Predictors of repeat Chlamydia trachomatis and/or Neisseria gonorrhoeae infections among African-American adolescent women
- Andrea Swartzendruber,
- Jessica M Sales,
- Jennifer L Brown,
- Teaniese Latham Davis,
- Ralph J DiClemente,
- Eve Rose
- Department of Behavioral Sciences and Health Education, Rollins School of Public Health, Emory University, Atlanta, Georgia, USA
- Correspondence to Dr Andrea Swartzendruber, Department of Behavioral Sciences and Health Education, Rollins School of Public Health, 1518 Clifton Road, Atlanta, GA 30322, USA;
- Received 15 February 2012
- Revised 7 November 2012
- Accepted 11 November 2012
- Published Online First 12 December 2012
Background Young African-American women have the highest rates of Chlamydia trachomatis and Neisseria gonorrhoeae in the USA. The objective was to identify baseline predictors of repeat chlamydia and/or gonorrhoea infections among African-American adolescent women.
Methods Sociodemographic, psychosocial and behavioural data were collected at baseline and every 6 months for 2 years from 701 African-American women (14–20 years) enrolled in an HIV prevention trial. Vaginal swabs were self-collected at each visit and assayed for chlamydia and gonorrhoea using DNA amplification. Among participants testing positive for chlamydia and/or gonorrhoea at baseline, logistic regression analyses assessed baseline predictors of repeat infection.
Results Of 618 (88%) participants with ≥1 follow-up assessment, 123 (20%) had a positive chlamydia and/or gonorrhoea test result at baseline; 49 (40%) had a repeat infection during the study period. Of those with a repeat infection, 30 (61%) were positive at one follow-up visit, 18 (37%) at two visits and 1 (2%) at three follow-up visits. Controlling for age and intervention condition, impulsivity (AOR: 1.71, p=0.018) was associated with an increased likelihood, and having a boyfriend (AOR: 0.21, p=0.006) was associated with a decreased likelihood of repeat infection.
Conclusions Repeat chlamydia and/or gonorrhoea infections are common among African-American adolescent women. Among young African-American women who test positive for chlamydia and/or gonorrhoea, tailored interventions for more impulsive adolescents and those not in a relationship may reduce risk of repeat infections. Given the high numbers of repeat infections after receipt of an evidence-based intervention, enhanced screening and treatment services for young men may be warranted.
In the USA, African-American women aged 15–19 years have the highest rates of both Chlamydia trachomatis and Neisseria gonorrhoeae among all age, race/ethnicity and gender categories.1 Figures 1 and 2 present the rates of chlamydia and gonorrhoea, respectively, among females 15–19 years in the USA in 2010, by race/ ethnicity.1 Studies have estimated that up to 53% of all chlamydia diagnoses are repeated infections and persons with repeat infections account for up to 48% of gonorrhoea diagnoses.2 ,3
Both chlamydia and gonorrhoea are associated with adverse reproductive health outcomes such as pelvic inflammatory disease, infertility, ectopic pregnancy and perinatal infections.4 Reinfection is associated with an increased risk of adverse reproductive health outcomes.4 Both chlamydia and gonorrhoea also facilitate HIV transmission.5 The annual direct medical cost of treating sexually transmitted infections (STIs), excluding HIV, and their sequelae is $8 billion.6 The time of diagnosis may be an important opportunity to prevent subsequent infections.2 Examining factors that predict repeat infection may help researchers and programmers develop and target interventions to reduce infection rates and their public health burden.
Individual and partner behaviour, and relationship and psychosocial factors influence STI risk. Individuals who have been adequately treated for chlamydia and/or gonorrhoea may become reinfected through a new infected partner or an existing partner who has not been adequately treated or has a newly acquired infection. Partner-level factors associated with STIs among adolescent women include older partner age, a partner with concurrent sexual partners, sex with a high or drunk partner and a partner recently released from incarceration.7–10 Psychosocial factors, such as stress and depression, are associated with increased sexual risk behaviour among adolescents.11 Among African-American adolescent women, heavy alcohol use has been shown to predict STI acquisition and sexual risk behaviour, including inconsistent condom use, multiple sexual partners and sex while high or drunk.12 Abuse and dating violence also impact STI risk. Evidence shows that young African-American women who experienced sexual violence have an increased risk for STIs, earlier sexual debut and a greater number of lifetime sex partners.13 ,14 Among adolescents, dating violence is associated with increased sexual risk behaviour and substance use.15 Adolescents who experience dating violence may be at increased risk of STIs through forced sex or reduced capacity negotiate safer sex practices.13 ,15 ,16 Among African-American adolescent women, a history of dating violence is associated with increased fear about the consequences of condom negotiation and fear of talking to a partner about pregnancy prevention.16
Despite marked disparities in the rates of chlamydia and gonorrhoea, we are not aware of studies that have explored predictors of repeat infection exclusively among young African-American women. The objective of this study was to identify baseline predictors of repeat chlamydia and/or gonorrhoea infections among African-American adolescent women enrolled in an HIV prevention trial who tested positive for chlamydia and/or gonorrhoea at baseline.
From July 2005 to June 2007, African-American adolescent females were recruited from three reproductive health clinics in Atlanta, Georgia, to participate in an HIV prevention trial. Potential participants were approached in clinic waiting areas by a female African-American recruiter who assessed study eligibility. Eligibility criteria included being aged 14–20 years and having had at least one episode of vaginal sex without a condom in the past 6 months. Individuals who were married, pregnant or attempting to become pregnant were excluded from participating. On a return visit to the clinic, written informed consent was obtained, with parental consent waived for individuals younger than 18 years. Participants then completed baseline assessments and were randomised to trial conditions. Of 1684 individuals screened, 745 met the eligibility criteria. Of the eligible adolescents, 94% (N=701) enrolled and were randomised (intervention condition, n=342; comparison condition, n=359). Participants were compensated $75 for travel and childcare to attend intervention sessions and complete assessments. This analysis reports on 618 (88%) participants who, in addition to the baseline assessment, completed one assessment or more during 2 years of post-intervention follow-up. The Emory University Institutional Review Board approved all study protocols.
The parent study was a randomised controlled supplemental treatment trial. As such, all participants received the same ‘primary’ treatment and, depending on study condition, received one of two supplemental treatments. The purpose of the trial was to assess whether supplemental treatment could enhance the maintenance of an efficacious HIV risk reduction intervention.17 All participants received a single 5-h group-delivered HIV risk reduction intervention based on HORIZONS, a two-session intervention shown to be efficacious among African-American women.18 Participants assigned to the intervention condition then received 12 brief phone calls, during which a health educator conducted an HIV risk assessment and provided individualised HIV prevention counselling. Participants assigned to the control condition also received 12 brief phone calls, but the health educator conducted a nutritional risk assessment and provided nutrition counselling. Participants received calls approximately every 8 weeks for 24 months following the primary treatment.
Data collection occurred at baseline and at 6, 12, 18 and 24 months post-intervention, and consisted of an audio computer-assisted self-interview (ACASI) and a self-collected vaginal swab. Participants provided a self-collected vaginal swab that was assayed for chlamydia and gonorrhoea, using the BDProbeTec ET Chlamydia trachomatis and Neisseria gonorrhoeae amplified DNA assays (Becton Dickinson and Company, Sparks, Maryland, USA).19 Participants with a positive STI test result received a directly observed single-dose antimicrobial treatment and risk-reduction counselling as per Centers for Disease Control and Prevention recommendations, and were encouraged to refer their sex partners for treatment. The County Health Department was notified of reportable STIs. After providing the vaginal swab specimen, participants completed the ACASI, which assessed sociodemographics, sexual history and psychosocial constructs associated with sexual risk behaviours.
Summary of main trial findings
There were no differences between study conditions in terms of sociodemographic characteristics, STIs or behavioural outcomes at the time of randomisation. Retention was similar, with 89.2% of individuals in the intervention condition and 87.2% in the comparison condition completing one or more follow-up assessments (p=0.414). There were statistically significant differences in the primary and secondary outcomes between the two arms through 18 months. Participants in the intervention arm had a lower proportion of incident chlamydia infections (7.0% vs 13.6%, p=0.009), a lower mean frequency of sex while high on drugs or alcohol (1.42 vs 2.42; p=0.009) and a higher mean percent of condom-protected sex acts (58% vs 52%; p=0.049).17
Repeat infection was defined as a positive chlamydia and/or gonorrhoea test following a negative result or documented treatment among individuals positive for chlamydia and/or gonorrhoea at baseline.
Sociodemographic factors considered as potential baseline predictors included age (years) and having a paid job (yes/no). Family aid was assessed with a four-item index. Participants respond yes/no to ‘In the past 12 months, did you or anyone you live with receive any money or services from: (a) welfare (including temporary assistance to needy families or supplemental security income; (b) food stamps; (c) Women, Infants and Childreni and (d) Section 8 housing (housing subsidies)?’
Cigarette smoking and substance use: Smoking was assessed with the question ‘Do you smoke cigarettes?’ (yes/no). Alcohol use in the 90 days prior to the baseline assessment was assessed with ‘In the past 90 days, [on] how many days have you used alcohol?’ and ‘How many alcoholic drinks do you usually have at one time?’ Participants who indicated they used alcohol on ≥1 days were categorised as users. Participants who indicated they usually had ≥3 drinks on one occasion were categorised as heavy alcohol users. Marijuana use prior to baseline assessment was assessed with ‘In the past 90 days, how many days have you used marijuana?’ Participants were dichotomised as having or not having used marijuana.
Sexual risk behaviours assessed at baseline included: age at first sex (years), lifetime number of vaginal sex partners (above or equal to the median number vs less than median) and self-reported STI history (yes/no). Sexual behaviours in the 90 days prior to baseline assessment included: (a) number of vaginal sex partners; (b) percent of condom use during vaginal sex, calculated by dividing the number of times the respondent had vaginal sex by the number of times condoms were used; (c) consistent condom use, defined as 100% use during vaginal sex (yes/no). We also assessed the participant's report of being drunk or high during sex (yes/no) as a potential baseline predictor. Participants’ responses were dichotomised based on the question ‘In the past 90 days, how many times did you have sex while high on alcohol or drugs?’; reports of ≥1 times were categorised as engaging in the behaviour. In addition, we explored whether hormonal contraceptive use (‘pills/ patch/depo’) at last sex (yes/no) predicted repeat infection.
Partner-level factors included: (a) currently having a boyfriend (yes/no), (b) participant's perception of whether her current boyfriend had sex with another woman during the relationship (yes/no), (c) currently having a casual partner (yes/no), (d) general age of partners (≥2 years older vs the same age or younger) and (e) sex in the past 6 months with a male partner who had recently been released from jail, prison or detention (yes/no) and (f) sex with a partner who was high or drunk during sex in the past 90 days (yes/no). Participants were asked ‘In the past 90 days, how many times did you have sex while your partner was high on alcohol or drugs?’; reports of ≥1 times were categorised as engaging in the behaviour.
Psychosocial constructs associated with sexual risk behaviour included perceived interpersonal stress, depressive symptomology, fear of condom negotiation, partner communication self-efficacy, impulsivity and sexual sensation-seeking. Perceived stress was assessed with a 14-item scale, in which respondents indicated their level of stress around various interpersonal situations and their overall stress level,which ranged from 1=‘no stress’ to 5=‘extreme stress’; a sample item was, ‘relationships with family members’ (α=0.88). Depression was measured with an eight-item scale (α=0.90);20 participants were asked how frequently during the past week they experienced depressive symptoms from 1=‘less than 1 day’ to 4=‘5–7 days’. A sample item was ‘I felt sad’. Participants with a score of 16 or greater were classified as having elevated depressive symptoms. Fear of consequences of condom negotiation was measured using a 7-item scale (α=0.88); a sample item was ‘I have been worried that if I talked about using condoms with my boyfriend or sex partner, he would threaten to leave me.’ Participants’ responses ranged from 1= ‘never’ to 5= ‘always’. Partner communication self-efficacy was assessed with a six-item index (α=0.82); a sample item was, ‘How hard is it for you to demand that he use a condom?’ Participants responded from 1=very hard to 4=very easy. Impulsivity was measured with a 15-item scale (α=0.82);21 a sample item was ‘I just react without thinking’. Participants’ responses ranged from 1= ‘never’ to 5=‘always’. A 10-item scale measured sexual sensation-seeking (α=0.74). The original nine-item scale was modified to include ‘Videotaping or photographing myself and my partner having sex is exciting.’22 Participants’ responses ranged from 1=‘strongly disagree’ to 4=‘strongly agree’. With the exception of depressive symptomology, all psychosocial measures used in regression models were standardised to mean zero and unit variance.
History of abuse: A history of emotional abuse was assessed with ‘Have you ever been emotionally abused? (threatened, called names, etc)’ (yes/no). Physical abuse was assessed with ‘Have you ever been physically abused? (hit, punched, kicked, slapped, etc)’ (yes/no). Sexual abuse was assessed with ‘Has anyone ever forced you to have vaginal sex when you didn't want to?’ (yes/no).
T-tests and χ2 statistics compared selected baseline characteristics among study participants according to completion of at least one post-intervention assessment. Descriptive statistics assessed the frequencies of repeat infections. Simple and multivariable logistic regression analyses identified baseline factors associated with repeat infection during the 2-year follow-up period among participants positive for chlamydia and/or gonorrhoea at baseline. Thus, the comparator groups are individuals who were positive for one or both STIs at baseline and at a subsequent visit, and individuals who were positive for one or both STIs at baseline only. First, simple logistic models assessed associations between potential baseline predictors and repeat infection. Factors significant at p<0.1 were entered together into a multivariable model, controlling for age and treatment assignment.
Table 1 compares selected baseline characteristics among study participants according to their completion of at least one post-intervention assessment. Participants excluded from the analysis because they only completed the baseline assessment were more likely to smoke (27.7% vs 14.9%, p=0.003), use alcohol heavily (21.7% vs 9.9%, p=0.001), report being high or drunk during sex in the past 90 days (39.8% vs 26.5%, p=0.012) and to have experienced sexual abuse (36.1% vs 22.3%, p=0.006). They also reported a greater proportion of protected sex acts in the past 90 days (38.2% vs 36.9%, p=0.023). Of participants who completed one or more post-intervention assessment (n=618), 19.9% (n=123) had a laboratory-confirmed diagnosis of chlamydia and/or gonorrhoea at baseline; 16.5% were positive for chlamydia, 6.5% were positive for gonorrhoea and 3.1% were positive for both chlamydia and gonorrhoea.
Frequency of repeat infections
Forty-nine participants (intervention condition n=30; comparison condition n=19) who were positive for chlamydia and/or gonorrhoea at baseline had a repeat infection during the study period (40.0%). Of those with a repeat infection, 30 (61.2%) were positive at a single follow-up visit, 18 (36.7%) at two visits and one (2.0%) at three follow-up visits.
Baseline predictors of repeat infection
Table 2 presents the unadjusted and adjusted associations between potential baseline predictors and repeat chlamydia and/or gonorrhoea infection among 123 participants who were positive for at least one of these infections at baseline. In both unadjusted and adjusted models, greater impulsivity (AOR: 1.69, 95% CI: 1.08 to 2.65, p=0.021) predicted an increased likelihood of repeat infection, and having a boyfriend (AOR: 0.21, 95% CI 0.07 to 0.62, p=0.005) predicted a decreased likelihood of repeat infection.
Repeat chlamydia and/or gonorrhoea infections were common among African-American adolescent women in this study; 40% who were positive at baseline experienced a repeat infection within 2 years of the intervention. Other studies among adolescents in the USA with a follow-up period of at least 20 months have reported chlamydia reinfection estimates of 18.0–23.3%.4 Gonorrhoea reinfection estimates of up to 40% within 20 months of follow-up have also been reported.4 Fortenberry et al23 found 41.1% of adolescent girls positive for gonorrhoea, chlamydia and/or trichomonas experienced a subsequent infection within 12 months. The high proportion of repeat infections in this study may suggest a reservoir of infection among asymptomatic males. Given that all participants received an evidence-based intervention as the primary treatment, enhanced screening and treatment efforts for young men may be warranted.
Greater impulsivity predicted repeat infection. Research suggests that individuals with poor impulse control may be more likely to engage in risky sexual behaviours or have sex with risky partners;24 ,25 however, to our knowledge, impulsivity has not previously been identified as a factor related to repeat chlamydia or gonorrhoea infections. Interestingly, sexual sensation-seeking, often related to impulsivity and sexual risk behaviours,22 ,24 ,25 was unrelated to repeat infection. Although impulsivity and sensation seeking are related and usually moderately correlated, they represent unique traits. The ability to differentiate between the two and understand how they independently relate to repeat infection is important when considering strategies to reduce repeat infections.
Having a boyfriend predicted a decreased likelihood of repeat infection. The increased risk of repeat chlamydia and/or gonorrhoea infection among individuals without a boyfriend could have been due to an increased likelihood of obtaining a new partner or of continued sex with an untreated partner. Among a sample of women aged at least 15 years, Niccolai et al26 found the population attributable risk of recurrent chlamydia infections was similar for sex with a new sex partner and sex with a partner not known to be treated.
Strengths and limitations
This study is subject to a number of limitations. Some of the factors assessed were subject to change over time and some were targeted by the intervention. In addition, the intervention was effective in reducing incident STIs. Despite these limitations, many types of potential factors were assessed, and two significant baseline predictors of repeat Chlamydia and/or gonorrhoea infection were identified. It is also possible that participants may have been diagnosed and treated between study visits or after the study ended. These findings are relevant only to the 2 years of post-intervention follow-up and could be subject to error to the degrees that participants were diagnosed and treated between study visits and their baseline characteristics varied from those who were not. The study did not achieve 100% retention; therefore, the proportion of participants who were categorised as experiencing a repeat infection may be overestimated or underestimated. Participants excluded from this analysis because they did not return for follow-up assessment were more likely to report heavy alcohol use, being high or drunk during sex and sexual abuse, which increase the likelihood of STI acquisition, and they were only slightly more likely to use condoms; therefore, our findings might underestimate repeat infections. Nevertheless, the particular strengths of this study are its duration of assessment and high retention rate. The number of cases available limited the ability to assess baseline predictors of repeated chlamydia and gonorrhoea infections separately. However, assessing these infections together is consistent with a public health approach in that both infections are prevalent in this population, are transmitted by unprotected sex with an infected partner and cause adverse reproductive health outcomes. Lastly, the possibility exists that treatment failures were misclassified as repeat infections; therefore, the findings may overestimate repeat infections, and the strength of the relationship between the baseline predictors and repeat infection may be underestimated.
These results confirm the need to rescreen young women diagnosed with chlamydia or gonorrhoea. The findings suggest that individual baseline factors influence repeat infection despite receipt of an efficacious intervention. Among young African-American women who test positive for chlamydia and/or gonorrhoea, assessing impulsivity and relationship status may allow for tailored interventions to help prevent repeat infections. For example, more impulsive adolescents may benefit by learning about the role of impulsivity and its potential impact on sexual decision-making as well as how to identify and use strategies to avoid situations where they are likely to behave impulsively. In addition, enhanced STI screening and treatment efforts among young men may be warranted, given the limitations of current interventions for female adolescents to reduce the number of repeat infections adequately.
Among African-American adolescent women, repeat Chlamydia trachomatis and/or Neisseria gonorrhoeae infections were common.
Greater impulsivity and not having a boyfriend at baseline predicted repeat infection during 2 years of follow-up.
Tailored interventions for more impulsive adolescents and those not in a steady relationship may reduce the risk of repeat infections.
Given high numbers of repeat infections after receipt of an evidence-based intervention, enhanced screening and treatment services for young men may also be warranted.
Contributors JMS: Drafting the manuscript and revising it critically for important intellectual content. JLB: Drafting the manuscript and revising it critically for important intellectual content. TPLD: Drafting the manuscript and revising it critically for important intellectual content. RJD: Acquisition of funding, study management, conceptualisation of design and analysis, drafting the article and revising it critically for important intellectual content and final approval of the manuscript for submission for publication. ER: Acquisition of funding, study management, drafting the article and revising it critically for important intellectual content.
Funding This work was supported by the National Institute of Mental Health grant number 5R01 MH070537-08 and Emory Center for AIDS Research grant number P30-A150409. Jennifer L Brown was supported by National Institute of General Medical Sciences grant number K12 GM000680. Jessica M Sales was supported by the National Institute of Mental Health grant number K01 MH085506.
Competing interests None.
Ethics approval Emory University.
Provenance and peer review Commissioned; externally peer reviewed.
↵i A federal assistance program that provides nutrition services to low-income pregnant, breastfeeding and postpartum women and to infants and children up to age 5 who are at nutritional risk