Paul, Charlotte MB, PhD; van Roode, Thea MSc; Herbison, Peter MSc; Dickson, Nigel FRACP
BOTH GENDER AND AGE influence the likelihood of acquisition of sexually transmitted infections (STIs) through effects on biologic susceptibility, sexual behavior, and partner choice. They also affect the likelihood that care will be sought and the ease with which infection will be recognized.1 All these factors are relevant to the interpretation of STI infection rates by gender and age.
Surveillance data, based on sexual health clinic attendance in developed countries such as New Zealand and Britain, show higher rates for most STIs (chlamydial infection, gonorrhoea, and genital warts) among women aged 15 to 19 compared with men, but the reverse at older ages.2,3 For instance, for chlamydial infection in both New Zealand and Britain, women have higher rates than men at 16 to 19 years and men have higher rates than women at 20 to 24 years.2,4 Nevertheless, there are limitations to these data because gender, and probably age, influence where people seek advice and treatment for STIs.5,6 As more men than women have their STIs diagnosed at sexual health clinics, clinic data are likely to underestimate the rates for women.
We had a unique opportunity to analyse repeated assessments of reported behaviors and STIs in the same population in a cohort study. Cohort studies that include information on all diagnosed STIs regardless of where people attend for care provide a valid method of comparing reported STI incidence rates by gender and age within a generation. Moreover, by including information collected repeatedly on sexual behavior and circumstances of diagnosis it is possible to gain insights into the effects of gender and age on the acquisition of STIs. Our aim was to examine how the incidence of self-reported STIs varies by gender and age and to elucidate factors that influence this variation. We did this by comparing the incidence rates of STIs from first coitus to age 21, from age 21 to 26, and from age 26 to 32 in a birth cohort.
Materials and Methods
The sample was enrolled in the Dunedin Multidisciplinary Health and Development Study, a longitudinal study of a cohort born in Dunedin, New Zealand between 1 April 1972 and 31 March 1973. They were first followed-up at 3 years of age when 1037 of 1139 eligible children were seen. Since then the sample was seen every 2 years until 15 years, then at 18, 21, 26, and 32 years. Compared with their age group in the country as a whole (at age 21), demographic features were similar, but the sample had a slightly higher level of educational achievement and fewer people of Maori ethnicity.7 The early history of the sample has been described by Silva and Stanton.8
At the 21-, 26-, and 32-year-old assessments, information was sought on sexual health and sexual activity through a computer-presented questionnaire. Ethical approval was obtained from the Otago Ethics Committee.
Questions on STIs were almost identical across assessments. Those who reported ever having had heterosexual intercourse or same sex sexual contact were asked at age 21 if they had ever had 1 or more STI, and at age 26 and age 32 if they had had 1 or more since the previous assessment. The conditions were identified from a list of the common STIs (chlamydia, non-specific urethritis (NSU), genital warts, herpes, gonorrhoea, trichomoniasis, syphilis and another STI). If the respondents indicated “another STI,” they could type the name or indicate what problems it caused. This analysis is restricted to bacterial or viral sexually transmitted infections, and the protozoal infection trichomoniasis. Candidiasis and scabies were excluded as they may not have been sexually transmitted.9 Those who did not know the name were included as having an “other STI.” Genital warts and/or herpes reported by people who had indicated the same infection at an earlier assessment were assumed to be recurrent infections and were excluded. They were also asked if the STI was detected when attending the doctor for symptoms, or for another reason (presumed opportunistic testing), or because a sexual contact had an STI.
The questions on sexual activity were based on those used in the 1990 British National Study of Sexual Attitudes and Lifestyles.10,11 These included questions on the number of sexual partners so far at age 21, in the past 5 years at age 26, and in the past 6 years at age 32, with separate questions for number of opposite sex intercourse partners as well as number of male anal intercourse partners for men. For men, a composite measure summing opposite sex intercourse and male anal intercourse partners was constructed. Additional measures considered were: the relative level of sexual experience of their last partner compared with themselves (fewer, more, or about the same number of sexual partners), whether they had concurrent sexual partners at the start of their last relationship, and the frequency of condom use in last year grouped as “always/usually” and “sometimes/never” (or “about half the time/sometimes/never” at the age 32 assessment).
STATA v10 was used for all analyses. Incidence rates (per 100 person-years) for STIs were calculated for each age period: up to age 21, between ages 21 and 26, and between ages 26 and 32, taking into account years since first coitus for all self-reported STIs, and for bacterial and viral STIs separately. For each age period, those who answered the questions on sexual behavior, and were sexually active, were included. First coitus was defined as the first of either heterosexual intercourse or same sex contact. Information on age at first heterosexual intercourse was sought only at age 21.
Incidence rate ratios (IRRs) and 95% confidence intervals (CIs) were calculated using poisson regression with years since first coitus as exposure and clustering those study members included at multiple assessments, to compare the total number of infections by time period for each gender. For those for whom age at first coitus was unavailable, we estimated years since first coitus based on the first assessment at which sexual contact was acknowledged. Those who were sexually active by age 21 but did not report an age at first intercourse were assumed active for 4 years for the period first coitus to age 21 (based on median age at first coitus). Those who became sexually active between the assessment at 21 and 26, or between the assessment at 26 and 32 were given an exposure time of 3 years for that period. Those who were not assessed at age 21 but attended a later assessment and reported sexual activity were considered active for that full later period. This was necessary for 87 men and 57 women and was primarily because of the higher response rates at the later assessments.
Similar analyses were done to compare total number of infections by gender for each time period, and again excluding those infections detected by screening. For all STI measures, these analyses were repeated adjusting for number of sexual partners. Additionally, when considering the analysis for all self-reported STIs, each of sexual experience of last partner, concurrency at the start of last relationship, condom use in last year, and same sex contact (for men) were considered as possible confounders using the appropriate measure from each time period. These additional measures were judged to be confounders where this further adjustment resulted in a 10% or more change in the point estimate of the IRR for gender (for the comparison by gender) or age period (for the comparison by age period), or if the CIs changed to include or exclude 1.0 resulting in a change of significance.12
Of the 1037 members of the cohort formed at age 3, 1020, 1019, and 1015 were believed to be alive at the 21, 26, and 32 year old assessments, respectively. At age 21, 91.7% (935) of the survivors (477 men and 458 women) completed the questions on their sexual experiences; at age 26, 94.8% (966) did so (490 men and 476 women); and at age 32, 94.5% (959) did so (487 men and 472 women).
At age twenty-one, 416 men and 417 women were sexually active, and completed the questions on STIs. Of these men, 36 (8.7%) reported 38 infections by that time; 73 women (17.5%) reported 89 infections. At age twenty-six, 460 men and 463 women were sexually active and completed the questions on STIs. Of these men, 63 (13.7%) reported 71 infections in that period; 63 women (13.6%) reported 70 infections. At age thirty-two, 472 men and 465 women were sexually active and completed the questions on STIs. Of these men, 43 (9.1%) reported 56 infections in that period; 37 women (8.6%) reported 40 infections. The commonest infections reported by men and women in all periods were chlamydial infection, genital warts, and genital herpes (Table 1). A description of the sexual behavior of these men and women over the 3 time periods is given in Table 2.
For men, the incidence rate for self-reported STIs was 2.0 per 100 person-years from first coitus to age twenty-one, 3.2 between age 21 to 26, and 2.0 between age 26 to 32. For women the rates over these time periods were 4.4, 3.0, and 1.4 per 100 person-years, respectively (Table 3).
The incidence rates for all self-reported STIs by number of sexual partners over these age periods are shown in Figure 1. Risk increased with number of partners, for both genders and most age periods. Within each stratum of number of partners save 0 to 1, the incidence rates for men peaked in the age period 21 to 26. For women with fewer than 10 sexual partners, there was a similar though much less marked peak in the age period 21 to 26, with rates across the 3 age periods similar to those for men. However, for women with 10 or more partners, the incidence rates were dramatically high for the youngest age period from first coitus to 21, decreasing slightly between 21 to 26 years, and dropping markedly by the age period 26 to 32.
Table 3 also displays the relationships between STI incidence and age period for men and women separately. Men in the age period 21 to 26 years had a statistically significantly higher incidence of all STIs compared with the younger age period (up to age 21) which increased after adjustment for number of partners to IRR = 1.9, 95% CI 1.3–2.8. Incidence rates dropped in the oldest age period (between ages 26 to 32) and were similar to those in the age period up to age 21; adjustment did not alter these findings (IRR = 1.1, 95% CI 0.70–1.9). Conversely for women, incidence rates decreased over time, with significantly lower rates in the age period 21 to 26 compared with first coitus to 21, though this difference was no longer significant after adjustment for number of sexual partners and condom use in the last 12 months (IRR = 0.79, 95% CI 0.56–1.1). Incidence rates decreased significantly further in the age period 26 to 32 years compared with the youngest period; this marked difference persisted after adjustment (IRR = 0.39, 95% CI 0.27–0.57). The general patterns for bacterial and viral STIs over these times were similar.
Table 4 displays the relationships between incidence of STIs and gender in 3 different age periods. There is evidence of a gender difference in STI incidence in the youngest age period (first coitus to 21), but not in the older age periods. In the age period up to age 21 years, incidence rates for all STIs were more than twice as high for women compared with men, a difference which was strengthened after adjustment for number of sexual partners (IRR = 2.6, 95% CI 1.8–4.0). No differences were found before or after adjustment in the older periods. This pattern held for bacterial and viral infections.
To investigate whether the patterns by gender and age could be explained by differences in opportunistic screening for STIs, we examined reasons for seeking health care (Table 5). Very few of the men’s diagnoses in the 2 youngest periods were made when “attending a health professional for another reason” (ie, could have been made by an opportunistic test). In contrast, for women this was the case for 22.5% of the infections diagnosed up to age 21. This partly explains the gender difference in the youngest age period. Nevertheless, women still had higher rates than men for the period first coitus to 21 (IRR = 2.0, 95% CI 1.3–3.1), then similar rates for the period 21 to 26 (IRR = 0.81, 95% CI 0.56–1.2) and the period 26 to 32 (IRR = 0.75, 95% CI 0.46–1.2). The patterns by age for men and women also remained the same when opportunistically diagnosed infections were removed from the analysis. For women, compared with the age period first coitus to 21 the rates decreased for the age period 21 to 26 (IRR = 0.90, 95% CI 0.62–1.3) and further for 26 to 32 (IRR = 0.45, 95% CI 0.29–0.69). For men, compared with the period first coitus to 21 the rates peaked for the age period 21 to 26 (IRR = 1.8, 95% CI 1.2–2.7) then dropped between 26 and 32 (IRR = 0.96, 95% CI 0.58–1.6).
We have shown, using data collected at 3 ages in a birth cohort, that adolescent women have significantly higher incidence rates of self-reported STIs than adolescent men. This difference remains after taking into account the number of partners and excluding diagnoses which may have been opportunistically detected. After adolescence, adjusted incidence rates increased significantly for men up to ages 21 to 26 and then fell, whereas for women, rates fell significantly throughout. Incidence of STIs increased with number of sexual partners, except for women in the age period 26 to 32 years. In this population, the commonest STIs were chlamydia, genital warts, and genital herpes.
A major strength of this study is that it is of a population-based birth cohort, so it is possible to examine age effects within the same people. Moreover, patterns by age between men and women can be compared directly. The cohort has a remarkably high retention rate, so participation bias is minimised. This is especially true in comparison with most studies of sexual behavior and self-reported STI where response rates are often in the region of 65% or less, and in which such bias is possible.13 Finally, detailed information on timing of first sexual intercourse and number of partners allowed adjustment for these aspects of sexual behavior. The results should be generalizable to industrialized countries with similar common STIs, as the cohort is in most respects representative of the New Zealand population. Although a comparison between this cohort and a national survey of sexual behavior showed that cohort members were more likely to report multiple sexual partners at age 18, this is likely to reflect the higher response rate and the methods of data collection which facilitated disclosure in our study.14 Furthermore, the occurrence of STIs up to age 21 was similar to US rates for a similar age and time period.15
One important limitation is that the data are based on self-reported diagnoses by a clinician. Self-report depends on access to diagnostic facilities for symptomatic cases (or their contacts) and to screening for asymptomatic cases. The accuracy of self-report is hard to gauge and it has been suggested that women may confuse cervical smear abnormalities with STI diagnoses.16 Although access to diagnostic facilities is good in New Zealand, access to screening has been variable and will have changed over time. Confusion with cytologic abnormalities is also likely to relate only to screen detected disease. Hence, most reliance can be placed on the results that exclude cases identified by opportunistic testing. Reassuringly these showed the same pattern as the overall results. Furthermore, most surveillance systems for STIs rely on diagnoses by clinicians and population surveys often rely on self-report. Both these data sources also have further limitations in examining patterns by gender and age, in particular that they are not able to account for generational differences.
A further limitation is that data on condom use and partner risk was incomplete and did not cover the whole period in which STI information was sought. Moreover, only limited information on patterns of sexual mixing was available. Hence, some residual confounding by aspects of sexual behavior is inevitable. The complexities of context, partners’ infection status, and the interdependency among sexual behaviors call for cautious interpretation of comparisons of STI rates.17 We have considered explanations for the observed differences with this caveat in mind.
Whether a person acquires an STI depends on biologic, behavioral, and epidemiologic factors. Further, whether the STI is recognized depends on the occurrence of symptoms, and whether testing is carried out in response to symptoms or undertaken on asymptomatic individuals. To interpret the incidence of self-reported STIs by gender and age requires consideration of all these factors related to acquisition and recognition.
Comparing women with men in the youngest age period, the much higher STI rate for women, after accounting for differences in testing and sexual behavior, implies either greater biologic susceptibility, greater prevalence of infection in sexual contacts, or a greater proportion of symptomatic infection for women. Current evidence suggests the latter is implausible.1 Greater biologic susceptibility may be because of the amount or vulnerability of the epithelial surface exposed (including cervical ectopy in adolescents) and the duration of exposure. This is the usual explanation for the higher efficiency of transmission from men to women in diseases that cause discharge, and hence for higher rates in women.1 Susceptibility may also be because of the body’s immunologic response. A further explanation, that there might be a greater prevalence of disease in sexual contacts of young women compared to men, was explored by adjusting for relative sexual experience of last partner, and for concurrent partnerships. These did not appreciably affect the incidence rate ratios over and above adjustment for number of sexual partners. From this evidence, biologic factors may be the most important determinants of the comparatively high rates of STIs among young women though an effect of prevalence of disease in sexual contacts cannot be ruled out because of the limitations of adjustment for confounding.
Comparing women and men in the older age periods, the similar rates of diagnosed STIs (adjusted for sexual behavior) were surprising, given the plausible biologic reasons for women’s greater susceptibility (apart from cervical ectopy). These similar rates were related to different ages of peak risk for men and women: for men, risk peaked at 21 to 26 years and then declined, whereas for women risk declined throughout. The decline in risk, after adjustment for sexual behavior, could represent residual confounding especially in relation to partners’ risk, but it could also suggest a decline in susceptibility.
The remarkably low rates for women between 26 to 32 years, and especially for those with 10 or more partners, was investigated further. They applied to both bacterial and viral infection. For genital herpes infections it is possible that the lower incidence at older ages for women could be because of previous infections (only first symptomatic infections were counted) or because of an increase in unrecognized infections. Our earlier analysis of serologically diagnosed herpes simplex virus 2 had shown that incident infections continued to increase with age.18 When we restricted our analysis to women who had reported no previous herpes infection, incidence rates of symptomatic disease still dropped with age (data not shown). On the other hand, comparison of self-reported herpes infections with serological diagnoses showed that, for women, the proportion that were clinically recognized dropped with age, though not significantly (data not shown). Hence, the decline in clinical herpes infections with age may be because of an increase in asymptomatic or unrecognized infections.
For bacterial STIs among women in the oldest age period (that were almost exclusively chlamydia), the most intriguing possibility for the lower incidence at older ages is that such women had greater exposure through more previous sexual partners, and had developed immunity to chlamydial infection. Prevalence of chlamydial infection in women has been shown to decrease with increasing age (though this analysis did not account for changes in partner numbers); and among sex workers the incidence of infection is significantly raised in adolescents compared with older women.19,20 A study of the natural history of chlamydial infection found that, at 4 years follow-up, 94% of women had cleared the infection.21 Rekart and Brunham have recently reviewed such evidence in seeking to understand the reason for increasing incidence of chlamydial infection in Canada.22 They propose that chlamydial infection which is not treated early leads to protective immunity, whereas early treatment may prevent its development, though the evidence for this explanation is contested.23,24
To explore whether older women with more partners might have protective immunity, we examined the previous sexual history of women aged 26 to 32. Women who reported 10 or more partners in that time period were significantly more likely to have had 10 or more sexual partners up to age 26 (82%), than those who reported fewer partners (40%), P <0.001. The proportions for men were similar (P <0.001). Thus our data, showing that older women who report more sexual partners were more likely to have had multiple partners in the past, and less likely to report a recent diagnosis of chlamydia, are consistent with previous infection which is not treated early leading to natural immunity.23 In New Zealand there has been no organized screening program for chlamydia so early treatment for asymptomatic infection will have been relatively uncommon. The fact that the same pattern was not seen for men, despite similar sexual histories, could be because of a different immune response by men, or to acquisition of immunity at a later age (not yet evident in our sample) as the peak risk for bacterial infections was older for men than for women. Further investigation of protective immunity to chlamydia and its effect on disease occurrence in both men and women is warranted.
In conclusion, the findings point to the period up to age 21 being a time of special risk for common STIs for women and of lower risk for men, confirming clinic-based surveillance data. In addition, the findings from age 21 to 32 point to changes in risk of STIs with age which are not accounted for by behavioral risk factors. It is not possible to determine definitively whether these differences represent unaccounted differences in sexual behavior, changes in the proportion with symptomatic disease, or changes in susceptibility.
1.Ehrhardt AA, Bolan G, Wasserheit JN. Gender perspectives and STD. In: Holmes KK, Mardh PA, Sparling PF, et al, eds Sex Transm Dis. New York: McGraw-Hill, 1999:117–128.
2.Sexually Transmitted Infections in New Zealand: Annual Surveillance Report 2004. Wellington: Institute of Environmental Science and Research Limited, April 2005.
3.Brown AE, Sadler KE, Tomkins SE, et al. Recent trends in HIV and other STIs in the United Kingdom: data to the end of 2002. Sex Transm Infect 2004; 80:159–166.
4.Health Protection Agency. SCIEH, ISD National Public Health Service for Wales, CDSC Northern Ireland and the UASSG. Renewing the focus. HIV and other sexually transmitted infections in the United Kingdom in 2002. London: Health Protection Agency 2003.
5.Dickson NP, Paul C, Herbison P. Where young people with multiple sexual partners access medical care: Implications for screening for chlamydial infection. Sex Transm Infect 1998; 74:445–447.
6.Fenton KA, Korovessis C, Johnson AM, et al. Sexual behaviour in Britain: reported sexually transmitted infections and prevalent genital Chlamydia trachomatis infection. Lancet 2001; 358:1851–1854.
7.Dickson N, Paul C, Herbison P, et al. The lifetime occurrence of sexually transmitted diseases among a cohort aged 21. NZ Med J 1996; 109:308–312.
8.Silva PA, McCann M. Introduction to the Dunedin Study. In: Silva PA, Stanton WR, eds. From child to adult: The Dunedin Multidisciplinary Health and Development Study. Auckland: Oxford University Press 1996:1–23.
9.Sobel JD. Vulvovaginal candidiasis. In: Holmes KK, Mardh P-A, Sparling PF, et al, eds Sex Transm Dis. New York: McGraw-Hill 1999:629–639.
10.Johnson AM, Wadsworth J, Wellings K, et al. Sexual attitudes and lifestyles. Oxford: Blackwell 1994:68–109.
11.Johnson AM, Wadsworth J, Elliott P, et al. A pilot study of sexual lifestyles in a random sample of the population of Great Britain. AIDS 1989; 3:135–141.
12.Greenland S. Modeling and variable selection in epidemiologic analysis. Am J Public Health 1989; 79:340–349.
13.Fenton KA, Mercer CH, Johnson AM, et al. Reported sexually transmitted disease clinic attendance and sexually transmitted infections in Britain: Prevalence, risk factors, and proportionate population burden. JID 2005; 191(suppl 1):S127–S137.
14.Paul C, Dickson N, Davis PB, et al. Heterosexual behaviour and HIV risk in New Zealand: data from a national survey. Aust J Public Health 1995; 19:13–18.
15.Laumann EO, Gagnon JH, Michael RT, et al. The social organisation of sexuality. Chicago: University of Chicago Press 1994:376–441.
16.Low N, McCarthy A, Macleod J, et al. Epidemiological, social, diagnostic and economic evaluation of population screening for genital chlamydial infection. Health Technol Assess 2007; 11.
17.Aral SO. Sexual risk behaviour and infection: epidemiological considerations. Sex Transm Infect 2004; 80:8–12.
18.Dickson N, van Roode T, Herbison P, et al. Risk of herpes simplex virus type 2 acquisition increases over early adulthood: Evidence from a cohort study. Sex Transm Infect 2007; 83:87–90.
19.Katz AR, Effler PV, Ohye RG, et al. Assessing age-related risk for gonococcal and Chlamydia infections among females in Hawaii, 2001: A comparison of morbidity rates with screening test positivity. Ambul Pediatr 2004; 4:188–191.
20.Pettifor AE, Turner AN, van Damme K, et al. Increased risk of Chlamydial and Gonococcal infection in adolescent sex workers in Madagascar. STD 2007; 34:475–478.
21.Molano M, Meijer CJL, Weiderpass E, et al. The natural course of Chlamydia trachomatis infection in asymptomatic Colombian women: A 5-year follow-up study. JID 2005; 191:907–916.
22.Brunham RC, Pourbohloul B, Mak S, et al. The unexpected impact of a Chlamydia trachomatis infection control program on susceptibility to reinfection. JID 2005; 192:1836–1844.
23.Rekart ML, Brunham RC. Epidemiology of Chlamydial infection: Are we losing ground? Sex Transm Inf 2008; 84:87–91.
24.Low N. Caution: chlamydia surveillance data ahead. Sex Transm Inf 2008; 84:80–81.