Chlamydia infection in general practiceBMJ 1999; 318 doi: https://doi.org/10.1136/bmj.318.7186.790 (Published 20 March 1999) Cite this as: BMJ 1999;318:790
- N R Hicks (), consultanta,
- M Dawes, general practitionerb,
- M Fleminger, general practitionerb,
- D Goldman, general practitionerb,
- J Hamling, general practitionerb,
- L J Hicks, general practitionerb
- aDirectorate of Public Health and Health Policy, Oxfordshire Health Authority, Oxford OX3 7LG
- bHollow Way Medical Centre, Oxford OX4 2NJ
- Correspondence to: Dr N R Hicks
- Accepted 2 September 1998
How common is chlamydia infection, and who should be investigated and treated for it? Is the net benefit of investigation worth the cost? At a recent discussion in our general practice it soon became apparent that our views and practices varied widely. Was there any evidence to help us reach a consensus? We resolved to try and find out.
Chlamydia infection is the commonest treatable sexually transmitted disease in the United Kingdom; it is most common in sexually active women aged under 20
Serological studies suggest that chlamydial infection may account for a large proportion of cases of tubal infertility and ectopic pregnancy
60-80% of genital chlamydia infections in women may be asymptomatic
In one randomised trial, screening high risk women and treating those found to be infected reduced the incidence of pelvic inflammatory disease by about half in 12 months
Access to the internet allows valid, relevant information to be identified and retrieved quickly—but appraising the evidence and deciding how best to reflect it in practice takes considerably longer
Ms A, a 20 year old secretary, was worried because she had had vaginal discharge and irritation for three days. The discharge was slight, clear, watery, and non-offensive, and she had no abnormal vaginal bleeding. Ms A had changed her sexual partner two months previously. Soon after this she had contracted genital thrush, which responded to topical clotrimazole. She uses a combined contraceptive pill and does not use condoms. Ms A has no other sexual partners, and thinks it unlikely her partner has. However, she has little knowledge of his previous sexual history.
The only noteworthy finding at vaginal examination was that Ms A's cervix bled easily when swabbed. A high vaginal swab was taken from the posterior fornix, and two swabs were taken from the endocervix and the urethra—a standard cotton swab and a plastic shafted chlamydia swab respectively. Ms A was prescribed doxycycline (200 mg for seven days) and metronidazole (400 mg three times daily for seven days).
A few days later the laboratory reported that chlamydia had been detected. Ms A was invited back to the surgery and was upset to be told that she might have had a sexually transmitted disease. She and her partner were referred to the local sexually transmitted diseases clinic for further investigation and follow up.
The case of Ms A prompted discussion in the practice about who we should investigate and treat for chlamydia. Of course, we all wanted to prevent our patients suffering avoidable morbidity—for example, pelvic inflammatory disease, infertility, and ectopic pregnancy—and we also wanted to use the NHS's scarce resources as wisely as possible. But some of us thought there was no place for chlamydia investigation in primary care, arguing that as chlamydia tests were expensive and insensitive, patients should be treated for chlamydia whenever this organism was suspected clinically. Others felt it important to obtain a microbiological diagnosis wherever possible, and, as chlamydial infection can be asymptomatic, thought we should be searching for asymptomatic cases—for example, among sexually active women attending for cervical smears or for contraceptive advice. However, none of us knew how common chlamydia was in our practice, nor were we certain that treating an asymptomatic infection reduced subsequent morbidity. We did not know the magnitude of any benefits and harms associated with proactive case finding or whether any net benefit would be worth the resources consumed.
Search for evidence
It is frequently written that the first step in evidence based practice is to turn the clinical problem into an answerable question.1 This proved more difficult than we first thought, as we wanted answers to several questions:
Is genital chlamydia an important cause of clinically importantmorbidity?
Does antibiotic treatment reduce subsequent morbidity in asymptomatic, sexually active women infected with chlamydia?
If so, is case finding in our population likely to be a cost effective way of reducing clinically important morbidity?
In the past, we might first have looked to standard textbooks for our answers. However, textbooks held in libraries are now more difficult to access from general practitioners' homes and surgeries than are online electronic databases. In addition, traditional textbooks are rarely written in a way that is sufficiently transparent to enable readers to determine how the authors reached their conclusions. More worryingly, the opinions expressed in textbooks may either be out of date even before publication or inconsistent with valid and relevant evidence.2 We therefore decided to look for answers to our questions using the information we could access from home or the surgery.
Categories of evidence
We thought our questions were unlikely to be original. Similar questions ought to have been addressed by anyone drafting evidence based guidelines. We also thought that a great many original research papers would have been published about chlamydia, and that it would be an inappropriate use of our time to attempt to obtain, read, and appraise every relevant article. We therefore decided to search for recent systematic reviews about chlamydia; evidence based guidelines about the detection and treatment of chlamydia; and randomised controlled trials of treatment or case finding, or both, and treatment of asymptomatic chlamydia published after the most recent systematic review or evidence based guideline that we retrieved. NRH offered to spend up to one hour searching at home for relevant material using a computer connected to the internet.
Where to search?
Before searching, NRH checked the one relevant textbook he had at home.3 It did not answer our questions. The next stops were Best Evidence4 and the Cochrane Library,5 searching with the single word “chlamydia.” The search of Best Evidence identified five articles, none of which had promising titles. The Cochrane Library search produced two completed Cochrane reviews, both about chlamydia in pregnancy, and three reviews listed on the Database of Abstracts of Reviews of Effectiveness. None of these looked relevant, so the next step was the internet.
The first stop was Bandolier's home page (www.jr2.ox.ac.uk/bandolier/). A search using “chlamydia” rapidly led to the full text of an article on treatment of chlamydia.6 This article noted that the Centers for Communicable Disease in the United States had recently changed its recommended treatment for chlamydial infection from oral doxycycline (100 mg twice daily for 7 days) to a single dose (1 g) of azithromycin.7 Bandolier had searched Medline for trials comparing doxycycline with azithromycin in the treatment of symptomatic and asymptomatic genital chlamydia and confirmed that both were effective treatments.
Bandolier had also looked at cost effectiveness, and it concluded that although azithromycin was more expensive, if compliance were better the higher drug costs would probably be “offset by lower costs associated with pelvic inflammatory disease, chronic pelvic pain, ectopic pregnancy, and tubal infertility.” Perhaps the most startling information in this article, however, was the epidemiological data: chlamydia is the commonest curable bacterial sexually transmitted disease in the United Kingdom and the organism is most likely to be isolated in sexually active women under the age of 20 years, with rates in excess of 350 per 100 000 population (0.35%).
Recent guidelines and systematic reviews
The next step was to search for recent guidelines and systematic reviews. Medline Express for 1996, 1997, and January-March 1998 was searched with WinSpirs. A search for the medical subject heading (MeSH) term “chlamydia” in all subheadings produced 813 articles. This search was then limited by publication type to “meta-analysis” (1 article), “guideline” (2), “review academic” (5), “review” (83), and “randomised controlled trial” (10). Browsing the titles showed that one of the guidelines had been prepared by the Canadian Task Force on the Periodic Health Examination,8 which has a reputation for using a rigorous approach based on systematic evidence to developing guidelines. The full text of the guideline is available free on the world wide web (www.cma.ca/cmaj/vol-154/1631.htm).
Canadian task force review
This text cited 201 references. In using a systematic approach to reviewing published reports and developing its recommendations it had considered many of the questions to which we sought answers. The text reported that in North America, as in Britain, chlamydia is the commonest sexually transmitted disease, and is two to three times as common as gonorrhoea. As in Britain, infection is most prevalent among sexually active women aged 15 to 19 years. In six Canadian studies, 1-25% of women tested were infected. Other risk factors and indicative signs were multiple sexual partners, a new partner in the previous year, no barrier method of contraception, low socioeconomic status, intermenstrual bleeding, cervical friability, and purulent cervical discharge. However, 60-80% of infections in women are asymptomatic. But what startled us most was the statement supported by 18 references that “serologic studies suggest that at least 64% of cases of tubal infertility, and 42% of ectopic pregnancies are attributable to chlamydial infection.” Unfortunately, the guideline authors did not comment on the strength of inference that could be drawn from these 18 papers.
Does case finding reduce morbidity?
But was there any evidence that case finding and treating asymptomatic infection reduced morbidity? The guideline authors could identify only one controlled study in a non-pregnant population. The reference cited a paper presented to the International Symposium on Human Chlamydial Infections, which was unlikely to be retrievable quickly and cheaply. However, scanning the 10 randomised trials that had been identified from Medline indicated that the same study had been published later in the New England Journal of Medicine.9 This article was readily retrieved the next day from the hospital library. (It could have been retrieved that night from online access to the BMA Library.)
The article reported a randomised controlled trial conducted in Seattle, in which 2607 asymptomatic women were randomised to an invitation to investigation for chlamydial infection (cervical swabs sent for enzyme linked immunosorbent assay and culture) and treatment if positive or to a control group given the “usual care.” Women were followed for 12 months.The chlamydial infection rate in control women was 7%. The rate of clinically defined pelvic inflammatory disease in the 12 months after randomisation was reduced by 56% (relative risk 0.44, 95% confidence interval 0.20 to 0.90) in the intervention group compared with the control women. The absolute risk reduction was 1.1%; 2% of the control group developed pelvic inflammatory disease during follow up compared with 0.9% of the women in the intervention group. In other words, in Seattle, 91 sexually active women aged 18-34 years had to be invited for investigation for chlamydia to prevent one case of pelvic inflammatory disease. The study did not provide information about the impact of case finding on tubal damage, ectopic pregnancy, or infertility. Nor did it comment on any harm that may have resulted—for example, from treating women whose results were falsely positive, from mistaken reassurance of women given false negative results, or from the social implications of being told that one has an asymptomatic sexually transmitted disease. Nevertheless, we thought that this was an important study.
Studies in British general practice
But was the prevalence of chlamydia likely to be as high as 7% among any readily identifiable groups of our patients? We wanted some prevalence data from British general practice. Among the reviews identified on Medline was an article with a promising title.10 This article cited nine prevalence studies of chlamydial infection undertaken in British general practices. The prevalence of infection ranged from 2% among asymptomatic women aged 15-40 attending for a cervical smear in Fife, Scotland, to 12% among women aged 16-44 requesting termination of pregnancy in inner city east London. The prevalence was also 12% among the mainly social class 3 women aged 19-58 attending for a cervical smear in an inner city Glasgow practice, and it was 11% in premenopausal women undergoing a speculum examination for any reason in a central London general practice.
What we learned and decided
Chlamydia is a much more important public health issue than any of us had suspected. We were all surprised at just how common it can be among young, sexually active women. We were also surprised that serological studies suggest that chlamydia may account for at least two thirds of tubal infertility and nearly half of ectopic pregnancies. Furthermore, we were impressed with the randomised controlled trial from Seattle which showed that in women in whom the prevalence of chlamydial infection was 7%, inviting them for investigation and treatment where necessary reduced the rate of pelvic inflammatory disease by half in 12 months. This suggested to us that much of the morbidity caused by chlamydia may be preventable.
The evidence that we have seen did not allow us to identify unequivocally a single best practice for deciding in whom and how to investigate and treat chlamydia. Given that there is substantial geographical variation in chlamydia prevalence, we think that it is unlikely that a case finding policy can be devised that is equally cost effective for all practices in the United Kingdom. However, the data do suggest that systematic case finding and treatment of chlamydia could reduce potentially life-ruining morbidity for appreciable numbers of British women. Thus, we have decided that in the absence of national guidance we want to discuss and agree a way forward with other local practices, genitourinary physicians, microbiologists, gynaecologists, family planning clinics, and the health authority. One important next step might be to measure the prevalence of chlamydia infection in selected groups of our practice populations.
What about Ms A? We now know that she had several risk factors for chlamydia—she was young, sexually active, had a new sexual partner, and had a friable cervix. We should therefore have suspected chlamydia infection. Until there is a local guideline, we have agreed that whenever we suspect chlamydia we will offer to take chlamydia swabs and treat with doxycycline—unless compliance may be a problem, in which case we will use azithromycin. We will also refer patients and their partners to the genitourinary medicine clinic.
Although the evidence we found did not answer our questions completely it was the best evidence that we could find. We could not escape making a decision about what to do for our patients, as to do nothing would be a decision in itself. Our search took less than an hour plus a 10 minute trip to the library. Reading and discussing the material we retrieved took rather longer, perhaps three hours. We learned a lot and made a decision about patient management that is based on the best evidence we could find. We think that our time was well spent.
Competing interests: NRH has received lecture fees for speaking at postgraduate education meetings for general practitioners sponsored by the manufacturers of azithromycin.