and 57% (34 out of 60) were women. The proportion of all cancers The seroprevalence of anti-HPV antibodies in controls was 10% comprising conjunctival carcinoma declined from 9% in those aged for HPV-16 (43 out of 418), 4% (16 out of 414) for HPV-18 15 – 24 years to 2% in those over the age of 45 years. Seven per cent and 6% (24 out of 414) for HPV-45. The corresponding results of cases and 5% of controls were born in Kampala, the remainder for those with conjunctival cancer were 21% (eight out of 39), being born outside the capital city (P=0.7) and, 41% of cases and 10% (four out of 39) and 5% (two out of 39) respectively. 23% of controls reported their current residence as being in Kampala However, after adjustment for age, sex, address, HIV status and (Table 1; P=0.13). The seroprevalence of anti-HIV-1 antibodies was personal income, there were no statistically significant associations 70% among cases and 15% among controls (Odds ratio [OR] 10.1, between the presence of anti-HPV-16, -18 and -45 antibodies and 95% confidence intervals [CI] 5.2 – 19.4; P50.001). The risk of the risk of conjunctival carcinoma. Results for each HPV subtype conjunctival carcinoma was significantly lower among those with a were also calculated according to a measure of the antibody titre: high personal income (OR 0.4, 95% CI 0.2 – 0.7; P50.001). For the optical densities at each level correspond to less than 0.2 for those who left home at ages 21+ years (including those who never negative, 0.270.39 for medium titre and 0.4 or above for high left), 15 – 20 years and 1 – 14 years, the odds ratio was 1.0 (reference titre. The numbers of cases and controls with anti-HPV antibodies group), 0.7 (0.4 – 1.5) and 0.4 (0.2 – 1.0) respectively (P =0.05). to subtypes -18 and -45 were too few to yield any significant trend Study participants were asked how long each week they spent culti- results. The results for anti-HPV-16 antibodies at each measure vating, 0 – 9 h, 10 – 19 h or 20+ h. The risk of conjunctival of titre were 1.0 (HPV-16 antibody negative, based on 31 cases carcinoma increased significantly with increasing time spent cultivat- and 375 controls), 0.7 (0.2 – 2.9; medium titre, based on four cases ing (ORs 1.0, 1.9 and 2.4 respectively; P =0.03). and 31 controls) and 6.3 (1.2 – 33.4; high titre, based on four cases trend ã 2002 Cancer Research UK British Journal of Cancer (2002) 87(3), 301 – 308 Epidemiology Epidemiology Conjunctival cancer in Uganda R Newton et al Table 2 Comparison of human papillomavirus antibodies (HPV types 16, were conducted on over 50 risk factors, greatly increasing the like- 18 and 45) and Kaposi’s sarcoma-associated herpesvirus (KSHV) antibodies lihood of significant results arising by chance alone. The role of between those with conjunctival cancer and those without HPV-16, -18 and -45 requires further evaluation, but KSHV does not appear to be associated with conjunctival carcinoma. Number of Odds ratio and 95% The study reported here is subject to the potential problems of cases/controls confidence intervals incomplete diagnostic verification and HIV testing. Laboratory veri- fication of cancer diagnosis (for example, by histology, cytology or HPV-16 Negative 31/375 1.0 blood chemistry) was available on 60% of cases (all of which were Positive 8/43 1.5 (0.5 – 4.3) invasive tumours) and 63% of controls. Typical of studies in devel- w heterogeneity (1df)=0.5; P=0.5 oping countries where laboratory services are limited, this HPV-18 proportion compares favourably with other cancer series reported Negative 35/398 1.0 from Africa (Bassett et al, 1995; Newton et al, 1996a; Wabinga et Positive 4/16 2.3 (0.5 – 9.6) 2 al, 2000). Another potentially important source of bias is the high w heterogeneity (1df)=1.2; P=0.3 proportion of adults who were not tested for HIV in the early HPV-45 stages of the study. However, the age and sex-specific seropreva- Negative 37/390 1.0 Positive 2/24 0.4 (0.1 – 2.1) lence rates of HIV infection among those tested is broadly similar w heterogeneity (1df)=1.5; P=0.2 to that found in studies of HIV seroprevalence in Uganda (STD/ HPV-16, -18 or -45 AIDS Control Programme, Uganda, 1997). Also, the introduction All negative 28/345 1.0 of a saliva test resulted in a dramatic reduction in refusals, but little Any positive 11/69 1.3 (0.5 – 3.2) change in HIV seroprevalence rates, suggesting that patients who w heterogeneity (1df)=0.4; P=0.6 were refusing to have a HIV test were unwilling to have blood KSHV taken, rather than unwilling to learn their HIV serostatus. The asso- Negative 17/196 1.0 ciation of conjunctival cancer with HIV infection has been reported Positive 15/188 0.9 (0.4 – 2.1) in a general paper on HIV and cancer among residents of Kampala, w heterogeneity (1df)=0.1; P=0.8 and the odds ratio was the same as that from the current study of Adjusted for age, sex, address, personal income and HIV-1 sero-status. all patients who were seen at hospitals in Kampala (Newton et al, 2001). A proportion of individuals in the study were also tested for antibodies to three HPV subtypes and to KSHV infection. This additional testing occurred in all those for whom enough stored and 12 controls; P =0.2). Only 15 people had anti-HPV antibo- sera were available and (as people are unaware of their own status trend dies to more than one tested HPV subtype (two cases and 13 with respect to infection), is therefore an unbiased sample. controls) and there was no significant excess risk of the tumour Another potential source of bias arises from the fact that most in these individuals, as compared to those who were considered (but not all) cases had been identified from out-patient clinics, to be negative for all three subtypes (OR 0.6, 95% CI 0.1 – 4.3). which tend to draw patients from the local area, whereas controls In relation to Kaposi’s sarcoma-associated herpesvirus, the seropre- had been hospital in-patients, who are from a wider area. Despite valence of anti-KSHV antibodies was 47% (15 out of 32) among adjusting for region of residence in the analyses, the possible cases and 49% (188 out of 384) among controls (OR 0.9, 95% impact of this on the results, is difficult to assess. CI 0.4 – 2.1; P=0.8). Case reports of squamous cell carcinoma of the conjunctiva in Further results are provided in Appendices 1 – 4. Results for human immunodeficiency virus (HIV) infected men in the USA other social and demographic factors, none of which was associated and France, coupled with a marked increase in the numbers of with an increased risk of conjunctival carcinoma, are shown in tumours being seen by ophthalmologists in at least two African Appendix 1. Factors, other than income, that might characterise centres (which mirrored the increases seen for Kaposi’s sarcoma), wealth and social status in Uganda, such as ownership of livestock led to the suggestion of an association with HIV (Winward and and travel away from home, are shown in Appendix 2. None of Curtin, 1989; Ateenyi-Agaba, 1995; Kestelyn et al, 1990; Kim et al, these variables were associated with conjunctival carcinoma. 1990; Denis et al, 1994). Among those with HIV, the lesions often Appendix 3 shows results for other possible exposures. Those affect young adults, in a fashion reminiscent of Kaposi’s sarcoma in who reported having had a blood transfusion in the past were less HIV-seropositive people (Ateenyi-Agaba, 1995; Kestelyn et al, 1990; likely to be cases (OR 0.4, 95% CI 0.2 – 0.9; P=0.02). Results for Waddell et al, 1996). This study, together with others from Africa sexual and reproductive variables among HIV seropositive people and the USA indicate about a 10-fold increased risk of the tumour only, are shown in Appendix 4. A self reported history of having in HIV infected, compared to HIV uninfected individuals (Ateenyi- either given or received gifts in exchange for sex was associated Agaba, 1995; Kestelyn et al, 1990; Goedert and Cote, 1995; Newton with an increased risk of the tumour (OR 3.5, 95% CI 1.2 – 10.4; et al, 1995, 2001; Waddell et al, 1996). Indeed, the spread of HIV P=0.03), although this may have been a chance finding as there in Uganda probably accounts for much of the approximately eight- was no statistically significant association with any other sexual fold increase in incidence of conjunctival carcinoma observed there or reproductive variable, including the reported number of lifetime since the 1960s (Wabinga et al, 2000). Using standard equations for sexual partners (P=0.4). case – control studies, the population attributable fraction is about 60%: almost two-thirds of cases would not occur in the absence of HIV infection (dos Santos Silva, 1998). However, the frequency of DISCUSSION conjunctival carcinoma is not such that it is yet a particularly Here, we report an investigation of the epidemiology and aetiology common manifestation of HIV disease (Piot et al, 1992). of conjunctival squamous cell carcinoma, examining over 50 risk The mechanism whereby HIV infection increases the risk of factors for the tumour. Conjunctival cancer affects relatively young conjunctival cancer is not clear. There is no evidence that it is individuals of both sexes and is strongly associated with HIV infec- directly oncogenic and the impact on cancer risk is most likely tion and poverty. In addition, the risk of the disease decreases with to be mediated via immunosuppression, as is the case for other decreasing age at which an individual leaves home and with HIV-associated cancers (Beral and Newton, 1998; Newton et al, increasing time spent cultivating, although both results are of 1999). It is not known if the risk of conjunctival squamous cell borderline statistical significance. Furthermore, significance tests carcinoma is increased in other immunosuppressed groups, such British Journal of Cancer (2002) 87(3), 301 – 308 ã 2002 Cancer Research UK Conjunctival cancer in Uganda R Newton et al as transplant recipients, because the tumour is very rare in parts of that the incidence of squamous carcinoma of the eye increases as the world where tissue transplantation occurs. However, there is a exposure to ambient solar ultraviolet radiation increases. Levels case report of a conjunctival cancer in a patient with malignant of solar radiation are higher towards the equator and the incidence lymphoma on immunosuppressive chemotherapy (Kushner and of conjunctival cancer increases by about 50% for each 108 decline Mushen, 1975). In general, HIV infection is thought to increase in latitude (Newton et al, 1996b; Sun et al, 1997). In addition, a cancer risk by facilitating the action of other oncogenic viruses, single case – control study reported that the risk of ocular surface such as KSHV, the principal cause of Kaposi’s sarcoma. In relation epithelial dysplasias is greatest in those reporting a past history to conjunctival cancer however, we found no evidence in these data of skin cancer, which is known to be caused by exposure to solar of an association with KSHV. ultraviolet radiation (Lee et al, 1994). Time spent cultivating may Several types of squamous carcinoma are associated with human indicate the time an individual spends in direct sunlight, and so papillomavirus (HPV) infection, most notably cancer of the uterine the increasing risk with increasing time cultivating may reflect cervix, induced primarily by HPV-16, -18 and others. Squamous exposure to ultraviolet radiation. This variable may also reflect carcinoma of the skin has also been associated with HPV-5 and exposure to dust or dirt, and it has been hypothesised that ocular -8 in immunosuppressed individuals (IARC, 1995). Evidence for trauma could facilitate development of the tumour (Templeton, an association between human papillomavirus and squamous cell 1973; Margo and Groden, 1986). The lower risk among people carcinoma of the conjunctiva is conflicting, although bovine papil- who leave home at an earlier age may reflect migration to towns lomaviruses are thought to cause conjunctival carcinomas in cattle for work, where exposure to solar ultraviolet radiation could be (IARC, 1995). The presence of human papillomavirus DNA (HPV; less. The extent to which the higher risk among those of low predominantly types 16, but also other types) in human ocular income is measuring the previously mentioned exposures is not surface squamous neoplasias, including invasive carcinomas, has clear. The fact that cases are less likely than controls to have had been reported in some studies, but not others (reviewed by a blood transfusion is probably an artefact of the control selection. Newton, 1996). This is the first study to look for an association Cases were generally seen as out-patients, whereas controls with between anti-HPV antibodies (types -16, -8 and -45) and the risk other cancers were generally hospital in-patients. The latter may of conjunctival carcinoma. The seroprevalence of antibodies against therefore be more likely to have had a transfusion as part of their HPV-18 and -45 was too low to make reliable conclusions. Results medical care. for anti-HPV-16 antibodies were suggestive of an association In summary, the current study gives independent support for among individuals with high titres, but the data presented here the strong epidemiological evidence that solar ultraviolet radiation are too few to draw valid conclusions and studies of larger is an important cause of squamous cell carcinoma of the conjunc- numbers of cases are required. tiva. Another established risk factor is HIV infection, although the The HPV assays used in this study are based on the expression mechanism whereby it increases the risk of conjunctival cancer is of L1 major capsid proteins of HPV-16, -18 and -45 in insect cells not clear. We find little evidence supporting the possible role of by using recombinant baculoviruses (Touze ´ et al, 1998). The result- sexually transmitted forms of HPV in the aetiology of conjunctival ing virus-like particles (VLPs), which appear similar to empty cancer, but larger studies are required. virions, can be used in serological studies to test for type specific immunological responses to viral capsid proteins, although there is evidence that a particular assay may cross react with related ACKNOWLEDGEMENTS HPV subtypes (Combita et al, 2002b). Presence of anti-VLP anti- bodies is an indicator of past and current infection (Kirnbauer et Support for this work was provided by Mulago Hospital and al, 1994; Le Cann et al, 1995; Wideroff et al, 1995; Dillner et al, Makerere Medical School, Kampala, Uganda, the Imperial Cancer 1996). The utility of such assays has been demonstrated in previous Research Fund (now Cancer Research UK), the United States studies of anti-HPV-16 antibodies in relation to the risk of cancer Centers for Disease Control and Prevention (interagency agreement of the uterine cervix (Lehtinen et al, 1996; Dillner et al, 1997; Shah with the Department of Veterans Affairs), the International Agency et al, 1997; Vonka et al, 1999; Hisada et al, 2001). Results from this for Research on Cancer, World Health Organisation, Lyon, France study on the relationship between anti-HPV-16 antibodies and the and the Ligue Contre le Cancer, France. J Ziegler was on second- risk of cancer of the uterine cervix are broadly comparable to those ment from the Department of Veterans Affairs and the reported before and will be the subject of a separate report. University of California, San Francisco, CA, USA, to the Interna- Although relatively rare everywhere, conjunctival carcinoma is tional Agency for Research on Cancer, France. L Bousarghin was more frequent in sub-Saharan Africa and other tropical areas than the recipient of a fellowship from the Region Centre, France. L in temperate countries, leading to the hypothesis that it may be Carpenter was on secondment from the Academic Department of associated with exposure to solar ultraviolet radiation (Templeton, Public Health, Oxford University, to the MRC Programme on 1973; Clear, 1979). Geographical studies support this premise, in AIDS, Entebbe, Uganda. REFERENCES Ateenyi-Agaba C (1995) Conjunctival squamous cell carcinoma associated Boshoff C, Gao SJ, Healy LE, Matthews S, Thomas AJ, Coignet L, Warnke RA, with HIV infection in Kampala, Uganda. Lancet i: 695 – 696 Strauchen JA, Matutes E, Kamel OW, Moore PS, Weiss RA, Chang Y Bassett MT, Chokunonga E, Mauchaza B, Levy L, Ferlay J, Parkin DM (1995) (1998) Establishing a KSHV+ cell line (BCP-1) from peripheral blood Cancer in the African population of Harare, Zimbabwe in 1990 – 92. Int J and characterising its growth in Nod/SCID mice. Blood 91: 1671 – 1679 Cancer 63: 29 – 36 Clear AS (1979) Solar keratosis, pterygium and squamous cell carcinoma of Beral V, Newton R (1998) Overview of the epidemiology of immunodefi- the conjunctiva in Malawi. Br J Ophthalmol 63: 102 – 109 ciency associated cancers. Monogr Natl Cancer Inst 23: 1– 6 Combita AL, Bravo MM, Touze ´ A, Orozco O, Coursaget P (2002a) Serologic Bosch FX, Manos MM, Munoz N, Sherman M, Jansen AM, Peto J, Schiffman response to Human oncogenic papillomavirus types 16, 18, 31, 33, 39, 58 MH, Moreno V, Kurman R, Shah KV The International Biological Study and 59 virus-like particles in Colombian women with invasive cervical on Cervical Cancer (IBSCC) Study Group (1995) Prevalence of human cancer. Int J Cancer 97: 796 – 803 papillomavirus in cervical cancer: a worldwide perspective. J Natl Cancer Inst 87: 796 – 802 ã 2002 Cancer Research UK British Journal of Cancer (2002) 87(3), 301 – 308 Epidemiology Epidemiology Conjunctival cancer in Uganda R Newton et al Combita AL, Touze A, Bousarghin L, Christensen ND, Coursaget P (2002b) Margo CE, Groden LR (1986) Squamous cell carcinoma of the cornea and Identification of two cross-neutralizing linear epitopes within the L1 major conjunctiva following a thermal burn of the eye. Cornea 5: 185 – 188 capsid protein of human papillomavirus. J Virol (in press). Newton R, Grulich A, Beral V, Sindikubwabo B, Ngilimana P-J, Nganyira A, Dean AD, Dean JA, Burton JH, Dicker RC (1990) Epi Info, Version 5: a word Parkin DM (1995) Cancer and HIV infection in Rwanda. Lancet i: 1378 processing, database and statistics program for epidemiology on micro- Newton R (1996) A review of the aetiology of squamous cell carcinoma of the computers. Centers for Disease Control, Atlanta, Georgia, USA. conjunctiva. Br J Cancer 74: 1511 – 1513 Denis P, Charpentier D, Roudier M, Lemaire JC, Levin N, Nguyea-Khoa JL Newton R, Ngilimana PJ, Grulich A, Beral V, Sindikubwabo B, Nganyira A, (1994) Conjunctival epidermoid carcinoma and human immunodeficiency Parkin DM (1996a) Cancer in Rwanda. Int J Cancer 66: 75 – 81 virus. J Fr Ophtalmol 17: 366 – 369 Newton R, Ferlay J, Reeves G, Beral V, Parkin DM (1996b) Incidence of Dillner JL, Kallings I, Brihmer C, Sikstrom B, Koskela P, Lehtinen M, Schiller squamous cell carcinoma of the eye increases with increasing levels of JT, Sapp M, Mardh PA (1996) Seropositivities to human papillomavirus ambient solar ultraviolet radiation. Lancet i: 1450 – 1451 type 16, 18 or 33 capsids and to Chlamydia trachomatis are markers of Newton R, Beral V, Weiss R (1999) Human immunodeficiency virus infec- sexual behaviour. J Infect Dis 173: 1394 – 1398 tion and cancer. In Cancer Surveys volume 33, Infections and Human Dillner JL, Lehtinen M, Bjorge T, Luostarinen T, Youngman L, Jellum E, Cancer, Newton R, Beral V, Weiss R (eds), pp 237 – 261. Cold Spring Koskela P, Gislefoss RE, Hallmans G, Paavonen J, Sapp M, Schiller JT, Harbor Laboratory Press Hakulinen T, Thoresen S, Hakama M (1997) Prospective seroepidemiolo- Newton R, Ziegler J, Beral V, Mbidde E, Carpenter L, Wabinga H, Mbulataiye gic study of human papillomavirus infection as a risk factor for invasive S, Appleby P, Reeves G, Jaffe H and the Uganda Kaposi’s Sarcoma Study cervical cancer. J Natl Cancer Inst 89: 1293 – 1299 Group (2001) A case-control study of Human Immunodeficiency Virus El Mehdaoui S, Touze ´ A, Laurent S, Sizaret PY, Rasschaert D, Coursaget P infection and cancer in adults and children residing in Kampala, Uganda. (2000) Gene transfer using recombinant rabbit hemorrhagic disease virus Intl J Cancer 92: 622 – 627 capsids with genetically modified DNA encapsidation capacity by addition Piot P, Kapita BM, Ngugi E, Mann JM, Colebunders RL, Wabitsch R (1992) of packaging sequences from L1 or L2 protein of human papillomavirus AIDS in Africa. A Manual for Physicains. WHO, Geneva type 16.. J Virol 74: 10332 – 10340 dos Santos Silva I (1998) Cancer Epidemiology Principles and Methods. IARC, Gao S-J, Kingsley L, Li M, Zheng W, Parravicini C, Ziegler JL, Newton R, Lyon Rinaldo CR, Saah A, Phair J, Detels R, Chang Y, Moore PS (1996) Seropre- Shah KV, Viscidi RP, Alberg AJ, Helzlsouer KJ, Comstock GW (1997) Anti- valence of Kaposi’s sarcoma-associated herpesvirus antibodies among bodies to human papillomavirus 16 and subsequent in situ or invasive Americans, Italians and Ugandans with and without Kaposi’s sarcoma. cancer of the cervix. Cancer Epidemiol Biomarkers Prev 6: 233 – 237 Nature Med 2: 925 – 928 STATA Corp (2001) Stata software, College Park, TX, USA Goedert JJ, Cote TR (1995) Conjunctival malignant disease with AIDS in STD/AIDS Control Programme HIV/AIDS Surveillance Report, March. USA. Lancet ii: 257 – 258 (1997) Ministry of Health, PO Box 8, Entebbe, Uganda Hisada M, van den Berg BJ, Strickler HD, Christianson RE, Wright WE, Sun EC, Fears TR, Goedert JJ (1997) Epidemiology of squamous cell conjunc- Waters DJ, Rabkin CS (2001) Prospective study of antibody to human tival cancer. Cancer Epidemiol, Biomark Preven 6: 73 – 77 papillomavirus type 16 and risk of cervical, endometrial and ovarian Templeton AC (1973) Tumours of the eye and adnexa. In Tumours of a cancers (United States). Cancer Causes Control 12: 335 – 341 Tropical Country: A survey of Uganda 1964 – 1968. Templeton AC (ed), IARC (1992) Monographs on the evaluation of carcinogenic risks to humans. pp 203 – 214 Recent Result Cancer Research Solar and Ultraviolet Radiation. IARC Monograph vol 55, Lyon Touze ´ A, el Mehdaoui S, Sizaret PY, Mougin C, Munoz N, Coursaget P IARC (1995) Monographs on the evaluation of carcinogenic risks to humans. (1998) The L1 major capsid protein of Human Papillomavirus Type 16 Human papillomaviruses.. IARC Monograph vol 64, Lyon variants affects the yield of virus-like particle produced in an insect cells Kestelyn P, Stevens AM, Ndayambaje A, Hanssens M, van de Perre P (1990) expression system. J Clin Microbiol 36: 2046 – 2051 HIV and conjunctival malignancies. Lancet 336: 51 – 52 Vonka V, Hamsikova E, Kanka J, Ludvikova V, Sapp M, Smahel M (1999) Kim RY, Seiff SR, Howes EL Jr, O’Donnell JJ (1990) Necrotizing scleritis Prospective study on cervical neoplasia IV. Presence of HPV antibodies. secondary to conjunctival squamous cell carcinoma in aquired immunode- Int J Cancer 80: 365 – 368 ficiency syndrome. Am J Opthalmol, 109: 231 – 233 Wabinga HR, Parkin DM, Wabwire-Mangen F, Nambooze S (2000) Trends Kirnbauer R, Hubbert NL, Wheeler CM, Becker TM, Lowy DR, Schiller JT in cancer incidence in Kyadondo County, Uganda, 196071997. Br J (1994) A virus-like particle enzyme-linked immunosorbent assay detects Cancer 82: 1585 – 1592 serum antibodies in a majority of women infected with human papilloma- Waddell KM, Lewallen S, Lucas SB, Ateenyi-Agaba C, Herrington CS, Liomba virus type-16. J Natl Cancer Inst 86: 494 – 499 G (1996) Carcinoma of the conjunctiva and HIV infection in Uganda and Kushner FH, Mushen RL (1975) Conjunctival squamous cell carcinoma Malawi. Br J Ophthalmol 80: 503 – 508 combined with malignant lymphoma. Am J Ophthalmol 80: 503 – 506 Wideroff L, Schiffman MH, Hoover R, Tarane TE, Nonnenmacher B, Le Cann P, Touze A, Enogat N, Leboulleux D, Mougin C, Legrand MC, Hubbert NL, Kirnbauer R, Greer CE, Lorincz AT, Manos MM, Glass Calvet C, Afoutou JM, Coursaget P (1995) Detection of antibodies against AG, Scott DR, Sherman ME, Buckland J, Lowy D, Schiller J (1995) Evalua- human papillomavirus (HPV) type-16 virions by ELISA using recombi- tion of sero-reactivity to human papillomavirus type 16 virus-like particles nant HPV 16 L1 capsids produced by recombinant baculovirus. J Clin in an incident case-control study of cervical neoplasia. J Infect Dis 172: Microbiol 33: 1380 – 1382 1425 – 1430 Lee GA, Williams G, Hirst LW, Green AC (1994) Risk Factors in the Devel- Winward KE, Curtin VT (1989) Conjunctival squamous cell carcinoma in a opment of Ocular Surface Epithelial Dysplasia. Ophthalmology 101: 360 – patient with human immunodeficiency virus infection. Am J Ophthalmol 364 107: 554 – 555 Lehtinen M, Dillner J, Knekt P, Luostarinen T, Aromaa A, Kirnbauer R, Ziegler JL, Newton R, Katongole-Mbidde E, Mbulataiye S, DeCock K, Wabin- Koskela P, Paavonen J, Peto R, Schiller JT, Hakama M (1996) Serologically ga H, Mugerwa J, Katabira E, Jaffe H, Parkin DM, Reeves G, Beral V for the diagnosed infection with human papillomavirus type 16 and risk for subse- Uganda Kaposi’s sarcoma study group (1997) Risk factors for HIV-asso- quent development of cervical carcinoma: nested case-control study. BMJ ciated Kaposi’s sarcoma in Uganda: a case-control study of 1026 Adults. 312: 537 – 539 AIDS 11: 1619 – 1626 British Journal of Cancer (2002) 87(3), 301 – 308 ã 2002 Cancer Research UK Conjunctival cancer in Uganda R Newton et al Appendix 1 Comparison of social and demographic factors between Appendix 2 Comparison of factors characterising wealth between people with and without conjunctival cancer those with and without conjunctival cancer Number of Odds ratio and 95% Number of Odds ratio and 95% a a cases/controls confidence intervals cases/controls confidence intervals Years at current address Number of rooms in house 510 27/220 1.0 1 – 2 31/306 1.0 10+ 33/778 0.8 3+ 29/696 0.8 2 2 w heterogeneity (1df)=0.8; P=0.4 w heterogeneity (1df)=0.4; P=0.5 Settlement type Electricity 1 – 99 houses 29/704 1.0 No 38/770 1.0 100+ houses 31/287 1.5 Yes 21/228 1.0 2 2 w heterogeneity (1df)=1.7; P=0.2 w heterogeneity (1df)=0.0; P=1.0 Travel time to market Own radio 530 mins 40/400 1.0 No 11/205 1.0 30+ mins 14/456 0.6 Yes 49/794 1.0 2 2 w heterogeneity (1df)=2.0; P=0.2 w heterogeneity (1df)=0.0; P=0.9 Age left school Own motorcar No school 7/270 1.0 No 56/890 1.0 515 years 25/289 1.4 Yes 4/109 0.5 16 – 19 years 20/262 1.2 w heterogeneity (1df)=1.8; P=0.2 20+ years 8/163 0.8 Own bicycle w trend (1df)=0.4; P=0.5 No 35/523 1.0 Education level Yes 25/478 1.0 No school 7/273 1.0 w heterogeneity (1df)=0.0; P=0.9 Primary 19/272 1.5 Own cows Secondary/tertiary 34/450 1.1 None 50/736 1.0 w trend (1df)=0.7; P=0.4 1+ 9/259 0.7 Religion w heterogeneity (1df)=0.7; P=0.4 Christian 55/882 1.0 Own pigs and goats Muslim 5/125 0.8 None 46/677 1.0 w heterogeneity (1df)=0.2; P=0.7 1+ 14/319 1.2 Occupation w heterogeneity (1df)=0.3; P=0.6 Cultivator 14/588 1.0 Own chickens Other 46/549 1.7 None 38/542 1.0 w heterogeneity (1df)=1.8; P=0.2 1+ 22/431 1.1 Number in house w heterogeneity (1df)=0.1; P=0.8 55 25/383 1.0 Travel away from home 4 5+ 35/621 1.2 7 nights per year w heterogeneity (1df)=0.3; P=0.6 No 53/877 1.0 Number of siblings Yes 7/105 1.3 55 32/567 1.0 w heterogeneity (1df)=0.2; P=0.7 5+ 28/438 0.8 Adjusted for age group, sex, address, personal income and HIV-1 sero-status; w heterogeneity (1df)=0.7; P=0.4 df=degrees of freedom. Birth order 1 – 2 25/467 1.0 3+ 35/538 0.9 w heterogeneity (1df)=0.1; P=0.7 Number sharing toilet 56 19/358 1.0 6+ 40/610 1.1 w heterogeneity (1df)=0.0; P=0.9 Adjusted for age group, sex, address, personal income and HIV-1 sero-status; df=degrees of freedom. ã 2002 Cancer Research UK British Journal of Cancer (2002) 87(3), 301 – 308 Epidemiology Epidemiology Conjunctival cancer in Uganda R Newton et al Appendix 3 Comparison of other exposures between those with and Appendix 4 Comparison of sexual and reproductive variables between without conjunctival cancer those with conjunctival cancer and those without, among HIV seropositive individuals only Number of Odds ratio and 95% Number of Odds ratio and 95% cases/controls confidence intervals cases/controls confidence intervals Use of shoes Never/rarely 26/619 1.0 Marital status Often/always 33/372 1.5 Single 1/10 1.0 w heterogeneity (1df)=1.5; P=0.2 Monogomous married 22/43 8.2 Age first use shoes (years) Polygamous married 9/22 9.8 515 36/423 1.0 Widower 4/26 3.3 15+ 23/532 1.1 Separated 3/10 4.2 2 2 w heterogeneity (1df)=0.1; P=0.7 w heterogeneity (1df)=6.6; P=0.16 Soil colour Age first sex Black 36/549 1.0 518 years 26/65 1.0 Other 16/343 1.0 18+ years 13/46 0.7 2 2 w heterogeneity (1df)=0.0; P=0.9 w heterogeneity (1df)=0.8; P=0.4 Time spent in water Number of children (hours per week) 0 – 2 13/23 1.0 51 hour 47/728 1.0 3 – 4 9/25 0.7 1+ hours 12/272 0.6 5 – 6 9/20 1.0 w heterogeneity (1df)=2.9; P=0.1 7+ 5/30 0.5 Tobacco use w trend (1df)=0.4; P=0.5 Never smoker 46/762 1.0 Children with different Ex-smoker 7/146 0.9 mothers or fathers Current smoker 7/91 1.1 No 13/40 1.0 w heterogeneity (1df)=0.0; P=1.0 Yes 21/53 1.4 Use of snuff w heterogeneity (1df)=0.5; P=0.5 No 58/977 1.0 No. sexual partners Yes 2/16 2.0 1 – 4 13/50 1.0 w heterogeneity (1df)=0.6; P=0.4 5 – 7 10/17 2.7 Drink home-brewed alcohol 8+ 16/40 1.8 No 36/533 1.0 w trend (1df)=0.8; P=0.4 About once/week 6/155 0.4 History sexual discharge 2 – 4/week 8/131 1.0 No 13/48 1.0 Most days 9/178 0.9 Yes 26/64 1.9 2 2 w heterogeneity (1df)=3.3; P=0.3 w heterogeneity (1df)=2.1; P=0.2 Blood transfusion Gifts for sex No 53/184 1.0 Never 29/101 1.0 Yes 7/183 0.4 Ever 10/11 3.5 2 2 w heterogeneity (1df)=5.6; P=0.02 w heterogeneity (1df)=5.0; P=0.03 Injection from traditional Use condoms healer Never 28/83 1.0 No 39/596 1.0 Ever 11/28 0.7 Yes 21/374 1.0 w heterogeneity (1df)=0.5; P=0.5 w heterogeneity (1df)=0.0; P=0.9 Circumcised (men) No 17/34 1.0 Odds ratios adjusted for age, sex, address, personal income and HIV-1 sero-status. Yes 0/7 0.0 w heterogeneity (1df)=1.3; P=0.3 Labial elongation (women) No 5/17 1.0 Yes 16/51 0.9 w heterogeneity (1df)=0.1; P=0.8 Adjusted for age, sex, address and personal income. British Journal of Cancer (2002) 87(3), 301 – 308 ã 2002 Cancer Research UK http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png British Journal of Cancer Springer Journals

Loading next page...
 
/lp/springer-journals/the-epidemiology-of-conjunctival-squamous-cell-carcinoma-in-uganda-NkUjAbrb4b

References (102)

Publisher
Springer Journals
Copyright
Copyright © 2002 by The Author(s)
Subject
Biomedicine; Biomedicine, general; Cancer Research; Epidemiology; Molecular Medicine; Oncology; Drug Resistance
ISSN
0007-0920
eISSN
1532-1827
DOI
10.1038/sj.bjc.6600451
Publisher site
See Article on Publisher Site

Abstract

British Journal of Cancer (2002) 87, 301 – 308 ã 2002 Cancer Research UK All rights reserved 0007 – 0920/02 $25.00 www.bjcancer.com The epidemiology of conjunctival squamous cell carcinoma in Uganda ,1 2 2 3 1 1 2 4 R Newton* , J Ziegler , C Ateenyi-Agaba , L Bousarghin , D Casabonne , V Beral , E Mbidde , L Carpenter , 1 5 2 2 6 7 7 3 G Reeves , DM Parkin , H Wabinga , S Mbulaiteye , H Jaffe , D Bourboulia , C Boshoff , A Touze ´ , 3 8 P Coursaget and the Uganda Kaposi’s Sarcoma Study Group 1 2 Cancer Research UK, Epidemiology Unit, Gibson Building, Radcliffe Infirmary, Oxford OX2 6HE, UK; Uganda Cancer Institute and Makerere University Medical School, Kampala, Uganda; Laboratoire de Virologie Mole ´culaire, INSERM EMIU 00-10 and USC INRA, Faculte ´ de Pharmacie, 37200 Tours, 4 5 France; MRC Programme on AIDS, Uganda Virus Research Institute, PO Box 49, Entebbe, Uganda; International Agency for Research on Cancer, 150 6 7 Cours Albert-Thomas, Lyon, France; Centers for Disease Control and Prevention, 1600 Clifton Road, Atlanta, Georgia, GA 30333, USA; Wolfson Institute of Medical Research, University College London, 46 Cleveland Street, London, UK As part of a larger investigation of cancer in Uganda, we conducted a case – control study of conjunctival squamous cell carcinoma in adults presenting at hospitals in Kampala. Participants were interviewed about social and lifestyle factors and had blood tested for antibodies to HIV, KSHV and HPV-16, -18 and -45. The odds of each factor among 60 people with conjunctival cancer was compared to that among 1214 controls with other cancer sites or types, using odds ratios, estimated with unconditional logistic regression. Conjunctival cancer was associated with HIV infection (OR 10.1, 95% confidence intervals [CI] 5.2 – 19.4; P50.001), and was less common in those with a higher personal income (OR 0.4, 95% CI 0.3 – 1.2; P50.001). The risk of conjunctival cancer increased with increasing time spent in cultivation and therefore in direct sunlight 2 2 (w trend=3.9, P=0.05), but decreased with decreasing age at leaving home (w trend=3.9, P=0.05), perhaps reflecting less exposure to sunlight consequent to working in towns, although both results were of borderline statistical significance. To reduce confounding, sexual and reproductive variables were examined among HIV seropositive individuals only. Cases were more likely than controls to report that they had given or received gifts for sex (OR 3.5, 95% CI 1.2 – 10.4; P=0.03), but this may have been a chance finding as no other sexual or reproductive variable was associated with conjunctival cancer, including the number of self-reported lifetime sexual partners (P=0.4). The seroprevalence of antibodies against HPV-18 and -45 was too low to make reliable conclusions. The presence of anti-HPV-16 antibodies was not significantly associated with squamous cell carcinoma of the conjunctiva (OR 1.5, 95% CI 0.5 – 4.3; P=0.5) and nor were anti-KSHV antibodies (OR 0.9, 95% CI 0.4 – 2.1; P=0.8). The 10-fold increased risk of conjunctival cancer in HIV infected individuals is similar to results from other studies. The role of other oncogenic viral infections is unclear. British Journal of Cancer (2002) 87, 301 – 308. doi:10.1038/sj.bjc.6600451 www.bjcancer.com ã 2002 Cancer Research UK Keywords: conjunctival carcinoma; HIV; HPV; KSHV; HHV-8; Uganda Squamous cell carcinoma of the conjunctiva is thought to be an carcinoma of the conjunctiva, because the tumour was relatively extreme form of a spectrum of clinical conditions, collectively frequent there, even before the onset of the HIV epidemic known as ‘ocular surface squamous neoplasias’, which range in (Templeton, 1973; Wabinga et al, 2000). In this report, we examine severity from mild dysplasia, to carcinoma in situ and, ultimately, the association of conjunctival tumours with over 50 possible risk to invasive carcinoma. Symptoms can range from none, to severe factors, including evidence of infection with HIV-1, HPV-16, -18 pain and visual loss. Lesions generally arise on exposed areas of and -45 and Kaposi’s sarcoma-associated herpesvirus (KSHV; the eye, particularly on the nasal side, and treatment involves local human herpesvirus type 8 [HHV-8]), using data from a case- excision, or in more severe cases, orbital clearance. Metastases are control study of cancer in adults (Ziegler et al, 1997; Newton et rare and the prognosis is usually favourable. al, 2001). Although relatively rare everywhere, conjunctival carcinoma is more frequent in parts of sub-Saharan Africa. Uganda offers a good MATERIALS AND METHODS setting in which to investigate the epidemiology of squamous cell Study subjects *Correspondence: R Newton; E-mail: rob_newton@cancer.org.uk The subjects included in these analyses were selected from a large The Uganda Kaposi’s Sarcoma Study Group includes named authors and V cross-sectional study of risk factors for cancer in Uganda, which Sembajwe, M Kalinaki, R Byansi, C Rwatooro, S Nambooze, B Tushimiere, included, in total, 2091 individuals with malignancies other than N Byabazaire (deceased), E Bitamazire, E Katabira, J Mugerwa (deceased), Kaposi’s sarcoma. Between August 1994 and February 1998, any D Tindyebura, J Whitworth, B Richardson, R Weiss and K de Cock adults (aged 15 years and over) with a provisional new diagnosis Received 8 February 2002; revised 19 April 2002; accepted 22 April 2002 of incident cancer were eligible for recruitment into the study from Epidemiology Epidemiology Conjunctival cancer in Uganda R Newton et al all the wards and out-patient clinics of the four main hospitals in culaire in Tours, France, for HPV testing and to University College Kampala, Uganda: Mulago (including the Uganda Cancer Institute), London, UK, for KSHV testing. Nsambya, Mengo and Rubaga. Further details of the methods can be found elsewhere (Ziegler et al, 1997; Newton et al, 2001). The HPV-16, -18 and -45 All assays were performed by a single subjects included in this report are restricted to 60 cases of conjunc- investigator (LB), who was unaware of each patient’s personal char- tival cancer and 1214 controls. Thirty-two people with eye cancers acteristics, and diagnosis. HPV VLPs were produced in Sf21 insect that were not specifically diagnosed clinically as being squamous cell cells using recombinant baculoviruses encoding the L1 gene of carcinomas, or had a pathological diagnosis of malignancy of uncer- HPV-16, -18, and -45, according to previously described procedures tain morphology were excluded. The controls comprised people (Touze ´ et al; 1998; Combita et al, 2002a,b). These subtypes were with other incident cancers, excluding those with cancer sites or chosen because they are known to be prevalent in tumour speci- types that are known to be associated with infection with HIV, mens from women with cancer of the uterine cervix in Uganda HPV or KSHV, or exposure to solar ultraviolet radiation – that (Bosch et al, 1995). Briefly, the HPV L1 genes were first amplified is, Kaposi’s sarcoma, non-Hodgkin’s lymphoma, Hodgkin’s disease, from an HPV DNA-positive biopsy using primers containing BglII cancers of the uterine cervix, anus and penis, and skin cancers sites. The amplified product was then inserted into pFastBacI to (IARC, 1992, 1995; Beral and Newton, 1998). The control group generate HPV L1 VLPs. VLPs were produced according to a proce- included men and women with cancers of the oral cavity (57), oeso- dure used for HPV-16 VLPs (Touze et al, 1998; El Mehdaoui et al, phagus (150), stomach (74), liver (103), breast (178), ovary (67), 2000). Sf21 cells, maintained in Grace’s insect medium supplemen- prostate (56), and other cancer sites or types (405). In addition, ted with 10% foetal calf serum (FCS) were infected with the 124 patients with a provisional diagnosis of cancer, but who subse- different recombinant baculoviruses at a m.o.i. of 10 and were incu- quently turned out to have benign tumours, were also included in bated for 72 h at 278C. Cells were harvested by centrifugation, re- the control group. suspended in PBS containing 0.5% NP40 and allowed to stand at room temperature for 30 min. Cell lysates were then centrifuged at 14 0006g for 15 min at 48C. The nuclear fractions were further Recruitment and questionnaire re-suspended in ice cold PBS and sonicated by three 15 s bursts at Interviewers (who were also trained HIV counsellors) approached 60% maximal power (Vibra Cell, Bioblock Scientific, Strasbourg, ward or clinic staff to gain permission to interview potential France). Fractions were then loaded on the top of a preformed CsCl recruits. If permission was granted, the patient was approached gradient and centrifuged at equilibrium in a Beckman SW28 rotor by one of four interviewers and invited to participate in the study. (20 h, 27 000 r.p.m., 48C). Gradient fractions were analysed for A HIV test was requested and appropriate counselling was given. density by refractometry and were tested for the presence of L1 Patients who had been HIV tested within a month of recruitment protein by ELISA. Immunoreactive fractions were finally pooled and had a medical certificate indicating the test result, were not re- and pelleted by ultracentrifugation in a Beckman SW 28 rotor tested. The patient was interviewed about social and demographic (3 h, 28 000 r.p.m., 48C). VLPs were resuspended in PBS (pH 7.4) factors, and sexual and reproductive history. Where possible, the and protein content was evaluated using the microBCA kit (Pierce, patient was interviewed by a counsellor of the same sex, and in Touzart et Matignon, France). Each preparation was tested for the their native language. Blood was drawn, primarily for the HIV test, presence of VLPs by electron microscopy. For this purpose, VLP but also for storage of serum and leukocytes. Interviewers reported preparations were applied to 400-mesh carbon-coated grids, nega- the HIV test results back to the patient, together with post-test tively stained with 1.5% uranyl acetate and then examined at a counselling. The study was approved by the Committee on Human nominal magnification of 50 000 with a Jeol 1010 electron micro- Research (VA Medical Centre and University of California, San scope. VLPs were then diluted in PBS and used in the following Francisco, CA, USA) and by the Uganda National Council for ELISA tests (Touze ´ et al, 1998; Combita et al, 2002a). Science and Technology. Flat-bottomed wells of 96-well microplates (Maxisorp, Nunc, Life Technologies, Eragny, France) were coated overnight at 48C with 200 – 400 ng of VLPs (test well) or 200 ng of BSA (control well) in Laboratory diagnoses PBS, pH 7.4. Each serum sample was tested twice against each of Histology Diagnoses of cancer were established by histology or the seven VLP types and BSA at the same time on the same plate. other laboratory investigation, where possible. Diagnoses made After washing with PBS, 0.1% Tween 20 and 200 ml of PBS contain- on clinical grounds alone, were reviewed by the investigators (C ing 1% newborn bovine serum (NBS, Sigma, St Quentin Favallier, A-A, JZ, EM and SM). Among the cases, 36 (60%) underwent France) were added (2 h at 378C). The blocking solution was histological review (all were invasive tumours) and the remainder replaced by 100 ml of sera diluted 1 : 20 in 5xPBS-10% NBS and were diagnosed clinically. Among controls, 63% of tumour diag- 2% Tween 20, and plates were incubated at 458C for 60 min. After noses were verified by laboratory investigation (histology, four washes, bound antibodies were detected with a goat anti-human cytology, blood chemistry or ultrasound examination). IgG immunoglobulin (diluted 1 : 5000) conjugated to horseradish peroxidase (Sigma). Following incubation at 458C for 1 h and four washes, 100 ml of a substrate solution containing ortho-phenylene- HIV-1 HIV-1 serostatus was determined using a single ELISA, the Cambridge Bioscience Recombigen enzyme linked immunosor- diamine and H O was added. After 30 min incubation, the 2 2 bent assay (Cambridge, MA, USA). Periodic laboratory quality reaction was stopped by addition of 100 mlof 4N H SO , and optical 2 4 control assays using blinded standards from the United States densities (OD) were read at 492 nm with an automated plate reader Centers for Disease Control (Atlanta, GA, USA) revealed test sensi- (BioRad, model 550). For each serum sample the background reac- tivity and specificity of 99%. In the initial months of the study, tivity found in the BSA coated wells was subtracted from the OD 40 – 50% of adults declined to have venepuncture, mostly because found in each of the HPV – VLP coated wells. Negative values were they were too ill. Therefore, a GACELISA saliva test for HIV was adjusted to zero. The cut-off values for positivity to HPV-16, -18 introduced as an option (with sensitivity and specificity similar and -45 were set up at 0.2 (OD test well minus OD control well) to the blood test), and the refusal rate dropped to 10%. In total, and those with a value of 0.4 or greater were considered to have a HIV test results were available on 57 cases and 826 controls. high titre of anti-HPV antibodies. In total, 457 samples were available However, for one case and seven controls the result was indetermi- for testing for antibodies against HPV-16 (39 cases and 418 controls). nate. Any remaining aliquots of sera were stored at 7808C and In addition, there were sufficient sera for 453 of those to be tested for were later shipped on dry ice to the Laboratoire de Virologie Mole ´ - antibodies against HPV-18 and -45 (39 cases and 414 controls). British Journal of Cancer (2002) 87(3), 301 – 308 ã 2002 Cancer Research UK Conjunctival cancer in Uganda R Newton et al Table 1 Distribution of region of birth, region of residence, tribe, nation- KSHV All assays were performed by a single investigator (DB), ality, HIV-1 sero-status, income, age left home and time spent cultivating who was unaware of each patient’s personal characteristics, and diag- among cases with conjunctival carcinoma and controls with other can- nosis. A body-cavity-related B-cell lymphoma cell, BCP-1, which is cers, in Uganda positive for KSHV and negative for Epstein – Barr virus, was used for an indirect immunofluorescence assay to detect IgG antibodies Number of Odds ratio and 95% against KSHV antigen (Gao et al, 1996; Boshoff et al, 1998). Latently cases/controls confidence intervals infected BCP-1 cells were fixed in 4% paraformaldehyde and were made permeable with 0.2% Triton X-100. Cells were re-suspended Region of birth Kampala 4/54 1.0 in phosphate-buffered saline and fixed on glass slides. The samples Rest of Uganda 59/982 1.2 were diluted 1 : 100 in phosphate-buffered saline with 3% foetal calf w heterogeneity (1df)=0.1; P=0.7 serum. The diluted serum was added to fixed BCP-1 cells and incu- Region of residence bated at 228C for 45 min. After the slides were washed in Kampala 25/272 1.0 phosphate-buffered saline with 3% foetal calf serum, rabbit antihu- Rest of Uganda 35/922 0.6 man IgG labelled with fluorescein isothiocyanate (Dako, High w heterogeneity (1df)=2.3; P=0.1 Wycombe, UK), diluted 1 : 40 in phosphate-buffered saline with Tribe 3% foetal calf serum, was added and the slides were incubated at Baganda 27/631 1.0 Other 33/560 1.4 228C for 20 min. The slides were then washed in phosphate-buffered w heterogeneity (1df)=1.3; P=0.3 saline without foetal calf serum and screened by ultraviolet micro- Nationality scopy for the nuclear stippling pattern characteristic of antibodies Ugandan 56/1132 1.0 against the latent nuclear antigen of KSHV encoded by orf73 (Gao Other 4/59 1.4 et al, 1996). In total, 416 samples were available for testing for anti- w heterogeneity (1df)=0.3; P=0.6 bodies against KSHV (32 cases and 384 controls). HIV-1 sero-status Negative 17/700 1.0 Positive 39/119 10.1 Statistical methods w heterogeneity (1df)=52.9; P50.001 Personal income (SH) Data were computerised by trained clerks using EPI-INFO-5 soft- 520,000 30/408 1.0 ware, and statistical analyses were conducted using STATA 7.0 20,000+ 29/552 0.4 (Dean et al, 1990; STATA Corp, 2001). The odds ratio (OR) for each w heterogeneity (1df)=9.5; P50.001 variable in cases versus controls, was estimated with unconditional Household income (SH) logistic regression. All odds ratios were adjusted for age (15 – 24, 530,000 33/485 1.0 25 – 34, 35 – 44 and 45+), sex, HIV-1 sero-status and personal income 30,000+ 22/419 0.6 (less than 20 000 Uganda Shillings per year or 20 000+ Shillings). w heterogeneity (1df)=1.9; P=0.2 Additional adjustment was made for region of residence (Kampala Age left home (years) or elsewhere), because most cases had been identified from out- 21+ (or never left) 22/330 1.0 patient clinics, which tend to draw patients from the local area, 15 – 20 26/375 0.7 1 – 14 11/292 0.4 whereas controls had been hospital in-patients, who are from a wider w trend (1df)=3.9; P=0.05 area (Ziegler et al, 1997). These factors were selected a priori, with the Time spent cultivating exception of personal income, which was included following preli- (Hours per week) minary analyses. HIV infection is sexually transmitted and is 0 – 9 9/241 1.0 strongly associated with many of the sexual and reproductive vari- 10 – 19 23/401 1.9 ables studied (data not shown). In order to avoid confounding, 20+ 26/348 2.4 these factors were examined among HIV seropositive individuals w trend (1df)=3.9; P=0.05 only (there were too few HIV seronegative people to justify analysis). Odds ratios adjusted for age group, sex, current region of residence, HIV-1 status All P values are two-sided. Note that numbers of cases and controls in and personal income. df=degrees of freedom. the tables do not always add to the total, because of missing values. RESULTS Among those with conjunctival cancer, 43% (26 out of 60) were men Table 2 shows the results for anti-HPV and KSHV antibodies. and 57% (34 out of 60) were women. The proportion of all cancers The seroprevalence of anti-HPV antibodies in controls was 10% comprising conjunctival carcinoma declined from 9% in those aged for HPV-16 (43 out of 418), 4% (16 out of 414) for HPV-18 15 – 24 years to 2% in those over the age of 45 years. Seven per cent and 6% (24 out of 414) for HPV-45. The corresponding results of cases and 5% of controls were born in Kampala, the remainder for those with conjunctival cancer were 21% (eight out of 39), being born outside the capital city (P=0.7) and, 41% of cases and 10% (four out of 39) and 5% (two out of 39) respectively. 23% of controls reported their current residence as being in Kampala However, after adjustment for age, sex, address, HIV status and (Table 1; P=0.13). The seroprevalence of anti-HIV-1 antibodies was personal income, there were no statistically significant associations 70% among cases and 15% among controls (Odds ratio [OR] 10.1, between the presence of anti-HPV-16, -18 and -45 antibodies and 95% confidence intervals [CI] 5.2 – 19.4; P50.001). The risk of the risk of conjunctival carcinoma. Results for each HPV subtype conjunctival carcinoma was significantly lower among those with a were also calculated according to a measure of the antibody titre: high personal income (OR 0.4, 95% CI 0.2 – 0.7; P50.001). For the optical densities at each level correspond to less than 0.2 for those who left home at ages 21+ years (including those who never negative, 0.270.39 for medium titre and 0.4 or above for high left), 15 – 20 years and 1 – 14 years, the odds ratio was 1.0 (reference titre. The numbers of cases and controls with anti-HPV antibodies group), 0.7 (0.4 – 1.5) and 0.4 (0.2 – 1.0) respectively (P =0.05). to subtypes -18 and -45 were too few to yield any significant trend Study participants were asked how long each week they spent culti- results. The results for anti-HPV-16 antibodies at each measure vating, 0 – 9 h, 10 – 19 h or 20+ h. The risk of conjunctival of titre were 1.0 (HPV-16 antibody negative, based on 31 cases carcinoma increased significantly with increasing time spent cultivat- and 375 controls), 0.7 (0.2 – 2.9; medium titre, based on four cases ing (ORs 1.0, 1.9 and 2.4 respectively; P =0.03). and 31 controls) and 6.3 (1.2 – 33.4; high titre, based on four cases trend ã 2002 Cancer Research UK British Journal of Cancer (2002) 87(3), 301 – 308 Epidemiology Epidemiology Conjunctival cancer in Uganda R Newton et al Table 2 Comparison of human papillomavirus antibodies (HPV types 16, were conducted on over 50 risk factors, greatly increasing the like- 18 and 45) and Kaposi’s sarcoma-associated herpesvirus (KSHV) antibodies lihood of significant results arising by chance alone. The role of between those with conjunctival cancer and those without HPV-16, -18 and -45 requires further evaluation, but KSHV does not appear to be associated with conjunctival carcinoma. Number of Odds ratio and 95% The study reported here is subject to the potential problems of cases/controls confidence intervals incomplete diagnostic verification and HIV testing. Laboratory veri- fication of cancer diagnosis (for example, by histology, cytology or HPV-16 Negative 31/375 1.0 blood chemistry) was available on 60% of cases (all of which were Positive 8/43 1.5 (0.5 – 4.3) invasive tumours) and 63% of controls. Typical of studies in devel- w heterogeneity (1df)=0.5; P=0.5 oping countries where laboratory services are limited, this HPV-18 proportion compares favourably with other cancer series reported Negative 35/398 1.0 from Africa (Bassett et al, 1995; Newton et al, 1996a; Wabinga et Positive 4/16 2.3 (0.5 – 9.6) 2 al, 2000). Another potentially important source of bias is the high w heterogeneity (1df)=1.2; P=0.3 proportion of adults who were not tested for HIV in the early HPV-45 stages of the study. However, the age and sex-specific seropreva- Negative 37/390 1.0 Positive 2/24 0.4 (0.1 – 2.1) lence rates of HIV infection among those tested is broadly similar w heterogeneity (1df)=1.5; P=0.2 to that found in studies of HIV seroprevalence in Uganda (STD/ HPV-16, -18 or -45 AIDS Control Programme, Uganda, 1997). Also, the introduction All negative 28/345 1.0 of a saliva test resulted in a dramatic reduction in refusals, but little Any positive 11/69 1.3 (0.5 – 3.2) change in HIV seroprevalence rates, suggesting that patients who w heterogeneity (1df)=0.4; P=0.6 were refusing to have a HIV test were unwilling to have blood KSHV taken, rather than unwilling to learn their HIV serostatus. The asso- Negative 17/196 1.0 ciation of conjunctival cancer with HIV infection has been reported Positive 15/188 0.9 (0.4 – 2.1) in a general paper on HIV and cancer among residents of Kampala, w heterogeneity (1df)=0.1; P=0.8 and the odds ratio was the same as that from the current study of Adjusted for age, sex, address, personal income and HIV-1 sero-status. all patients who were seen at hospitals in Kampala (Newton et al, 2001). A proportion of individuals in the study were also tested for antibodies to three HPV subtypes and to KSHV infection. This additional testing occurred in all those for whom enough stored and 12 controls; P =0.2). Only 15 people had anti-HPV antibo- sera were available and (as people are unaware of their own status trend dies to more than one tested HPV subtype (two cases and 13 with respect to infection), is therefore an unbiased sample. controls) and there was no significant excess risk of the tumour Another potential source of bias arises from the fact that most in these individuals, as compared to those who were considered (but not all) cases had been identified from out-patient clinics, to be negative for all three subtypes (OR 0.6, 95% CI 0.1 – 4.3). which tend to draw patients from the local area, whereas controls In relation to Kaposi’s sarcoma-associated herpesvirus, the seropre- had been hospital in-patients, who are from a wider area. Despite valence of anti-KSHV antibodies was 47% (15 out of 32) among adjusting for region of residence in the analyses, the possible cases and 49% (188 out of 384) among controls (OR 0.9, 95% impact of this on the results, is difficult to assess. CI 0.4 – 2.1; P=0.8). Case reports of squamous cell carcinoma of the conjunctiva in Further results are provided in Appendices 1 – 4. Results for human immunodeficiency virus (HIV) infected men in the USA other social and demographic factors, none of which was associated and France, coupled with a marked increase in the numbers of with an increased risk of conjunctival carcinoma, are shown in tumours being seen by ophthalmologists in at least two African Appendix 1. Factors, other than income, that might characterise centres (which mirrored the increases seen for Kaposi’s sarcoma), wealth and social status in Uganda, such as ownership of livestock led to the suggestion of an association with HIV (Winward and and travel away from home, are shown in Appendix 2. None of Curtin, 1989; Ateenyi-Agaba, 1995; Kestelyn et al, 1990; Kim et al, these variables were associated with conjunctival carcinoma. 1990; Denis et al, 1994). Among those with HIV, the lesions often Appendix 3 shows results for other possible exposures. Those affect young adults, in a fashion reminiscent of Kaposi’s sarcoma in who reported having had a blood transfusion in the past were less HIV-seropositive people (Ateenyi-Agaba, 1995; Kestelyn et al, 1990; likely to be cases (OR 0.4, 95% CI 0.2 – 0.9; P=0.02). Results for Waddell et al, 1996). This study, together with others from Africa sexual and reproductive variables among HIV seropositive people and the USA indicate about a 10-fold increased risk of the tumour only, are shown in Appendix 4. A self reported history of having in HIV infected, compared to HIV uninfected individuals (Ateenyi- either given or received gifts in exchange for sex was associated Agaba, 1995; Kestelyn et al, 1990; Goedert and Cote, 1995; Newton with an increased risk of the tumour (OR 3.5, 95% CI 1.2 – 10.4; et al, 1995, 2001; Waddell et al, 1996). Indeed, the spread of HIV P=0.03), although this may have been a chance finding as there in Uganda probably accounts for much of the approximately eight- was no statistically significant association with any other sexual fold increase in incidence of conjunctival carcinoma observed there or reproductive variable, including the reported number of lifetime since the 1960s (Wabinga et al, 2000). Using standard equations for sexual partners (P=0.4). case – control studies, the population attributable fraction is about 60%: almost two-thirds of cases would not occur in the absence of HIV infection (dos Santos Silva, 1998). However, the frequency of DISCUSSION conjunctival carcinoma is not such that it is yet a particularly Here, we report an investigation of the epidemiology and aetiology common manifestation of HIV disease (Piot et al, 1992). of conjunctival squamous cell carcinoma, examining over 50 risk The mechanism whereby HIV infection increases the risk of factors for the tumour. Conjunctival cancer affects relatively young conjunctival cancer is not clear. There is no evidence that it is individuals of both sexes and is strongly associated with HIV infec- directly oncogenic and the impact on cancer risk is most likely tion and poverty. In addition, the risk of the disease decreases with to be mediated via immunosuppression, as is the case for other decreasing age at which an individual leaves home and with HIV-associated cancers (Beral and Newton, 1998; Newton et al, increasing time spent cultivating, although both results are of 1999). It is not known if the risk of conjunctival squamous cell borderline statistical significance. Furthermore, significance tests carcinoma is increased in other immunosuppressed groups, such British Journal of Cancer (2002) 87(3), 301 – 308 ã 2002 Cancer Research UK Conjunctival cancer in Uganda R Newton et al as transplant recipients, because the tumour is very rare in parts of that the incidence of squamous carcinoma of the eye increases as the world where tissue transplantation occurs. However, there is a exposure to ambient solar ultraviolet radiation increases. Levels case report of a conjunctival cancer in a patient with malignant of solar radiation are higher towards the equator and the incidence lymphoma on immunosuppressive chemotherapy (Kushner and of conjunctival cancer increases by about 50% for each 108 decline Mushen, 1975). In general, HIV infection is thought to increase in latitude (Newton et al, 1996b; Sun et al, 1997). In addition, a cancer risk by facilitating the action of other oncogenic viruses, single case – control study reported that the risk of ocular surface such as KSHV, the principal cause of Kaposi’s sarcoma. In relation epithelial dysplasias is greatest in those reporting a past history to conjunctival cancer however, we found no evidence in these data of skin cancer, which is known to be caused by exposure to solar of an association with KSHV. ultraviolet radiation (Lee et al, 1994). Time spent cultivating may Several types of squamous carcinoma are associated with human indicate the time an individual spends in direct sunlight, and so papillomavirus (HPV) infection, most notably cancer of the uterine the increasing risk with increasing time cultivating may reflect cervix, induced primarily by HPV-16, -18 and others. Squamous exposure to ultraviolet radiation. This variable may also reflect carcinoma of the skin has also been associated with HPV-5 and exposure to dust or dirt, and it has been hypothesised that ocular -8 in immunosuppressed individuals (IARC, 1995). Evidence for trauma could facilitate development of the tumour (Templeton, an association between human papillomavirus and squamous cell 1973; Margo and Groden, 1986). The lower risk among people carcinoma of the conjunctiva is conflicting, although bovine papil- who leave home at an earlier age may reflect migration to towns lomaviruses are thought to cause conjunctival carcinomas in cattle for work, where exposure to solar ultraviolet radiation could be (IARC, 1995). The presence of human papillomavirus DNA (HPV; less. The extent to which the higher risk among those of low predominantly types 16, but also other types) in human ocular income is measuring the previously mentioned exposures is not surface squamous neoplasias, including invasive carcinomas, has clear. The fact that cases are less likely than controls to have had been reported in some studies, but not others (reviewed by a blood transfusion is probably an artefact of the control selection. Newton, 1996). This is the first study to look for an association Cases were generally seen as out-patients, whereas controls with between anti-HPV antibodies (types -16, -8 and -45) and the risk other cancers were generally hospital in-patients. The latter may of conjunctival carcinoma. The seroprevalence of antibodies against therefore be more likely to have had a transfusion as part of their HPV-18 and -45 was too low to make reliable conclusions. Results medical care. for anti-HPV-16 antibodies were suggestive of an association In summary, the current study gives independent support for among individuals with high titres, but the data presented here the strong epidemiological evidence that solar ultraviolet radiation are too few to draw valid conclusions and studies of larger is an important cause of squamous cell carcinoma of the conjunc- numbers of cases are required. tiva. Another established risk factor is HIV infection, although the The HPV assays used in this study are based on the expression mechanism whereby it increases the risk of conjunctival cancer is of L1 major capsid proteins of HPV-16, -18 and -45 in insect cells not clear. We find little evidence supporting the possible role of by using recombinant baculoviruses (Touze ´ et al, 1998). The result- sexually transmitted forms of HPV in the aetiology of conjunctival ing virus-like particles (VLPs), which appear similar to empty cancer, but larger studies are required. virions, can be used in serological studies to test for type specific immunological responses to viral capsid proteins, although there is evidence that a particular assay may cross react with related ACKNOWLEDGEMENTS HPV subtypes (Combita et al, 2002b). Presence of anti-VLP anti- bodies is an indicator of past and current infection (Kirnbauer et Support for this work was provided by Mulago Hospital and al, 1994; Le Cann et al, 1995; Wideroff et al, 1995; Dillner et al, Makerere Medical School, Kampala, Uganda, the Imperial Cancer 1996). The utility of such assays has been demonstrated in previous Research Fund (now Cancer Research UK), the United States studies of anti-HPV-16 antibodies in relation to the risk of cancer Centers for Disease Control and Prevention (interagency agreement of the uterine cervix (Lehtinen et al, 1996; Dillner et al, 1997; Shah with the Department of Veterans Affairs), the International Agency et al, 1997; Vonka et al, 1999; Hisada et al, 2001). Results from this for Research on Cancer, World Health Organisation, Lyon, France study on the relationship between anti-HPV-16 antibodies and the and the Ligue Contre le Cancer, France. J Ziegler was on second- risk of cancer of the uterine cervix are broadly comparable to those ment from the Department of Veterans Affairs and the reported before and will be the subject of a separate report. University of California, San Francisco, CA, USA, to the Interna- Although relatively rare everywhere, conjunctival carcinoma is tional Agency for Research on Cancer, France. L Bousarghin was more frequent in sub-Saharan Africa and other tropical areas than the recipient of a fellowship from the Region Centre, France. L in temperate countries, leading to the hypothesis that it may be Carpenter was on secondment from the Academic Department of associated with exposure to solar ultraviolet radiation (Templeton, Public Health, Oxford University, to the MRC Programme on 1973; Clear, 1979). Geographical studies support this premise, in AIDS, Entebbe, Uganda. REFERENCES Ateenyi-Agaba C (1995) Conjunctival squamous cell carcinoma associated Boshoff C, Gao SJ, Healy LE, Matthews S, Thomas AJ, Coignet L, Warnke RA, with HIV infection in Kampala, Uganda. Lancet i: 695 – 696 Strauchen JA, Matutes E, Kamel OW, Moore PS, Weiss RA, Chang Y Bassett MT, Chokunonga E, Mauchaza B, Levy L, Ferlay J, Parkin DM (1995) (1998) Establishing a KSHV+ cell line (BCP-1) from peripheral blood Cancer in the African population of Harare, Zimbabwe in 1990 – 92. Int J and characterising its growth in Nod/SCID mice. Blood 91: 1671 – 1679 Cancer 63: 29 – 36 Clear AS (1979) Solar keratosis, pterygium and squamous cell carcinoma of Beral V, Newton R (1998) Overview of the epidemiology of immunodefi- the conjunctiva in Malawi. Br J Ophthalmol 63: 102 – 109 ciency associated cancers. Monogr Natl Cancer Inst 23: 1– 6 Combita AL, Bravo MM, Touze ´ A, Orozco O, Coursaget P (2002a) Serologic Bosch FX, Manos MM, Munoz N, Sherman M, Jansen AM, Peto J, Schiffman response to Human oncogenic papillomavirus types 16, 18, 31, 33, 39, 58 MH, Moreno V, Kurman R, Shah KV The International Biological Study and 59 virus-like particles in Colombian women with invasive cervical on Cervical Cancer (IBSCC) Study Group (1995) Prevalence of human cancer. Int J Cancer 97: 796 – 803 papillomavirus in cervical cancer: a worldwide perspective. J Natl Cancer Inst 87: 796 – 802 ã 2002 Cancer Research UK British Journal of Cancer (2002) 87(3), 301 – 308 Epidemiology Epidemiology Conjunctival cancer in Uganda R Newton et al Combita AL, Touze A, Bousarghin L, Christensen ND, Coursaget P (2002b) Margo CE, Groden LR (1986) Squamous cell carcinoma of the cornea and Identification of two cross-neutralizing linear epitopes within the L1 major conjunctiva following a thermal burn of the eye. Cornea 5: 185 – 188 capsid protein of human papillomavirus. J Virol (in press). Newton R, Grulich A, Beral V, Sindikubwabo B, Ngilimana P-J, Nganyira A, Dean AD, Dean JA, Burton JH, Dicker RC (1990) Epi Info, Version 5: a word Parkin DM (1995) Cancer and HIV infection in Rwanda. Lancet i: 1378 processing, database and statistics program for epidemiology on micro- Newton R (1996) A review of the aetiology of squamous cell carcinoma of the computers. Centers for Disease Control, Atlanta, Georgia, USA. conjunctiva. Br J Cancer 74: 1511 – 1513 Denis P, Charpentier D, Roudier M, Lemaire JC, Levin N, Nguyea-Khoa JL Newton R, Ngilimana PJ, Grulich A, Beral V, Sindikubwabo B, Nganyira A, (1994) Conjunctival epidermoid carcinoma and human immunodeficiency Parkin DM (1996a) Cancer in Rwanda. Int J Cancer 66: 75 – 81 virus. J Fr Ophtalmol 17: 366 – 369 Newton R, Ferlay J, Reeves G, Beral V, Parkin DM (1996b) Incidence of Dillner JL, Kallings I, Brihmer C, Sikstrom B, Koskela P, Lehtinen M, Schiller squamous cell carcinoma of the eye increases with increasing levels of JT, Sapp M, Mardh PA (1996) Seropositivities to human papillomavirus ambient solar ultraviolet radiation. Lancet i: 1450 – 1451 type 16, 18 or 33 capsids and to Chlamydia trachomatis are markers of Newton R, Beral V, Weiss R (1999) Human immunodeficiency virus infec- sexual behaviour. J Infect Dis 173: 1394 – 1398 tion and cancer. In Cancer Surveys volume 33, Infections and Human Dillner JL, Lehtinen M, Bjorge T, Luostarinen T, Youngman L, Jellum E, Cancer, Newton R, Beral V, Weiss R (eds), pp 237 – 261. Cold Spring Koskela P, Gislefoss RE, Hallmans G, Paavonen J, Sapp M, Schiller JT, Harbor Laboratory Press Hakulinen T, Thoresen S, Hakama M (1997) Prospective seroepidemiolo- Newton R, Ziegler J, Beral V, Mbidde E, Carpenter L, Wabinga H, Mbulataiye gic study of human papillomavirus infection as a risk factor for invasive S, Appleby P, Reeves G, Jaffe H and the Uganda Kaposi’s Sarcoma Study cervical cancer. J Natl Cancer Inst 89: 1293 – 1299 Group (2001) A case-control study of Human Immunodeficiency Virus El Mehdaoui S, Touze ´ A, Laurent S, Sizaret PY, Rasschaert D, Coursaget P infection and cancer in adults and children residing in Kampala, Uganda. (2000) Gene transfer using recombinant rabbit hemorrhagic disease virus Intl J Cancer 92: 622 – 627 capsids with genetically modified DNA encapsidation capacity by addition Piot P, Kapita BM, Ngugi E, Mann JM, Colebunders RL, Wabitsch R (1992) of packaging sequences from L1 or L2 protein of human papillomavirus AIDS in Africa. A Manual for Physicains. WHO, Geneva type 16.. J Virol 74: 10332 – 10340 dos Santos Silva I (1998) Cancer Epidemiology Principles and Methods. IARC, Gao S-J, Kingsley L, Li M, Zheng W, Parravicini C, Ziegler JL, Newton R, Lyon Rinaldo CR, Saah A, Phair J, Detels R, Chang Y, Moore PS (1996) Seropre- Shah KV, Viscidi RP, Alberg AJ, Helzlsouer KJ, Comstock GW (1997) Anti- valence of Kaposi’s sarcoma-associated herpesvirus antibodies among bodies to human papillomavirus 16 and subsequent in situ or invasive Americans, Italians and Ugandans with and without Kaposi’s sarcoma. cancer of the cervix. Cancer Epidemiol Biomarkers Prev 6: 233 – 237 Nature Med 2: 925 – 928 STATA Corp (2001) Stata software, College Park, TX, USA Goedert JJ, Cote TR (1995) Conjunctival malignant disease with AIDS in STD/AIDS Control Programme HIV/AIDS Surveillance Report, March. USA. Lancet ii: 257 – 258 (1997) Ministry of Health, PO Box 8, Entebbe, Uganda Hisada M, van den Berg BJ, Strickler HD, Christianson RE, Wright WE, Sun EC, Fears TR, Goedert JJ (1997) Epidemiology of squamous cell conjunc- Waters DJ, Rabkin CS (2001) Prospective study of antibody to human tival cancer. Cancer Epidemiol, Biomark Preven 6: 73 – 77 papillomavirus type 16 and risk of cervical, endometrial and ovarian Templeton AC (1973) Tumours of the eye and adnexa. In Tumours of a cancers (United States). Cancer Causes Control 12: 335 – 341 Tropical Country: A survey of Uganda 1964 – 1968. Templeton AC (ed), IARC (1992) Monographs on the evaluation of carcinogenic risks to humans. pp 203 – 214 Recent Result Cancer Research Solar and Ultraviolet Radiation. IARC Monograph vol 55, Lyon Touze ´ A, el Mehdaoui S, Sizaret PY, Mougin C, Munoz N, Coursaget P IARC (1995) Monographs on the evaluation of carcinogenic risks to humans. (1998) The L1 major capsid protein of Human Papillomavirus Type 16 Human papillomaviruses.. IARC Monograph vol 64, Lyon variants affects the yield of virus-like particle produced in an insect cells Kestelyn P, Stevens AM, Ndayambaje A, Hanssens M, van de Perre P (1990) expression system. J Clin Microbiol 36: 2046 – 2051 HIV and conjunctival malignancies. Lancet 336: 51 – 52 Vonka V, Hamsikova E, Kanka J, Ludvikova V, Sapp M, Smahel M (1999) Kim RY, Seiff SR, Howes EL Jr, O’Donnell JJ (1990) Necrotizing scleritis Prospective study on cervical neoplasia IV. Presence of HPV antibodies. secondary to conjunctival squamous cell carcinoma in aquired immunode- Int J Cancer 80: 365 – 368 ficiency syndrome. Am J Opthalmol, 109: 231 – 233 Wabinga HR, Parkin DM, Wabwire-Mangen F, Nambooze S (2000) Trends Kirnbauer R, Hubbert NL, Wheeler CM, Becker TM, Lowy DR, Schiller JT in cancer incidence in Kyadondo County, Uganda, 196071997. Br J (1994) A virus-like particle enzyme-linked immunosorbent assay detects Cancer 82: 1585 – 1592 serum antibodies in a majority of women infected with human papilloma- Waddell KM, Lewallen S, Lucas SB, Ateenyi-Agaba C, Herrington CS, Liomba virus type-16. J Natl Cancer Inst 86: 494 – 499 G (1996) Carcinoma of the conjunctiva and HIV infection in Uganda and Kushner FH, Mushen RL (1975) Conjunctival squamous cell carcinoma Malawi. Br J Ophthalmol 80: 503 – 508 combined with malignant lymphoma. Am J Ophthalmol 80: 503 – 506 Wideroff L, Schiffman MH, Hoover R, Tarane TE, Nonnenmacher B, Le Cann P, Touze A, Enogat N, Leboulleux D, Mougin C, Legrand MC, Hubbert NL, Kirnbauer R, Greer CE, Lorincz AT, Manos MM, Glass Calvet C, Afoutou JM, Coursaget P (1995) Detection of antibodies against AG, Scott DR, Sherman ME, Buckland J, Lowy D, Schiller J (1995) Evalua- human papillomavirus (HPV) type-16 virions by ELISA using recombi- tion of sero-reactivity to human papillomavirus type 16 virus-like particles nant HPV 16 L1 capsids produced by recombinant baculovirus. J Clin in an incident case-control study of cervical neoplasia. J Infect Dis 172: Microbiol 33: 1380 – 1382 1425 – 1430 Lee GA, Williams G, Hirst LW, Green AC (1994) Risk Factors in the Devel- Winward KE, Curtin VT (1989) Conjunctival squamous cell carcinoma in a opment of Ocular Surface Epithelial Dysplasia. Ophthalmology 101: 360 – patient with human immunodeficiency virus infection. Am J Ophthalmol 364 107: 554 – 555 Lehtinen M, Dillner J, Knekt P, Luostarinen T, Aromaa A, Kirnbauer R, Ziegler JL, Newton R, Katongole-Mbidde E, Mbulataiye S, DeCock K, Wabin- Koskela P, Paavonen J, Peto R, Schiller JT, Hakama M (1996) Serologically ga H, Mugerwa J, Katabira E, Jaffe H, Parkin DM, Reeves G, Beral V for the diagnosed infection with human papillomavirus type 16 and risk for subse- Uganda Kaposi’s sarcoma study group (1997) Risk factors for HIV-asso- quent development of cervical carcinoma: nested case-control study. BMJ ciated Kaposi’s sarcoma in Uganda: a case-control study of 1026 Adults. 312: 537 – 539 AIDS 11: 1619 – 1626 British Journal of Cancer (2002) 87(3), 301 – 308 ã 2002 Cancer Research UK Conjunctival cancer in Uganda R Newton et al Appendix 1 Comparison of social and demographic factors between Appendix 2 Comparison of factors characterising wealth between people with and without conjunctival cancer those with and without conjunctival cancer Number of Odds ratio and 95% Number of Odds ratio and 95% a a cases/controls confidence intervals cases/controls confidence intervals Years at current address Number of rooms in house 510 27/220 1.0 1 – 2 31/306 1.0 10+ 33/778 0.8 3+ 29/696 0.8 2 2 w heterogeneity (1df)=0.8; P=0.4 w heterogeneity (1df)=0.4; P=0.5 Settlement type Electricity 1 – 99 houses 29/704 1.0 No 38/770 1.0 100+ houses 31/287 1.5 Yes 21/228 1.0 2 2 w heterogeneity (1df)=1.7; P=0.2 w heterogeneity (1df)=0.0; P=1.0 Travel time to market Own radio 530 mins 40/400 1.0 No 11/205 1.0 30+ mins 14/456 0.6 Yes 49/794 1.0 2 2 w heterogeneity (1df)=2.0; P=0.2 w heterogeneity (1df)=0.0; P=0.9 Age left school Own motorcar No school 7/270 1.0 No 56/890 1.0 515 years 25/289 1.4 Yes 4/109 0.5 16 – 19 years 20/262 1.2 w heterogeneity (1df)=1.8; P=0.2 20+ years 8/163 0.8 Own bicycle w trend (1df)=0.4; P=0.5 No 35/523 1.0 Education level Yes 25/478 1.0 No school 7/273 1.0 w heterogeneity (1df)=0.0; P=0.9 Primary 19/272 1.5 Own cows Secondary/tertiary 34/450 1.1 None 50/736 1.0 w trend (1df)=0.7; P=0.4 1+ 9/259 0.7 Religion w heterogeneity (1df)=0.7; P=0.4 Christian 55/882 1.0 Own pigs and goats Muslim 5/125 0.8 None 46/677 1.0 w heterogeneity (1df)=0.2; P=0.7 1+ 14/319 1.2 Occupation w heterogeneity (1df)=0.3; P=0.6 Cultivator 14/588 1.0 Own chickens Other 46/549 1.7 None 38/542 1.0 w heterogeneity (1df)=1.8; P=0.2 1+ 22/431 1.1 Number in house w heterogeneity (1df)=0.1; P=0.8 55 25/383 1.0 Travel away from home 4 5+ 35/621 1.2 7 nights per year w heterogeneity (1df)=0.3; P=0.6 No 53/877 1.0 Number of siblings Yes 7/105 1.3 55 32/567 1.0 w heterogeneity (1df)=0.2; P=0.7 5+ 28/438 0.8 Adjusted for age group, sex, address, personal income and HIV-1 sero-status; w heterogeneity (1df)=0.7; P=0.4 df=degrees of freedom. Birth order 1 – 2 25/467 1.0 3+ 35/538 0.9 w heterogeneity (1df)=0.1; P=0.7 Number sharing toilet 56 19/358 1.0 6+ 40/610 1.1 w heterogeneity (1df)=0.0; P=0.9 Adjusted for age group, sex, address, personal income and HIV-1 sero-status; df=degrees of freedom. ã 2002 Cancer Research UK British Journal of Cancer (2002) 87(3), 301 – 308 Epidemiology Epidemiology Conjunctival cancer in Uganda R Newton et al Appendix 3 Comparison of other exposures between those with and Appendix 4 Comparison of sexual and reproductive variables between without conjunctival cancer those with conjunctival cancer and those without, among HIV seropositive individuals only Number of Odds ratio and 95% Number of Odds ratio and 95% cases/controls confidence intervals cases/controls confidence intervals Use of shoes Never/rarely 26/619 1.0 Marital status Often/always 33/372 1.5 Single 1/10 1.0 w heterogeneity (1df)=1.5; P=0.2 Monogomous married 22/43 8.2 Age first use shoes (years) Polygamous married 9/22 9.8 515 36/423 1.0 Widower 4/26 3.3 15+ 23/532 1.1 Separated 3/10 4.2 2 2 w heterogeneity (1df)=0.1; P=0.7 w heterogeneity (1df)=6.6; P=0.16 Soil colour Age first sex Black 36/549 1.0 518 years 26/65 1.0 Other 16/343 1.0 18+ years 13/46 0.7 2 2 w heterogeneity (1df)=0.0; P=0.9 w heterogeneity (1df)=0.8; P=0.4 Time spent in water Number of children (hours per week) 0 – 2 13/23 1.0 51 hour 47/728 1.0 3 – 4 9/25 0.7 1+ hours 12/272 0.6 5 – 6 9/20 1.0 w heterogeneity (1df)=2.9; P=0.1 7+ 5/30 0.5 Tobacco use w trend (1df)=0.4; P=0.5 Never smoker 46/762 1.0 Children with different Ex-smoker 7/146 0.9 mothers or fathers Current smoker 7/91 1.1 No 13/40 1.0 w heterogeneity (1df)=0.0; P=1.0 Yes 21/53 1.4 Use of snuff w heterogeneity (1df)=0.5; P=0.5 No 58/977 1.0 No. sexual partners Yes 2/16 2.0 1 – 4 13/50 1.0 w heterogeneity (1df)=0.6; P=0.4 5 – 7 10/17 2.7 Drink home-brewed alcohol 8+ 16/40 1.8 No 36/533 1.0 w trend (1df)=0.8; P=0.4 About once/week 6/155 0.4 History sexual discharge 2 – 4/week 8/131 1.0 No 13/48 1.0 Most days 9/178 0.9 Yes 26/64 1.9 2 2 w heterogeneity (1df)=3.3; P=0.3 w heterogeneity (1df)=2.1; P=0.2 Blood transfusion Gifts for sex No 53/184 1.0 Never 29/101 1.0 Yes 7/183 0.4 Ever 10/11 3.5 2 2 w heterogeneity (1df)=5.6; P=0.02 w heterogeneity (1df)=5.0; P=0.03 Injection from traditional Use condoms healer Never 28/83 1.0 No 39/596 1.0 Ever 11/28 0.7 Yes 21/374 1.0 w heterogeneity (1df)=0.5; P=0.5 w heterogeneity (1df)=0.0; P=0.9 Circumcised (men) No 17/34 1.0 Odds ratios adjusted for age, sex, address, personal income and HIV-1 sero-status. Yes 0/7 0.0 w heterogeneity (1df)=1.3; P=0.3 Labial elongation (women) No 5/17 1.0 Yes 16/51 0.9 w heterogeneity (1df)=0.1; P=0.8 Adjusted for age, sex, address and personal income. British Journal of Cancer (2002) 87(3), 301 – 308 ã 2002 Cancer Research UK

Journal

British Journal of CancerSpringer Journals

Published: Jul 23, 2002

There are no references for this article.