64 Peter Greenhouse Bristol Sexual Health Centre, Bristol, UK Sexually transmitted infections (STIs) make a substantial and often under‐appreciated contribution to gynaecological and obstetric morbidity, because their covert nature and transmission efficiency ensure widespread distribution throughout the entire sexually active population. Pelvic infection (Chapter 45), tubal factor infertility (Chapter 51), ectopic pregnancy (Chapter 43), genital ulceration, vulvo‐vaginal discharge (Chapter 58), genital malignancy (Chapters 60 and 61), premature labour (Chapter 28) and neonatal infection (Chapter 19) have been widely researched and described in relation to STIs. Yet the role of infections, particularly chlamydia, in such common gynaecological problems as abnormal uterine and heavy menstrual bleeding due to endometritis, and right iliac fossa pain in teenage girls due to salpingo‐appendicitis, is controversial and requires further elucidation. Given the above, a sexual contact history – so often neglected in the past – is as important a part of routine gynaecological work‐up as the menstrual or contraceptive history. Risk assessment cannot reliably predict the likelihood of infection nor its location in vagina, cervix, rectum or oropharynx, as no STI causes easily recognizable signs or symptoms in more than one‐quarter of women, while most are obvious in less than one‐tenth, and the partner’s (or partners’) contact history and infection status is usually unknown. A thorough account of STI in women requires an entire textbook [1–3], so this chapter focuses principally on those aspects of bacterial and cutaneous viral STI which have the greatest relevance to daily gynaecology practice, excluding blood‐borne viruses, while emphasizing the anatomical, hormonal and immunological factors which render women more vulnerable to infection by, and present differently with, the same conditions in men. Women are typically more vulnerable than heterosexual men to acquisition of STI as, in any sexual transaction, infections are more likely to spread to the biologically receptive partner. The rate of spread (R0) of any infectious disease depends on a combination of transmission efficiency (β), rate of change of contact or partner (c) and duration of infectiousness (D), expressed in its simplest form as R0 = βcD [4]. Transmission efficiency of STIs to women depends on the infecting organism load and a combination of the amount and force of frottage or penetrative friction trauma and receptor tissue vulnerability, the latter being exacerbated by poor oestrogenization during use of progestogen‐only contraception [5], breastfeeding or menopause, and the former by performance‐enhancing drug use and non‐consensual activity. Vulnerability, particularly to HIV, is greater in the rectum than in vagina or pharynx and can be diminished at all sites or substantially prevented by lubrication and condom use. Individual immune response and organism biology determines duration of infectiousness, which is typically short – a few weeks – for epidemic infections that are either briefly overtly symptomatic, such as urethral gonorrhoea in men or where there is transient high viraemia at seroconversion such as with HIV, or longer (months or years) for endemic STIs that remain covert in both sexes. Thus well over 50% of women who have been sexually active with more than one partner will have acquired and/or transmitted endemic STIs such as herpes simplex virus (HSV), human papillomavirus (HPV) or Chlamydia trachomatis without ever displaying any signs or symptoms of infection, nor sustaining any significant harm. The latter two conditions will, in most cases, have been cleared spontaneously within 1 or 2 years (in many cases, sooner) by the woman’s normal immunological response [6,7]. Individual genetics and frequency of exposure, allied to hormonal and immunological status, typically define the minority of women who will develop symptoms and suffer adverse sequelae. Most analysis of epidemiological trends reflects incidence of infections diagnosed and is dependent on appropriate provision of services, contraceptive choice, clinic attendance patterns, reporting systems and advances in diagnostic technology. The UK’s system of free, confidential sexual health clinics has provided consistent methodology in clinical diagnosis and collection of statistics [8]. Reported cases of gonorrhoea in women over the past 70 years (Fig. 64.1) offers an important historical perspective of how subsequent generations have experienced different rates of infection. Plotted against teenage fertility it acts as a barometer of sexual mores and partner change rates, which is directly relevant to past and future provision of gynaecological care throughout women’s life stages. Fig. 64.1 England and Wales annual cases of gonorrhoea 1945–2015 (England only from 2010) and conception rates 1969–2014 [10]. Extrapolated from Public Health England [8] and Health Protection Agency [9] with corrections. The immediate post‐war ‘baby boom’ and its associated peak of infections was followed by a rapid fall in gonorrhoea (and syphilis) with the introduction of penicillin. The late 1950s to 1960s rise was due to sexual liberalization, demographic change and the introduction of the combined oral contraceptive pill (COC). Women who were teenagers in the mid‐1970s – who are now recently menopausal – experienced the highest ever recorded rates of infection, coinciding with the greatest number of COC prescriptions issued and resulting in the lowest conception rates thereto. The effect of the HIV/AIDS education campaigns of the mid to late 1980s produced a dramatic fall in all STIs which lasted throughout the 1990s, so that women born between 1970 and 1980 had the lowest recorded rates of STIs (and greatest paranoia about infection) of any generation before or since, but had higher teenage conception rates as they used condoms in preference to pills. Women born since 1990 have intermediate infection rates due to less meticulous condom use, but are the first generation to grow up with freely accessible internet pornography, triggering a rise in coercion, acceptance and practice of rectal sex [11], and mobile dating apps offering a more efficient means of finding and changing new partners quickly. Yet they also benefited from widespread use of long‐acting reversible contraceptives, and consequently had much lower pregnancy rates [10]. Rates of STIs such as gonorrhoea and chlamydia are unsurprisingly highest in women in their late teenage years and early twenties, peaking before those of young men, at the time when partner change rates are greatest and requiring repeated testing to aim for control of infections [12]. An individual’s sexual behaviour and infection risk are substantially shaped by their demographic context and geographical setting [13]. Trends in STIs are broadly similar across northern European countries with similar sexual mixing patterns and the typical Western societal norm of serial monogamy with occasional infidelity. As the dating app‐enabled ‘half night stand’ culture takes over, control of infections will become more challenging. Infection rates are already higher across entire populations where the religious cultural norm is a double‐standard morality of male machismo and female virginity, as in southern Europe. STI rates are highest where poverty, drug use, economic necessity, ethnic tradition, or flight from conflict dictate that there will be large numbers of itinerant men having concurrent sexual partnerships [14] and wherever women’s social, sexual and reproductive rights are least valued or respected [15]. Any significant genital symptom, such as irritation, discharge, bleeding pattern alteration, pelvic pain or dyspareunia, should prompt enquiry and investigation as to whether the problem is due to change in hormonal milieu or a recently acquired infection, or a combination of both, given their equal contribution to these aspects of women’s health (Fig. 64.2). The importance of this integrated approach [16] is particularly evident in women with recurrent pelvic pain, many of whom present to GPs and sexual health services and are presumed to have pelvic infection, when they actually have endometriosis [17,18]. These women receive antibiotics when they need hormones or surgery. The reverse occurs in women with abnormal uterine bleeding caused by chlamydial endometritis (see below): they present to gynaecologists or contraception clinics and get hormones or surgery when they need antibiotics. Fig. 64.2 Is it hormones or infection (endometriosis or endometritis) or both? Consideration of the complex interactions between a woman’s hormonal status – whether due to cyclical variation, use of contraception, pregnancy, breastfeeding or life‐stage changes – and their effects on systemic and local genital tract immunity is especially important for the management of recurrent genital infections (see below) and their associated symptoms. Any medical, or psychological, condition with a repeating and precisely timed cyclical fluctuation must have a hormonal cause or trigger, and might thus be amenable to hormonal treatment. Whilst biological plausibility for the latter is strong, the formal evidence base is weak due to lack of targeted research, which tends to focus purely on the infection while ignoring the contribution of luteal phase hormonally‐induced immunosuppression. This approach, and the reality of ubiquitous covert STI, requires that sexual history‐taking should blend seamlessly into any routine gynaecology work‐up by following on logically from menstruation details and contraceptive use. Special attention should be paid to onset of symptoms in relation to timing of hormonal or partner change. Questions do not need to be as detailed or intrusive as those typically taken in sexual health clinics [19] or following sexual assault. All that is initially required, after briefly stressing confidentiality and relevance of the questions to assisting diagnosis, is to enquire as to whether the woman is, or has been, sexually active, whether or not in a sexual relationship and, if so for how long, the interval since any previous sexual contact, be it recent or in the distant past, and whether she has had any post‐coital bleeding or technical problems with sex. By keeping questions brief, relevant and non‐judgemental, sufficient useful detail should be gleaned without undue embarrassment. Clinicians should beware of making an automatic presumption of heterosexuality given the fourfold rise (to 8.7%) in experience of same‐sex relationships among younger women in the most recent UK national survey [11]. Once rapport and trust in confidentiality have been established, this also represents an ideal opportunity to enquire routinely about sexual coercion or intimate partner violence [20]. Given the high prevalence and relevance of this problem, particularly during pregnancy [21], this should be considered an essential competence for any gynaecologist who supports women’s human rights [22]. Relevant diagnostic tests should be taken and are discussed in Summary box 64.2. Treatment with appropriate antivirals, antifungals or antibiotics should ideally await the return of laboratory results and follow published national guidelines [23], which are regularly updated according to alterations in practice and resistance patterns, and are certain to have changed for some infections well before the next edition of this textbook (see Summary box 64.3). Immediate treatment may need to be given, either presumptively depending on the urgency of the situation, or precisely if there is the luxury of access to accurate near‐patient diagnostics [24]. Unfortunately, immediate treatment gives less time for consideration of, and counselling about, the most neglected area of STI care in gynaecology, namely that of partner notification and treatment to prevent reinfection. Broaching the subject of sexual acquisition of infection and the need to treat the partner requires skill, discretion, privacy and accurate take‐home information, and should not be rushed given the potential for misunderstanding and heightened risk of intimate partner violence [25]. Given these complexities, the minimum duty of care of a gynaecologist should be to advise abstention from intercourse until the partner or contact is treated and to seek immediate further assistance and support from health advisors and clinicians in the local sexual health clinic [26] or a nurse counsellor in their own department specially tasked with this remit. Local protocols should be established to facilitate this support, to handle confusion as results are frequently discordant and to assist the process of patient‐delivered partner therapy, wherever this is part of the routine standard of care [27]. Summary box 64.1 contains some points of discussion to help reduce stigma and aid understanding. Much of our understanding of STI in women has changed with the introduction of highly sensitive molecular diagnostic technology, starting with polymerase chain reaction (PCR) in the mid‐1990s and other nucleic acid amplification tests (NAATs) by the mid‐2000s, gradually replacing the traditional methods of microscopy and culture for gonorrhoea and Trichomonas, viral culture for herpes, enzyme immunoassay (EIA) for chlamydia and in situ hybridization for HPV [28]. For these organisms, the increase in sensitivity is of a magnitude between 50 and 100% greater than previously, amplifying the epidemiological fluctuations of the past century (see Fig 64.1) and suggesting that there was far more than the recorded fourfold greater number of gonorrhoea cases among women 40 years ago than nowadays. It also emphasizes the ubiquity of endemic infections, increasing the observed proportion of symptomless carriage to over 90% for most STIs in both men and women, and has prompted a thorough re‐evaluation or clarification of research into disease associations such as those of chlamydia with premature labour (Chapter 28) and HPV with genital cancers (see Chapter 61). The benefits of the ability to detect a positive signal from, in theory, a single organism or even a fragment thereof has revolutionized genital sampling. Previously, endocervical swabs with high organism copy numbers were essential for detecting chlamydia and gonorrhoea. Now self‐taken vulvo‐vaginal swabs are equal to, or even more sensitive than, those taken by the clinician from the cervix [29]. This substantially reduces the need for speculum examination except in cases of post‐coital or other unexplained bleeding, intractable discharge, cytology sampling and colposcopy or intrauterine procedures. Where on‐site vaginal microscopy is available, women with discharge can also take their own slides for Gram staining and/or wet mount, thus fewer women are having genital examinations in sexual health clinics [30]. Indeed, without access to immediate microscopy, it is difficult to understand how gynaecologists’ management of vaginal discharge can be any better than the best‐guess care of a competent GP or community contraception clinic [31]. A quick and simple low‐tech solution is advocated by Donders et al. [32], using wet mount microscopy to diagnose bacterial vaginosis (BV), aerobic vaginitis, candidiasis and trichomoniasis for immediate treatment, although this lacks sensitivity for the latter two organisms. The latest developments have brought laboratory‐quality analysis to the bedside or outpatient clinic with the ability to produce reliable results in less than 2 hours. Near‐patient testing for chlamydia and gonorrhoea has been shown to reduce unnecessary antibiotic treatment [33], with the added benefit of allowing partners to be notified quicker, thereby reducing onward transmission. This also has potential, as yet unresearched, to improve management of the acute abdomen, particularly in young women with right iliac fossa pain, by identifying or excluding STI before surgery is contemplated or antibiotics are started. A further recent advance is the ability to test for several organisms simultaneously. As STIs frequently coexist, a vaginal swab sample routinely tested for some 8–12 organisms circumvents the a priori problem of knowing or guessing what to test for while apparently sacrificing very little in the way of sensitivity, and thus fewer unsuspected infections should be missed [34]. Unfortunately, some commercially available multiplex assays include organisms (e.g. Mycoplasma hominis, Ureaplasma parvum) which most would consider commensal or which have yet to be classified as genuinely pathogenic in women, resulting in unnecessary antibiotic treatment in some women seen privately or outside the UK. The downside of test hypersensitivity is the potential for sample contamination from surfaces in clinics [35] and the problem of persistence of dead organisms for days or weeks after successful treatment, making timing of tests of cure problematical, particularly for chlamydia and gonorrhoea (see section on individual organisms). Minimum routine STI screening in women consists of a vulvo‐vaginal swab for a combined chlamydia and gonorrhoea NAAT, and bloods for syphilis and HIV [36] (see Summary box 64.2). Clinicians are accustomed to antenatal screening for the latter two infections. This should now extend to routine gynaecological care, given the consequences of missing these treatable hidden conditions, and is essential in cases of genital ulceration or wart‐like lesions, and in areas where HIV prevalence is greater than 1 in 500 of the population [37]. STIs affect all aspects of pregnancy, from subfertility, implantation failure, ectopic pregnancy, early and mid‐trimester miscarriage, premature delivery and stillbirth to fetal anomaly and neonatal infection. Yet the relatively low prevalence of epidemic STIs in developed countries means that the most severe complications are rarely encountered in obstetric practice, and the consequences of endemic STIs may pass unnoticed. The most substantial effect of STIs on obstetric outcome is found in developing countries, particularly sub‐Saharan Africa, where untreated infection rates are highest and cofactors such as poverty, poor nutrition and other endemic infectious diseases such as malaria and tuberculosis exacerbate the problem. In this setting, even after controlling for other variables, the risk of preterm birth is doubled by gonorrhoea and Trichomonas, quadrupled by chlamydia, and increased sevenfold if BV is diagnosed before 16 weeks’ gestation [38]. Syphilis, with an overall antenatal prevalence of between 4 and 15%, affects some 2 million African pregnancies annually, of which 1.6 million remain untreated, resulting in half a million infant deaths [39]. It is associated with one‐quarter of preterm births and half of all stillbirths [38]. Recent advances in near‐patient testing using simple rapid combined syphilis and HIV kits produced much higher rates of positive diagnoses, and dramatically improved the proportion of syphilis cases treated from 51.1 to 95.2% [40]. A measure of the value of treating STIs can be estimated from pragmatic trials in Rakai, Uganda of a one‐off antibiotic regimen (cefixime 400 mg, azithromycin 1 g with metronidazole 2 g) given at around 28 weeks’ gestation. In the late 1990s, this combination therapy would have been simultaneously effective against gonorrhoea, chlamydia and trichomoniasis, and should have rendered syphilis non‐infectious. Despite non‐treatment of partners and the risk of reinfection, the relative risks of low birthweight and neonatal death were both significantly reduced, at 0.68 and 0.83 respectively [41]. Based on the above, any effort to improve screening and treatment of STIs in pregnancy in any geographical setting should be rewarded with reductions in preterm birth and neonatal death, and increases in mean gestational age and mean birthweight, although these may be difficult to measure or prove depending on absolute disease prevalence. The effects of STIs on pregnancies in developed countries is considered in the sections on individual diseases, and is further discussed in Chapter 13. ‘The effect of chlamydial infection in women of reproductive age is overwhelmingly underestimated’ [42]. Chlamydia trachomatis is the most important STI affecting women’s health because of its ubiquity and causal association with adverse reproductive sequelae. As some 5–10% of young women are found to be infected in screening campaigns [8,12] it is probable that at least one‐third to half will have been exposed to chlamydia in a lifetime. It is an obligate intracellular bacterium with a unique quasi‐viral 48‐hour life cycle, transforming between its extracellular infectious state as an elementary body (EB) and the intracellular reticulate body (RB), which uses cellular material to reproduce. Tissue damage is caused partly by this process, but principally by inducing an exaggerated cell‐mediated immune response in a minority of genetically predisposed individuals [43]. Some RBs can convert to persistent non‐replicative forms that retain long‐term viability and are thereby less susceptible to clearance with antibiotics, leading to so‐called heterotypic resistance, where treatment may fail with high organism load [44]. Chlamydia is almost entirely transmitted by direct sexual contact because the organisms do not grow on or infect external skin. It can only multiply in specific types of tissues, including the columnar epithelium of the endocervix, urethra, rectum, endometrium, peritoneum, conjunctiva and pharynx, and ciliated epithelium of the fallopian tubes, nasopharyngeal sinuses and bronchi. Pharyngeal carriage of chlamydia in women genitally infected is around 12% with almost all asymptomatic [45], but oral sex is considered a relatively inefficient route of chlamydial transmission. As with most other STIs in women, highest rates are found in the late teens and early twenties due to greatest rates of partner change, with a substantial decline from the peak at age 18–20 identified in sexual health clinic testing and by the UK national chlamydia screening programme [8,12]. Unsurprisingly, total recorded case numbers increased and overall positivity rates fell as the screening campaign rolled out to a wider proportion of the population. A better perspective on recent chlamydial epidemiology can be gleaned from Swedish data collected since the early 1980s and derived not from population screening but from widespread, low clinical threshold, diagnostic testing [46]. The sustained fall in incidence from 1986 – which was a major impetus for the UK screening initiative – occurred not because of the testing programme, but due to HIV education campaigns reducing partner change rates and increasing condom use. Since HIV became treatable in 1996, chlamydial infection rates have risen consistently. There is increasing concern as to whether widespread screening and treatment has any realistic effect on disease prevalence, given the rate of population mixing [47]. Repeatedly exposing large numbers of individuals to antibiotics for an infection which harms only a minority contributes to the generation of resistant strains of other organisms, such as gonorrhoea and mycoplasma (see below). For many years it was presumed that chlamydial infection would remain present and identifiable in an infected individual until such time as they were given appropriate antibiotic treatment. It has now been shown that the majority of adolescent and adult women will eventually clear the infection from the lower genital tract via their own immune system with no apparent long‐term ill effects but that a minority will retain active or quiescent chlamydial infection within their endometrium, ovarian surface and/or fallopian tubes regardless of whether the organisms are identifiable by swabs taken from the cervix [48]. The rate of spontaneous clearance in a cohort of initially infected and untreated adult women of mixed ages has been demonstrated as 50%, 80% and 95% clear after 1, 2 or 3 years respectively [49]. It has also been shown that the speed of clearance increases proportionally with increasing age so that women in their thirties are likely to have cleared the infection within a few weeks or months [50]. This explains why women in their late teens appear to have more chlamydia than those in their mid‐twenties, despite the latter’s increased cumulative exposure to infection. It also explains how couples are often found to be discordant for infection, with a different speed of clearance between partners causing misunderstanding and possible recrimination [26,27]. The clinical spectrum of chlamydial infection in women runs from the asymptomatic millions, who clear their infections spontaneously without suffering tubal damage [51], to the moribund few hospitalized with acute pelvic infection (see Chapter 45). Between these extremes are a range of conditions such as endometritis, menorrhagia, ectopic pregnancy and tubal factor infertility variously referred to as ‘subclinical salpingitis’ [52], ‘atypical PID’ [53] or ‘subclinical PID’ [54] in which the role of chlamydial infection is covert, usually unrecognized and therefore untreated. Chlamydia is one of several infections contributing to cervical friability and overt mucopurulent cervicitis [55] (Fig. 64.3), which can result in increased vaginal discharge and post‐coital bleeding (PCB). Yet both the physical sign of cervicitis and the microscopic finding of leucocytes are non‐specific, having a very low positive predictive value for chlamydia, but being a strong negative predictor for absence of endometritis [56]. Management of PCB in the UK is inconsistent [57], and requires definitive exclusion or treatment of chlamydia, gonorrhoea and, probably, Mycoplasma genitalium before any ablative surgical treatment is contemplated for benign conditions such as ectropion. Fig. 64.3 Chlamydial mucopurulent cervicitis. Overt chlamydial mucopurulent discharge with endocervical gland oedema showing follicular cervicitis. A similar appearance is seen with gonorrhoea, but most chlamydia is covert showing little or no such obvious signs. Source: photograph by Peter Greenhouse FRCOG, 1992. (See also colour plate 64.3) Any woman with upper genital tract infection (e.g. salpingitis, pelvic peritonitis, perihepatitis, periappendicitis) must, by definition, have a concomitant endometritis. Yet the extent to which chlamydial or other STI‐related endometritis contributes to abnormal uterine bleeding (AUB) or heavy menstrual bleeding (HMB) is as yet unknown because of the paucity of specifically targeted research. Some 29% of women with recent‐onset irregular intermenstrual bleeding (IMB) while taking COC pills were found to have chlamydia using a highly sensitive direct immunofluorescence test, compared with 11% who had vaginal discharge or new partners, and only 6% among new contraception clinic attenders [58]. Thus, second only to poor compliance with COC, chlamydia is the next commonest cause of IMB in young women. Single‐dose azithromycin (1 g) was shown to produce a 50% reduction in abnormal bleeding in a proof‐of‐concept study in predominantly African‐American women attending a public sexual health clinic [59]. While histological evidence of plasma‐cell endometritis reduced from 38% to 4% after treatment (P <0.001), this was not precisely linked to each case of successful resolution of bleeding. Analysis of historical endometrial biopsies from women with unexplained AUB, who had been neither tested nor treated for chlamydia at the time of biopsy, found that some 58% had PCR evidence of chlamydia and that infection was strongly correlated (P <0.001) with the finding of macrophages, lymphocytes and plasma cells, leading the authors to make this section’s opening quote on the unrecognized importance of chlamydial infection in women [42]. If this work were to be repeated prospectively and results confirmed, it would suggest that C. trachomatis infection could be sufficient to account for the majority of ‘idiopathic’ or unknown causes of AUB‐E or AUB‐I in the PALM‐COEIN classification (see Chapter 48). Thus in the initial investigation of AUB and HMB, routine checking of pipelle or hysteroscopy samples for both plasma‐cell endometritis and chlamydia, with appropriate pre‐insertion treatment, might reduce the rate of persistently heavy or irregular bleeding patterns after levonorgestrel‐releasing intrauterine system (LNG‐IUS) insertion. According to Moritz in 1912, ‘The causative relationships between adnexal disease and appendicitis still form a dark area in gynaecology’ [60]. The incidence of this commonest surgical emergency is significantly greater in women than in men only between the ages of 16 and 21, with a doubling in the excess rate of operation and trebling of ‘negative’ histology [61] coinciding with the sharp peak [8] in incidence of the commonest STI at ages 18–20. The coincident finding of chlamydia and serosal plasma cell periappendicitis [62] suggests that chlamydial ‘salpingo‐appendicitis’ might explain this diagnostic dilemma and its ubiquity [63], necessitating chlamydia screening of all acute cases of right iliac fossa pain in young women regardless of where they present [64]. Accurate point‐of‐care STI tests could enhance acute management in emergency departments. If specific antibiotic regimens for chlamydial pelvic inflammatory disease (PID), such as one including high‐dose ofloxacin (400 mg b.d. for 14 days), were used in future trials in this age group, this could improve the hitherto equivocal outcomes achieved [65] when comparing antibiotics with surgery, but this rationale has yet to be explored. The association between STI and perihepatitis was first described in 1920 [66], and was originally thought to be caused by gonorrhoea. Curtis frequently found bridal veil and/or violin string adhesions between liver and diaphragm (Fig. 64.4) when performing pelvic clearance for end‐stage gonococcal salpingitis [67]. The acute presentation of right hypochondrial pain, limiting deep inspiration, with positive Murphy’s sign indistinguishable from that of acute cholecystitis, in association with gonococcal peritonitis was described by Fitz‐Hugh [68]. Since the first isolation of chlamydia from the liver surface [69], very few cases of the Fitz‐Hugh–Curtis syndrome have been described without evidence of chlamydial infection [70] or high anti‐chlamydial antibody titres [71], because of the frequent coexistence of gonorrhoea with chlamydia. About 20% of PID cases have some degree of right upper quadrant pain that, in most cases, resolves rapidly with appropriate anti‐chlamydial treatment, and requires adhesiolysis only in the rarest cases of persistent diaphragmatic restriction. Symptoms may present acutely without any apparent pelvic pain or other physical signs, but the association is strong enough such that any right hypochondrial pain in a young sexually active woman must be assumed to be chlamydial in origin until proven otherwise, because of the relative rarity of cholecystitis in this age group. Very rarely, women may present with pain over the left lobe of the liver or with perisplenitis. Fig. 64.4 Chlamydial perihepatitis/Fitz‐Hugh–Curtis syndrome. Bridal veil and violin‐string adhesions between liver and diaphragm due to chlamydial infection, causing right hypochondrial pain and restriction of respiration. Note normal gallbladder. Source: photography by Peter Greenhouse FRCOG, 1986. (See also colour plate 64.4) There is a wide variety of opinion as to the proportion of women who will eventually develop tubal damage following chlamydial infection, and on the effectiveness of screening and treatment in disease prevention. Estimates range from the widely accepted 1–2% [72] to 17% in a recent thorough assessment which suggested that for every 1000 chlamydial infections in women aged 16–44 years, there would be approximately 171 episodes of PID, 73 of salpingitis, two ectopic pregnancies and five women with tubal factor infertility (TFI) by age 44 years [73]. Despite these predictions, prospective randomized studies of chlamydia screening to prevent pelvic infection report widely differing results, dependent on patient selection and data interpretation. The Seattle PID prevention study found a relative risk of PID of 0.44 (95% CI 0.2–0.9) in screened and treated women of whom 7% had chlamydia compared with an unscreened population [74]. Yet the London Prevention of Pelvic Infection (POPI) trial found a small but non‐significantly reduced number of cases in those screened and immediately treated, but the majority (79%) of incident PID cases occurred in women originally found to be uninfected [75]. As discussed in Chapter 45, the principal determinants of reproductive tract damage due to chlamydia are repeated exposure to and severity of infection [76], delayed treatment of symptomatic disease [77] and individual anti‐chlamydial cell‐mediated immune responses, which are predetermined by single nucleotide polymorphisms [43] conferring enhanced susceptibility on a minority of women. These responses should theoretically become apparent at first and thus youngest exposure to infection. In contrast, some women are genetically protected against tubal damage. Those with the CCR5 Δ32 gene deletion, which also prevents infection by HIV, are four times less likely to have tubal pathology among subfertile women with anti‐chlamydial IgG antibodies [78]. This raises the future prospect of being able to predict a woman’s risk of infertility from genotype and chlamydial serology response [79], an approach which could be particularly useful in the investigation of ‘unexplained’ infertility [80]. The strongest link between TFI and chlamydial cell‐mediated immune response is seen in a linear increase in anti‐chlamydial antibody titre (CAT) with severity of tubal damage in subfertile women, when measured by the whole‐cell inclusion immunoflourescence test [81]. As CAT is logarithmic, this fits a dose–response effect. Almost all of these women were negative on cervical NAAT screening and it is not yet known if women with high CAT would benefit from prolonged antibiotic therapy [80], though this could be an option in those with proven endometritis. Anti‐chlamydial heat shock protein (HSP)‐60 antibodies have been more widely studied, having highest titres in women with acute perihepatitis and salpingitis [71], and also being highly predictive of TFI when combined with markers of anti‐chlamydial humoral immunity [82]. Higher chlamydial HSP‐60 titres are also associated with significantly greater risk of failure of assisted reproduction techniques (ART). Embryo implantation is more likely to fail [83], and early spontaneous miscarriages are more common [84]. As human HSP‐60 has a similar structure to that of chlamydial HSP‐60, the postulated immunopathological mechanism is either one of chronic covert endometritis preventing implantation, or a possible direct effect on early embryo development [84]. Increased expression of endometrial prostaglandin receptors has separately been proposed as a risk factor in chlamydia‐infected women with recurrent miscarriage [85]. On a hopeful note, a significantly higher live‐birth rate has been demonstrated in women whose chronic endometritis improved after a combination of antibiotic treatments, most of which included anti‐chlamydials [86]. Chlamydia probably contributes to the majority of ectopic pregnancies despite rarely being found on vaginal or endocervical swabbing at the time of presentation, and the range of estimated risk is a matter of dispute (see Chapter 43). Chlamydia causes suboptimal oocyte transport by partial deciliation and reduced ciliary beat frequency [87] in women who have not already suffered complete fimbrial enclosure. Over 85% of women with ectopic pregnancy have markedly raised CAT compared with women having live births [80,88]. In an ecological study, annual reduction in chlamydial incidence was directly related to a linear fall in ectopic pregnancy: among 20–24 year olds, extension of the plot line transected the zero point of the abscissa, suggesting that if there were no chlamydia in this age group, there would be little or no ectopic pregnancy [89]. The PCR finding of reverse‐transcription chlamydial mRNA in fallopian tubes of women with ectopic pregnancy suggests that viable, metabolically active chlamydial infection remains present and contributing to tubal damage long after initial infection [90]. Taken together, these points suggest that young women with ectopic pregnancies may warrant extended anti‐chlamydial treatment even if lower genital tract swabs are negative. The first described example and strongest evidence of individual genetic response to chlamydia is that of HLA‐B27 tissue type [91] conferring susceptibility to the triad of sexually acquired reactive arthritis (SARA), uveitis/conjunctivitis and urethritis (formerly known as Reiter’s syndrome), which appears commoner in men than women [92]. SARA is an asymmetric multi‐joint ‘sterile’ inflammation of synovium, fascia and tendons typically affecting hands, lower limbs and sacro‐iliac joints, where chlamydia can be found in synovial membrane biopsy [93]. Joint involvement can occur in the absence of other features and isolated sacro‐iliitis is probably frequently missed in women [94]. Treatment is by standard or extended anti‐chlamydial regimens [94,95] in conjunction with anti‐inflammatory drugs, and the condition is best managed jointly between sexual health clinicians and rheumatologists because of the high recurrence rate [94]. Azithromycin becomes highly concentrated in cartilage, but it is not yet proven if this regimen confers special benefit [95]. Through the 1980s and 1990s, studies of chlamydia in pregnancy using relatively insensitive EIA testing consistently reported higher rates of spontaneous preterm labour and preterm birth (PTB) in chlamydia‐infected women, but none reached statistical significance. No meta‐analysis was performed, and it was assumed that chlamydial infection was not a major cause of PTB. Studies using early chlamydial DNA detection tests suggested a twofold to threefold increased risk of PTB if chlamydia was detected at 24 weeks [96]. In the treatment of women with BV in pregnancy (see Chapter 28), only systemic regimens containing recognized anti‐chlamydial antibiotics (erythromycin, clindamycin) were successful in reducing PTB [97], explaining the failure of metronidazole‐only regimens. Accidental treatment of chlamydial infection is probably the hitherto unexplained reason for the marked success of the only randomized antenatal trial of doxycycline (against presumed bacteriuria) yielding an odds ratio of 0.25 for PTB versus placebo [98]. Taking these facts together suggests that greater importance should be attached to antenatal screening for chlamydia [99], despite the problems of evaluating the relevance of treating an infection which can resolve spontaneously between antenatal visits in older women [49]. This is not recommended in the UK National Institute for Health and Care Excellence (NICE) guidelines [100], which merely suggest directing women under 25 years old to the National Chlamydia Screening Programme. This analysis failed to take account of the largest antenatal natural history study of untreated chlamydia, which showed adjusted odds ratios of 4.35 and 2.66 for PTB in infected women under 32 and under 35 weeks, with chlamydia accounting for 14.9% and 7.4% of premature deliveries at these gestations respectively [101]. Mother‐to‐child transmission of chlamydia occurs in up to 60–70% of infants born vaginally [102]. Vaginal and rectal infection in newborn girls is more likely to occur in breech deliveries, with a quoted 20% transmission efficiency [103]. Chlamydial ophthalmia neonatorum is much more common than that caused by gonorrhoea, but the incubation period is longer at around 1–2 weeks, with all infections presenting within 30 days [104]. Chlamydial pneumonitis can occur any time from 1 to 3 months after birth, with the majority presenting between 8 and 11 weeks [105]. Although recent UK guidelines [106] still recommend equal first‐line treatment of uncomplicated chlamydia with doxycycline 100 mg b.d. for 7 days or azithromycin 1 g stat, there is increasing concern about the inadequacy of the latter regimen when assessed by test of cure (TOC) using NAAT [107,108]. Doxycycline is some 20% more effective than single‐dose azithromycin for rectal infection [109], and as the rectum is a preferential site for chlamydial carriage even in women who have never had rectal sex [110], an alternative is needed. Azithromycin has a very long intracellular half‐life, remaining above the MIC90 for chlamydia for 10 days, and is especially concentrated in uterus, fallopian tubes and bladder [111]. Yet its serum half‐life is less than 24 hours, which explains its potential failure in high organism load infections [44], requiring an extended regimen prolonged over at least three and preferably five days [112]. Extended azithromycin using a total of 1.5 g minimum up to 3 g maximum (e.g. 1 g stat followed by 500 mg for 4 days) should almost certainly be used in preference to single‐dose therapy when there is any suggestion of endometritis. This oral regimen was extrapolated from a successful PID monotherapy trial [113]. A further alternative for uncomplicated chlamydia is ofloxacin 200 mg b.d. (or 400 mg o.d.) for 7 days, but this must not be given to professional athletes because of the rare side effect of tendon rupture [106]. This new understanding of treatment inadequacy sheds light on, and may call into question, recommendations against antibiotic prophylaxis prior to uterine instrumentation to prevent PID, particularly before IUD insertion [114]. This analysis found no overall reduction in PID, yet a small but significant benefit with less women re‐attending for pain or bleeding, in studies using a suboptimal azithromycin regimen (500 mg stat) or a homeopathic dose of doxycycline (200 mg stat). The current recommendation against routine prophylaxis [115] should be backed up by trials of an appropriately dosed anti‐chlamydial regimen (i.e. extended azithromycin), particularly for young women requesting emergency IUD contraception. This is another example where near‐patient high‐quality diagnostics could lead to better‐targeted antibiotic husbandry. There are less therapeutic options for treatment of chlamydia in pregnancy or breastfeeding, as both doxycycline and ofloxacin are contraindicated, but are not apparently harmful in the interval between conception and first missed period. Azithromycin 1 g is recommended as first line [106] because of its safety [116] and superior tolerability to the alternative of erythromycin [117]. Amoxicillin 500 mg t.d.s. for 7 days is an alternative [118], but is not used outside pregnancy because of theoretical concerns of inducing latency. The importance of ensuring clearance of infection prior to delivery mandates a TOC, which must be performed no less than 4 weeks [119] to 6 weeks [106] after treatment has finished, because the persistence of dead chlamydial antigen can give a false‐positive NAAT if taken sooner [107]. Assistance from sexual health colleagues is particularly useful in this scenario to handle sensitive relationship issues and ensure correct management of partners [106]. Further guideline‐based regimens for upper genital tract chlamydia are considered in Chapter 45. Gonorrhoea is the second commonest bacterial STI, with less than one‐tenth the prevalence of chlamydia in the UK [8], producing a similar range of gynaecological morbidity. It is more often symptomatic because of its more direct mode of pathogenesis, but still remains unrecognized in some 70–80% of women. Gonorrhoea epidemiology, discussed in detail in Fig. 64.1, is more sensitive to changes in sexual behaviour than chlamydia. Even when cases were at their lowest in the UK in the early to mid‐1990s, Black British or Black Caribbean compared with White ethnicity was associated with a 12‐fold or eightfold greater risk of gonorrhoea or chlamydia respectively [120]. This was ascribed to much higher partner change rates and concurrency of relationships. This pattern is seen worldwide in other populations depending on these parameters and prevailing socioeconomic, cultural and other coercive factors. Neisseria gonorrhoeae is a Gram‐negative diplococcus which adheres via surface pili to columnar epithelial cells of moist mucous membranes of the urethra, endocervix, rectum, pharynx and conjunctiva [121]. Transmission occurs by direct inoculation of warm secretions onto a surface with near neutral pH. Although it does not infect the vagina in adolescent and premenopausal adult women, gonorrhoea causes vulvovaginitis in prepubertal girls as the hypo‐oestrogenic vaginal epithelium is thin with a neutral pH and is therefore easily inflamed, with most infections being symptomatic [122]. Gonococci produce a peptidoglycan‐stimulating cytotoxic tumour necrosis factor (TNF)‐α [123] and a lipopolysaccharide enzyme that facilitates direct spread of the organism particularly into glandular tissues, producing damage and potential abscess formation and contributing significantly to the more severe presentation of acute gonococcal salpingitis (see Chapter 45). The principal lower genital tract symptoms, if present, arise from an endocervicitis manifesting as excess vaginal discharge – which can often remain unnoticed because of normal cyclical changes – and increased cervical friability which predisposes to PCB [124]. Gonorrhoea is nowadays much less likely to be missed as a cause of discharge and PCB because of the routine use of dual chlamydia/gonorrhoea NAAT, such as the Aptima Combo 2 [125]. Overt gonococcal urethritis is rarely seen in women compared with men, and urethral infection is now rarely tested for other than in hysterectomized women. When previously researched using culture it was commonly found in some 80% of women with endocervical infection [1], many of whom had only mild and largely ignored dysuria. Both rectal and pharyngeal gonorrhoea are asymptomatic in around 95% of men and women. The pharynx represents a major reservoir for oral sex transmission of gonorrhoea, and fellatio is the principal route of gonorrhoea spread in those who use condoms meticulously for other forms of sex [126]. Control of pharyngeal gonorrhoea is particularly important as most antibiotic resistance develops at this site [127]. Rarely, gonorrhoea infects paraurethral tissues including Skene’s glands [3], and was classically recognized as the commonest cause of Bartholin’s abscesses, although chlamydia also contributes to a lesser quantity of acute and chronic Bartholin’s inflammation [128]. Disseminated gonococcal infection (DGI) may occur in up to 3% of cases [129]. It usually presents without obvious genital symptoms in women, commencing with a flitting polyarthralgia, tendon pain and pyrexia. Painful pustular lesions occur in hands and feet, with infectious micro‐emboli lodging in the peripheral circulation. The main feature is one or two hot swollen joints (septic arthritis), most commonly the knee and wrist, with organisms identifiable from aspiration of synovial fluid [130]. DGI is more common in women than men and often presents during pregnancy, probably because the combination of lowered immunity, increased vascular permeability and hyperdynamic circulation may facilitate haematogenous spread. Rare complications of DGI are gonococcal meningitis and endocarditis [129]. Treatment requires higher and longer doses of the first‐line agents [121]. Gonorrhoea is too sporadic for trial research in a European setting regarding PTB, but much of this infection will remain unrecognized in pregnancy unless vaginal discharge is copious and dual chlamydia/gonorrhoea testing is used [131]. A new diagnosis in pregnancy raises acutely sensitive issues of partner treatment [132], mutual mistrust and support needed in case of intimate partner violence. As with chlamydia, correct treatment and TOC is essential [121] to reduce the risk of PTB and prevent ophthalmia neonatorum. This was first managed with silver nitrate prophylaxis by Credé over a century ago [133], but remained the principal cause of blindness, apart from syphilis, in the UK up to the 1920s. Symptoms of discharge and eyelid oedema typically develop within 2–5 days, but occasionally present up to 2 weeks after delivery [134], and must be treated immediately as corneal scarring can occur within 24 hours [135]. Current UK [121] and European [136] guidelines recommend ceftriaxone as the only first‐line antibiotic remaining to treat gonorrhoea effectively (including in pregnancy), with a current 99.2% cure rate [127]. Ceftriaxone 500 mg i.m. stat needs to be given with azithromycin 1 g [121] or 2 g [136] stat, regardless of chlamydia test results, to reduce the chance of resistance developing to the primary drug. After initial diagnosis by NAAT, a swab must be taken for culture before treatment is given. Gonorrhoea is best managed in conjunction with a sexual health service to provide this essential antibiotic sensitivity checking, partner notification and follow‐up TOC. Timing of TOC must be delayed by 2 weeks to allow for clearance of residual gonococcal DNA [137]. In the case of allergy or other contraindication to ceftriaxone, choice of antibiotic is determined by the culture result on advice from the microbiologist based on local resistance patterns. Spectinomycin 2 g i.m. stat is one of the second‐line recommendations, but this is ineffective against pharyngeal gonorrhoea [138]. The most pressing problem of gonorrhoea control is the likelihood that the infection will become untreatable within the next decade [139]. The organism can evade host immune responses by having a widely variable surface structure and rapidly evolves resistance mechanisms to each new antibiotic within a few years of its introduction. Antibiotic concentrations are higher in the genitals and pelvis than in the pharynx, where gonorrhoea mixes with commensal organisms, particularly N. meningitidis, from which it acquires resistance genetic information. Recommended antibiotic dosage regimens will increase as resistance ‘drift’ increases [140]. Complete resistance to ceftriaxone is currently rare, but first occurred in pharyngeal infection of a Japanese female sex worker [141]. The pharynx was the site of all resistant strains in the last European‐wide report [142]. High‐level azithromycin resistance in the UK was first reported among heterosexuals [143], probably occurring due to widespread use of the drug in the chlamydia screening programme, with cases of undiagnosed gonorrhoea being missed and suboptimally treated. The first dual failure of ceftriaxone and azithromycin was recently reported from the UK [144]. While awaiting newly developed antibiotics, future management of multidrug‐resistant gonorrhoea will require reuse of combinations of older antibiotics, guided by culture results and newer molecular techniques [140], necessitating ever closer working between gynaecologists, sexual health clinicians and microbiologists to achieve therapeutic success. For obvious ethical reasons, no modern data are available on the long‐term course or duration of untreated gonorrhoea in women but, in the worst‐case scenario, the possibility of untreatable infection may prompt a return to heroic surgical intervention for chronic disease. Reading from pre‐antibiotic era accounts of optimal choice of timing for pelvic clearance of ‘burnt out’ gonococcal salpingitis, it is probable that almost all infection will have spontaneously cleared within 2 years, in many cases much sooner [67,145]. Mycoplasma genitalium (MGen) belongs to a genus of the smallest known free‐living organisms, having DNA and RNA but no cell wall. It is considered to be an emerging [146] and overlooked STI [147] as it has been shown to cause some 30% of non‐gonococcal urethritis in men [148], yet a thorough review could not conclusively implicate M. genitalium as a cause of female urethritis [146]. An equivalent contribution to other lower or upper genital tract infection in women, over and above that of gonorrhoea and/or chlamydia, has been less widely researched because of its relatively recent discovery and the limitations of diagnostic expertise confined to a small number of centres. Mycoplasma genitalium infects genital and respiratory mucosal surfaces, and other members of the genus (e.g. M. hominis, Ureaplasma parvum) have been found to be commensal in these sites. Although progesterone is a prerequisite for establishment of infection in laboratory animals, nothing is known of the interaction of hormones and M. genitalium in humans. Mycoplasma genitalium is associated with mucopurulent cervicitis, with one group showing a 3.3‐fold greater risk [149]. Yet the difficulties of standardizing a definition of cervicitis hamper accurate comparisons between studies [146], and characterizing its role in vaginal discharge has remained elusive for similar methodological reasons. Since the original report implicating raised M. genitalium antibody titres in acute PID [150], substantial work has been published linking the immune response to M. genitalium in women with TFI [151]. Despite this, another major study showed a lack of association of M. genitalium antibodies with PID and ectopic pregnancy [152]. Mycoplasma genitalium has been found in significantly greater proportions in women with endometritis, but its low overall prevalence in women with severe PID suggests a minor role only [153]. A recent meta‐analysis suggests that the association is statistically significant [154], but might be only marginally clinically relevant due to the low pooled odds ratio (POR) of 2.14 (95% CI 1.31–3.49). This suggests that only a small minority of infected women are susceptible to significant damage while most are unaffected. These issues are also considered in Chapter 45. In the antenatal setting, a large London study found a low (0.7%) prevalence of M. genitalium, suggesting it was ‘unlikely to be an important risk factor in adverse pregnancy outcome in healthy women in the community’ [155]. While an authoritative review found considerable evidence for the role of M. hominis in postpartum and post‐abortal sepsis, there was difficulty differentiating the role of M. genitalium from that of BV in PTB [156]. The aforementioned meta‐analysis [154] reached an alternative conclusion, suggesting a POR of 1.89 (95% CI 1.25–2.85) for M. genitalium in PTB, but this is considerably less than the OR of 4.35 (95% CI 1.3–15.2) for chlamydia in the Rotterdam study [101]. The greatest problem in management of M. genitalium is the inadequacy of treatment to clear infection, and the finding that some 40–50% of the 15% who fail treatment with azithromycin 1 g will have developed chromosomally mediated resistance at the 23S site, necessitating alternative treatment with moxifloxacin [157]. The European guidelines recommend extended azithromycin regimens for 3 or 5 day as first‐line choice [157]. Yet as most people are never tested for M. genitalium, macrolide resistance is already endemic at around 40% in countries such as the UK and Norway because of widespread exposure to single‐dose azithromycin, but not in Sweden where doxycycline is used in preference to treat chlamydia [158]. The latest PID treatment guidelines have introduced specific anti‐mycoplasmal regimens for the first time (see Chapter 45), but routine screening for M. genitalium is not recommended by the CDC [159], and few UK sexual health centres currently offer testing. However, over the next decade, the true relevance of M. genitalium in obstetrics and gynaecology, and the implications of its increasing resistance to macrolides in populations with widespread exposure to single‐dose azithromycin treatment of chlamydial infection, will become clearer as new diagnostic techniques are standardized and used routinely in sexual health clinic practice and in cross‐disciplinary research.
Sexually Transmitted Infections
Epidemiology
Integrated approach to STIs in gynaecology
Sexual history‐taking
Partner management
Diagnostic advances
Near‐patient and multiplex testing
Routine STI testing
STI in pregnancy
Lessons from Africa
Chlamydia
Longevity of infection and spontaneous clearance
Spectrum and sequelae of chlamydial infection
Chlamydia and abnormal bleeding: unanswered questions
Chlamydia, periappendicitis and perihepatitis
Chlamydial PID and tubal factor infertility
Failure of assisted reproduction techniques
Chlamydia and ectopic pregnancy
Chlamydia and arthritis
Chlamydia in pregnancy
Chlamydia treatment and prophylaxis (see also Summary box )
Gonorrhoea
Disseminated gonococcal infection: arthritis and septicaemia
Gonorrhoea in pregnancy
Treatment of gonorrhoea (see also Summary box 64.3)
Untreated or untreatable gonorrhoea
Mycoplasma genitalium
Mycoplasma genitalium in pregnancy
Treatment of Mycoplasma genitalium (see also Summary box 64.3)
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
