Jacques Pepin addresses the important
question of parenteral transmission of HIV-1 in Leopoldville, and criticises
aspects of our article (Sousa_et_al,2010,PLoS_ONE
5(4):e9936, http://www.plosone.org/article/info:doi/10.1371/journal.pone.0009936), in which we modelled the initial heterosexual
spread of HIV-1 in that city.1 Here, we respond directly to his
criticisms.
He writes that we used a genital ulcer
disease(GUD)-related risk ratio of HIV-1 transmission of up to 430. He derives
the latter from the measured and published 43% risk incurred by an
uncircumcised man exposed to a HIV-1-infected index woman with a genital ulcer,
which is 430 times the 0.1% baseline.2,3 However, it is misleading
to represent this parameter as equivalent to a GUD risk ratio of 430, as Pepin
does, because the measured 43% risk derived from two concomitant cofactors,
GUD and lack of circumcision, the latter increasing the risk from 4% to 43%.3
In addition, for most other situations involving GUD, the cofactor effect was
much lower in our simulations, consistent with the studies that estimated them
on a per act basis.3,4
The GUD-related risk ratios of around 3-5 cited by Pepin
are based on studies that asked participants whether they had GUD at any time within a long period (e.g.,
the last 12 months), a procedure that strongly underestimates the per act GUD effect.2
Pepin questions our postulated 10 weeks average ulcer
duration. Chancroid's ulcer lasts 10 weeks on average.5 On average,
syphilis' chancre lasts 12 days, and the secondary stage lasts 3.6 months, often
relapsing.6 A modelling study postulated 4 weeks of high sexual
infectiousness during syphilis secondary stage.7 We feel it is reasonable
to assume 4-6 weeks high HIV-1 infectiousness due to primary syphilis or
mucosal lesions associated with secondary syphilis. Lymphogranulomavenereum (LGV)
causes short initial genital ulcers but, in some women, it causes
genitoanorectal syndrome, involving years-lasting genital ulceration.8,9
Given that syphilis, chancroid, and LGV were the most common GUDs in early 20th
century Leopoldville, we consider our choice of ulcers averaging 10 weeks
appropriate.
As Pepin acknowledges, we previously reported low GUD prevalence
in 1950s Leopoldville/Kinshasa.1 We agree with him that substantial parenteral
transmission of HIV-1 may have occurred in Leopoldville/Kinshasa.
1.Sousa_JDd,Müller_V,Lemey_P,Vandamme_A-M,PLoS_ONE(2010);5(4):e9936.
http://www.plosone.org/article/info:doi/10.1371/journal.pone.0009936.
2.Powers_KA_et_al,Lancet_Infect_Dis(2008);8:553-563.
3.Cameron_DW_et_al,Lancet(1989);2(8660):403-407.
4.Hayes_et_al,J_Trop_Med_Hyg(1995);98:1-8.
5.Korenromp_EL_et_al,Sex_Transm_Inf(2002);78:55-63.
6.Garnett_GP_et_al,Sex_Transm_Dis(1997);24:185-200.
7.Oxman_GL_et_al,Sex_Transm_Dis(1996);23:30-39.
8.Woodward_JA,In:_Nelson_Al_&_Woodward_JA(Eds.),Current Clinical Practice:
Sexually Transmitted Diseases,1997,Totowa,NJ:Humana_Press.
9.Kurz_L,British_J_Obstretics_Gynaecology(1913);23:353-388.
Conflict of Interest:
None declared
Jacques Pepin addresses the important question of parenteral transmission of HIV-1 in Leopoldville, and criticises aspects of our article (Sousa_et_al,2010,PLoS_ONE 5(4):e9936, http://www.plosone.org/article/info:doi/10.1371/journal.pone.0009936), in which we modelled the initial heterosexual spread of HIV-1 in that city.1 Here, we respond directly to his criticisms.
He writes that we used a genital ulcer disease(GUD)-related risk ratio of HIV-1 transmission of up to 430. He derives the latter from the measured and published 43% risk incurred by an uncircumcised man exposed to a HIV-1-infected index woman with a genital ulcer, which is 430 times the 0.1% baseline.2,3 However, it is misleading to represent this parameter as equivalent to a GUD risk ratio of 430, as Pepin does, because the measured 43% risk derived from two concomitant cofactors, GUD and lack of circumcision, the latter increasing the risk from 4% to 43%.3 In addition, for most other situations involving GUD, the cofactor effect was much lower in our simulations, consistent with the studies that estimated them on a per act basis.3,4
The GUD-related risk ratios of around 3-5 cited by Pepin are based on studies that asked participants whether they had GUD at any time within a long period (e.g., the last 12 months), a procedure that strongly underestimates the per act GUD effect.2
Pepin questions our postulated 10 weeks average ulcer duration. Chancroid's ulcer lasts 10 weeks on average.5 On average, syphilis' chancre lasts 12 days, and the secondary stage lasts 3.6 months, often relapsing.6 A modelling study postulated 4 weeks of high sexual infectiousness during syphilis secondary stage.7 We feel it is reasonable to assume 4-6 weeks high HIV-1 infectiousness due to primary syphilis or mucosal lesions associated with secondary syphilis. Lymphogranulomavenereum (LGV) causes short initial genital ulcers but, in some women, it causes genitoanorectal syndrome, involving years-lasting genital ulceration.8,9 Given that syphilis, chancroid, and LGV were the most common GUDs in early 20th century Leopoldville, we consider our choice of ulcers averaging 10 weeks appropriate.
As Pepin acknowledges, we previously reported low GUD prevalence in 1950s Leopoldville/Kinshasa.1 We agree with him that substantial parenteral transmission of HIV-1 may have occurred in Leopoldville/Kinshasa.
1.Sousa_JDd,Müller_V,Lemey_P,Vandamme_A-M,PLoS_ONE(2010);5(4):e9936. http://www.plosone.org/article/info:doi/10.1371/journal.pone.0009936.
2.Powers_KA_et_al,Lancet_Infect_Dis(2008);8:553-563.
3.Cameron_DW_et_al,Lancet(1989);2(8660):403-407.
4.Hayes_et_al,J_Trop_Med_Hyg(1995);98:1-8.
5.Korenromp_EL_et_al,Sex_Transm_Inf(2002);78:55-63.
6.Garnett_GP_et_al,Sex_Transm_Dis(1997);24:185-200.
7.Oxman_GL_et_al,Sex_Transm_Dis(1996);23:30-39.
8.Woodward_JA,In:_Nelson_Al_&_Woodward_JA(Eds.),Current Clinical Practice: Sexually Transmitted Diseases,1997,Totowa,NJ:Humana_Press.
9.Kurz_L,British_J_Obstretics_Gynaecology(1913);23:353-388.
Conflict of Interest:
None declared