SEARCH

SEARCH BY CITATION

Keywords:

  • KSHV/HHV-8;
  • Kaposi's sarcoma;
  • HIV;
  • Uganda

Abstract

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

As part of a larger investigation of cancer in Uganda, we conducted a case-control study of Kaposi's sarcoma in human immunodeficiency virus-1 (HIV)-seronegative adults presenting at hospitals in Kampala. Cases comprised 117 HIV-seronegative patients with Kaposi's sarcoma and controls comprised 1,282 HIV-seronegative patients with a provisional diagnosis of cancer other than Kaposi's sarcoma. Study participants were interviewed about social and lifestyle factors, tested for HIV and, if there was sufficient sera, for antibodies against Kaposi's sarcoma-associated herpesvirus (KSHV or human herpesvirus 8 [HHV8]), using an immunofluorescent assay. Independent effects of these factors were identified using unconditional logistic regression, after adjusting for age group (<30, 30–44, 45+) and sex. Antibody status for KSHV was available for 68% (80) of cases and for 45% (607) of controls. Among cases, 78% (91) were male and 57% (66) were over the age of 35. Cases were more likely than controls to be from tribal groups other than the Baganda (p = 0.05), to have higher household incomes (p = 0.003), to have left their home region at younger ages (p < 0.001), to own goats or pigs (p = 0.02) and to rarely or never use shoes (p < 0.001). Similar results were obtained when analyses were restricted to cases and controls with anti-KSHV antibodies. The seroprevalence of KSHV was 79% (63/80) in those with Kaposi's sarcoma as compared to 50% (302/607) in those without (χ2 heterogeneity (1 df) = 21.0; p < 0.001) and the risk of the tumour increased with increasing anti-KSHV antibody titres (χ2 trend (1 df) = 29.7; p < 0.001). The risk of Kaposi's sarcoma is clearly linked to antibody status for KSHV, but it would seem that in Uganda other factors are also important in the development of the tumour. © 2002 Wiley-Liss, Inc.

Before the human immunodeficiency virus (HIV) epidemic, Kaposi's sarcoma had a much greater geographical variation in incidence than most other malignancies. It was as common in parts of Africa as colon cancer is in Europe and the United States, representing up to 9% of all cancers in men.1, 2, 3, 4 Despite this, relatively little is known about the epidemiology of Kaposi's sarcoma in Africa among HIV-seronegative people.5 Since the onset of the HIV epidemic, the incidence of Kaposi's sarcoma has increased markedly in Uganda such that it is now the most common cancer in men and second only to cancer of the uterine cervix in women.6, 7 The epidemiology of Kaposi's sarcoma among HIV-seropositive Ugandan adults has been described elsewhere.8 The aim of our study was to identify risk factors for Kaposi's sarcoma among HIV-seronegative adults in Uganda, using a case-control study. A proportion of cases and controls were tested for antibodies against Kaposi's sarcoma-associated herpesvirus (KSHV). The sero-epidemiology of KSHV among the controls is described in Newton et al. on pages 226–232.9

MATERIAL AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Study participants

The participants included in our analyses were selected from a large study of risk factors for cancer in Uganda, which included, in total, 2,091 patients with malignancies other than Kaposi's sarcoma and 669 patients diagnosed with Kaposi's sarcoma. From August 1994 to February 1998, adults (aged 15 years and over) with a provisional new diagnosis of cancer were eligible for recruitment into the study from all the wards and out-patient clinics of the 4 main hospitals in Kampala, Uganda: Mulago (including the Uganda Cancer Institute), Nsambya, Mengo and Rubaga. Further details of our study design can be found elsewhere.8, 9, 10 The subjects included in our report are restricted to 117 HIV-seronegative patients with Kaposi's sarcoma (cases) and 1,282 HIV-seronegative patients with other cancers (controls). This latter group included people with cancers of the oral cavity (35), oesophagus (91), stomach (40), liver (68), skin (39), breast (120), uterine cervix (343), ovary (37), prostate (33), penis (22), eye (32), other cancers (222) and lymphomas (66). The control group also included 134 people who subsequently turned out to have benign lesions.

Recruitment and questionnaire

Interviewers (4 in total, who were also trained HIV counselors) approached ward or clinic staff to gain permission to interview potential recruits. If permission was granted, the patient was approached, invited to participate in our study and, if they consented, was interviewed about sociodemographic and lifestyle factors, including sexual and reproductive history. Where possible, the interview was conducted by a counselor of the same sex and in their native language. An HIV test was performed on serum or saliva and counseling provided (see Newton et al., pages 226–232, for further details of the HIV testing).10 Patients who had been, HIV tested within a month of recruitment and had a medical certificate indicating the test result were not retested. Blood was drawn, primarily for the HIV test, but also for storage of serum and leukocytes. Interviewers reported the HIV test results back to the patient, together with post-test counseling. Our study was approved by the Committee on Human Research (VA Medical Centre and University of California, San Francisco, California) and by the Uganda National Council for Science and Technology.

Laboratory diagnoses

Diagnoses of cancer were established by histology or other laboratory investigation, where possible. Diagnoses made on clinical grounds alone were reviewed by the investigators (J.Z., E.M. and S.M.). In 90% of those with Kaposi's sarcoma and in 63% of those with other cancers, the diagnosis was verified by laboratory investigation (e.g., histology, cytology, blood chemistry or ultrasound examination). HIV-1 serostatus was determined using a single ELISA, either the Cambridge Bioscience Recombigen enzyme-linked immunosorbent assay (Cambridge, MA) or a GACELISA saliva test (Murex, Dartford, UK). Periodic laboratory quality-control assays using blinded standards from the United States Centers for Disease Control (Atlanta, GA) revealed test sensitivity and specificity for both assays of 99%.

Remaining sera were stored at −80°C and later shipped on dry ice to University College London for KSHV testing. Details of the testing procedure are described elsewhere.9, 11, 12, 13, 14 Briefly, a single investigator (D.B.) performed the immunofluorescence assays using a BCP-1 cell line infected with KSHV but not Epstein Barr virus (EBV). Results were graded by fluorescent intensity and are reported here as negative, 1 plus, 2 plus and 3 plus, indicating absent, weak, moderate and strong fluorescent signals, respectively. The fluorescent signal intensity was related to anti-KSHV antibody titre—the median titres were 1:800 for low-signal intensity (1+), 1:25,600 for medium-signal intensity (2+) and 1:51,200 for high-signal intensity (3+). Further details of the scoring mechanism for signal intensity are available in Newton et al. on pages 226–232.9

Anti-KSHV antibody status was available for 68% (80) of cases and for 45% (607) of controls, for whom a stored blood sample was available for testing. This latter group is the subject of Newton et al. on p. 226–232,9 and included men (201) and women (406) with cancers of the oral cavity (21), oesophagus (36), stomach (19), liver (41), skin (21), breast (58), cervix (140), ovary (18), prostate (11), penis (10), eye (18) and non-Hodgkin's lymphoma (28), Hodgkin's disease (19), other cancer sites or types (104) and patients with a provisional diagnosis of cancer who subsequently were diagnosed with a nonmalignant condition (63).

Statistical methods

Data were computerised by trained clerks using EPI-INFO software, and statistical analyses were conducted using STATA.15, 16 An unconditional logistic regression model was used to estimate the odds ratio (OR) for Kaposi's sarcoma associated with each variable, adjusted for age group (<30, 30–44, 45+), sex and other factors as noted. All p-values are 2-sided. Analyses of risk factors for Kaposi's sarcoma in HIV-seronegative people in Uganda were then repeated, restricting to KSHV-seropositive cases and controls. Exclusion of patients with cancer of the uterine cervix (the most numerous tumour in the control group) made no material difference to the results and these data are not shown. Note that numbers of cases and controls in the tables do not always add up to the total because of missing values.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Of the patients with Kaposi's sarcoma, 78% (91/117) were male and 17% (20) were aged 15–24 years, 26% (31) aged 25–34, 8% (9) aged 35–44, 14% (16) aged 45–54 and 35% (41) aged 55+ years. Among controls, 33% (416) were male and the percentage in each age group was 9% (109), 13% (170), 22% (280), 21% (264) and 36% (458), respectively.

Among those for whom KSHV antibody status was available (80 cases and 607 controls), the seroprevalence of KSHV was 79% (63/80) in cases and 50% (302/607) in controls (χ2 heterogeneity (1 df) = 21.0; p < 0.001). In controls, the seroprevalence of KSHV did not differ significantly by cancer site or type (χ2heterogeneity (14 df) = 15.4; p = 0.4). The odds of Kaposi's sarcoma increased significantly with increasing anti-KSHV antibody titre (χ2trend (1 df) = 29.7; p < 0.001; Table I).

Table I. Risk of Kaposi's Sarcoma According to Fluorescent Signal Intensity of Anti-KSHV Antibodies Among HIV-Seronegative People in Uganda
KSHV (median titre)Kaposi's sarcoma (n = 80)Controls (other diseases) (n = 607)Odds ratios (95% CI)1
  • 1

    Odds ratios adjusted for age group and sex.

Negative (<1:100)173051.0
+(1:800)241672.3 (1.2–4.7)
++(1:25,600)271055.7 (2.8–11.6)
+++(1:51,200)12307.1 (2.8–17.8)
   χ2 trend = 29.7, p < 0.001

The association between risk of Kaposi's sarcoma and 46 other variables is shown in Tables II–VI. After adjustment for age group and sex only, 12 factors were significantly associated with the tumour at the 5% level. The risk of Kaposi's sarcoma was higher among people who lived outside of Kampala and the surrounding area (p = 0.04) and among tribes other than the Baganda (p = 0.01) and lower among Muslims than among Christians (OR = 0.4, 95% CI 0.2–0.9; p = 0.01; Table II). There was a trend of increasing risk of Kaposi's sarcoma with increasing household income (χ2 trend (1 df) = 18.1; p < 0.001; Table II). The risk was also higher among people who left home at an earlier age (p < 0.001), but lower among people who lived further from a market (p = 0.05; Table III). People who owned goats or pigs (p < 0.001) and who traveled away from home for more than 7 nights per year (p = 0.008) also appeared at increased risk (Table IV). Kaposi's sarcoma was more frequent among people who rarely or never used shoes than among people who usually or always wore shoes (OR = 3.1, 95% CI 2.0–4.7; p < 0.001) and was also more frequent among people who regularly drank alcohol (p = 0.003) and among people who reported having had an injection from a traditional healer (p = 0.04; Table V). Those who reported having had a blood transfusion in the past were less likely to have Kaposi's sarcoma (p = 0.001). The tumour did not appear to be associated with any of the sexual and reproductive factors studied (Table VI).

Table II. Risk Factors for Kaposi's Sarcoma Among HIV-Seronegative People in Uganda
 No. of cases of Kaposi's sarcoma/controls (117/1,203)Odds ratios (95% confidence intervals)1
  • 1

    Odds ratios adjusted for age group and sex. df, degrees of freedom.

  • 2

    Mukono and Mpigi are districts adjacent to Kampala.

Region of birth  
 Kampala9/691.0
 Mukono, Mpigi225/3580.8 (0.3–1.8)
 Rest of Uganda81/8280.9 (0.4–2.0)
  χ2 heterogeneity (2 df) = 0.6; p = 0.7
Region of residence2  
 Kampala19/2371.0
 Mukono, Mpigi25/4011.0 (0.5–1.9)
 Rest of Uganda73/6401.7 (1.0–3.0)
  χ2 heterogeneity (2 df) = 6.5; p = 0.04
Tribe  
 Baganda42/6531.0
 Other75/6261.7 (1.1–2.5)
  χ2 heterogeneity (1 df) = 6.2; p = 0.01
Nationality  
 Ugandan109/1,2021.0
 Other8/781.5 (0.7–3.3)
  χ2 heterogeneity (1 df) = 0.9; p = 0.3
Household income (shillings)  
 <20,00026/4301.0
 20,000–69,99945/3982.7 (1.6–4.7)
 70,000+41/3323.2 (1.9–5.6)
  χ2 trend (1 df) = 18.1; p < 0.001
Religion  
 Christian108/1,1051.0
 Muslim7/1560.4 (0.2–0.9)
  χ2 heterogeneity (1 df) =6.3; p = 0.01
Occupation  
 Cultivator51/7061.0
 Other65/5651.0 (0.6–1.5)
  χ2 heterogeneity (1 df) = 0.1; p = 0.8
Table III. Social and Demographic Factors in Relation to the Risk of Kaposi's Sarcoma
 No. of cases of Kaposi's sarcoma/controlsOdds ratios (95% confidence intervals)1
  • 1

    Odds ratios adjusted for age group and sex. df, degrees of freedom.

Age left home (years)  
 16+ (or never)45/7151.0
 <1669/5122.9 (1.9–4.4)
  χ2 heterogeneity (1 df) = 24.9; p < 0.001
Years at current address  
 <825/2431.0
 8–2232/4180.9 (0.5–1.6)
 23+56/5841.0 (0.6–1.9)
  χ2 trend (1 df) = 0.0; p = 0.9
Settlement type  
 1,000+ houses12/741.0
 100–999 houses27/2620.6 (0.3–1.3)
 10–99 houses73/8060.7 (0.3–1.4)
 <10 houses2/910.2 (0.0–0.8)
  χ2 trend (1 df) = 1.8; p = 0.2
Travel time to market  
 <30 mins64/5031.0
 30+ mins40/5470.6 (0.4–1.0)
  χ2 heterogeneity (1 df) = 4.0; p = 0.05
Age left school  
 No school19/3491.0
 <16 years38/3961.3 (0.7–2.4)
 16+ years56/4931.1 (0.6–2.0)
  χ2 trend (1 df) = 0.0; p = 0.9
Education level  
 No school17/3481.0
 Primary57/5951.3 (0.7–2.4)
 Secondary29/2431.2 (0.6–2.5)
 University/trade11/592.0 (0.8–4.7)
  χ2 trend (1 df) = 1.4; p = 0.2
No. in house  
 <546/4431.0
 5+69/8140.9 (0.6–1.3)
  χ2 heterogeneity (1 df) = 0.5; p = 0.5
No. of siblings  
 0–450/5231.0
 5+64/7340.9 (0.6–1.4)
  χ2 heterogeneity (1 df) = 0.3; p = 0.6
Birth order (no. of older siblings)  
 1–250/5791.0
 3–429/3480.9 (0.6–1.6)
 5+36/3291.2 (0.7–1.9)
  χ2 trend (1 df) = 0.5; p = 0.5
No. sharing toilet  
 <641/4401.0
 6–935/4060.8 (0.5–1.4)
 10+36/3700.8 (0.5–1.4)
  χ2 trend (1 df) = 0.3; p = 0.6
Table IV. Some Factors Characterising Wealth in Relation to Kaposi's Sarcoma
 No. of cases of Kaposi's sarcoma/controlsOdds ratios (95% confidence intervals)1
  • 1

    Odds ratios adjusted for age group and sex. df, degrees of freedom.

No. of rooms in house  
 1–235/3841.0
 3+80/8701.3 (0.8–2.0)
  χ2 heterogeneity (1 df) = 0.9; p = 0.3
Electricity  
 No91/9741.0
 Yes24/2720.9 (0.5–1.5)
  χ2 heterogeneity (1 df) = 0.2; p = 0.6
Own radio  
 No28/2811.0
 Yes87/9710.8 (0.5–1.3)
  χ2 heterogeneity (1 df) = 0.7; p = 0.4
Own motorcar  
 No100/1,1171.0
 Yes15/1341.2 (0.7–2.2)
  χ2 heterogeneity (1 df) = 0.4; p = 0.5
Own bicycle  
 No42/6171.0
 Yes73/5371.4 (0.9–2.1)
  χ2 heterogeneity (1 df) = 2.3; p = 0.1
Own cows  
 None76/9081.0
 1+38/3381.2 (0.8–1.9)
  χ2 heterogeneity (1 df) = 1.0; p = 0.3
Own pigs or goats  
 None51/7891.0
 1+64/4622.4 (1.6–3.6)
  χ2 heterogeneity (1 df) = 17.4; p < 0.001
Own chickens  
 None49/6031.0
 1+65/6151.5 (1.0–2.2)
  χ2 heterogeneity (1 df) = 3.3; p = 0.07
Travel away from home > 7 nights per year  
 No83/1,0901.0
 Yes31/1411.9 (1.2–3.1)
  χ2 heterogeneity (1 df) = 7.0; p = 0.008
Table V. Other Exposures in Relation to Kaposi's Sarcoma
 No. of cases of Kaposi's sarcoma/controlsOdds ratios (95% confidence intervals)1
  • 1

    Odds ratios adjusted for age group and sex. df, degrees of freedom.

Use of shoes  
 Half the time–always48/7311.0
 Rarely or never67/5123.1 (2.0–4.7)
  χ2 heterogeneity (1 df) = 27.8; p < 0.001
Age first use shoes (years)  
 <1656/5371.0
 16+53/6500.9 (0.5–1.3)
  χ2 heterogeneity (1 df) = 0.5; p = 0.5
Time spent cultivating per week (hr)  
 0–919/2271.0
 10–1956/6041.6 (0.9–2.7)
 20+40/4151.4 (0.7–2.5)
  χ2 trend (1 df) = 0.6; p = 0.5
Soil colour  
 Red15/1691.0
 Black64/6790.8 (0.5–1.6)
 Brown21/2330.8 (0.4–1.9)
 Other10/731.3 (0.6–3.3)
  χ2 heterogeneity (3 df) = 1.6; p = 0.7
Time spent in water per week  
 <1 hr72/8691.0
 1+ hr43/3851.4 (0.9–2.1)
  χ2 heterogeneity (1 df) = 2.0; p = 0.2
Tobacco use  
 Never smoker87/9981.0
 Ex-smoker12/1450.6 (0.3–1.0)
 Current smoker15/1100.8 (0.4–1.4)
  χ2 trend (1 df) = 1.7; p = 0.2
Use of snuff  
 No115/1,2331.0
 Yes0/13
  
Drink alcohol  
 No45/6901.0
 About once/week28/2192.6 (1.5–4.4)
 2–4/week23/1611.8 (1.0–3.2)
 Most days19/1811.1 (0.6–1.9)
  χ2 trend (1 df) = 13.9; p = 0.003
Blood transfusion  
 No109/1,0251.0
 Yes5/2270.3 (0.1–0.7)
  χ2 heterogeneity (1 df) = 11.9; p = 0.001
Injection from traditional healer  
 No60/7551.0
 Yes52/4691.5 (1.0–2.3)
  χ2 heterogeneity (1 df) = 4.3; p = 0.04
Table VI. Sexual and Reproductive Variables in Relation to Kaposi's Sarcoma
 No. of cases of Kaposi's sarcoma/controlsOdds ratios (95% confidence intervals)1
  • 1

    Odds ratios adjusted for age group and sex. df, degrees of freedom.

Marital status  
 Single20/1091.0
 Ever married95/1,1441.1 (0.6–2.2)
  χ2 heterogeneity (1 df) = 0.1; p = 0.7
Age first sex  
 <18 years56/7661.0
 18+ years56/4481.3 (0.8–2.0)
  χ2 heterogeneity (1 df) = 1.3; p = 0.3
No. of children  
 0–215/1461.0
 3–415/1761.2 (0.6–2.8)
 5–615/1941.2 (0.5–2.9)
 7+46/5901.2 (0.5–2.6)
  χ2 trend (1 df) = 0.1; p = 0.8
Children with different mothers or fathers  
 No40/5771.0
 Yes44/4811.0 (0.6–1.6)
  χ2 heterogeneity (1 df) = 0.0; p = 0.9
No. of lifetime sexual partners  
 1–225/3361.0
 3–417/2880.7 (0.4–1.5)
 5–722/1951.2 (0.6–2.2)
 8–1422/1650.9 (0.5–1.8)
 15+24/1940.6 (0.3–1.2)
  χ2 trend (1 df) = 1.1; p = 0.3
History of sexual discharge  
 No49/6491.0
 Yes61/5641.1 (0.7–1.8)
  χ2 heterogeneity (1 df) = 0.3; p = 0.6
Gifts for sex  
 Never95/1,0951.0
 Ever14/1181.2 (0.6–2.2)
  χ2 heterogeneity (1 df) = 0.3; p = 0.6
Used condoms  
 Never88/1,0531.0
 Ever21/1621.0 (0.5–1.7)
  χ2 heterogeneity (1 df) = 0.0; p = 0.9
Circumcised (men)  
 No68/3121.0
 Yes21/871.1 (0.6–1.9)
  χ2 heterogeneity (1 df) = 0.1; p = 0.8
Labial (genital) elongation (women)  
 No9/2311.0
 Yes17/6130.7 (0.3–1.6)
  χ2 heterogeneity (1 df) = 0.8; p = 0.4

Those risk factors from Tables II–VI that were significantly associated with Kaposi's sarcoma at the 1% level are summarised in Table VII. After adjustment for age, sex and each other, only 5 factors remained independently associated with the tumour. People with Kaposi's sarcoma were more likely than those without to be from tribal groups other than the Baganda (p = 0.05), had significantly higher household incomes (p = 0.003), left their home region at younger ages (p < 0.001) and were more likely to own goats or pigs (p = 0.02) and to rarely or never wear shoes (p < 0.001). When analyses were restricted to 63 cases and 302 controls who were known to have antibodies against KSHV, the results did not materially change (Table VII).

Table VII. Summary of Risk Factors for Kaposi's Sarcoma in HIV-Seronegative Ugandans
 Odds ratios (95% confidence intervals) adjusted for age, sex and all other variables shownOdds ratios (95% confidence intervals) adjusted for age, sex and all other variables shown, restricted to people with anti-KSHV antibodies1
  • Variables were selected on the basis of significance at the 1% level.

  • 1

    Based on analyses of 63 cases and 302 controls, who had anti-KSHV antibodies.

Tribe  
 Baganda1.01.0
 Other1.6 (1.0–2.6)1.8 (0.8–3.9)
 χ2 heterogeneity (1 df) = 4.0; p = 0.05χ2 heterogeneity (1 df) = 1.9; p = 0.2
Household income (shillings)  
 <20,0001.01.0
 20,000–69,9992.2 (1.2–3.9)2.4 (0.9–6.1)
 70,000+2.5 (1.4–4.6)3.5 (1.3–9.4)
 χ2 trend (1 df) = 9.1; p = 0.003χ2 trend (1 df) = 6.4; p = 0.01
Religion  
 Christian1.01.0
 Muslim0.5 (0.2–1.3)0.1 (0.0–1.1)
 χ2 heterogeneity (1 df) = 2.4; p = 0.1χ2 heterogeneity (1 df) = 5.1; p = 0.02
Age left home (years)  
 16+ (or never)1.01.0
 <162.8 (1.8–4.5)2.8 (1.3–5.7)
 χ2 heterogeneity (1 df) = 19.9; p < 0.001χ2 heterogeneity (1 df) = 7.7; p = 0.006
Own pigs or goats  
 None1.01.0
 1+1.7 (1.1–2.8)1.4 (0.7–2.9)
 χ2 heterogeneity (1 df) = 5.6; p = 0.02χ2 heterogeneity (1 df) = 0.8; p = 0.4
Travel away from home >7 nights per year  
 No1.01.0
 Yes1.4 (0.8–2.3)1.8 (0.8–4.0)
 χ2 heterogeneity (1 df) = 1.3; p = 0.3χ2 heterogeneity (1 df) = 1.8; p = 0.2
Use of shoes  
 Half the time–always1.01.0
 Rarely or never2.6 (1.6–4.2)3.1 (1.5–6.7)
 χ2 heterogeneity (1 df) = 15.5; p < 0.001χ2 heterogeneity (1 df) = 9.4; p = 0.002
Drink alcohol  
 No1.01.0
 About once/week2.0 (1.1–3.6)1.8 (0.6–5.2)
 2–4/week1.5 (0.8–2.9)1.7 (0.6–4.5)
 Most days1.1 (0.6–2.2)1.1 (0.4–3.2)
 χ2 trend (1 df) = 0.4; p = 0.5χ2 trend (1 df) = 0.1; p = 0.7

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The primary aim of our study was to identify risk factors for Kaposi's sarcoma in HIV-seronegative Ugandan adults. About 50 potential risk factors were examined and so there is a strong possibility that some of the associations identified were due to chance. For this reason, in the final analysis of risk factors for Kaposi's sarcoma (Table VII), only those that were significant at the 1% level were considered.

Among HIV-seronegative people in Uganda, Kaposi's sarcoma is clearly associated with KSHV and the risk of the tumour increased with increasing anti-KSHV antibody titre, as noted previously in South Africa.14 A number of other factors were associated with Kaposi's sarcoma among HIV-seronegative people, including male sex, being from tribes other than the Baganda, having a higher income, leaving home at an earlier age, owning goats or pigs and going barefoot. In Newton et al. on pages 226–232,9 we analyse the risk factors for anti-KSHV antibodies in the same population of HIV-seronegative Ugandan adults without Kaposi's sarcoma. We show that the prevalence of anti-KSHV antibodies rises with age but is equal in males and females and so the male predominance of Kaposi's sarcoma cannot be explained by differences in the prevalence of KSHV but rather on the basis of exposure to some as yet unknown cofactor. Indeed, apart from a lower prevalence of KSHV among the Baganda tribe, none of the risk factors associated with Kaposi's sarcoma, such as male sex, household income, age left home and going barefoot, are associated with the presence of anti-KSHV antibodies in this population.9 The only possible exception is ownership of pigs or goats, although this result was not statistically significant for KSHV and could be a chance finding. Furthermore, when analyses were restricted to those with anti-KSHV antibodies, the results did not materially change. The independent effects we have identified as risk factors for Kaposi's sarcoma in Uganda probably, therefore, relate to factors acting in addition to KSHV.

The apparent protective effect of having had a blood transfusion in the past is probably an artefact arising from the type of controls used. Among HIV-seronegative people, Kaposi's sarcoma can be a relatively indolent tumour, whereas some of the cancers among controls, such as cancer of the uterine cervix, present late and involve significant blood loss. The controls are therefore more likely than cases to include patients who had a blood transfusion as part of their medical care.

Relative affluence is a risk factor for Kaposi's sarcoma both in HIV-seronegative and -seropositive Ugandans.8 In contrast, the prevalence of anti-KSHV antibodies in Uganda was slightly lower, but not significantly so, in those with more education.9 Although the results were of only borderline statistical significance, similar findings have been reported from South Africa.14 If true, the interpretation of this finding is difficult; however, if high social status protects an individual from early infection with KSHV, it could imply that the age at which infection occurs affects the subsequent risk of Kaposi's sarcoma. This hypothesis is reminiscent of the effect of infection in adult life with the Epstein Barr virus (a related gamma herpesvirus) in relation to the risk of infectious mononucleosis.

It has been suggested that fine soil particles might pass through the skin of barefoot individuals and block the lymphatic system, causing local immunosuppression in the lower limbs, thus explaining the characteristic distribution of Kaposi's sarcoma lesions in HIV-uninfected people.17 The higher risk of Kaposi's sarcoma associated with going barefoot is not incompatible with this hypothesis, although the exact nature of the relevant exposure remains unclear. Nor is relative affluence incompatible with going barefoot, as wealthy farmers with large land holdings are active cultivators. Moreover, farming in urban areas is prevalent in Uganda.18 The role of other risk factors for Kaposi's sarcoma in Uganda remains unexplained.

In summary, in addition to infection with KSHV, sex, age, income, tribe, age left home, pig or goat ownership and going barefoot are risk factors for Kaposi's sarcoma in HIV-seronegative people in Uganda. Of these, only age and tribe have been found to be associated with anti-KSHV antibodies. It would seem that, in addition to KSHV infection, other factors are important in the development of Kaposi's sarcoma in Uganda.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The authors would like to acknowledge the support of Mulago Hospital and Makerere Medical School, Kampala, Uganda. We thank Prof. I. Kakonde and Prof. F. Mmiro (Mulago Hospital), Dr. R. Moser (deceased) (Rubaga Hospital), Dr. P. Okong (Nsambya Hospital) and Dr. C. Kohere (Mengo Hopital) for access to their patients. We thank the many patients who volunteered for our study.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES
  • 1
    Oettlé AG. Geographical and racial differences in the frequency of Kaposi's sarcoma as evidence of environmental or genetic causes. Acta Unio Int Contra Cancrum 1962; 18: 33063.
  • 2
    Templeton AC. Kaposi's sarcoma. In: SommersSC, RosenPP, eds. Pathology annual. New York: Century-Crofts, 1981. 31536.
  • 3
    Hutt MSR. Classical and endemic form of Kaposi's sarcoma. A review. Antibiot Chemother 1984; 32: 127.
  • 4
    Cook-Mozaffari P, Newton R, Beral V, Burkitt DP. The geographical distribution of Kaposi's sarcoma and of lymphomas in Africa before the AIDS epidemic. Br J Cancer 1998; 78: 15218.
  • 5
    McHardy J, Williams EH, Geser A, de-Thé G, Beth E, Giraldo G. Endemic Kaposi's sarcoma in the West Nile District of Uganda. Int J Cancer 1984; 33: 20312.
  • 6
    Wabinga HR, Parkin DM, Wabwire-Mangen F, Mugerwa JW. Cancer in Kampala, Uganda, in 1989–91: changes in incidence in the era of AIDS. Int J Cancer 1993; 54: 2636.
  • 7
    Wabinga HR, Parkin DM, Wabwire-Mangen F, Nambooze S. Trends in cancer incidence in Kyadondo County, Uganda, 1960–1997. Br J Cancer 2000; 82: 158592.
  • 8
    Ziegler JL, Newton R, Katongole-Mbidde E, Mbulataiye S, DeCock K, Wabinga H, Mugerwa J, Katabira E, Jaffe H, Parkin DM, Reeves G, Beral V, for the Uganda Kaposi's sarcoma study group. Risk factors for HIV-associated Kaposi's sarcoma in Uganda: a case-control study of 1026 adults. AIDS 1997; 11: 161926.
  • 9
    Newton R, Ziegler J, Bourboulia D, Casabonne D, Beral V, Mbidde E, Carpenter L, Reeves G, Parkin DM, Wabinga H, Mbulaiteye S, Jaffe H, et al. The sero-epidemiology of Kaposi's sarcoma-associated herpesvirus in adults with cancer in Uganda. Int J Cancer 2003; 103: 226232.
  • 10
    Newton R, Ziegler J, Beral V, Mbidde E, Carpenter L, Wabinga H, Mbulataiye S, Appleby P, Reeves G, Jaffe H and the Uganda Kaposi's Sarcoma Study Group. A case-control study of Human Immunodeficiency Virus infection and cancer in adults and children residing in Kampala, Uganda. Int J Cancer 2001; 92: 6227.
  • 11
    Gao S-J, Kingsley L, Li M, Zheng W, Parravicini C, Ziegler JL, Newton R, Rinaldo CR, Saah A, Phair J, Detels R, Chang Y, et al. Seroprevalence of Kaposi's sarcoma-associated herpesvirus antibodies among Americans, Italians and Ugandans with and without Kaposi's sarcoma. Nature Med 1996; 2: 9258.
  • 12
    Boshoff C, Gao SJ, Healy LE, Matthews S, Thomas AJ, Coignet L, Warnke RA, Strauchen JA, Matutes E, Kamel OW, Moore PS, Weiss RA, et al. Establishing a KSHV+ cell line (BCP-1) from peripheral blood and characterising its growth in Nod/SCID mice. Blood 1998; 91: 16719.
  • 13
    Davidovici B, Karakis I, Bouboulia D, Ariad S, Zong J-C, Benharroch D, Dupin N, Weiss R, Hayward G, Sarov B, Boshoff C. Seroepidemiology and molecular epidemiology of Kaposi's sarcoma-associated herpesvirus among Jewish population groups in Israel. J Natl Cancer Inst 2001; 93: 194202.
  • 14
    Sitas F, Carrara H, Beral V, Newton R, Reeves G, Bull D, Patel M, Ruff P, Bezwoda WR, Hale M, Jentsh U, Retter E, et al. The seroepidemiology of HHV-8/KSHV in a large population of black cancer patients in South Africa. N Engl J Med 1999; 340: 186371.
  • 15
    Dean AD, Dean JA, Burton JH, Dicker RC. Epi Info, version 5: a word processing, database and statistics program for epidemiology on micro-computers. Atlanta, GA: Centers for Disease Control, 1990.
  • 16
    STATA. Stata software. College Park, TX: STATA, 2001.
  • 17
    Ziegler JL. Endemic Kaposi's sarcoma and local volcanic soils. Lancet 1994; 342: 134851.
  • 18
    Maxwell D, Zziwa S. Urban farming in Africa. The case of Kampala. Nairobi: African Center for Technology Studies, 1992.