Older HIV-infected individuals present late and have a higher mortality: Brighton, UK cohort study

Data from the Health Protection Agency (HPA) show that about one-quarter of people with HIV infection in the UK remain undiagnosed and 50% of adults diagnosed in 2010 had a CD4 cell count below 350/mm³ at diagnosis [1]. Initiating therapy with a CD4 cell count below 200 cells/mm³ is associated with a substantially greater risk of disease progression and death than earlier initiation [2]. Similarly, individuals entering the Strategies for the Management of Antiretroviral Therapy (SMART) study who were either treatment-naive or who had not been on therapy for the previous 6 months, and who deferred treatment until their CD4 cell count was below 250 cells/mm³ had a more than 4 fold higher risk of opportunistic infections and serious non-AIDS clinical events compared to patients who started treatment with CD4 cell count above 350 cells/mm³[3]. The SMART study [4] and other cohort studies [5, 6] have shown that there is a continual gradient of increased risk of both death and disease progression associated with lower CD4 cell counts and no specific clear threshold at which risk increases.

The most frequent definition of late presentation for treatment is a CD4 cell count below 200 cells/mm³, although other thresholds have ranged from 50 to 350 cells/mm³[1, 7–18]. This has sometimes been combined with whether an individual presented with an AIDS-defining diagnosis [19, 20]. However, current UK guidelines recommend initiation of antiretroviral therapy (ART) in all patients with a CD4 cell count of less than 350 cells/mm³[21] and the European consensus is now that late presentation for care refers to persons presenting with CD4 cell count below 350 cells/mm³ or those presenting with an AIDS-defining event regardless of CD4 count [20, 22]. Two recent studies have reported on late presentation using this European consensus definition. In one of these studies, 73% of CD4 cell count at diagnosis were missing and had to be imputed [23] while the other study examined late presentation over a shorter period of time compared to our study [24].

In this study, we examined factors associated with late presentation (CD4 cell count < 350 cells/mm³), and mortality in a large UK cohort comprising predominantly of men who have sex with men (MSM) over a 14-year period.

In our study, 49% of those diagnosed with HIV were late presenters according to the new consensus definition for late presentation and nearly one-quarter presented with advanced HIV infection. The frequency of late presentation is similar to that recently reported from Germany and New Zealand of 49.5% and 50% respectively using the new threshold of CD4 <350 [23, 24]. In this UK cohort, factors associated with late presentation were being older than 50 years, male heterosexual, African ethnicity and being diagnosed in the earlier calendar periods. The trend suggests HIV infection is increasingly being diagnosed early in the whole cohort and this was largely due to increased HIV testing in men who have sex with men as described in another cohort [26] and in those less than 50 years of age. The risk factors identified in this study corroborate findings from similar studies on late diagnosis [10, 11, 14, 18, 27].

We showed a progressive improvement in earlier diagnosis with time in those diagnosed under the age of 50 years but strikingly the likelihood of late presentation in those over 50 years of age remained constant over this 14-year period. The association of older age with late presentation of HIV has also been identified in other studies [11, 13, 14, 19, 28] that used other CD4 cell count thresholds for late presentation. However we know that individuals older than 50 years could have lower CD4 cell counts as a result of the normal ageing process even in the absence of HIV infection, so would be more likely to be considered “late presenters” when they acquire HIV infection.

A study from our group has shown that individuals older than 60 years were less likely to be offered an HIV test even in situations where an opt-out approach was adopted [29].

The failure of physicians to consider the possibility of HIV infections in these patients and confusion between symptoms of opportunistic infections and those of common co-morbid conditions associated with ageing further delays HIV diagnosis [30–32].

The association of African ethnicity with late presentation is consistent with other studies [11, 14] and may reflect the fact that these groups of individuals are often marginalized and hard to reach. This may be due to lack of knowledge about HIV infection or reduced access to medical services [33]. Nearly all the Africans in our study acquired HIV infection through heterosexual intercourse, which could be due to sexual relations within or outside of the UK. A recent HPA report states that an increasing proportion of heterosexual transmission are occurring within the UK [34].

Trends suggest that early diagnosis of HIV infection is improving in the whole cohort. From 1996-2001, 55% of patients were diagnosed late using the current CD4 threshold, but this had dropped to 44% in the 2006-2010 calendar periods. This still represents a high burden of late presentation. Some studies have shown that many late presenters had suffered an HIV related illness, which had been missed in an earlier contact with a healthcare giver [9, 35].

Being diagnosed with HIV with a CD4 cell count <350 and being older than 50 years of age were independently associated with increased mortality.

Late presentation in our cohort is associated with a tripling of mortality. This is the first study reporting a tripling of mortality with the higher CD4 count threshold for late presentation.

The association of late presentation with increased mortality has also been described in other studies both in the UK [18, 36, 37] and in other countries such as Spain [38], the Netherlands [39] and France [40]. These studies differ from ours in that they have used different CD4 cell count thresholds for defining late presentation most commonly a CD4 count less than 200 cells/mm³. Individuals diagnosed late are usually more likely to present with multiple HIV-related complications as well as have a higher risk of immune reconstitution inflammatory syndrome (IRIS), with a likely associated increase in mortality.

The strengths of this study lie in the utilization of the new European consensus definition for late presentation as this would facilitate regional comparisons within Europe and allow investigation of temporal trends after interventions [22], and the duration of the study period (14 years) which allowed a robust estimate of trends as well as large sample size. The study however has some limitations. The age at seroconversion was missing for the majority of the cohort; hence age at cohort entry (presentation) has been used. Age at entry is the sum of age at seroconversion and time since seroconversion [41, 42], both of which are strong predictors of disease progression and death. In our study it was not possible to separate their two effects without knowledge of the date of seroconversion. We did not separate the effect of background mortality which is higher in older individuals from that associated with the additional effect of HIV infection. However, in individuals within the same age group, those presenting with lower CD4 cell count had a higher short-term mortality and this was more likely to be attributable to the effect of HIV itself rather than the effect of ageing of the cohort. In the author’s experience, some individuals with “missing” CD4 count at presentation (significantly more African and heterosexual risk group in this cohort) were those admitted through the emergency department in Brighton because of presentation with advanced HIV. Their CD4 counts, although in the hospital system did not always follow through to the clinic database which is different. We showed that these individuals have a higher mortality (unknown CD4 category), hence mortality would have been underestimated in those with CD4 <350 cells/mm³. Some deaths were coded as unspecified especially in those who died outside healthcare facilities as further information was lacking. This could have resulted in misclassification of some causes of deaths. Finally, the majority of the patients in this cohort were Caucasian MSM, hence the findings cannot be generalised to a cohort that is diverse in both ethnicity and sexual orientation; however this does not undermine the study validity.

As a result of the high mortality associated with late diagnosis, strategies must be put in place to encourage HIV testing. The CDC recommend routine testing in all healthcare settings for patients aged 13–64 years, unless the local HIV prevalence is known to be less than 0.1% [43].

In addition to recommending routine testing in genitourinary medicine and antenatal clinics, 2008 UK guidelines for HIV testing also advocate that tests should be offered to all adults registering in general practice and to all general medical admissions patients in areas where diagnosed HIV prevalence is greater than 2 per 1,000 among 15 to 59 year olds [44]. The UK guidelines have the advantage of not having any age restrictions but restricting HIV testing to only general medical admissions in hospitals within areas of high prevalence has been shown to miss the majority of HIV infections in one study [29].

HIV testing should take an ‘opt-out’ approach. Higher rates of HIV testing are achieved with such ‘opt-out’ approaches than opt-in approaches [45, 46]. Identifying high-risk groups who continually refuse an HIV test and addressing the reasons for this could also encourage earlier diagnosis. One of the mains reasons for refusing a test is the stigma associated with HIV infection. To overcome the issue of stigma, we need to avoid the approach of offering an HIV test to only individuals perceived to be at high risk and ‘normalize’ the test by offering it to everyone alongside a battery of other routine investigations. Early HIV diagnosis could result in reduction of onward transmission as individuals aware of their HIV infection would be more likely to modify their sexual behaviours [47]. Furthermore, the use of HAART can also reduce an individual’s infectiousness by reducing plasma and genital HIV viral loads [48, 49]. The findings of this study re-enforce the need for a universal HIV testing policy. General practitioners and hospital physicians should always be aware of the possibility of HIV infection in their patients irrespective of their age and there should be a policy that enables HIV test to be offered routinely alongside other investigations.

This paper showed that HIV late presentation is very common and does not appear to be improving in older adults. This impacts individual health as documented by increased mortality in this study. The consequent increase in hospitalizations also results in increased health care costs. Most HIV transmissions are from individuals unaware of their status.

Prompt HIV diagnosis could be achieved through a policy that enables HIV test to be offered routinely alongside other investigations.