Do condoms work?

Consistently used condoms provide significant protection against HIV, pregnancy and sexually transmitted infections (STIs).

The degree of protection they offer against HIV and STIs is significantly better than any other single prevention method, taken in isolation, other than sexual abstinence or complete mutual monogamy between two people who have tested negative for HIV.

Despite this, the use and promotion of condoms continue to be targets for controversy and criticism, and sexual abstinence and monogamy are often promoted as superior alternatives.

While condoms offer useful and vital protection, they have also become associated with promiscuity and infidelity.

One widely quoted remark of this nature came from Ugandan President Yoweri Museveni who, at the Fifteenth International AIDS Conference in Bangkok in 2004, advocated for HIV prevention based on “optimal relationships based on love and trust instead of institutionalised mistrust, which is what the condom is all about…I think of condoms as an improvisation, not a solution”. (See http://kaisernetwork.org/health_cast/uploaded_files/museveni_speech071204.pdf for the text of Museveni’s speech.) Museveni later complained of being misunderstood and signed an article in The Lancet saying that condoms formed a valuable part of HIV prevention.¹

Pope Benedict XVI reignited this controversy when, in March 2009, he commented that AIDS in Africa “cannot be overcome…through the distribution of condoms, which can even increase the problem".²

Moral questions about condom use are not within the remit of this resource, but questions of fact are, and condoms’ ability to stop HIV is periodically questioned by people opposed to their use on religious or moral grounds. Therefore questions of condom efficacy have to be addressed and misapprehensions corrected.

In one of the most highly publicised statements, in October 2003, the President of the Vatican's Pontifical Council for the Family, Cardinal Alfonso Lopez Trujillo, said: "The AIDS virus is roughly 450 times smaller than the spermatozoon. The spermatozoon can easily pass through the 'net' that is formed by the condom. These margins of uncertainty...should represent an obligation on the part of the health ministries and all these campaigns to act in the same way as they do with regard to cigarettes, which they state to be a danger."³

These statements are quite simply untrue.

Finding out the degree to which condoms protect against HIV is important both for HIV-negative people who want to protect themselves against HIV, and HIV-positive people who want to avoid transmitting it. Knowing how well they protect against other STIs is important for sexual health in general and may be particularly important for people with HIV, who may be more vulnerable to the effects of certain STIs.

The main findings of studies we look at in more detail below are as follows:

• Condoms used 100% of the time, though not necessarily 100% perfectly (i.e. with usual rates of breakage and slippage) provide protection of 80 to 85% against HIV (uncertainty range: 76 to 93%). In other words, for every 100 cases of HIV infection that would happen without condom use, about 15 (range: 7 to 24) would happen when condoms are used consistently.

• Condoms offer a similar degree of protection against gonorrhoea.

• They offer protection in the range of about 50 to 66% against syphilis, though this depends on factors such as the location of primary syphilis lesions.

• They provide protection against chlamydia and trichomoniasis of a similar or somewhat smaller degree, with different studies showing protection rates varying from 85 to 26%.

• Against genital herpes (HSV-2), estimates of efficacy range considerably; studies are hampered by the fact that people with herpes are only intermittently symptomatic and/or infectious. The best estimate we have is that using condoms more than three-quarters of the time halves the chance of acquiring HSV-2, and may reduce the chances of genital infection with the cold sore virus (HSV-1) too.

• One study has demonstrated that consistent condom use offers women significant protection against HPV infection by men (in the region of 73%). Another has found that condom use helps to prevent HPV infection progressing to cervical or penile cancer in both women and men.

These degrees of protection may be lower than some readers expect, and rates of 98% reliability are still sometimes quoted for condoms. These are based upon observations of their use in contraception: studies have shown that 98% of women relying on condoms as their sole form of contraception remain pregnancy free if condoms are used perfectly, meaning that they are used consistently and correctly at every act of sexual intercourse.⁴

However, because they are not always used correctly even if they are used consistently, studies have found efficacy rates of 85 to 87% when young women use condoms as their sole form of contraception.⁵ This contraceptive efficacy compares with an estimated:

• 92.5% for contraceptive pills

• 96.5% for implants and intra-uterine

• 99% for the pill and condoms used together

• 99.8% for vasectomy

but only

• 77% for the rhythm method (calendar-based

• 75% for withdrawal

• 72% for spermicides.

Condoms are, however, the only method on that list that has been shown to protect against STIs as well as pregnancy.

Laboratory studies and product testing have shown that reputable condoms tested in the laboratory are completely impermeable to micro-organisms as small as viruses.⁶ Research has also found that during vaginal intercourse condoms break less than 2% of the time⁷ and during anal intercourse less than 4% of the time.⁸

However, the same studies show that condoms come off the penis altogether 3 to 5% of the time but may slip down (but not off) up to 13% of the time. In these circumstances, it is easy to see why condoms sometimes fail, even in consistent users.

In addition, however, people are not consistent in their use of condoms, and may not even be consistent when they claim to be, or think they are. Research early on in the epidemic showed that 40 to 70% of men who claimed they use condoms 100% of the time in fact did not use them for every act of intercourse.⁹

Twenty years later, 51 to 66% of women taking part in a microbicide study reported 100% condom usage at different time points during the trial, but the researchers calculated from inconsistency in their answers that the actual proportion who maintained 100% condom use was 25%. Women were much less likely to report inconsistent use of condoms than never using them: over the course of the study, 46% of women said they used condoms ‘always’, 48% ’never’ and only 6% ’sometimes’.¹⁰

For the reasons described above, there is a convention to use two different words when describing the effect of prevention interventions. The efficacy of an intervention is how well it works in a scientific trial or when people use it as indicated, i.e. consistently; its effectiveness is how well it actually works to prevent disease or infection in a given population, given actual levels of use.

Studies of condom efficacy have therefore largely contrasted HIV and STI incidence or prevalence in people who claim 100% consistent use against people who use them inconsistently or not at all. Because these studies involve private behaviours that investigators cannot observe directly, it is difficult to determine accurately whether an individual is a condom user and whether condoms are used consistently and correctly.

The next problem is deciding what kind of study provides truly reliable evidence. It would be unethical to mount a randomised trial of condom use because the control group would have to stop using them altogether. The evidence we have is based on three types of trials, and each has potential weaknesses.

For efficacy against HIV and other chronic STIs, studies of the incidence of HIV (or HSV or HPV) in monogamous serodiscordant couples provides the best evidence. These can be done in individuals whose characteristics are known and can be controlled for, and if the relationship truly is monogamous then infections by acute STIs and from outsiders can be ruled out.

One disadvantage is that condom use in long-term relationships, even in serodiscordant couples, is relatively rare. Another is that the HIV-positive partner will be chronically infected and so will not have the very high viral load characteristic of acute HIV infection. Thirdly, in long-term serodiscordant relationships, studies have shown that the HIV-negative partner can acquire a degree of immunity to their partner’s HIV. For these reasons, HIV transmission within long-term serodiscordant relationships, especially heterosexual ones, may be rarer than it is between casual sex partners. For all these reasons, large studies may be needed to establish differences in HIV (and HSV and HPV) incidence between condom users and non-users.

Another kind of study is to conduct a prospective cohort study, looking at differences in HIV incidence between two groups of people according to their usage of condoms. This cannot be a randomised controlled study, but participants’ HIV and STI incidence can be related to their condom use either at baseline or preferably (because behaviours change over time) by means of regular questionnaires and monitoring. There is opportunity for qualitative research too, contrasting attitudes and drivers of behaviour between people who become infected with HIV or other STIs and those who do not. Condom efficacy against acute STIs can also be measured, if people have multiple partners, or their partners do.

The weaknesses of this kind of study include the fact that condom use cannot be corroborated by partners, so self-report is likely to be even more unreliable. Because behaviours change over time, it can be challenging to decide whether people really are ‘consistent’ users and the pool of consistent users will shrink over time; and, like all cohort studies, results are prone to be confounded by participant characteristics that were not monitored. A study that measures HIV incidence in condom users and non-users will be confounded, for instance, if one group has substantially fewer sexual partners than the other.

For this reason and because HIV seroconversion even in high-risk populations is a relatively uncommon event, prospective cohort studies have to be large and can be quite costly.

A third kind of study is to conduct a retrospective cohort study, asking people about their condom use and contrasting HIV and STI prevalence in users and non-users. Retrospective cohort studies are subject to greater limitations that prospective ones. Participants’ recall of behaviour is often inaccurate; the studies may rely on medical records that may omit significant demographic and medical information; and it can be difficult to determine retrospectively people’s degree of STI exposure.

For all these reasons, measuring the efficacy of condoms (or indeed other established prevention methods and strategies such as serosorting) can be challenging. Nonetheless, a number of carefully conducted studies have demonstrated that consistent condom use is a highly effective means of preventing HIV transmission.

When it comes to STIs other than HIV, most epidemiologic studies of these are characterised by methodological limitations, and thus, the results across them vary widely - ranging from demonstrating no protection to demonstrating substantial protection. However, we now have enough evidence to demonstrate that condoms offer at least some and in some cases excellent protection against most STIs.

Given that condoms have been promoted as the first line of defence against HIV since the beginning of the epidemic, at least in the developed world, it is perhaps surprising that a really rigorous review establishing their efficacy against HIV and STIs was not conducted till June 2000,¹¹ when the US National Institute of Allergy and Infectious Diseases (NIAID) conducted a review of the evidence for their efficacy, spurred on partly by a political climate in the US which at the time was turning against the promotion of condoms and contraception, and towards abstinence and monogamy as the favoured method of protecting against STIs and pregnancy.

The US Centers for Disease Control (CDC) had previously issued a fact sheet¹² that stated that: “several studies provide compelling evidence that latex condoms are highly effective in protecting against HIV infection when used for every act of intercourse”. After the NIAID review, they withdrew this fact sheet and issued another¹³ that also stated that: “Epidemiologic studies that are conducted in real-life settings, where one partner is infected with HIV and the other partner is not, demonstrate conclusively that the consistent use of latex condoms provides a high degree of protection.” This fact sheet also stated that condoms could protect against gonorrhoea, chlamydia and trichomoniasis, but that there was not enough evidence to say whether they were effective against genital ulcer diseases such as syphilis, herpes and HPV infection.

It also prefaced this with the following warning, in bold print:

“The surest way to avoid transmission of sexually transmitted diseases is to abstain from sexual intercourse, or to be in a long-term mutually monogamous relationship with a partner who has been tested and you know is uninfected. For persons whose sexual behaviors place them at risk for STDs, correct and consistent use of the male latex condom can reduce the risk of STD transmission. However, no protective method is 100 percent effective, and condom use cannot guarantee absolute protection against any STD.”

This rewording was interpreted at the time as a move away from the promotion of condoms and an attempt to appease the pro-abstinence lobby, but it is an accurate statement of the protection condoms offer.

The NIAID review first determined the risks of exposure to semen due to condom breakage and found that this, given that breakage is quite rare, was a low risk: about one chance per 166 sex acts using a condom.

It also reviewed patterns of condom use amongst people in the US. It found that condom use amongst women aged from 15 to 45 had increased from 12% in 1982 to 20% in 1995.¹⁴ It found that 33% of young men aged 15 to 19 reported using condoms in 1988 and that that proportion had gone up to 45% in 1995.¹⁵ The Youth Risk Behavior Survey, an annual survey conducted by CDC, also found that condom use had increased from 55 to 62% in young men between 1991 and 1997 and from 38 to 51% in young women.¹⁶ However, these studies did not report rates of slippage and breakage.

HIV is the only STI for which formal meta-analyses of condom efficacy have been published, though only for heterosexual couples, and NIAID evaluated what was then the most recent analysis, by Davis and Weller.¹⁷ This analysis, in which only longitudinal or cohort studies were included, used the following criteria to select studies related to condom use and HIV/AIDS prevention:

• The sample only included serodiscordant, sexually active, heterosexual couples

• HIV status was determined by serology (so that exposure to HIV was known)

• Data collection included self report about condom use

• The study design afforded longitudinal follow-up of HIV uninfected partner.

Davis and Weller found 12 studies that met these criteria. The meta-analysis noted the direction of transmission (male-to-female, female-to-male, and unstated) and date of study enrolment. Condom usage was classified into the following three categories: always (100% use), sometimes, and never.

Among participants who reported always using condoms, the summary estimate of annual HIV/AIDS incidence from the 12 studies was 0.9 seroconversions per 100 person years (0.9%). Among those who reported never using condoms, the summary estimate of annual HIV/AIDS incidence from the seven studies was 6.7%. Overall, Davis and Weller estimated that condoms provided an 85% reduction in HIV/AIDS transmission risk when infection rates were compared in ‘always’ versus ‘never’ users, though the confidence intervals meant that the ‘true’ efficacy could be as low as 60% or as high as 96%.

To cite one of the 12 studies in more detail,¹⁸ researchers looked at Italian serodiscordant couples in which the male partner was HIV positive. Among 171 couples who used condoms consistently, three women (2%) became infected with HIV, whereas among 134 couples who used condoms inconsistently or not at all, 16 women (12%) became infected - a relative risk of 6.6 indicating a condom efficacy of about 84%. Annual HIV incidence was 7.2% among women not using condoms and 1.1% among consistent condom users. This study also found that reported anal sex raised the risk to women by another 40%, but only if they did not use condoms.

Davis and Weller subsequently published another meta-analysis in 2003,¹⁹ this time of 14 studies. These included 13 cohorts of ‘always’ users that yielded an annual HIV incidence estimate of 1.14%, with a confidence interval ranging from 0.56 to 2.04%. There were 10 cohorts of ‘never’ users. The studies with the longest follow-up time, consisting mainly of studies of partners of haemophiliac and transfusion patients, yielded an HIV incidence estimate of 5.75% with a confidence interval of 3.16% to 9.66%. Overall effectiveness, the proportionate reduction in HIV seroconversion with condom use, was approximately 80%, a downward revision from 1999.

The only other meta-analysis of condom efficacy that satisfied NIAID’s criteria was in 1997.²⁰ In this analysis by Pinkerton two sets of studies were evaluated employing two different standards of ‘consistent’ use. For the less rigorous standard, in which ten studies compared ‘users’ with ‘non-users’, the estimated condom efficacy rate was 79%. For the more rigorous standard, nine studies compared ‘always’ users with ‘never or inconsistent’ users (note this is different from Davis and Weller, who compared ‘always’ with ‘never’ users).

In these nine studies there were only four seroconversions reported among 277 (1.4%) steady partners of HIV-positive men and women who consistently used condoms. In contrast, when condoms were used inconsistently or not at all, 171 of 867 (19.7%) sexual partners seroconverted. This led to an efficacy estimate of 93% effective when condoms are used consistently as opposed to inconsistently.

This is about the highest standard of proof we can expect from studies of condom efficacy. Taking Davis and Weller and Pinkerton together, one can say that the best efficacy estimates we have for the use of condoms in preventing HIV are:

• ‘Almost always’ versus ’never’: 79%

• ‘Always’ versus ‘never’: 80 to 85%

• ‘Always’ versus ’sometimes: 93%.

One fact that at first sight seems puzzling is that a number of studies of condom efficacy report that inconsistent use of condoms is in some cases worse than not using them at all. In many studies, as soon as attempted condom use falls off from 100% then the degree of protection they offer rapidly declines.

To take one study from Rakai, Uganda: Ahmed found that 79% of a large sample of people never used condoms,²¹ 16.5% did so irregularly, and only 4.4% always did. He found that annual HIV incidence in non-users was 1.7% a year and in consistent users 0.4% a year – a 76% reduction in risk, concordant with other studies. But he found that HIV incidence in inconsistent users was 2.1% - in other words, larger than that in non-users (though the difference was not statistically significant).

Consistent condom users had half as many cases of gonorrhoea or chlamydia as non-users – again, broadly in line with other studies. But inconsistent users had 44% more gonorrhoea and chlamydia than non-users – and this was statistically significant.

How can sometimes using condoms be worse than never using them? The confounder which distorts these figures is sexual risk behaviour. A high proportion of people who say they ‘never’ use condoms are likely to be people in monogamous relationships who are at no or little risk of HIV or STIs. Conversely people who ‘sometimes’ use condoms are likely to be people who sometimes have non-monogamous or casual encounters, and may be less expert at using condoms than people who always use them.

The same paradox applies in studies of anal sex – see the next section.

What about condoms’ efficacy against HIV in sex between men? Or rather, in anal sex, as this is the transmission behaviour in question?

There are innumerable behavioural studies of gay men’s use of condoms – who uses them, who does not, why they don’t and how they can be persuaded to.

But there has been only one small analysis of the extent to which using condoms actually prevents HIV infection in people who have anal sex, compared with people who do not use condoms. This may be because the figures for vaginal sex are simply extrapolated to anal sex; it may also be because, in gay men at least, a lot of HIV transmission happens in casual situations where the HIV serostatus of partners cannot be assessed, and so the degree of HIV exposure risk are difficult to quantify.

A small review of condom efficacy and anal sex²² found two studies amongst gay men and one amongst women that gave some indication of the relative effectiveness of condom use in anal sex.

The one in women followed seroconversions amongst 256 serodiscordant heterosexual couples and did ask whether they had anal intercourse.²³ There were no seroconversions in 124 couples who always used condoms for vaginal and for anal sex and 12 in people who used condoms inconsistently. Anal intercourse was already a minority behaviour and unprotected anal intercourse even rarer, and the researchers could not directly compare seroconversion rates between women who used condoms for anal sex and ones who did not.

As a second best they took the high HIV seroconversion rate amongst the eight women who practised unprotected anal sex with their partners – two seroconverted, so this was 25% - with the lower rate of 9% (six out of 65 women) who practised condom use inconsistently but did not have unprotected anal sex. The difference between the seroconversion rates in these two groups was non-significant (p=0.15), but the second group would include women who did not have anal intercourse at all, so there was no direct comparison.

The only large longitudinal study of condom efficacy in gay men was published back in 1989.²⁴ Its findings were affected by the sexual risk behaviour paradox we noted above. The Multi-AIDS Cohort Study (MACS), the oldest HIV cohort study in the world, found that, amongst 2914 initially HIV negative gay men, the six-month incidence rate in men who claimed 100% condom use was 0.7% and amongst men who never used condoms it was 2.9%. This yields a condom efficacy of ‘never’ versus ‘always’ of approximately 70%, somewhat lower than that seen in studies in heterosexuals, probably because anal sex is less forgiving of occasional exposure.

However the six-month incidence in men who ‘sometimes’ used condoms was 4.3%, meaning that men who seroconverted were 79% more likely to sometimes use condoms than not to use them at all. Again, this is probably because men who never used condoms were likely to include monogamous men and ones who had less anal sex.

The only later data we have relating HIV incidence among gay men to condom use come from retrospective studies of gay men diagnosed with HIV who were asked about their condom use. In a 2006 study, gay men at the main HIV clinic in Seattle were asked about whether they had consistently used condoms and also whether they had tried to ‘serosort’,²⁵ i.e. restricted unprotected sex to men they knew or were sure were HIV-negative.

The rate of new HIV diagnosis among men who attempted always to use condoms was 1.5%; among patients who had unprotected sex but tried only to do it with same-status partners it was 2.6%; and among men who had unprotected sex regardless it was 4.1%.

Adjusting for the number of partners (though not for their HIV status), attempted consistent condom use was 76% effective in preventing new HIV infections. This is a retrospective epidemiological study with nothing like the same degree of rigour as the studies of HIV serodiscordant couples, but, like the MACS study, it does yield an estimate of condom efficacy somewhat but not hugely lower than the lower bounds of condom efficacy noted in the Weller and Davis and Pinkerton meta-reviews.

Against the other STIs, the NIAID review panel could not find any studies with sufficient methodological rigour to even attempt to estimate efficacy rates for condoms. However, in 2004, the World Health Organization (WHO), benefiting from new study evidence, issued its own review²⁹ in which it was able at least to hazard a guess at condom efficacy against all STIs except human papillomavirus (HPV), the genital wart virus.