Seroprevalence of syphilis among HIV-infected individuals in Addis Ababa, Ethiopia: a hospital-based cross-sectional study.

OBJECTIVE: To determine the prevalence of syphilis and its risk factors among people with HIV at a hospital in Ethiopia.
DESIGN: A hospital-based cross-sectional study.
SETTING: This study was conducted at one of the largest public hospitals in Addis Ababa , Ethiopia. PARTICIPANTS: A consecutive 306 HIV-positive patients were recruited prospectively from January to March 2010. For comparative purposes, 224 HIV-negative consecutive attendees at the voluntary counselling and testing centre in the same period were also included. Participants under 15 years of age and treated for syphilis and with a CD4 T-cell count below 50 cells/mm(3) were excluded. OUTCOME MEASURES: Blood samples and data on sociodemographic and risk factors for syphilis were collected. Sera were screened for syphilis using rapid plasma reagin (RPR) test, and those positives were retested using Treponema pallidum haemagglutination assay (TPHA) test.
RESULTS: The seroprevalence of syphilis among HIV-infected individuals was 9.8% compared with 1.3% among HIV-uninfected individuals, OR 8.01 (95% CI 2.4 to 26.6; p=0.001). A comparable rate of syphilis was found among men (11%) and women (8.9%) with HIV infection. Syphilis prevalence non-significantly increased with age, with the highest rate in 40-49 years of age (16.9%). Except a history of sexually transmitted infections, which was associated with syphilis OR 2.25 (95% CI 1.03 to 4.9; p=0.042), other risk factors did not raise the odds of infection.
CONCLUSIONS: The high prevalence of syphilis among people with HIV infection highlights the need to target this population to prevent the transmission of both infections. Screening all HIV-infected people for syphilis and managing those infected would have clinical and epidemiological importance.

To determine the prevalence of syphilis among HIV-infected people.

To compare the prevalence of syphilis by HIV status.

To assess the risk factors for syphilis in HIV-infected people.

High prevalence of syphilis among HIV-positive individuals was observed.

Syphilis prevalence is significantly higher among HIV positives than among HIV-negative people.

Syphilis prevalence is not significantly influenced by age and gender.

This hospital-based cross-sectional study provided preliminary data that would support future research.

The study did not use stronger statistical power to detect the differences in risk factors of syphilis status.

No clinical features of syphilis-positive patients were assessed.

Introduction

Sexually transmitted infections (STIs) are the major public health problems in most parts of the world. Based on the WHO estimate, STIs and their complications are among the top five disease categories for which adults seek health care in developing countries.1 Syphilis is one of the most important STIs, caused by the bacterium Treponema pallidum.2 It has been estimated that, annually, about 12 million new infections occur worldwide; of which, almost two-thirds are in sub-Saharan Africa and south/southeast Asia.3 Unprotected sex, blood transfusion, needle sharing and vertical transmission from mother to the child are major modes of syphilis transmission.2 4

Syphilis, as a cause of ulcerative genital lesions, presents site for HIV entry and shading. Moreover, by activating immune cells and raising viral load, syphilis could facilitate HIV transmissibility.5 On the contrary, concurrent HIV infection may adversely affect the natural history, clinical manifestations and treatment response of syphilis.3 6

In Ethiopia, studies reported syphilis prevalence ranging from 1% to 10.9% in diverse risk groups such as pregnant women, blood donors, street dwellers and elderly people.7–10 Moreover, according to the antenatal care (ANC)-based sentinel surveillances, syphilis prevalence increased from 1.8% in the year 2003 to 2.7% in 2005, and then stabilised at 2.3% in 2007 and 2009. The rates of syphilis-HIV coinfection among ANC attendees had also been rising from 4.1% in the year 2003 to 4.9% in 2005 and 5.3% in 2007, but dropped to 3.9% in 2009.11–14 However, because of the limitations that the aforementioned risk groups consist of smaller size of HIV-infected individuals, and the sex and age composition of ANC attendees is limited to female gender and reproductive age group, the generated data may not reflect the true picture of syphilis among HIV-infected population. Therefore, this study was conducted to determine the prevalence and risk factors of syphilis among HIV-infected clients at St Paul's General Specialized Hospital.

Methods and materials

This cross-sectional study was conducted at St Paul's General Specialized Hospital, Addis Ababa from January to March 2010. The hospital is among the largest public hospitals in Ethiopia and provides HIV voluntary counselling and testing (VCT) as a routine service. Clients who are tested HIV positive are registered at the antiretroviral therapy (ART) clinic and assessed for their disease status. Clinical and immunological assessments (CD4 T cell count) at enrollment and at 6 monthly follow-up visits help determine patients’ eligibility for ART. Those receiving ART are also monitored for clinical progress on a regular basis. Services including HIV counselling and testing, clinical and immunological assessments as well as ART are provided free of charge. HIV-infected patients are not routinely screened for syphilis and only those with clinical indications are tested.

Consecutive HIV-infected individuals with and without ART status, and who had immunological and biochemical testing were recruited prospectively. Clients tested HIV negative at the VCT centre during the study period were also recruited for comparative purposes. In total, 306 HIV positive and 224 HIV negative clients were considered for analysis. In either HIV serogroups, participants less than 15 years of age, and those who took syphilis treatment were excluded, as reactive non-treponemal test result may not remain after treatment. In HIV-positive clients, those found with a CD4 T cell count below 50 cells/mm3 were excluded from the study owing to the unreliability of serological tests in a state of severe immunosuppression.

Counsellor nurses interviewed the study participants using structured questionnaire on sociodemographic and other risk factors such as history of blood transfusion, unsafe injection, multiple sexual partners, STIs, and syphilis family history. Blood samples were collected and screened for syphilis using the non-treponemal serologic test, rapid plasma reagin (RPR) test (Human, Germany). Sera found to be positive by RPR tests were further tested using treponemal test, modified T pallidum haemagglutination assay (TPHA) (Syphicheck-WB, Qualpro Diagnostics, India). Laboratory testing was carried out according to the directions of the manufacturers and all tests were run against the positive and negative controls. Only those samples positive by both RPR and TPHA were considered to have syphilis infection.

The study was approved by the Ethics Review Committee of Aklilu Lemma Institute of Pathobiology, Addis Ababa University and the St Paul's Hospital management body. Participation was entirely voluntary, and written consent was obtained from the study participants. Any information obtained during the study was kept with utmost confidentiality. Syphilis screening was performed free of charge, and those tested positive were managed by the physicians.

Data entry and analysis was performed using SPSS V.16. Results were summarised using descriptive statistics. Pearson's χ2 test was used to evaluate differences between proportions; χ2 for linear trend was also calculated using Epi Info V.7. Binary logistic regression analysis was used to assess the effect of sociodemographic and other risk factors on syphilis seropositivity. The OR was used as a measure of association.

Results

Of 312 HIV-positive and 228 HIV-negative individuals approached during the study period, 6 and 4 individuals were excluded owing to refusal to participate, and insufficient serum sample and incomplete questionnaire, respectively. Thus, 306 HIV positive and 224 HIV negative clients were considered for analysis. A total of 188 (61.4%) participants with HIV received ART and the rest were ART naïve (38.6%). Majority of HIV-infected participants were urban dwellers (95.4%) and married (53.3%; table 1). HIV-infected respondents had a mean age 35.8 years (SD = 8.7, range 19–73 years) compared with 28.2 years (SD=9.8, range 15–73 years) in HIV non-infected groups. The male to female ratios in participants with and without HIV infection were 0.71 : 1 and 0.96 : 1, respectively.

Table 1

Syphilis infection in relation to sociodemography in HIV-positive and HIV-negative individuals at St Paul's Hospital, 2010

	HIV positive			HIV negative
Characteristics	Number (%) tested	Number (%) positive for syphilis	Crude OR (95% CI)	Number (%) tested	Number (%) positive for syphilis	Crude OR (95% CI)
Residence
Rural	14 (4.6)	0		18 (8)	0
Urban	292 (95.4)	30 (10.3)	–	206 (92)	3 (1.5)	–
Sex
Female	179 (58.5)	16 (8.9)	1	114 (50.9)	1 (0.9)	1
Male	127 (41.5)	14 (11)	1.26 (0.59 to 2.69)	110 (49.1)	2 (1.8)	2.1 (0.18 to 23.4)
Age (years)
<19	2 (0.7)	0	–	52 (23.2)	1 (1.9)	1.84 (0.11 to 30.1)
20–29	65 (21.2)	4 (6.2)	1	95 (42.4)	1 (1.1)	1
30–39	156 (51)	13 (8.3)	1.39 (0.41 to 4.42)	47 (21)	0	–
40–49	59 (19.3)	10 (16.9)	3.11 (0.92 to 10.5)	19 (8.5)	1 (5.3)	5.2 (0.31 to 87.4)
>50	24 (7.8)	3 (12.5)	2.18 (0.45 to 10.5)	11 (4.9)	0	–
Marital status
Single	60 (19.6)	5 (8.3)	1.41 (0.39 to 5.1)	146 (65.2)	3 (2.1)	–
Married	163 (53.3)	20 (12.3)	2.2 (0.79 to 6)	60 (26.8)	0
Divorced/widowed	83 (27.1)	5 (6)	1	18 (8)	0
Religion
Orthodox	228 (74.5)	24 (10.5)	2.1 (0.47 to 9.1)	170 (75.9)	3 (1.8)	–
Protestant	41 (13.4)	4 (9.8)	1.9 (0.32 to 10.9)	25 (11.2)	0
Muslim	37 (12.1)	2 (5.4)	1	29 (12.9)	0
Educational status
Illiterate	41 (13.4)	9 (22)	4.78 (0.96 to 23.8)	13 (5.8)	0
Primary school	95 (31)	8 (8.4)	1.56 (0.32 to 7.74)	51 (22.8)	0
Secondary school	134 (43.8)	11 (8.2)	1.52 (0.32 to 7.19)	124 (55.4)	2 (1.6)	–
Certificate and above	36 (11.8)	2 (5.6)	1	36 (16.1)	1 (2.8)
Occupation
Government employee	41 (13.4)	3 (7.3)	1	29 (12.9)	1 (3.4)	1
Private employee	82 (26.8)	8 (9.8)	1.37 (0.34 to 5.46)	71 (31.7)	1 (1.4)	0.4 (0.02 to 6.62)
Housewife	63 (20.6)	6 (9.5)	1.33 (0.31 to 5.66)	21 (9.4)	0	–
Student	5 (1.6)	0	–	41 (18.3)	1 (2.4)	0.7 (0.04 to 11.67)
Merchant	35 (11.4)	4 (11.4)	1.63 (0.34 to 7.86)	19 (8.5)	0
Housemaid	11 (3.6)	3 (27.3)	4.75 (0.81 to 27.9)	7 (3.1)	0
No work	69 (22.5)	6 (8.7)	1.21 (0.29 to 5.11)	36 (16.1)	0
Ethnicity
Amhara	156 (51)	14 (9)	1	117 (52.2)	2 (1.7)	1
Oromo	87 (28.4)	9 (10.3)	1.2 (0.49 to 2.83)	64 (28.6)	1 (1.6)	0.9 (0.08 to 10.3)
Others	63 (20.6)	7 (11.1)	1.3 (0.49 to 3.3)	43 (19.2)	0	–

The prevalence of syphilis infection was 9.8% in HIV positive participants compared to 1.3% in HIV negative participants; OR 8.01 (95% CI 2.4 to 26.6, p=0.001). The distribution of syphilis was similar among HIV-infected clients with and without ART (11.2% vs 7.6%, respectively; p=0.31). Sera reactive by RPR test were more likely found TPHA positive among HIV positives (54.5%) than in HIV-negatives (10%, p<0.001; table 2).

Table 2

Syphilis serological tests in HIV-positive and HIV-negative individuals at St Paul's Hospital, 2010

			HIV positive
			ART users		ART naïve		Total		HIV negative
Syphilis test	Total tested	Number (%) of positive	Tested	+ve (%)	Tested	+ve (%)	Tested	+ve (%)	Tested	+ve (%)
RPR	530	85 (16)	188	36 (19.1)	118	19 (5.6)	306	55 (18)	224	30 (13.4)
TPHA	85	33 (38.8)	36	21 (58.3)	19	9 (47.4)	55	30 (54.5)	30	3 (10)
Syphilis seropositivity	530	33 (6.2)	188	21 (11.2)	118	9 (7.6)	306	30 (9.8)	224	3 (1.3)

ART, antiretroviral therapy; RPR, rapid plasma reagin; TPHA, Treponema pallidum haemagglutination; +ve, positive.

Syphilis occurred exclusively among urban dwellers in either of the HIV serogroups. Seropositivity of syphilis was comparable between men (11%) and women (8.9%) with HIV infection. Syphilis prevalence seems to increase with increasing age, with the highest rate in the age range 40–49 years (16.9%), though χ2 for linear trend analysis showed no statistical significance (χ2=2.46, p=0.117). A decreasing rate of syphilis was observed with increasing educational level, where illiterate HIV-positive participants (22%) had higher odds of infection compared with those having at least a certificate (5.6%); OR 4.78 (95% CI 0.96 to 23.8, p=0.056). Similarly, the association between occupation and syphilis was marginally non-significant where housemaids (27.3%) were affected compared with government employees (7.3%); OR 4.75 (95% CI 0.81 to 27.9, p=0.085; table 1).

The exposure of HIV-infected and HIV-non-infected participants to various risk factors of syphilis is summarised in table 3. Except syphilis family history, which occurred in a comparable rate in either of the HIV serogroups, other risk factors such as history of blood transfusion (10.5%), having multiple sexual partners (36.9%) and unsafe injection (12.7%) and a history of STIs (45.4%) were more frequently reported by HIV-infected participants. However, it was only a history of STIs, which was significantly associated with syphilis among HIV-infected participants; OR 2.25 (95% CI 1.03 to 4.9, p=0.042).

Table 3

Syphilis infection in relation to syphilis risk factors in HIV-positive and HIV-negative individuals at St Paul's Hospital, 2010

	HIV positive			HIV negative
Characteristics	Number (%) tested	Number (%) positive for syphilis	Crude OR (95% CI)	Number (%) tested	Number (%) positive for syphilis	Crude OR (95% CI)
Blood transfusion
No	274 (89.5)	28 (10.2)	1	216 (96.4)	3 (1.4)	–
Yes	32 (10.5)	2 (6.2)	0.59 (0.13 to 2.58)	8 (3.6)	0
Multiple sexual partner
No	193 (63.1)	19 (9.8)	1	195 (87.1)	2 (1)	1
Yes	113 (36.9)	11 (9.7)	0.99 (0.45 to 2.16)	29 (12.9)	1 (3.4)	3.45 (0.3 to 39.2)
Unsafe injection
No	267 (87.3)	27 (10.1)	1	219 (97.8)	3 (1.4)	–
Yes	39 (12.7)	3 (7.7)	0.71 (0.21 to 2.57)	5 (2.2)	0
Syphilis family history
No	281 (91.8)	26 (9.3)	1	200 (89.3)	2 (1)	1
Yes	25 (8.2)	4 (16)	1.87 (0.60 to 5.86)	24 (10.7)	1 (4.2)	4.3 (0.38 to 49)
STIs
No	167 (54.6)	11 (6.6)	1	182 (81.2)	2 (1.1)	1
Yes	139 (45.4)	19 (13.7)	2.25 (1.03 to 4.9)	42 (18.8)	1 (2.4)	2.2 (0.19 to 24.8)

STIs, sexually transmitted infections.

Discussion

This study showed that the prevalence of syphilis among HIV positives was 9.8%, with no significant difference between those receiving ART (11.2%) and ART naives (7.6%). The finding appears to be compatible with rates of syphilis–HIV coinfection among street dwellers (7.9%)8 and elderly people (6%)9 in northwest Ethiopia (Gondar) and in Nigeria (14%).15 However, contrasting our result, the coinfection rate was lower among ANC attendees in Ethiopia (3.9%)14 and higher among sexually transmitted disease (STD) clinic attendees in Argentina (59.7%).16 The observed inconsistencies may be because of the composition of the investigated subpopulation, where ANC attendees, for instance, have apparently lower risk of syphilis compared with STD clinic attendees. In view of the adverse impact syphilis has to facilitate the transmission of coexisting HIV, intervention measures targeting this particular risk group has greater importance to prevent both infections.

In the present study, syphilis was significantly associated with HIV infection, where HIV-infected individuals had about eightfold higher risk of syphilis compared with HIV-non-infected people. This result was in line with findings that revealed the existence of association between HIV and syphilis in different localities and subpopulations. A consistent twofold increase in syphilis–HIV coinfection rates among ANC attendees11–14 and fourfold among street dwellers in Ethiopia,10 as well as eightfold in HIV-infected population in Nigeria15 may be because of the fact that HIV and syphilis shares routes of transmission. These reports also indicated the varying strength of association between HIV and syphilis in diverse risk groups. However, none of these studies pointed out whether syphilis and HIV were contracted concurrently or one infection preceded another to explain the causal nature of such epidemiologic synergy between HIV and syphilis.

The seroprevalence of syphilis was not significantly affected by gender in either HIV serogroups, similar to findings elsewhere.10 15 However, Griemberg et al16 reported men had a higher risk of HIV, syphilis and syphilis–HIV coinfection compared with women. This report is also in contrast to the established higher rate of HIV among women in our region,17 which may be because of the difference in risk behaviour by gender in various geographical regions. We also found increasing syphilis prevalence with age among HIV-infected individuals, with the highest rate reported in the age group 40–49 years (16.9%), followed by age group above 50 years (12.5%), though no statistically significant linear trend was observed. A raising syphilis prevalence with age was consistently reported by others,8 10 14 15 which might be because of the increased risk of exposure to syphilis with time. Moreover, our data showed that illiterate and housemaid HIV-infected participants were disproportionately affected by syphilis, which point the significance of education to prevent syphilis transmission.

In Ethiopia, where HIV and syphilis has strong association, and transmission of the former is primarily through heterosexual exposure,17 people with multiple sexual partners would obviously be at higher risk of contracting syphilis as well. Of course, the significance of such risk behaviour to influence syphilis prevalence was documented in our context, where having more than two sexual partners increased odds of syphilis infection sixfold compared with those with no sexual partner.10 However, the lack of association between a history of multiple sexual partners and syphilis in our study deserves further investigation for possible explanation. Syphilis prevalence was about twofold higher among HIV-infected participants who reported a history of STIs compared with those with no history of STIs.

Findings in this study need to be interpreted in light of its methodological limitations. First, absence of association between various risk factors and syphilis might be because of the fact that the study did not use stronger statistical power to detect the differences. Second, the reduced sensitivity of non-treponemal tests in primary as well as late latent syphilis and the potential for false-negative results owing to prozone reactions might lead to underestimation of syphilis infection rate. Moreover, the limitation of possible false-positive reaction with non-treponemal and treponemal tests needs to be given attention, as positive results may not necessarily indicate disease activity. Finally, this study overlooked the importance of including clinical data, which would have been a good opportunity to describe the clinical presentation of syphilis among HIV-infected patients.

In conclusion, this study showed high prevalence of syphilis among HIV-infected people compared with HIV-non-infected people. Thus, intervention measures targeting HIV-infected individuals would have paramount importance to prevent transmission of syphilis as well as HIV. As part of this effort, screening all HIV-infected people for syphilis and managing those infected is critically needed. Further studies using a longitudinal design with stronger statistical power would reliably investigate the possible interaction between HIV and syphilis.