The annual number of new diagnoses of HIV continues to rise in the UK1 and elsewhere in the developed world.2 3 Although most new diagnoses are of chronic infections, the number of recently acquired infections in men who have sex with men (MSM) is thought to be either stable4 or increasing.5 Of those individuals diagnosed with having chronic infection, a significant proportion are diagnosed “late” and such advanced immunosuppression represents the most common cause of HIV related mortality in the UK.6

An estimated 50–90% of individuals with primary HIV infection (PHI) are thought to develop a symptomatic seroconversion illness.7 8 However, these symptoms are non-specific in nature and may mimic many common febrile illnesses.8 9 Correct diagnosis at this stage of HIV infection is important: it may be the only chance to diagnose HIV in an individual prior to developing advanced disease; it allows consideration of early therapeutic intervention, which may represent a unique window;10 and it offers a significant opportunity to prevent onward transmission at a time of increased infectiousness.11 12 The extent to which PHI presents to healthcare providers and the degree to which it is recognised in the UK is unknown. We report a study investigating missed opportunities for diagnosing PHI.

METHODS

Patients with recent HIV infection (RHI) between 2003 and 2005 were identified from the HIV clinical database. RHI was defined by one or more of the following: a previous negative HIV antibody test within 18 months, an evolving Western blot or antibody response, or by application of a serological testing algorithm for recent HIV seroconversion (STARHS).13

At initial presentation to the HIV clinic, the following data were collected: recall of any seroconversion symptoms (for example, fever, rash, sore throat14 15); timing of the seroconversion illness and any self reported illness prior presentation to healthcare providers.

For those individuals who did not self report healthcare presentation, we analysed laboratory records to determine whether any blood tests had been requested by healthcare providers (for example, general practitioner (GP), accident and emergency department (A&E)) at the time of seroconversion symptoms before presentation to the HIV clinic.

RESULTS

Of 356 individuals diagnosed with having anti-HIV antibody positive during the study period, 117 (33%) were classified as RHI. Clinic notes were accessible for 108 (92%). Of these, 103 (95%) were male and 93 (86%) were MSM. RHI was identified by application of the STARHS assay in 60 (56%), a previous negative antibody test within 18 months in 11 (10%), an evolving serological response in 3 (3%) or by combination of the above in 34 (31%).

Altogether, 76 of the 108 patients (70%) reported symptoms of seroconversion, of whom 91% were MSM. The most common symptoms reported were fever (55%), rash (37%), sore throat (33%), diarrhoea (28%), swollen glands (21%), arthralgias (12%), headache (11%) and flu-like symptoms (46%). Of the 76 symptomatic individuals, 40 (53%) accessed health care at the time. Of these, 21 (52%) patients were correctly diagnosed with having PHI at first presentation: 12 (57%) by genitourinary medicine (GUM); 5 (24%) by other hospital teams (4 by general medical teams); and 4 (19%) by their GP.

Altogether, 19 of 40 patients (48%) had a prior presentation to health care at the time of seroconversion, yet a diagnosis of PHI was missed (details in table 1); 17 (89%) of these were MSM. Fifteen (79%) had presented to their GP, three (16%) to A&E and one (5%) to GUM. Of the 19 “missed” cases, 13 were identified as missed because the patient reported prior presentation to health care with symptoms suggestive of seroconversion. Diagnoses the patients recall being given at these consultations included bacterial infection (n = 4), tonsillitis (n = 2), glandular fever (n = 1), viral infection (n = 1), thrombocytopenia of unknown cause (n = 1) and in four cases no diagnosis was given. Six of the 19 missed cases were identified as such because laboratory records showed blood tests requested by healthcare providers at the time of the reported seroconversion symptoms. In one of these six cases a HIV test had been performed. This patient had presented to GUM where a HIV test (a fourth generation combined antibody/antigen test) was negative at the time of presentation but the patient was not retested until a subsequent presentation 3 months later.

DISCUSSION

This study shows that the diagnosis of PHI is often missed, and highlights two points at which this occurs. First, almost half (47%) of the symptomatic individuals did not seek medical attention when they were experiencing classical seroconversion symptoms. Second, in almost half (48%) of the symptomatic individuals that presented to health care the diagnosis of PHI was not made; the majority of these missed diagnoses were in primary care (79%). Despite the fact that MSM represents the major risk category for acquiring HIV infection in the UK,1 our results show that PHI remains largely undiagnosed in this group.

Comparison with previous studies

This is the first UK study of its kind. Melzer et al previously showed that PHI may not always be diagnosed by healthcare providers: they anonymously tested the sera of 238 attendees with “symptoms consistent with seroconversion” and found two cases of undiagnosed PHI.16 Our findings are also consistent with two North American studies that showed that PHI is infrequently considered as the diagnosis when symptomatic patients first present to health care.8 9

The other principal finding in our study was that only 53% of the symptomatic patients accessed health care at the time of their seroconversion illness. This is in contrast to Schacker et al,8 who showed in a cohort of 46 patients with PHI (of whom the majority were MSM) that 94% sought health care when they had symptoms. Of note, that study was completed before the advent of highly active antiretroviral treatment and in an entirely different healthcare system. Furthermore, our findings are supported by a recent questionnaire study in North America, which showed that although most HIV negative men could identify symptoms of PHI relatively few would access health care for such symptoms.17

Limitations

Our study used patient’s recall of prior presentation to health care and therefore reporting may not be completely accurate. The effect of this reporting bias is highlighted by identification of further missed opportunities using laboratory records (table 1, cases 14–19). In this study, a significant proportion of individuals did not access health care through the duration of their PHI. This includes 32 individuals who did not report seroconversion symptoms, as well as 36 symptomatic individuals who did not access health care at the time. We were unable to comment on the specific reasons why these individuals did not access health care initially or the factors leading them eventually to present. In addition, further research is needed to explore strategies to detect those individuals who remain undiagnosed in the community. From a provider perspective, it was not possible to access primary healthcare records because of patient confidentiality. Hence we could not clarify if healthcare professionals had considered HIV as a diagnosis, completed a risk assessment (including knowledge of sexual orientation) or if HIV testing was offered and refused at time of presentation. Of relevance, a survey of over 13 000 MSM showed that only 28% believed that their GPs had knowledge of their sexuality.18 Our study was conducted in an urban centre with a large white MSM community and, therefore, may not be representative of those populations faced by many primary care physicians. However, our findings show that in two instances the diagnosis of PHI was missed in heterosexual females, which suggests high degree vigilance is needed for diagnosis of PHI in all populations.

Implications

Strategies to improve these outcomes need to consider both patient and provider. From a patient perspective, the likelihood of future diagnosis could be improved by encouraging at-risk groups (for example, MSM) to access health care when they experience symptoms of seroconversion or following high risk exposure. From a provider perspective, training for non-HIV/GUM specialists should include recognition of the acute seroconversion illness. In the current era, reducing stigma and enabling discussion of potential risk factors for HIV acquisition is likely to facilitate the diagnosis of PHI. In this study, a combination of awareness of sexuality with recognition of classical symptoms of PHI may have enabled the diagnosis in 16/19 (84%) of missed cases that had presented to health care.

In one missed case (case 19, table 1), the diagnosis of PHI was not made even though a HIV test had been performed. Although the latest fourth generation combined antibody/antigen test provides improved sensitivity for detecting PHI, those presenting very early after onset of symptoms may still be missed. In such instances, where PHI is strongly suspected, follow up HIV testing 1–2 weeks later is recommended.

As HIV testing becomes increasingly “normalised”,19 barriers to performing HIV tests in non-GUM settings will hopefully be overcome. Given that individuals with PHI are more infectious and contribute disproportionately to onward transmission,11 12 a significant opportunity exists to improve both individual and public health.