Secondary Logo

Journal Logo


Estimating the Magnitude of STD Cofactor Effects on HIV Transmission

How Well Can it Be Done?


Author Information
  • Free


OBSERVATIONAL STUDIES have consistently shown the presence of sexually transmitted diseases (STDs) to correlate with increased rates of sexual HIV transmission. Biological and clinical studies suggest that this association is in part caused by cofactor effects of STDs on HIV transmission. STDs may enhance the infectivity of HIV-positive patients, because they increase shedding of the HIV virus in the genital tract (the infectivity cofactor effect). Also, STDs may increase the susceptibility to HIV infection, because they disrupt the mucosal integrity of the genital tract and increase the presence and activation of HIV susceptible leukocytes in the genital tract due to inflammation (the susceptibility cofactor effect). 1–6

If STD cofactor effects are strong and STDs are highly prevalent, STD control, e.g., via improvement of STD management, can be a strategy for HIV prevention. The effectiveness of STD treatment for HIV prevention has recently been tested in two trials in sub-Saharan Africa, with apparently contrasting outcomes. 7,8

While community-randomized trials can provide evidence of the role of STDs in HIV transmission in a particular population, their implications for prevention policy in other populations (external validity) are not straightforward. The population-attributable fraction (PAF) of HIV incidence which is due to STD cofactors varies between populations, 9 for example with the type of sexual network and the stage of the HIV epidemic. 10 Because of this, similar STD reductions may impact HIV incidence differently in different populations, and this needs consideration in the evaluation of cofactor effects and the implementation of interventions. For example, the lack of impact of STD reductions on HIV incidence in the mass treatment trial in Rakai, Uganda 7 does not disprove the existence of cofactor effects, nor the possible effectiveness of mass treatment in populations with a larger PAF of STD in HIV spread. 11,12

Besides external validity, other problems can complicate the interpretation of trial outcomes. Lack of impact on HIV may point to the nonexistence of (strong) STD cofactor effects, but might as well simply be due to a failure to reduce STD rates. Conversely, if a trial showed a large reduction in HIV incidence but only small reductions in STD rates (as did the trial of syndromic case management in Mwanza, Tanzania 8,13), the impact of the intervention on HIV is not easily explained from a reduced STD cofactor burden, and other causal pathways cannot be excluded. 14

Prior to organizing STD intervention trials, and as a complementary approach to these trials, researchers have tried to infer the importance of STDs in HIV transmission by estimating the magnitude of biological cofactor effects. Using cofactor magnitudes as input in epidemiologic transmission models, the proportion of HIV transmission attributable to STDs, and the impact of STD control relative to alternative HIV prevention strategies can be predicted for various settings. 10,12,15–17 Provided good estimates of STD cofactor values are available, this approach would obviate the need to repeat effectiveness trials of every alternative STD control strategy in every new target population.


Cofactor values are commonly estimated from longitudinal observational studies monitoring HIV transmission in subjects or couples who had or did not have STDs during or at the start of an observation interval. (Often, the acquisition of STDs is, however, measured retrospectively, as otherwise it would have to be treated.) Associations between HIV transmission and STD presence in these studies can be expressed in different ways: as odds ratios of cumulative transmission throughout the duration of follow-up intervals, as hazard rate ratios based on a survival analysis, or as relative risks per single sexual contact. 18 The interpretation of these cofactor measures is extremely difficult, as they are subject to a multitude of possible biases which may inflate or deflate their value. 3,18–22

Factors Inflating STD Cofactor Estimates

Table 1 lists factors contributing to a positive association between STDs and HIV, which may positively inflate cofactor estimates.

Table 1:
Factors Enhancing the Association Between HIV and STDs, Which Can Inflate STD Cofactor Estimates

The primary cause of spurious associations between STDs and HIV is their common mode of acquisition, through sexual contact with an infected partner. Since individuals with many or promiscuous partners are at high risk for both STDs and HIV, associations between STD and HIV incidence are expected even in the absence of biological interactions. 3,18,19 In order to reduce confounding resulting from underlying same risk factors, cofactor estimates are commonly statistically adjusted for the presence of known risk factors, such as the number of partners of the subject at risk, e.g., by multivariate logistic regression or Mantel-Haenszel stratification. Such adjustment typically reduces the magnitude of the risk estimate, 23–26 as would be expected in case of confounding. It may however not fully resolve confounding, because STDs, HIV and risk behavior cluster not only in study subjects, but also in the (unknown) partners of study subjects, the partners of their partners, i.e., their sexual network at large. Associations between STDs and HIV caused by common underlying risk behavior are particularly strong because of heterogeneity in risk behavior (for example, few individuals have many partners while the majority has few) and assortative mixing (high-risk individuals tend to choose high-risk partners). 27,28 Thus, having at least one unprotected contact in the form of an STD source partner may indicate a relatively large probability of unprotected contacts with other high-risk partners, which may confer a risk of HIV on top of that from the STD source partner.

Consider a cohort of HIV-susceptibles, some of whom experience an STD episode during the observation period, and some of whom do not. A subject experiencing an STD must have had an STD-infected partner during follow-up, who has him-/herself earlier acquired this STD by risk contact and is therefore relatively likely to harbor HIV as well. Consequently, an STD-exposed subject is at increased risk of being exposed to HIV from the STD-infected partner, relative to subjects who do not experience STDs. By enhancing the association between STD and HIV acquisition, this effect thus inflates susceptibility cofactor estimates. Importantly, since the window period until diagnostic detectability of STD (typically 1 week 29) is shorter than the window period until seroconversion for HIV (between 2–3 months 30,31), the acquisition of HIV and STDs in a single sexual contact from an HIV- and STD-positive partner may in longitudinal studies appear as if the STD infection preceded HIV acquisition.

Nonbehavioral risk factors that enhance either the infectivity with or the susceptibility to STDs and HIV simultaneously also contribute to the association between STDs and HIV. Lack of circumcision, for instance, probably contributes to the association between ulcerative STDs and HIV by increasing the susceptibility to, and possibly the infectivity with those STDs and HIV in males independently. 32–36 This effect may indirectly also inflate cofactor estimates in females, if male circumcision causes STDs and HIV to cluster in their male partners and consequently in the women themselves. Other nonbehavioral common risk factors include general immune and health status, nutritional/vitamin A status, hormonal contraception and young age in females. 37–39 By causing heterogeneity among individuals in the fragility of the genital mucosa and hence in susceptibility to and infectivity with HIV and STDs, these and other unknown factors will inflate STD cofactor estimates, unless they could be adjusted for in both study subjects and partners.

In the relation between HIV infectivity and STDs, besides common risk factors, “reverse causation,” i.e., HIV predisposing a person to having STDs, is a source of bias. A factor inducing such reverse causation is HIV-related immunosuppression. Immunosuppression increases the presence and duration of STDs in the later, highly infectious stages of HIV/AIDS disease, due to a worse treatment response, prolonging of the episode duration if left untreated, and an increased occurrence of (recurrent) ulcers. 3,40–43 Evidence for this can be gleaned from the fact that treatment of HIV patients for STD coinfection reduces the level of HIV shedding, but commonly not to the low levels of non–STD-infected HIV patients. 44–47 It has also been suggested on theoretical grounds that HIV might increase the susceptibility to STD infection by altering the host response, 3 although empirical support for this is lacking.

Finally, clustering between different STDs, which is plausible because of common underlying risk factors and has been observed empirically, 32,48–50 enhances the association between HIV and any single STD. Unless adjusted for, clustering may cause the effect on HIV transmission of each single cofactor to be overestimated. Clustering between a cofactor STD and a noncofactor STD may cause the noncofactor STD to be erroneously perceived as a cofactor itself. This effect may for instance play a role in observed associations between bacterial vaginosis (BV) and HIV, 51–54 because BV has been shown to associate independently with trichomoniasis and other nonulcerative STDs. 52,54–56

A special case of STD clustering occurs in sexual couples. Due to the high transmission efficiency of classical STDs (in the order of 10–30% per sexual contact 29,57), STDs in HIV-discordant couples are often present in both partners during at least part of the follow-up period. Estimates of the susceptibility cofactor can then be inflated by an additional enhancement of HIV transmission by the infectivity cofactor, and vice versa.

The likelihood and extent of inflation due to concurrent STDs depend on the interaction between coexistent cofactors, which has not been studied empirically. Their effects can either multiply (e.g., if genital ulcer disease [GUD] increases HIV transmission 5-fold and chlamydia 3-fold, their combination increases HIV transmission 15-fold), add up (e.g., the combination increases HIV transmission 7-fold), or saturate (e.g., chlamydia on top of GUD does not further increase HIV transmission). If cofactors multiply, the concurrent STDs can increase HIV transmission equally in GUD-positive and GUD-negative individuals, and, in case the concurrent STDs cluster with GUD, the STD cofactor effects then enhance the association between HIV and GUD. If the cofactor effects add up or saturate, however, the net effect on the risk estimate for the STD of interest is not obvious and depends on the degree of clustering between STDs in the study population. In the extreme case of no clustering between STDs and saturating or additive cofactors, the concurrent STDs might increase HIV transmission more in GUD-negative individuals than in GUD-positive individuals, thus deflating the risk estimate for GUD.

If cofactors result from different mechanisms, multiplication of their effects is biologically plausible. This may, for example, be the case if an ulcer in an HIV-negative person created a portal of entry for HIV, while a concurrent chlamydia infection increased the presence of HIV-susceptible inflammatory cells in the genital region. One same STD in both partners, e.g., genital ulcers which can create blood-blood contact, may also have a combined cofactor effect of at least the product of the individual cofactor effects (Heiner Grosskurth, personal communication, October 1999). For cofactor effects resulting from the same underlying mechanisms, saturation would seem more likely. An example could be coinfection of an HIV patient with chancroid and syphilis. Individually, these infections enhance the infectivity with HIV, 6 but their combined effect probably saturates at some point.

Factors Deflating STD Cofactor Estimates

Other mechanisms may dilute observed associations between STDs and HIV and bias cofactor estimates downward (Table 2).

Table 2:
Factors Weakening the Association Between HIV and STDs, Which Can Deflate STD Cofactor Estimates

Nondifferential misclassification of STD status will dilute any association, whether positive or negative. Misclassification is particularly likely for self-reported symptoms of nonulcerative STDs in females, which often cause only mild, aspecific symptoms that go unnoticed. 29 Also for laboratory-diagnosed infections, getting a complete track of STD occurrence during follow-up is often difficult. Poor sensitivity and specificity of diagnostic tests may cause misclassification. The time interval between follow-up visits in cofactor studies is typically longer than the duration of STD episodes, so that infections may appear and resolve in between two subsequent visits and hence go unnoticed.

Differential STD misclassification, by contrast, can either dilute or enhance STD/HIV associations. As an example of enhancement, study participants experiencing frequent STD episodes may be more likely to be classified as STD-exposed than participants experiencing less frequent episodes, because the former are better at symptom recognition and their chances of having at least one episode coinciding with a sampling moment are larger. In that case, the true difference in the number of STD episodes between the “positive” and “negative” group would be larger than apparent, leading to overestimation of per-episode or per-contact cofactor effects.

As with STDs, nondifferential misclassification of HIV status at the beginning and end of follow-up intervals, due to a delay in seroconversion after HIV infection of about two months 30,31 deflates positive associations.

Finally, if STD patients abstain from having sex when symptomatic, actual associations would be lower than the theoretical cofactor effect, because the number of STD-enhanced HIV exposures is small.


The above summary of determinants of STD/HIV associations elucidates the complexity of estimating cofactor magnitudes from observational studies. While confounding due to common risk factors can be reduced by adjusting for as many subject attributes as possible, confounding due to sexual network effects is, however, virtually impossible to control for completely. Even though the majority of studies taken as evidence for the existence of significant cofactor effects can be credited for using designs and statistical analyses that reduce confounding as much as possible, 21,23,25,32,48,58,59 none of these can be free of all bias.

Theoretically, the ideal design for a cofactor study would be a trial randomizing (matched pairs of) individuals into two groups, of which one would be infected with STDs and the other not. Both individuals in a pair would then be exposed either in a controlled way and at a controlled frequency to the same HIV-positive partner(s), or uncontrolled to a randomized set of HIV-positive partners. By controlling and/or randomizing both subjects and partners, distortions by heterogeneity in susceptibility to and infectivity with STDs and HIV are avoided. Controlling the mode and frequency of contact with HIV-positive partner(s) should ensure homogeneity in exposure. Obviously, such human experiments are for ethical and practical reasons never performed; imagining them however illustrates how observational studies inevitably fall short of the standards.

The impossibility of observational studies avoiding all these biases would not be of such importance if this could cause only small deviations. However, this is not true, as we will illustrate with data from an often quoted quantitative example.

The cofactor effect of GUD on female susceptibility was estimated 21 from a study of HIV-negative prostitutes in Nairobi who were followed for, on average, 21 months. 59 In this cohort, of 68 prostitutes who reported one or more GUD episodes during follow-up, 72% seroconverted for HIV, while 55% of 49 prostitutes without GUD seroconverted. To avoid biases in the estimates due to the relatively long duration of follow-up during only part of which the STD would be really present, the cofactor estimation focused on the effect per single sexual contact. Per-contact cofactors were derived from the observed cumulative risks, using estimates of the numbers of sexual contacts with HIV-positive client-partners during follow-up and the duration of STD episodes. 21 As summarized in Table 3, it was estimated that GUD enhances the per-contact risk of HIV acquisition by the women by a factor 23. 21

Table 3:
Quantitative Illustration of the Effect of Confounding on a Per-contact Cofactor Estimate for Genital Ulcer Disease (GUD) on Susceptibility to HIV*

In this calculation, HIV prevalence was assumed to be equal in clients of GUD-exposed and unexposed prostitutes. Because of clustering between the presence of HIV and STDs and between HIV exposure and STDs, which has for instance been observed in studies in Uganda 60 and Cameroon, 24 the prevalence of HIV might, however, be higher among the clients of GUD-exposed prostitutes. In addition, STDs- and HIV- exposed prostitutes may have more frequent client contact. In a subcohort of the same Nairobi population, HIV-positive prostitutes had a 22% higher client contact frequency than HIV-negative prostitutes. 43 If we assume that because of these effects the number of HIV exposures was 20% lower among GUD-unexposed prostitutes relative to GUD-exposed prostitutes, the cofactor estimate reduces from 23-fold to 12-fold (Table 3).

Table 3 also shows the possible bias if the effect of GUD were confounded with that of lack of circumcision in clients. Based on circumcision rates of Nairobi males of about 91% in the general population 61 and from 73% to 81% among STD clinic attenders, 32,62,63 we estimated that about 20% of clients of GUD-exposed prostitutes, and 10% of clients of GUD-unexposed prostitutes were uncircumcised. If we assume that the per-contact infectivity with HIV is 3-fold higher in uncircumcised males, 32–34,36 the cofactor estimate would be 14.

In a third scenario, we considered that the GUD-exposed prostitutes suffered not only a susceptibility cofactor effect due to their own GUD, but, preceding this, also a male infectivity cofactor effect during the client contact in which they acquired GUD. We estimated that GUD-exposed prostitutes, who each experienced on average 2.5 GUD episodes during follow-up, 21 had on average 2.5 contacts with HIV-positive GUD source clients during follow-up, during which a cofactor effect of 5 applied; no such effect was assumed for GUD-unexposed prostitutes. This changed the cofactor estimate from 23 to 20.

Finally, as an illustration of the overall effect that residual confounding may have, we considered these three confounders in combination. In this combined scenario, the per-contact cofactor estimate for GUD on female susceptibility consistent with the Nairobi data was only 3 (Table 3).

In contrast to factors inflating cofactor estimates (Table 1), factors deflating cofactor estimates (Table 2) are unlikely to have played a large role in this example. Misclassification of ulcer incidence would seem uncommon in this group of prostitutes, who because of intensive and regular counseling on symptom recognition and health seeking were believed to report to the clinic for the majority of ulcer episodes. 21,59 The long follow-up (21 months) relative to the HIV window period renders misclassification of HIV status during the first and last months of the observation interval relatively unimportant. A reduced frequency of sexual contact during GUD was however possible, and indeed taken into account in the original analysis, under the estimated number of HIV exposures during the presence of GUD. 21


By considering all determinants of associations between STDs and HIV, we showed that available study designs and methods of statistical correction are not sufficient to resolve all biases that can occur when estimating the magnitude of STD cofactor effects from observational studies. The most important source of confounding is the common mode of acquisition, and hence the multiple common risk factors, for STDs and HIV (Table 1). In particular, ignorance of sexual network effects can result in considerable overestimation of cofactor effects. Our quantitative example hereof, in which the consideration of three plausible confounders reduced a per-contact cofactor estimate for GUD on female susceptibility from 23 to 3 (Table 3), by no means exhausts all possible levels and combinations of distortions. In more heterogeneous study populations, inflation is probably more severe than in the Nairobi prostitute cohort, where heterogeneity in risk behavior was likely present in the client partners but probably not so much among the prostitutes themselves. 59

In contrast to confounding, which besets all observational studies, a subset of cofactor studies is affected by STD misclassification (Table 2). The dilution of cofactor estimates that may result from this can be reduced by increasing the frequency of follow-up visits, as has been done in recent studies. 64 The availability of more reliable STD diagnostic tests may in future reduce this problem further.

Within the restriction of observational studies, a preferred study design may be the follow-up of monogamous HIV-discordant couples. 11,36,60,65 In such cohorts, partner attributes such as circumcision status and the level of immunosuppression can be taken into account, and heterogeneity in exposure to HIV and STDs is relatively limited. Couple studies, however, have disadvantages as well. These relationships may not be representative for all partnerships. For example, stable relationships in which HIV has not yet been transmitted from the HIV-positive partner may disproportionally include recently initiated relationships, or partners who for some reason are relatively noninfectious. Furthermore, the validity of this design depends critically on whether the HIV-negative partners behave truly monogamously during follow-up, and the HIV-positive partners do not.

A further step can be the reanalysis of observational studies using dynamic individual-based transmission models that simulate the clustering between STDs and HIV due to same underlying risk factors, patterns in sexual behavior and STD cofactor effects. Projected STDs/HIV associations in simulated individuals and couples can be compared with classical cofactor estimates by processing them statistically as commonly done with actual data, iteratively fitting them against those empirical estimates by varying the underlying model input cofactor values. By simulating a specific cohort study with respect to its design, the sexual network in which it was based and the statistical analysis, cofactor magnitudes can thus be indirectly estimated from these data, accounting for both commonly considered confounders and sexual network effects. Network simulations have provided important theoretical insights into the determinants of STDs/HIV associations. 18,66,67 It must however be borne in mind that the added value of model-based cofactor estimates depends critically on the correctness of representation of the sexual network and of the quantitative effect of all other sources of bias. If the network or some confounders are not well known, model-based estimates may not be better than simpler estimates.

Previous overestimation of STD cofactor effects may be one of the reasons for the disappointing lack of impact of STD prevalence reductions on HIV incidence in the STD treatment trial in Rakai. 7 The contrasting outcomes of the Mwanza trial, which showed a large impact of improved syndromic STD management on HIV incidence with apparently very limited STD reductions 8,13 have been discussed elsewhere. 68–70 Like intervention expectations, estimates of the fraction of HIV transmission attributable to STDs calculated on the basis of classical cofactor estimates may as well be inflated. 6,11,48,71,72 Also, previous model projections, which have assumed cofactor effects of up to a 100-fold per sexual contact 10,12,16,17 may have exaggerated the importance of STDs in HIV spread and prevention.

Besides biasing the overall value of cofactor estimates, the limitations of observational studies have other implications. The strength of distortions differs between populations, depending on the extent of heterogeneity in sexual behavior and the local prevalence of additional confounders, such as AIDS-induced immunosuppression. This complicates the comparison between studies and populations of PAFs based on such population-dependent risk estimates. For instance, the PAF of STDs in HIV transmission has been estimated as larger in the Mwanza trial population as compared to the Rakai trial population. This was however for a considerable part due to higher risk estimates in Mwanza, and not only to a higher fraction of HIV seroconverters exposed to STDs. 11,72 If the larger risk estimates in Mwanza related mainly to more (residual) confounding than in Rakai, e.g., due to the different follow-up scheme or to stronger clustering in Mwanza between STDs and other risk factors, the site comparison would be flawed and the site difference in PAFs overestimated.

Furthermore, as some confounders apply only to subsets of STDs, it is unclear whether the observed strength of associations with HIV of different STDs directly reflect their relative cofactor strengths. For example, the relatively strong association between HIV and ulcerative STDs as compared to nonulcerative STDs may in part result from the strong association of ulcerative STDs with lack of circumcision and HIV-related immunosuppression.

In conclusion, given the difficulty in estimating the magnitude of STD cofactor effects and the absence of solid and consistent evidence from STD intervention trials, it remains uncertain how much STD treatment can contribute to HIV prevention. Clearly, additional community-based intervention trials would help clarify this issue. As yet, for evidence-based HIV prevention policy, condom promotion and other forms of (targeted) primary prevention, which directly prevent the transmission of HIV as well as STDs, remain the safest bet.

However, even if STD cofactor effects on HIV transmission would be weaker than previously thought, improving STD management remains an important component of HIV prevention programs. Any association between STDs and HIV, whether causal or not, indicates that STD patients are at high risk of contracting and transmitting HIV. The education, counseling, condom provision and contact tracing that is part of comprehensive STD management can therefore be an effective means of targeting these HIV prevention strategies to those most in need.


1. Moss GB, Kreiss JK. The interrelationship between human immunodeficiency virus infection and other sexually transmitted diseases [Review]. Med Clin North Am 1990; 74: 1647–1660.
2. Laga M, Nzila N, Goeman J. The interrelationship of sexually transmitted diseases and HIV infection: implications for the control of both epidemics in Africa [Review]. AIDS 1991; 5: S55-63.
3. Wasserheit JN. Epidemiological synergy: interrelationships between human immunodeficiency virus infection and other sexually transmitted diseases [Review]. Sex Transm Dis 1992; 19: 61–77.
4. Mostad SB, Kreiss JK. Shedding of HIV-1 in the genital tract [Editorial review]. AIDS 1996; 10: 1305–1315.
5. Vernazza PL, Eron JJ, Fiscus SA, Cohen MS. Sexual transmission of HIV: infectiousness and prevention. AIDS 1999; 13: 155–166.
6. Fleming DT, Wasserheit JN. From epidemiological synergy to public health policy and practice: the contribution of other sexually transmitted diseases to sexual transmission of HIV infection. Sex Transm Inf 1999; 75: 3–17.
7. Wawer MJ, Sewankambo NK, Serwadda D, et al. Control of sexually transmitted diseases for AIDS prevention in Uganda: a randomised community trial. Lancet 1999; 353: 525–535.
8. Grosskurth H, Mosha F, Todd J, et al. Impact of improved treatment of sexually transmitted diseases on HIV infection in rural Tanzania: randomised controlled trial. Lancet 1995; 346: 530–536.
9. Vittinghoff E, Padian NS. Attributable risk of exposures associated with sexually transmitted disease. J Infect Dis 1996; 174 (Suppl 2): S182–187.
10. Robinson JN, Mulder DW, Auvert B, Hayes RJ. Proportion of HIV infections attributable to other sexually transmitted diseases in a rural Ugandan population: simulation model estimates. Int J Epidemiol 1997; 26: 180–189.
11. Gray RH, Wawer MJ, Sewankambo NK, et al Relative risks and population attributable fraction of incident HIV associated with symptoms of sexually transmitted diseases and treatable symptomatic sexually transmitted diseases in Rakai district, Uganda. AIDS 1999; 13: 2113–2123.
12. Korenromp EL, Van Vliet C, Grosskurth H, et al. Model-based evaluation of single-round mass STD treatment for HIV control in a rural African population. AIDS 2000; 14: 573–593.
13. Mayaud P, Mosha F, Todd J, et al Improved treatment services significantly reduce the prevalence of STDs in rural Tanzania: results of a randomised controlled trial. AIDS 1997; 11: 1873–1880.
14. Hudson CP. Community-based trials of sexually transmitted disease treatment: repercussions for epidemiology and HIV prevention. Bull World Health Organ 2001; 79: 48–58.
15. Stover J, O’Way P. Impact of interventions on reducing the spread of HIV in developing countries: computer simulation applications. Afr J Med Pract 1995; 2: 110–120.
16. Robinson JN, Mulder DW, Auvert B, Hayes RJ. Modelling the impact of alternative HIV intervention strategies in rural Uganda. AIDS 1995; 9: 1263–1270.
17. Bernstein RS, Sokal DC, Seitz ST, Auvert B, Stover J, Namaara W. Simulating the control of a heterosexual HIV epidemic in a severely affected East-African city: computational modeling of single vs combined intervention strategies with the iwgAIDS and simulAIDS models. Interfaces 1998; 28.
18. Boily MC, Anderson RM. Human immunodeficiency virus transmission and the role of other sexually transmitted diseases: measures of association and study design. Sex Transm Dis 1996; 23: 312–332.
19. Mertens TE, Hayes RJ, Smith PG. Epidemiological methods to study the interaction between HIV infection and other sexually transmitted diseases [Review]. AIDS 1990; 4: 57–65.
20. Nagelkerke NJD, Kimani J, Moses S, Plummer FA. Exploring confounding in nested case-control studies [letter]. AIDS 1995; 9: 1202–1203.
21. Hayes RJ, Schulz KF, Plummer FA. The cofactor effect of genital ulcers on the per-exposure risk of HIV transmission in sub-Saharan Africa [Review]. J Trop Med Hyg 1995; 98: 1–8.
22. Dickerson MC, Johnston J, Delea TE, White A, Andrews E. The causal role for genital ulcer disease as a risk factor for transmission of human immunodeficiency virus: an application of the Bradford Hill criteria. Sex Transm Dis 1996; 5: 429–440.
23. Telzak EE, Chiasson MA, Bevier PJ, et al. HIV-1 seroconversion in patients with and without genital ulcer disease: a prospective study. Ann Intern Med 1993; 119: 1181–1186.
24. Weir SS, Feldblum PJ, Roddy RE, Zekeng L. Gonorrhea as a risk factor for HIV acquisition. AIDS 1994; 8: 1605–1608.
25. Nelson KE, Eiumtrakul S, Celentano D, et al. The association of herpes simplex virus type 2 (HSV-2), Haemophilus ducreyi, and syphilis with HIV infection in young men in northern Thailand. J Acquir Immune Defic Syndr Hum Retrovirol 1997; 16: 293–300.
26. Nopkesorn T, Mock PA, Mastro TD, et al. HIV-1 subtype E incidence and STD in a cohort of military conscripts in northern Thailand. J Acquir Immune Defic Syndr Hum Retrovirol 1998; 18: 372–379.
27. Garnett GP, Anderson RM. Factors controlling the spread of HIV in heterosexual communities in developing countries: patterns of mixing between different age and sexual activity classes. Philos Trans R Soc Lond B Biol Sci 1993; 342: 137–159.
28. Morris M. Sexual networks and HIV. AIDS 1997; 11 (suppl A): S209–S216.
29. Holmes KK, Mardh P-A, Sparling PF, et al. Sexually transmitted diseases, 2nd ed. New York: McGraw-Hill, 1990.
30. Horsburgh R Jr, Jason J, Longini IM, et al. Duration of HIV infection before detection of antibody. Lancet 1989; 8664: 637–640.
31. Willerford DM, Bwayo JJ, Hensel M, et al. Human immunodeficiency virus infection among high-risk seronegative prostitutes in Nairobi. J Infect Dis 1993; 167: 1414–1417.
32. Cameron DW, Simonsen JN, D’Costa LJ, et al. Female to male transmission of human immunodeficiency virus type 1: risk factors for seroconversion in men [see comments]. Lancet 1989; 2: 403–407.
33. Jessamine PG, Plummer FA, Ndinya Achola JO, et al. Human immunodeficiency virus, genital ulcers and the male foreskin: synergism in HIV-1 transmission [Review]. Scand J Infect Dis Suppl 1990; 69: 181–186.
34. Moses S, Bailey RC, Ronald AR. Male circumcision: assessment of health benefits and risks. Sex Transm Infect 1998; 74: 368–373.
35. Lavreys L, Rakwar JP, Thompson ML, et al. Effect of circumcision on incidence of HIV type I and other STDs: a prospective cohort study of trucking company employees in Kenya. J Infect Dis 1999; 180: 330–336.
36. Quinn TC, Wawer MJ, Sewankambo N, et al. Viral load and heterosexual transmission of human immunodeficiency virus type 1. Rakai Project Study Group [see comments]. N Engl J Med 2000; 342: 921–929.
37. Clemetson DBA, Moss GB, Willerford DM, et al. Detection of HIV DNA in cervical and vaginal secretions: prevalence and correlates among women in Nairobi, Kenya. JAMA 1993; 269: 2860–2864.
38. Mostad SB, Overbaugh J, DeVange DM, et al. Hormonal contraception, vitamin A deficiency and other risk factors for shedding of HIV-1 infected cells from the cervix and vagina. Lancet 1997; 350: 922–927.
39. John GC, Nduati RW, Mbori-Ngacha D, et al. Genital shedding of human immunodeficiency virus type 1 DNA during pregnancy: association with immunosuppression, abnormal cervical or vaginal discharge, and severe vitamin A deficiency. J Infect Dis 1997; 175: 57–62.
40. Nzila N, Laga M, Thiam MA, et al. HIV and other sexually transmitted diseases among female prostitutes in Kinshasa. AIDS 1991; 5: 715–721.
41. Wald A, Corey L, Handsfield HH, Holmes KK. Influence of HIV infection on manifestations and natural history of other STDs. Annu Rev Public Health 1993; 14: 19–42.
42. Ghys PD, Diallo MO, Ettiègne-Traoré V, et al. Genital ulcers associated with human immunodeficiency virus-related immunosuppression in female sex workers in Abidjan, Ivory Coast. J Infect Dis 1995; 172: 1371–1374.
43. Kaul R, Kimani J, Nagelkerke NJD, et al. Risk factors for genital ulcerations in Kenyan sex workers: the role of HIV-1 infection. Sex Transm Dis 1997; 24: 387–392.
44. Moss GB, Overbaugh J, Welch M, et al. HIV DNA in urethral secretions in men: association with gonococcal urethritis and CD4 cell depletion. J Inf Dis 1995; 172: 1469–1474.
45. Cohen MS, Hoffman IF, Royce RA, et al. Reduction of concentration of HIV-1 in semen after treatment of urethritis: implications for prevention of sexual transmission of HIV-1. Lancet 1997; 349: 1868–1873.
46. Ghys PD, Fransen K, Diallo MP, et al. The associations between cervicovaginal HIV shedding, sexually transmitted diseases and immunosuppression in female sex workers in Abidjan, Cote d’Ivoire. AIDS 1997; 11: F85–F93.
47. Rotchford K, Strum AW, Wilkinson D. Effect of coinfection with STDs and of STD treatment on HIV shedding in genital-tract secretions: systematic review and data synthesis. Sex Transm Dis 2000; 27: 243–248.
48. Laga M, Manoka A, Kivuvu M, et al. Non-ulcerative sexually transmitted diseases as risk factors for HIV–1 transmission in women: results from a cohort study. AIDS 1993; 7: 95–102.
49. Mbizvo MT, Machekano R, McFarland W, et al. HIV seroincidence and correlates of seroconversion in a cohort of male factory workers in Harare, Zimbabwe. AIDS 1996; 10: 895–901.
50. Dada AJ, Ajayi AO, Diamondstone L, Quinn TC, Blattner WA, Biggar RJ. A serosurvey of Haemophilus ducreyi, syphilis, and herpes simplex virus type 2 and their association with human immunodeficiency virus among female sex workers in Lagos, Nigeria. Sex Transm Dis 1998; 25: 237–242.
51. Cohen CR, Duerr A, Pruithithada N, et al. Bacterial vaginosis and HIV seroprevalence among female commercial sex workers in Chiang Mai, Thailand. AIDS 1995; 9: 1093–1097.
52. Sewankambo N, Gray RH, Wawer MJ, et al. HIV-1 infection associated with abnormal vaginal flora morphology and bacterial vaginosis. Lancet 1997; 350: 546–550.
53. Taha TE, Hoover DR, Dallabetta GA, et al. Bacterial vaginosis and disturbances of vaginal flora: association with increased acquisition of HIV. AIDS 1998; 12: 1699–1706.
54. van de Wijgert JHHM, Mason PR, Gwanzura L, et al. Intravaginal practices, vaginal flora disturbances, and acquisition of STD in Zimbabwean women. J Infect Dis 2000; 181: 587–594.
55. Zenilman JM, Fresia A, Berger B, McCormack WM. Bacterial vaginosis is not associated with circumcision status of the current male partner. Sex Transm Infect 1999; 75: 347–348.
56. Taha TE, Gray RH, Kumwenda NI, et al. HIV infection and disturbances of vaginal flora during pregnancy. J Acquir Immune Defic Syndr Hum Retrovirol 1999; 20: 52–59.
57. Brunham RC, Plummer FA. A general model of sexually transmitted disease epidemiology and its implications for control. Med Clin North Am 1990; 74: 1339–1352.
58. Holmberg SD, Stewart JA, Gerber AR, et al. Prior HSV-2 infection as a risk factor for HIV infection. JAMA 1988; 259: 1048–1050.
59. Plummer FA, Simonsen JN, Cameron DW, et al. Cofactors in male-female sexual transmission of human immunodeficiency virus type 1. J Infect Dis 1991; 163: 233–239.
60. Carpenter LM, Kamali A, Ruberantwari A, Malamba SS, Whitworth JAG. Rates of HIV-1 transmission within marriage in rural Uganda in relation to the HIV sero-status of the partners. AIDS 1999; 13: 1083–1089.
61. Hunter DJ, Maggwa BN, Mati JK, Tukei PM, Mbugua S. Sexual behavior, sexually transmitted diseases, male circumcision and risk of HIV infection among women in Nairobi, Kenya. AIDS 1994; 8: 93–99.
62. Simonsen JN, Cameron DW, Gakinya MN, et al. Human immunodeficiency virus infection among men with sexually transmitted diseases: experience from a center in Africa. N Engl J Med 1988; 319: 274–278.
63. Nasio JM, Nagelkerke NJD, Mwatha A, Moses S, Ndinya-Achola JO, Plummer FA. Genital ulcer disease among STD clinic attenders in Nairobi: association with HIV-1 and circumcision status. Int J STD AIDS 1996; 7: 410–414.
64. Rakwar J, Lavreys L, Thompson ML, et al. Cofactors for the acquisition of HIV-1 among heterosexual men: prospective cohort study of trucking company workers in Kenya. AIDS 1999; 13: 607–614.
65. De Vincenzi I, for the European study group on heterosexual transmission of HIV: a longitudinal study of HIV transmission by heterosexual partners. New Engl J Med 1994; 331:341–346.
66. Koopman JS, Longini IM, Jacques JA, et al. Assessing risk factors for transmission of infection. Am J Epidemiol 1991; 133: 1199–1209.
67. Robinson NJ. The association of HIV-1 and other sexually transmitted diseases, and its relevance to intervention programmes in rural Uganda: a simulation modelling exercise [PhD dissertation]. London, UK: London School of Hygiene and Tropical Medicine, 1994.
68. Dik J, Habbema F, de Vlas SJ. Impact of improved treatment of sexually transmitted disease on HIV infection [letter; comment]. Lancet 1995; 346:1157–8; discussion 1159–1160.
69. Hudson CP. Syndromic management for STDs: back to the drawing board [Review]. Int J STD AIDS 1999; 10: 423–434.
70. Grosskurth H, Gray R, Hayes R, Mabey D, Wawer M. Control of STDs for HIV prevention: understanding the implications of the Mwanza and Rakai trials [Viewpoint paper]. Lancet 1999; 355: 1981–1987.
71. Otten MW Jr, Zaidi AA, Peterman TA, Rolfs RT, Witte JJ. High rate of HIV seroconversion among patients attending urban STD clinics. AIDS 1994; 8: 549–553.
72. Orroth KK, Gavyole A, Todd J, et al. Syndromic treatment of sexually transmitted diseases reduces the proportion of incident HIV infections attributable to these diseases in rural Tanzania. AIDS 2000; 14: 1429–1437.
© Copyright 2001 American Sexually Transmitted Diseases Association