• Gonorrhoea
• screening
• diagnosis
• sexually transmitted diseases
• molecular technique
• testing

The current low incidence of gonorrhoea, new options for gonococcal diagnosis and increased pressure to reduce costs suggest that we need to review our current approach to testing. If confirmatory testing is carried out, then an accurate diagnosis of gonorrhoea can be made even in areas of low prevalence, although once the population prevalence drops below around 1% even this approach starts to become unreliable. Screening of all patients for gonorrhoea should be retained in sexual health clinics, and offered to high-risk individuals (such as sexual contacts of patients with gonorrhoea) in all healthcare settings. The low marginal costs support the addition of gonorrhoea testing to those attending a national chlamydia screening programme, unless the local prevalence of gonorrhoea is very low.

Rates of gonorrhoea in the UK continue to fall, down over 10% year on year and at a 10-year low,1 while there is increasing pressure to reduce service costs and deliver healthcare more cost effectively. For those with genital discharge, epididymo-orchitis or pelvic inflammatory disease there can be little argument that tests for gonorrhoea are appropriate, but can they be justified for those without symptoms requesting a sexual health screen? As gonorrhoea becomes an increasingly rare disease we need to step back and review screening to determine the most appropriate approach. The value of routine screening for gonorrhoea has been questioned even in dedicated sexually transmitted infection (STI) clinics, where it has been suggested that the resources currently being used for this purpose might be better employed to improve contraceptive services.2 A recent audit of sexual health clinic attendees in the UK suggests that over a third of asymptomatic men, and almost half of men who have sex with men, are now screened for gonorrhoea using urine nucleic acid amplification tests (NAATs).3 Should the increasing use of NAATs affect our decision on who to screen? In particular, as increasing numbers of patients are offered testing for chlamydia through national screening programmes there is an opportunity to also test for gonorrhoea, but is this appropriate?

A major problem with screening asymptomatic patients is in interpreting the test result in a low-prevalence population. The relationship between test specificity and positive predictive value means that tests become increasingly unreliable with falling rates of infection. NAATs offer a clear advantage in sensitivity over gonococcal culture and will detect probably 20–30% more infections, and their specificity can be as high as 98%.4 5 But if we consider a hypothetical clinic population which has a gonorrhoea rate of 1% then this translates into a positive predictive value of only 33%. In other words, two out of three patients attending who test positive will not have gonorrhoea. This is a concept which most patients (and possibly a few healthcare professionals) find difficult to grasp. Gonococcal culture is recommended for those who are NAAT positive to enable antibiotic sensitivity testing to be performed, but the lower sensitivity of culture means that it cannot be relied upon to confirm whether the ‘NAAT-positive infection’ is a true positive or not.5 The solution to this problem is to routinely use a second confirmatory test, preferably one that is directed against a new target—for example, a second NAAT from a different manufacturer. Using the earlier scenario we had a patient with a 1 in 3 chance of infection after a single gonorrhoea test. Because of this higher pre-test probability of infection, adding a second test which has a specificity of at least 95% will translate into a positive predictive value of >90%, which reaches the acceptable level outlined in recently published UK national standards.6

A second concern is cost—if the test for gonorrhoea was free (and accurate) we would certainly offer it widely. Almost all of the currently available NAATs combine gonorrhoea with chlamydia testing, so the test itself is almost free. There are still additional costs for the laboratory and clinical service associated with maintaining quality control, processing the results to ensure that positive samples are identified and appropriate action taken, and in providing negative results in a timely manner to patients. However, if a system is already in place for chlamydia then the marginal costs of adding gonorrhoea are likely to be low, particularly as electronic systems to automate the process become increasingly available. The question therefore is not whether screening for gonorrhoea is cost effective when performed in isolation, but whether the marginal costs of adding a gonorrhoea test to existing chlamydia testing can be justified.

An additional refinement might be to limit the offer of screening to those who are at higher risk of having gonorrhoea. Screening of such a high-risk population has the potential to identify asymptomatic individuals, reduce the burden of infection within a population and reduce potentially serious complications, especially in women. Various demographic and behavioural factors for infection are well recognised, including younger age, ethnicity, area of residence, previous gonorrhoea, number/concurrency/newness of partners, men who have sex with men and condom use.7 Restricting the availability of gonorrhoea screening to individuals in these risk groups would increase the ‘yield’ of positive results but at the expense of missing around 20% of infections.8 Indeed the predictive value of risk factors has largely been derived from STI clinic populations using culture-based tests for gonorrhoea, and their accuracy in identifying asymptomatic individuals with low bacterial loads detected by a NAAT remains unproved.9 Therefore before advocating targeted screening of the wider population we need more information on how effective it would be in identifying untreated gonorrhoea, and how it would translate into reducing the prevalence of the disease and its complications.

The preferences and expectations of patients must also be a factor in deciding whether, even in the absence of symptoms, a test for gonorrhoea should be available.10 There has been little formal assessment of what patients expect when they request a STI ‘check-up’ but awareness of gonorrhoea is high in the general population and it would be reasonable to assume that patients generally expect their ‘check-up’ to include testing for gonorrhoea regardless of symptoms. At an individual level patients should be offered an explanation about all the tests that are being offered, and retain the right to decline testing.

An increasing number of people are now attending for chlamydia testing as part of the UK national screening programme. Should they also be offered testing for gonorrhoea? It is extremely unlikely that the low prevalence of gonorrhoea, and its widely varying prevalence in different demographic subgroups, would justify a population screening programme in its own right. Patients with gonorrhoea (where infection is related to contact with core groups) are not the same as those who are infected with chlamydia (where the distribution of infection appears more diffuse). Nonetheless, the two risk groups do overlap to a large degree11 and since the marginal cost of adding a gonorrhoea test is low and, with appropriate confirmatory testing, the accuracy of the result is generally high, its addition is probably justified,12 even though the evidence base is currently lacking.13 However, in very low prevalence populations (gonorrhoea prevalence <1%) the positive predictive value remains low even after confirmatory testing and cautious interpretation of a positive gonorrhoea result would still be required. Some chlamydia screening programmes are already offering joint testing (eg, http://www.ruclear.co.uk/, accessed 21 August 2010), although explicit additional consent for the gonorrhoea test is required before it is carried out.14 Also, caution is required when urine samples are taken and tested for gonorrhoea in women attending the screening programme because sensitivity is reduced when urine is tested in comparison with vulval or cervical specimens.15

After taking these issues into account a number of recommendations can be made on screening for gonorrhoea in asymptomatic individuals:

• Regardless of setting, a test for gonorrhoea should be performed in all those who request an STI screen, preferably using a NAAT with confirmatory testing to minimise false-positive results.

• High-risk groups should be offered a test for gonorrhoea. This would include contacts of a patient with gonorrhoea wherever they present, and all those presenting to a sexual health clinic, but we have insufficient evidence to target other general population groups at present.

• It may be appropriate to offer a test for gonorrhoea to people being tested for chlamydia within a national screening program but only after obtaining specific consent.

STI rates vary greatly over time. We may be approaching a trough for gonorrhoea prevalence in 2010 but a future resurgence of the infection is possible, and even probable, which demands a dynamic approach to testing strategies. Testing for gonorrhoea has never been more sensitive, easier or more acceptable to patients and offers us the opportunity to control its spread more effectively than ever before.