Hagen Frickmann and Norbert G. Schwarz equally contributed to this work.
Corresponding Author Hagen Frickmann, Department of Tropical Medicine at the Bernhard Nocht Institute, German Armed Forces Hospital of Hamburg, Bernhard Nocht Str. 74, 20359 Hamburg, Germany. Tel.: +49 40 6947 2862; Fax: +49 40 6947 2859; E-mail: email@example.com
Objectives Peripartal transmission of human immunodeficiency virus (HIV) and Treponema pallidum, the causative agent of syphilis, leads to severe consequences for newborns. Preventive measures require awareness of the maternal infection. Although HIV and syphilis testing in Madagascar could be theoretically carried out within the framework of the national pregnancy follow-up scheme, the required test kits are rarely available at peripheral health centres. In this study, we screened blood samples of pregnant Madagascan women for HIV and syphilis seroprevalence to estimate the demand for systemic screening in pregnancy.
Methods Retrospective anonymous serological analysis for HIV and syphilis was performed in plasma samples from 1232 pregnant women that were taken between May and July 2010 in Ambositra, Ifanadiana, Manakara, Mananjary, Moramanga and Tsiroanomandidy (Madagascar) during pregnancy follow-up. Screening was based on Treponema pallidum haemagglutination tests for syphilis and rapid tests for HIV, with confirmation of positive screening results on line assays.
Results Out of 1232 pregnant women, none were seropositive for HIV and 37 (3%) were seropositive for Treponema pallidum.
Conclusions Our findings are in line with previous studies that describe considerable syphilis prevalence in the rural Madagascan population. The results suggest a need for screening to prevent peripartal Treponema pallidum transmission, while HIV is still rare. If they are known, Treponema pallidum infections can be easily, safely and inexpensively treated even in pregnancy to reduce the risk of transmission.
La transmission lors de l'accouchement du virus l'immunodéficience humaine (VIH) et de Treponema pallidum, l'agent causal de la syphilis, conduit à des conséquences graves pour les nouveau-nés. Les mesures préventives exigent la conscientisation sur l'infection maternelle. Bien que le dépistage du VIH et de la syphilis à Madagascar puisse théoriquement être fait dans le cadre du schéma national de suivi de la grossesse, les kits des tests requis sont rarement disponibles dans les centres de santé périphériques. Dans cette étude, nous avons criblé des échantillons de sang de femmes malgaches enceintes, pour la séroprévalence du VIH et de la syphilis, afin d'estimer les besoins pour le dépistage systématique dans la grossesse.
Un dépistage sérologique rétrospectif anonyme pour le VIH et la syphilis a été réalisé sur des échantillons de plasma de 1232 femmes enceintes, prélevés entre mai et juillet 2010 à Ambositra, Ifanadiana, Manakara, Mananjary, Moramanga et Tsiroanomandidy (Madagascar) pendant le suivi de la grossesse. Le dépistage a été basé sur des tests d'hémagglutination de Treponema pallidum (TPHA) pour la syphilis et les tests rapides pour le VIH, avec la confirmation des résultats positifs par les tests sur bandelettes réactives.
Aucune des 1232 femmes enceintes n’étaient séropositives pour le VIH; 37 (3%) étaient séropositives pour Treponema pallidum.
Nos résultats sont en accord avec des études précédentes qui décrivent une prévalence considérable de la syphilis dans la population rurale malgache. Les résultats suggèrent un besoin de dépistage pour prévenir la transmission de Treponema pallidum lors de l'accouchement, tandis que le VIH reste encore rare. Lorsque les infections à Treponema pallidum sont détectées, elles peuvent être facilement traitées, en toute sécurité et à moindre coût, même dans la grossesse, pour réduire le risque de transmission.
La transmisión durante el parto del VIH y de Treponema pallidum, el agente causal de la sífilis, tiene consecuencias severas en los recién nacidos. Las medidas preventivas requieren del conocimiento de la infección materna. Aunque en Madagascar las pruebas para el VIH y la sífilis podrían hacerse, en teoría, dentro del marco del programa nacional de seguimiento de las embarazadas, los kits de pruebas necesarios rara vez están disponibles en los centros sanitarios periféricos. En este estudio hemos evaluado muestras de sangre de mujeres de Madagascar embarazadas en busca de la seroprevalencia del VIH y sífilis, para calcular la demanda de una evaluación sistémica durante el embarazo.
Se realizó un análisis serológico, anónimo y retrospectivo, para el VIH y sífilis, en muestras de plasma de 1232 mujeres embarazadas tomadas entre Mayo y Julio del 2010 en Ambositra, Ifanadiana, Manakara, Mananjary, Moramanga, y Tsiroanomandidy (Madagascar), durante el seguimiento del embarazo. La evaluación se basó en pruebas de hemoaglutinación de Treponema pallidum (TPHA) para sífilis y pruebas rápidas para VIH, con confirmación de los resultados positivos con tiras.
De 1232 mujeres embarazadas ninguna era VIH seropositiva; 37 (3%) eran seropositivas para Treponema pallidum.
Nuestros hallazgos están en línea con estudios previos que describen una prevalencia de sífilis importante entre la población rural de Madagascar. Los resultados sugieren la necesidad de realizar las pruebas necesarias para prevenir la transmisión durante el parto de Treponema pallidum, mientras que la infección por VIH aún es rara. Si es conocida, la infección por Treponema pallidum puede tratarse de forma fácil, segura y barata, incluso durante el embarazo, reduciendo así el riesgo de transmisión.
Sexually transmitted diseases (STDs) including human immunodeficiency virus (HIV) infections and syphilis, which is caused by Treponema pallidum, are prevalent in Madagascar (Harms et al. 1994a,b; Behets et al. 1996). Neglect of the threat of STDs and of condom use facilitates STD acquisition in Madagascar (Andrianasolo et al. 2011), associated with underage sex (9–36%), sex with multiple partners (3–20%) and a low rate of laboratory testing (0–29%) among people at risk for acquiring HIV (WHO 2010: 303–357). Neglect of condom use is reported by 60–70% during casual sexual encounters (Khan et al. 2008), by 20–50% during professional sex (Xueref et al. 2003; WHO 2010: 341). Being married in association with occasional, non-professional sex work is a striking risk factor for STD acquisition, next to poor education, young age (Harijaona et al. 2009) and ambiguous discrimination between client and lover status of sexual partners (Stoebenau et al. 2009). Only 6% of university students in Antananarivo and 7% of men in rural Madagascan areas report consistent condom use during sexual intercourse (Leutscher et al. 2003; Rahamefy et al. 2008), while nearly half of the men questioned have sex with multiple partners (Leutscher et al. 2003).
Consequently, STD infections were diagnosed in 37.5% of women and 26.8% of men in rural Madagascar in 2005. STD acquisition rates were highest in women aged 14–24 years (Leutscher et al. 2005), posing a high risk of vertical transmission.
Peripartal transmission of HIV or Treponema pallidum leads to severe consequences for newborns. Implementation of preventive and therapeutic procedures, including screening and specific therapy, can reduce the risk of transmission (Casalini et al. 2001). HIV and syphilis testing in Madagascar should be offered within the framework of the national pregnancy follow-up scheme as the assessment of HIV and syphilis in pregnancy is demanded by the Madagascar Action Plan (World Bank 2007) and recommended by the WHO (2002). However, the required test kits are rarely available at peripheral health centres (centres de santé de base), which accords with Madagascar being among the 10 most inequitable countries regarding access to maternal health interventions (Barros et al. 2012) and previously detailed inadequacies of the national health system (Harimanana et al. 2011). Low-threshold screening programmes are usually restricted to sentinel analyses on which available surveillance data are based (UNAIDS/WHO Epidemiological Fact Sheets on HIV and AIDS, 2008 Update).
The aim of the study was to identify serological parameters indicating the prevalence of HIV and syphilis infections in pregnant women at selected sites in Madagascar to estimate the demand for systematic HIV and syphilis screening in pregnancy.
Materials and methods
Origin of the analysed samples
In the course of a study to identify malaria parasites in asymptomatic pregnant women, 1244 (EDTA) plasma samples were collected between May 2010 and July 2010 in six intermediated-sized locations with 18 000–36 000 inhabitants (see Figure 1 in the Results section): The two coastal cities of Mananjary (28 000 inhabitants) and Manakara (36 000 inhabitants), Ifanadiana (18 000 inhabitants), which lies on the ascending road from Mananjary and Manakara to the highlands, the two highland cities of Tsiroanomandidy (27 000 inhabitants) and Ambositra (30 000 inhabitants) and Moramanga (29 000 inhabitants), a highland city that lies on the road between the capital Antananarivo and Toamasina. The study sites were chosen to include sites at different elevations for malaria parasite assessment. Furthermore, a chikungunya/dengue seroscreening was initiated after a meeting at the WHO country office because of a previous chikungunya epidemic in the Manajary region (Schwarz et al. 2012).
The sample period at each site was 1 week. Blood was collected from all pregnant women presenting for routine pregnancy screenings to the local health centre who gave consent to their participation in the study. All participants signed an agreement to participate in the study. Ethical clearance for an anonymous retrospective screening for HIV and syphilis in the residual plasma samples as detailed above was obtained from the Ethical Committee of the University of Antananarivo, Madagascar.
Frozen (−20 °C) residual EDTA plasma samples of 1232 pregnant women presenting for routine pregnancy follow-up were screened for anti-HIV antibodies, HIV p24 antigen and anti-Treponema pallidum antibodies.
Determine™ HIV-1/2 Ag/Ab Combo rapid tests (Alere Ltd, Stockport, UK) with a sensitivity of 100% and a specificity of 99.49% for African samples (Alere Ltd) was used for HIV screening. These fourth-generation tests, screening for both p24 antigen and antibodies against HIV, show better sensitivity for antibody responses than ‘antibody-only’-based third-generation rapid tests in case of established HIV infections (Chetty et al. 2012). The sensitivity regarding early infections without antibody response is, however, poor with no more than 0–2% antigen detection (Chetty et al. 2012; Kilembe et al. 2012; Rosenberg et al. 2012). Positive screening results were confirmed using the line assay INNO-LIA™ HIV I/II Score (INNOGENETICS®, Ghent, Belgium) according to the diagnostic procedure of our institutes. If only antigen but no antibodies had been detected in the screening test, a subsequent PCR (Cobas® TaqScreen MPX Test; Roche, Basel, Switzerland) would have been added at a well-equipped German blood transfusion service.
To screen for syphilis, Treponema pallidum haemagglutination tests (TPHA) with a sensitivity of 98.7% and a specificity of 95.4% (Oxoid, Basingstoke, UK) were used. In an independent evaluation, the sensitivity was 92% for primary syphilis (35/38), 100% for secondary syphilis (10/10), early latent syphilis (28/28) and congenital syphilis (2/2), respectively, while specificity was 100% with no cross-reactions with samples from patients with leptospirosis (8), infectious mononucleosis (7), hepatitis (9), diabetes mellitus (11), rheumatoid arthritis (13), leprosy (11), tuberculosis (9), HIV/AIDS (12), systemic lupus erythematosus (4), rheumatic fever (3), as well as from elderly patients (9), pregnant women (29) and blood donors (164) (Rodriguez et al. 2002). Positive results were confirmed or rejected by line assays ‘recom line’ for IgG (Mikrogen GmbH, Neuried, Germany). An individual was counted as syphilis positive if the TPHA screening test was positive and the detected band pattern in the line assay provided at least an indeterminate result.
The age-dependent syphilis seroprevalence was compared for pregnant women in the age groups under 20 years of age, between 20 and 30 years of age and more than 30 years of age. Analysis was performed by chi-squared testing. Significance was accepted at P <0.05. Risk factors for the acquisition of STDs had not been assessed.
Among the 1244 included, sufficient plasma for HIV and syphilis tests was available for 1232 pregnant women. We investigated 190 samples from Mananjary, 251 from Manakara, 192 from Ifanadiana, 198 from Moramanga, 199 from Ambositra and 202 from Tsiroanomandidy. The right-skewed age distribution was similar in all study locations. Overall the median age was 25.3 years (range: 12.5–50.3).
HIV screening with Determine™ HIV-1/2 Ag/Ab Combo rapid tests led to two reactive results for antibodies but not for p24 antigen. Consecutive confirmation tests of these two samples with INNO-LIA™ HIV I/II Score line assays remained negative, so the diagnosis of HIV could not be confirmed in any sample (0% HIV prevalence).
Forty-five of 1232 samples tested positive in the TPHA screening. Confirmation tests with ‘recom line’ IgG-line assays confirmed this result for 37 samples (3% seroprevalence), while eight screening results could not be confirmed and were interpreted as ‘false positive’. In detail, anti-Treponema pallidum-IgG was positive in 34 instances and indeterminate in further three instances.
Syphilis seroprevalences varied between 1.5% and 4.6% among the study locations. Seroprevalences were 2.6% (5 of 190) in Mananjary, 3.2% (8 of 251) in Manakara, 4.2% (8 of 192) in Ifanadiana, 4.6% (9 of 198) in Moramanga, 1.5% (3 of 199) in Ambositra and 2.0% (4 of 202) in Tsiroanomandidy (Figure 1). The age of the corresponding pregnant women was available for 1226 out of the 1236 analysed samples. Syphilis prevalences were by trend slightly higher (4.2%, n = 15) in 359 women over 30 than in 258 women under 20 years (2.3%, n =6) (P =0.21, chi-squared test) and 609 women between 20 and 30 years (2.6%, n =16, P =0.19, chi-squared test; Figure 2).
The study demonstrated complete absence of HIV seroresponse and a moderate Treponema pallidum seroresponse of about 3% in more than 1200 Madagascan pregnant women presenting for routine pregnancy follow-up.
An anonymous study design was demanded for ethical considerations because STDs in general and HIV infections in particular are still associated with discrimination in Madagascar, even in medical settings (Andrianasolo et al. 2011). Specific risk factors for the acquisition of STDs could not be analysed. The residual samples were taken from a study without any focus on sexual behaviour, so no respective practices were recorded. Expectations of future analysis for STDs could not have attracted or deterred any of the pregnant women regarding participation in pregnancy follow-up, because neither the pregnant women nor the investigators knew about or anticipated these analyses at the time of presentation. A differential selection bias is thus unlikely. Moramanga and Ifanadiana, the places with the highest measured syphilis seroprevalence, are comparatively large cities on arterial roads from the capital to the coastline. Haulage traffic might pose a risk factor for STDs there.
The study was focused solely on prevalence data, so the analysis of syphilis activity markers such as IgM, TPHA quantification or lipoid antibody tests such as cardiolipin agglutination or Venereal Disease Research Laboratory test was deliberately avoided. The quantities of residual material would have been insufficient for more sophisticated serodiagnostic testing. In addition, serological evidence of active syphilis infections without the possibility to provide therapy because of the need for anonymization would have led to ethical conflicts.
Syphilis seroprevalence increased slightly with increasing age. Syphilis prevalence in each age group of our study population is still considerably lower than in Madagascan risk groups, as 18%Treponema pallidum seropositivity was detected among 316 registered female sex workers in Toliary in 1998 (Xueref et al. 2003). Four per cent of women from Antananarivo without history of sex work but with genital discharge syndrome were seropositive for syphilis in 2001. Young age, multiple sex partners and poor education were independent risk factors (Behets et al. 2001). Considering that our samples were taken from non-symptomatic pregnant women, a syphilis seroprevalence of 3% in this group suggests a further increase of syphilis infections in the last 10 years.
General HIV seroprevalence in Madagascar rose from 20/100 000 in 1989 to about 30/100 000 in 1992 to 70/100 000 in 1995 (Zeller et al. 1997; Ravaoarimalala et al. 1998). The 2010 UNAIDS report estimated that the total number of people living with HIV in Madagascar was 24 000 in 2009. With the Madagascan population of around 22 million this leads to a prevalence of 1.1 per 1000 (WHO 2010: 180). An HIV seropositivity of 0.9% was found in rural Madagascar in the 15–49 age group in 2005 (Leutscher et al. 2005). Our data suggest that HIV seroprevalence has at least not increased in non-risk groups since that time. The reasons for the differences in HIV spread on the two sides of the Mozambique channel remain unclear as specific risk factors are present in Madagascar (Leutscher et al. 2003; Khan et al. 2008; Rahamefy et al. 2008; WHO 2010: 303–357; Andrianasolo et al. 2011). Future epidemiological studies might focus on the differential distribution of high-risk sexual practices such anal intercourse or dry sex (i.e. sexual intercourse after deliberate drying of the female partner’s vagina to increase friction-associated sensations associated with an increased risk of bleeding).
As HIV prevalence is still low in the assessed pregnant non-risk group, universally available routine screening for HIV during pregnancy follow-up is not a priority issue. In contrast, our data suggest that the establishment of an easily accessible syphilis routine screening for pregnant women in combination with a test-and-treat strategy should be considered for Madagascar regarding the relevant risk of peripartal syphilis transmission to foetuses and the comparatively easy and inexpensive treatment of active syphilis infections in pregnancy (Casalini et al. 2001).
We are grateful to Auguste Rakotondramanana and Jean de Dieu H. Rafalimanana, Laurent Vergnes, Sarah Müssener and Annett Michel for excellent technical assistance.