Delayed Presentation of HIV Among Older Individuals: A Growing Problem

Summary

Late presentation for care is a major impediment to prevention and effective treatment of HIV infection. Older individuals are at increased risk for late presentation, represent a growing proportion of all those with late presentation, and may require interventions tailored to their age group. We provide a summary of the worldwide literature published between 2016–21 (reporting data from 1984–2018) quantifying the association of age with delayed presentation. Using the most common definitions of late presentation and older age from these earlier studies, we update this work with data from the International epidemiology Databases to Evaluate AIDS (IeDEA) consortium focusing on data from 2000 to 2019 encompassing 4 continents. Finally, we consider how late presentation among older individuals might be more effectively addressed as electronic medical records become widely adopted.

Introduction

The successful scale-up of effective antiretroviral therapy (ART) has supported the long-term survival of people with HIV infection (PWH). More people are living with HIV than ever before and this population is aging(1–4). Globally, between 2015 and 2020, UNAIDS estimated that the total number of PWH over the age of 50 years increased from 5.4 million to 8.1 million (aidsinfo.unaids.org). In this four-part series co-sponsored by The Lancet HIV and The Lancet Healthy Longevity, we explore pressing issues facing those aging with HIV in the era of ART. In this article, we begin by addressing risk of delayed presentation for ART, subsequent articles consider 1) evidence for and against accentuated biologic ageing compared with people without HIV infection, 2) how health systems might adapt to an ageing population of PWH, and finally 3) the syndemic of stigma particular to those aging with HIV.

In many settings, as the prevalence of HIV among older individuals has grown the number of new infections in this age group has increased. For example, between 2015 and 2019 in the United States, the overall prevalence of PWH increased by 8% and incident infections decreased by 4%(5). In contrast, we saw a 40% increase (289,900 to 407,100) in prevalence and 15% increase (2700 to 3100) in incidence among those 50 years and older – the largest increases of any age group(5). This is likely due to intra-generational and cross generational unprotected sexual activity(6, 7).

Large scale population based statistics on HIV incidence in older age groups are limited outside the United States but some data is available from South Africa. By the end of 2013, 14% (6304/44909) of PWH in care were ≥50 years(8). Among 84,078 patients starting antiretroviral therapy from 2004 to 2013, the proportion of those ≥50 years increased from 6% (290/4999) in 2004 to 10% (961/9657) in 2012–13(8). Another study tested in 2010 and retested in 2015 a cohort of 1,360 individuals aged 40 or more years in 2015 (6). HIV prevalence increased from 21% to 23% corresponding to 33 incident infections (0.49 infections per 100 Person Years); only those 80 or more years of age experienced no new infections(6).

Twelve years ago, we used data from the United States and Canada to compare CD4 cell count and AIDS-defining conditions at presentation for HIV care among those under 50 and those 50 years of age and older(9). Older individuals had lower CD4 counts and a higher prevalence of AIDS-defining conditions at diagnosis, and these gaps between younger and older at presentation persisted over calendar time despite decreases in new diagnoses among both groups(9). Now that an even larger proportion of individuals living with HIV are 50 years and older worldwide, we revisit the relationship between age and delayed presentation for care globally with a review of recent literature, data analyses from the International epidemiology Databases to Evaluate AIDS (IeDEA) network, and a consideration of what might be done to decrease new HIV infections and delayed presentation for care among older individuals.

Review of Recent Literature (2016–2021)

We conducted a structured review of recent literature (see Search Strategy and Table 1). These studies were conducted in North and South America, Europe, Africa, Middle East, Asia, and Australia and include observations from 1984 through 2018. Most studies (n=32) defined late presentation as having a CD4 count of <350 cells/μL or an AIDS diagnosis at or near the time of presentation for care. Although these studies document improvements in recent years, delayed presentation remains a significant global issue in HIV care. In many settings, approximately half of those newly diagnosed with HIV infection have CD4 counts below 350 cells/μL at presentation and the proportion is even higher in lower- and middle-income countries.

Table 1.

Structured Review of Delayed Presentation Publications Last Five Years

1st Author	Definition of Late Presenter	New Diagnoses (n)	Late Presenters (n, %)		Years	Location	Age Variable (Yrs)	Odds Ratio	95% CI
Gardner	AIDS	1385	422	31%	2009–14	USA (single site)	65+ vs. 25–44	1.3	0.6	2.6
Nduaguba	AIDS	77844	30359	39%	1996–2001	USA (multisite)	60+ vs<30	4.0	3.3	4.7
Nduaguba	AIDS	77844	30359	39%	2002–2007	USA (multisite)	60+ vs<30	4.6	3.9	5.4
Kwobah	CD4<100	10533	2421	23%	2010–11	Kenya (multisite)	>24 vs. <19	1.6	1.0	2.6
Kadam	CD4<200	659	264	40%	2011–15	India (single site)	continuous variable	na
Honge	CD4<200	3720	1810	48%	2005–13	West Africa (single site)	50+ vs.<30	1.5	1.1	2.0
Senard	CD4<200	186	49	26%	2012–13	France (single site)	continuous variable	1.1	1.0	1.1
Taborelli	CD4<200 or AIDS	16601	7720	47%	1999–13	Italy (multisite) Italians	50+ vs.35–49	1.5	1.4	1.7
Taborelli	CD4<200 or AIDS	4152	2831	68%	1999–14	Italy (multisite) Non- Italians	50+ vs.35–49	0.9	0.7	1.2
Tang	CD4<200 or AIDS	528234	179700	34%	2006–14	China (multisite)	55+ vs.15–24	2.9	2.9	3.0
Mohammadi	CD4<350	4402	2562	58%	1987–2016	Iran (158 sites)	50+ vs.<30	3.6	2.6	4.8
Rava	CD4<350	14876	6635	45%	2004–18	Spain (46 sites)	50+ vs. <30	2.8	2.5	3.1
Ribeiro	CD4<350	356	218	59%	2017	Brazil (single site)	continuous variable	1.0	1.0	1.1
Bath	CD4<350	2469	1342	54%	2008–14	England (multisite)	55+ vs. 16–19	3.5	1.6	7.7
Cuzin	CD4<350	1421	625	44%	2014–15	France (10 sites)	>47 vs. <29	1.9	1.4	2.5
MacCarthy	CD4<350	1970	698	61%	2010	Brazil (3 sites)	45+ vs.18–44	1.7	1.1	2.5
Hu	CD4<350	519	188	38%	2011–14	China (8 cities)	40+ vs.18–24	3.1	1.8	5.5
Gullon	CD4<350	316	158	50%	2007–14	Spain (single site)	>38 vs. <38	2.2	1.3	3.7
Schafer	CD4<350	165	105	64%	2009–11	Germany (single site)	mean age LP 41 vs 32	na
Miranda	CD4<350 or AIDS	907	459	51%	1984–2017	Portugal (single site)	>56 vs. <30	2.9	1.5	5.9
Jablonowska	CD4<350 or AIDS	1522	682	45%	2016–17	Poland (13 sites)	per decade	1.5	1.4	1.7
Robles	CD4<350 or AIDS	3842	2793	73%	2012–17	Panama (multisite)	>65 vs. 18–24	2.9	1.7	5.0
Palacios-Baena	CD4<350 or AIDS	205	102	50%	2014–18	Spain (single site)	32+ vs. <32	3.4	1.9	6.1
Muelas Fernandez	CD4<350 or AIDS	74	33	45%	2013–18	Spain (single site)	40+ vs. <40	2.6	1.0	6.9
Karaosmanoglu	CD4<350 or AIDS	1673	826	49%	2003–16	Turkey (single site)	>50 vs ≤50	1.8	*	*
Krueger	CD4<350 or AIDS	1644585	Na	na	2013–16	USA (multisite)	45+ vs. 25–44	1.7	*	*
Siwak	CD4<350 or AIDS	3972	2288	58%	2000–15	Poland (14 sites)	60+ vs.<20	5.2	1.9	14.0
Hu	CD4<350 or AIDS	45118	31673	70%	2012–16	China (multisite)	>50 vs. 15–30	1.5	1.4	1.6
Zhonghua	CD4<350 or AIDS	293187	200503	68%	2009–17	China (multisite)	60+ vs. 18–29	2.3	2.3	2.4
Wilton	CD4<350 or AIDS	1819	1476	54%	1999–2013	Canada (multisite)	50+ vs.<30	2.8	1.9	4.1
Lin	CD4<350 or AIDS	436	82	19%	2000–14	Australia (single site)	mean age LP 45 vs 39	na
Rao	CD4<350 or AIDS	474	356	75%	2012–13	India (single site)	≤50 vs.≤25	4.2	1.3	13.2
Darcis	CD4<350 or AIDS	687	302	44%	2006–17	Belgium (single site)	10 year increments	1.3	1.1	1.5
Wojcik-Cichy	CD4<350 or AIDS	412	259	63%	2009–16	Poland (single site)	10 year increments	1.8	1.4	2.4
Johnson	CD4<350 or AIDS	401	307	77%	2013–16	Sudan (single site)	34+ vs. <34	na
Jin	CD4<350 or AIDS	7073	2949	42%	2011–15	China (multisite)	60+ vs. 0–19	2.2	1.5	3.1
Gesesew	CD4<350 or AIDS	4900	3268	67%	2003–15	Ethiopia (single site)	50+ vs. 15–24	0.4	0.3	0.6
Fomundam	CD4<350 or AIDS	8138	4817	59%	2014–15	South Africa (35 sites)	50+ vs. <50	1.9	*	*
Levy	CD4<350 or AIDS	356	118	33%	2010–15	Israel (single site)	>50 vs. <50	2.4	1.1	5.0
Raffetti	CD4<350 or AIDS	19391	10471	54%	1985–2013	Italy (multisite)	55+ vs.<25	7.5	6.1	9.2
Brannstrom	CD4<350 or AIDS	575	334	58%	2009–12	Sweden (12 sites)	>50 vs.<30	4.0	2.1	7.6
Kesselring	CD4<500	702	442	63%	2013–15	Canada (multisite)	continuous variable	1.0	1.02	1.05

Older age was variably defined, sometimes as young as “35 years or older”, but older age (usually defined as ≥50-years-old) was consistently associated with delayed presentation. Relative risk (typically measured using adjusted odds ratios but in some cases we calculated unadjusted odds ratios from data provided) for delayed presentation associated with older compared to younger individuals (variably defined as <35 or <20 years) ranged from 1.1–7.4. The most common odds ratios were from 1.5–4.

Only one study that considered the role of age in late presentation concluded that older individuals were at decreased risk of late presentation for care. Gesewew et al. studied 4,900 people presenting for care at a single site in Southwestern Ethiopia and found that, compared to those 15–24 years of age, those 50 years and older were less likely to experience a delayed presentation (HR 0.4; 95% CI 0.3–0.6)(10). Another study conducted in Italy separated Italians from non-Italians and found that, compared to those 35–49 years of age, Italians >50 years of age were at increased risk (HR 1.5; 95% CI 1.4–1.7) but non-Italians 50 years of age were not (HR 0.9; 95% CI 0.7–1.2)(11).

(12)Some of these studies considered whether there had been opportunities for earlier diagnosis and whether these differed by age(12–16). These opportunities were variably defined from as broad as “any prior medical encounter” to very specific as “diagnosis with an AIDS defining condition”. These studies documented more “missed opportunities” among older individuals.

IeDEA Data

To add a more recent and standardized accounting of delayed presentation for HIV care around the world, we present data from the International epidemiology Databases to Evaluate AIDS (IeDEA) Global Consortium. IeDEA harmonizes data on care and treatment of people with HIV from seven international regional data centers including four in Africa, and one each in Asia-Pacific (which includes an Australia sub-cohort), Central/South America (also includes Mexico, Haiti, Honduras), and North America (United States and Canada). Each region contributed aggregated data from adults (≥18 years old) to the Epidemiology and Biostatistics Core of the North American AIDS Cohort Collaboration on Research and Design (NA-ACCORD), the North American region of IeDEA, where the figures presented were created. Cohorts in most regions have an ongoing process of adding all individuals presenting for HIV care, with the exception of cohorts in Southern Africa and the Asia Pacific. In the Southern African IeDEA region, participants enter into observation at ART initiation which may occur after presentation for HIV care; this region did not contribute to data visualizations of those presenting for HIV care. Asia-Pacific data combine two approaches to cohort enrollment – selectively enrolling patients to replace participants who died, were transferred, or were lost to follow-up (including all Australian sub-cohort sites), or enrolling all patients seen at the site. The results presented may not be representative of all persons in HIV care in the specified regions of the world as the IeDEA regional cohorts are observational and do not employ sampling strategies for representativeness. Additional information regarding selection of participants for enrollment into the IeDEA regional cohorts, the adoption of the Treat All guidelines, and the changes in CD4 testing that influence the results presented can be found in Supplement Table 1 (pages 1–3 of the supplement) and a recent global IeDEA study (17).

Three study populations were defined. First, the population of individuals observed to present for HIV care at an IeDEA-contributing clinical care site was restricted to those who did not have prior evidence of an HIV care visit, a history of antiretroviral therapy, or a suppressed HIV viral load. Second, the population of individuals observed to be in HIV care in any calendar year from 2000 to the most recent data available for the region was restricted to those who were receiving ART, had a CD4 or HIV RNA measurement, or had evidence of an HIV care encounter. Third, the study population of individuals presenting for HIV care were further restricted to those who were observed to initiate ART at, or after, presentation for care.

Age was measured from year of birth. Sex was defined as sex at birth. The CD4 count closest to the date of presentation for HIV care measured within +/− 12 months and no more than 7 days after ART start was selected for this analysis. For the CD4 at ART initiation, we used a window of 12 months prior through 7 days post ART start to select the closest measurement for this analysis.

Histograms were created for each region to visualize the age distribution at presentation for care, and in the most recent complete calendar year of data available among those who were in HIV care. A kernel density smoothing bandwidth of 2.0 was used to visualize the age distribution histograms. We quantified the difference between the observed medians and the kernel density median estimate (which is not necessarily equivalent to the observed medians). Animated age distributions that visualize these changing age distributions over the last two decades can be found on the IeDEA YouTube Channel (https://www.youtube.com/channel/UC9cfdwlBl944eQ9AGj1EOkw). The proportion of adults presenting for HIV care was estimated within age groups (<50, 50–64, and 65+ years) among the total presenters for HIV care.

In 2013, the World Health Organization recommended viral load testing (and not CD4 testing) to monitor virologic failure after ART initiation(17–20). In 2018, the President’s Emergency Plan for AIDS Relief (PEPFAR) reduced their support for CD4 testing to prioritize viral load monitoring(19). IeDEA has previously shown a decline in pre-ART CD4 testing after adoption of Treat All policies that is steeper in low- and middle-income countries than in high-income countries(17). Trends in median and interquartile range of CD4 count at presentation for care and at ART initiation were stratified by age at presentation for care (< and ≥50 years) to the calendar year through which complete data were available in each region.

Adults presenting for HIV care who had a CD4 count <350 cells/μL at presentation for care were considered “late presenters.” The proportion of late presenters was estimated within each age category (<50, 50–64, and 65+ years) for those presenting for care in the most recent complete calendar year of data available.

The most recent, complete calendar years of data contributed by each IeDEA region were as follows: North America: 2018; Central and South America and the Caribbean: 2019; Central Africa: 2019; East Africa: 2019; West Africa: 2017; Southern Africa: 2017; Asia-Pacific: 2019 (Australia sub-cohort: 2016).

Age in IeDEA Regions

The proportion of adults in HIV care who are ≥50-years-old is substantial throughout IeDEA regions ranging from a low of 17% in Southern Africa to 50% in North America (Figure 1). The proportions of women and men in care who are ≥50-years-old are similar in North America and the Central and South America and Caribbean regions; however, there is a lower proportion of older women in care (compared to men) in the African and Asia-Pacific (including the Australian sub-cohort) regions.

Figure 1.

Age at presentation for HIV care and age among all those in HIV care in IeDEA regions (A) North America region (2018). (B) Central America, South America, and the Caribbean region (2019). (C) Central Africa region (2019). (D) East Africa region (2019). (E) West Africa region (2017). (F) Asia-Pacific region (2019). (G) Asia-Pacific region Australia subcohort (2016). (H) Southern Africa region (2017). In the IeDEA Southern Africa regional cohort, participants were observed from ART initiation (not from presentation for HIV care); age at ART initiation is believed to be reflective of age at presentation for HIV care as of 2017 when the Treat All guidelines were adopted in Southern Africa. The age at presentation for HIV care in the Australia subcohort of IeDEA Asia-Pacific in 2016 is not presented because the median age at presentation for HIV care was based on a relatively small subpopulation (<20 participants) of those presenting for HIV care. Participants were recruited to replenish the Australian subcohort in 2016. ART=antiretroviral therapy. IeDEA=International epidemiology Databases to Evaluate AIDS.

A concerning proportion of adults were ≥50-years-old at initial presentation for care: 24% in the North America region; 11% in Central and South America and the Caribbean; 13% in Central Africa; 12% in East Africa; 19% in West Africa;16% in Asia (excluding Australia). The proportion of older adults (≥50-years) initiating ART in Southern Africa was 8% in the Treat All era (Figure 1 and Table 2). Differences in the proportion presenting for care at older ages (≥50-years-old) in women vs. men also varied by region: 32% vs. 22% in the North America region; 16% vs. 10% in Central and South America and the Caribbean; 12% vs. 15% in Central Africa; 10% vs. 15% in East Africa; 17% vs. 26% in West Africa; 15% vs. 16% in Asia (excluding Australia); and 7% vs. 9% at ART initiation in the Treat All era in Southern Africa.

Table 2:

Age at presentation for HIV care or ART initiation, IeDEA regions

IeDEA Region	Range of the number presenting for care or ART initiation	% <50 at presentation or ART initiation	% 50–64 at presentation or ART initiation	% 65+ at presentation or ART initiation
Presenting for care
North America (2018)	1,000–1,500	76%	21%	3%
Central and South America & the Caribbean (2019)	1,000–1,500	89%	10%	1%
Central Africa (2019)	1,500–2,000	87%	11%	2%
East Africa (2019)	10,000–10,500	88%	10%	2%
West Africa (2017)	1,000–1,500	81%	17%	2%
Asia-Pacific (2019)	500–1,000	84%	14%	1%
Initiating ART (in the Treat All era)
Southern Africa (2017)	22,000–22,500	93%	7%	1%

Footnotes:

In the IeDEA Southern Africa regional cohort, participants are observed at ART initiation (as opposed to at presentation for HIV care) and then followed forward in time; age at ART initiation is believed to be reflective of age at presentation for HIV care as of 2017 when the “Treat All” guidelines were adopted in Southern Africa.

Estimates of age at presentation for HIV care are not presented for the Australia sub-cohort of the IeDEA Asia-Pacific region. Participants were recruited to replenish the Australian sub-cohort in 2016; the median age at presentation for HIV care is based on a relatively small sub-population (<20 participants) of those presenting for HIV care at participating IeDEA clinics. Presenting estimates in these age groups would involve subgroups of <5, which breaches confidentiality arrangements.

While the differences vary by IeDEA region, in nearly every region, PWH ≥50-years-old are presenting with lower CD4 counts than their younger counterparts (Figure 2). Even more concerning, in many regions (Central and South America and the Caribbean, Central Africa, East Africa, and Asia-Pacific Region), the gaps are widening over time as the average CD4 count at presentation rises in younger adults presenting for care.

Figure 2.

Median CD4 cell count per year by age group at presentation for HIV care and at ART initiation in IeDEA regions (A) North America region (2018). (B) Central America, South America, and the Caribbean region (2019). (C) Central Africa region (2018). (D) East Africa region (2019). (E) West Africa region (2017). (F) Asia-Pacific region (2019). (G) Asia-Pacific region Australia subcohort (2016). (H) Southern Africa region (2017). The total number of people contributing to the estimates in each year, by age (18–49 years and ≥50 years) are noted for the calendar years labelled on the x axis. The plots for Central Africa end in 2018 due to the decrease in CD4 cell count measured in 2019 (the last complete calendar year of data available from the Central Africa region). Estimates of CD4 cell count at presentation for HIV care are not available for Southern Africa. The Southern Africa regional cohort observes participants starting from ART initiation; age at ART initiation is believed to be reflective of age at presentation for HIV care as of 2017, when the Treat All guidelines were adopted in Southern Africa. In the Australia subcohort of the IeDEA Asia-Pacific region, participants were recruited into the clinical cohort to replenish the cohort; the median age at presentation for HIV care was based on a subpopulation (<20 participants) of those presenting for HIV care at participating IeDEA clinics between 2000 and 2016. Breaks in the line representing CD4 cell count at ART initiation among those older than 50 years old shows that no individuals older than 50 years initiated ART in that year. The y axis for CD4 count is different for the Australian subcohort compared with the other regions. ART=antiretroviral therapy. IeDEA=International epidemiology Databases to Evaluate AIDS.

Finally, recent IeDEA data (Table 3) support findings from the structured review of the literature (Table 1). Compared to those less than 50 years of age, those ≥50-years-old are substantially more likely to experience late presentation for care. In most regions, the majority of those ≥50-years-old present late to care.

Table 3:

Late presentation (CD4 <350 cells/μL) for HIV care, by age, in the most recent complete calendar year of data available, IeDEA regions

IeDEA Region	Range of the number of late Presenters (CD4<350 cells/μL)	% of <50 years old who were late presenters	% of 50-to-64-year-olds who were late presenters	% of 65+ years-old who were late presenters
Presenting for HIV care
North America (2018)	500–1,000	38%	42%	47%
Central and South America & the Caribbean (2019)	1–500	49%	61%	60%
Central Africa (2019)	1–500	52%	57%	25%
East Africa (2019)	1,500–2,000	54%	67%	50%
West Africa (2017)	500–1,000	63%	62%	64%
Asia-Pacific (2019)	1–500	69%	81%	75%
Initiating ART (in the Treat All era)
Southern Africa (2017)	4,500–5,000	55%	62%	50%

Footnotes:

Estimates of CD4 at presentation for HIV care is not available for Southern Africa. In the IeDEA Southern Africa regional cohort, participants are observed at ART initiation (as opposed to at presentation for HIV care) and then followed forward in time; age at ART initiation is believed to be reflective of age at presentation for HIV care as of 2017 when the “Treat All” guidelines were adopted in Southern Africa. Estimates of CD4 at presentation for HIV care are not presented for the Australia sub-cohort of the IeDEA Asia-Pacific region. Participants were recruited to replenish the sub-cohort in 2016; the median age at presentation for HIV care is based on a relatively small sub-population (<20 participants) of those presenting for HIV care at participating clinics. Presenting estimates would involve subgroups of <5, which breaches confidentiality arrangements.

Discussion

Increasingly, older people are presenting for HIV care. Some of these individuals were recently infected, but a disproportionate number of them have experienced a substantial delay in diagnosis. While it is known that CD4 counts decline with age among uninfected individuals(21), these disparities in CD4 count at presentation are unlikely to be explained by the biology of aging alone. This is especially true since the gap appears to be widening in much of the world as the CD4 count at presentation is increasing at a faster rate among younger adults who are often targeted for test-and-treat strategies. Further, a natural decline in CD4 counts and the phenomenon of accentuated aging with HIV (paper 2 in series “Biologic Ageing in PWH) only underscores the need for earlier diagnosis and treatment for older individuals.

We are concerned that a troubling cycle may be developing. The world’s population is experiencing increased life expectancy in general, increasing the absolute number of older individuals(22). With increased life expectancy, older individuals are continuing to enjoy sexual activity(7, 24, 25)(7, 23, 24) which may be both intra and cross generational(6, 7). Many older individuals also continue to use alcohol and other substances(25, 26). Substance use, age-associated erectile dysfunction, and women being beyond child-bearing age all contribute to inconsistent use of condoms(27, 28), increasing opportunities for HIV transmission. This is concerning because we know that older PWH have delayed presentation for HIV treatment compared with younger PWH, prolonging the period in which they may expose others to infection. Delayed presentation also decreases their ability to benefit from early ART initiation(8, 29). Increased HIV incidence among older individuals further increases prevalence and the cycle continues. It is time to tailor language and mediums of communication to reach older individuals more effectively with HIV prevention, diagnosis, and treatment interventions.

We need to implement interventions specifically targeting older individuals. Many of these interventions require health system if not national government involvement (Paper 3 in series, “How health systems can adapt to an ageing population of PWH”). No single intervention will fix this problem. Each country and health system will need to consider which of these interventions are most cost effective in their setting:

• Expansion of universal HIV screening

• HIV self-testing

• Routine clinical discussion of sexual health and substance use

• Improved recognition and response to HIV indicator conditions

• Use of electronic decision support to prompt and facilitate HIV testing

• Discussion of pros/cons of PrEP among older adults at-risk for HIV

We discuss each of these in turn recognizing that their feasibility will need to be determined based upon local resource constraints.

Expansion of universal HIV screening

Universal screening has the advantage that it does not require identification of risk and compliance can be easily assessed. Cost-effectiveness studies, using a QALY threshold of $50,000, suggest that screening is justified in any population with a threshold of ≥0.1% undiagnosed HIV prevalence (30–32). Recent work that considered more effective and durable antiretroviral therapy, adoption of test and treat strategies, and a $100,000 QALY standard found routine testing to be cost-effective at diagnostic rates ≥0.01% (33). This threshold is met (or surpassed) among those ≥65-years-old in many settings. For example, in South Africa the prevalence of HIV in those 50 and more years (7.1%) easily justifies universal screening(34), yet only 54% of those 50 and more years old reported ever testing for HIV compared to 78% of those 25–49 years of age (34). Further, the cost of HIV screening continues to decrease which could lower the threshold for universal screening in the future. Yet, (36)United States Centers for Disease Control and Prevention guidelines for one-time universal screening remain restricted to those between 13 and 64 years(35).

(31–33)(34)It is time to remove age restrictions on universal screening. When screening regardless of age was implemented in the United States Veterans Health Administration in 2009, new HIV diagnoses were established in 0.14% of 210,957 tested from 2009–12 compared to 0.46% of 89,652 tested from 2006–9 under risk-based testing(36). Overall, those ≥65-years-old did not cross the threshold (65–74 years: 0.07% (95% CI 0.02 – 0.09%) and those ≥75 years: 0.02% (95% CI 0.01 – 0.03%))(36). However, corresponding with societal inequities, some populations are at greater risk than others; there are circumstances where universal screening of those ≥65-years-old is justified. The investigators found that rates of new diagnoses among Black patients aged 65–74 and ≥75 years were 0.16% (95% CI 0.07–0.24%) and 0.09% (0.00–0.19%), respectively. Ten years ago, based on a 0.1% diagnostic threshold, universal screening would have been justified among Black veterans in care and came close to being justified among all veterans in care aged 65–74 years(36, 37). What we would see now if the study was repeated is not known. It is time to find out.

(36)There are special reasons to shift away from risk-based testing for older individuals. By making HIV testing the default, it would be less stigmatizing(38). In many countries, older individuals are not viewed by health care providers, nor do they see themselves, as “at risk”. They may also be concerned that their privacy will not be protected making them less likely to request testing or to present where testing is provided(38). Further, while all sexual minorities face challenges in having frank discussions of risk behavior with their providers, older sexual minorities face the combined stigma of age and sexual minority status(39). Finally, prior studies have convincingly demonstrated the value of “normalizing” HIV testing(40) possibly by including HIV testing as part of an array of tests for common age-associated illnesses.

HIV Self Testing

Nearly 40% of new HIV infections are transmitted by people who don’t know that they are infected in the United States and proportions may be higher in countries where testing is less accessible(35). However, stigma, fear of isolation from friends and family, and poor HIV health literacy is particularly strong among older people with HIV(38) (Paper 4 “Aging as a PWH” in the series). Further, older individuals are more likely to have established linkages to care for other chronic conditions (Paper 3 How health systems might adapt to an ageing population of PWH). While these pre-existing conditions may make it more likely that physicians will misattribute signs of HIV infection it may also mean that linkage to care is less challenging for older individuals.

Making self-testing more readily available might be particularly helpful for older individuals by empowering them to first learn their diagnosis and then choose where to seek care(41). This is particularly true for older individuals who are concerned about privacy and/or are sexual minorities(38, 42). Research has begun to identify ideal characteristics of HIV self-tests(43) and, in Agincourt, South Africa, home testing is already available (38). Similarly, expanding point-of-care accessibility for testing in resource constrained settings makes sense, so long as a clear linkage to care is possible(44).

Routine clinical discussion of sexual health and substance use

Guidelines recommend annual testing for anyone with active risk behaviors (35), but providers are often unaware of ongoing substance use or risky sex among their older patients and rarely ask (45, 46). They are particularly uncomfortable discussing risky sexual behaviors with older people who are sexual minorities(39). One study characterized primary care physician’s response to HIV testing among older patients as, “unnecessary and laughable.” Quoting one provider as saying, “older patients are mostly monogamous, so they are low risk, hence low priority…”(47).

Yet older individuals continue to be sexually active, some with multiple intra and cross generational partners(7) and many continue to use alcohol and other substances with multiple implications for their health and well-being including their risk of HIV infection(23, 26, 48). As lifespan has extended, so has sexual healthspan and ongoing sexual activity into older age(48, 49). In South Africa this is particularly true for men who report continuing to have sex with their wives and with younger unmarried women(7). Further, the cohort of individuals currently aging in upper- and middle-income countries commonly used alcohol and other substances in earlier decades of life and many continue to use these substances as they age, especially alcohol, tobacco, marijuana, and cocaine(26). Injection drug use also occurs but is less common than non-injection use among older individuals.

Providers may feel inhibited about discussing sex with their older patients, but HIV risk is only one of many reasons why providers should ask older patients about their sexual health(23, 24, 48, 49). Older men and women experience challenges to continuing sexual activity including erectile dysfunction for men and vaginal dryness for women, both of which are addressable problems. Erectile dysfunction may make use of a condom difficult if not impossible(27, 28). Further, most welcome discussion of their sexuality with their providers but prefer that the provider raise the issue(23, 48, 49). This provides a nonthreatening and non-stigmatizing means of asking about sexual risk behaviors and HIV status of their partners as well.

Similarly, there are compelling reasons why providers should also ask older patients about alcohol(25) and other substance use. Unhealthy alcohol use is increasingly common among older individuals(26) and has critically important health implications including risk of cancer(50), liver disease(25), metabolic disease(51), interaction with prescription medications(52), risk of falls and fractures(53), and cognitive decline(54). Non-injection drug use including alcohol use increases disinhibition and leads to risky behaviors including sex with multiple partners and unprotected intercourse(55, 56). When disinhibition is combined with erectile dysfunction and a perceived lack of concern regarding pregnancy, condoms are rarely employed. Individuals in New York City using heroin or cocaine were equally likely to test positive for HIV infection whether their use was via injection or other means(57). Along with multiple sexual partners and injection drug use, non-injection drug use, including unhealthy alcohol use, should be considered a risk for HIV infection.

Improved recognition and response to HIV indicator conditions

One approach to earlier detection and treatment of HIV infection has been the use of indicator conditions(58–61). The underlying premise is that certain conditions should be considered indications for HIV testing, regardless of disclosed risk behaviors. These conditions fall into three general categories: indicators of risk behaviors that may be undisclosed, indications of early symptomatic HIV disease, and possible indicators of advanced HIV disease. Identified indicators of undisclosed risk behaviors include viral hepatitis and any sexually transmitted infections. Indicators of possible early symptomatic HIV disease include persistent flu like symptoms, a single episode of bacterial pneumonia, herpes zoster, lymphocytopenia, thrombocytopenia and cervical or vulvar dysplasia (CIN2+ or VIN2+). Indicators of possibly advanced HIV disease include cervical cancer, unexplained neuropathy, weight loss, or dementia—while these should always trigger HIV testing, they often occur ten years after initial infection. Tuberculosis also indicates advanced disease but may occur much earlier.

Unfortunately, indicator conditions that might trigger HIV testing among younger individuals may be attributed to other causes in older individuals. Ten years ago, we conducted a study using the US national Veterans Administration data demonstrating that veterans already in care prior to their HIV diagnosis were no more likely to be diagnosed early in the course of their disease as those newly entering VA care(62). Further, only a minority of these patients had an indicator condition prior to their diagnosis. Recently there has been renewed interest in the use of triggers and these studies have confirmed and extended our findings. These studies underscore that trigger conditions are more common among older individuals, but less commonly prompt HIV testing in this age group(58–61).

Use of electronic decision support to prompt and facilitate HIV testing

There is a practical problem with all the HIV testing strategies we have discussed. All these strategies require individuals who are not focused on HIV or its treatment to consider the possibility of HIV infection, obtain the test, and act on the results.

For many primary care and specialty providers in higher-income countries throughout North America, Europe, and Australia/New Zealand, few things are further from their clinical focus. Even in countries with higher HIV prevalence and greater general awareness, providers may not consider testing older individuals who they deem to be at lower risk. In this context, 20 years of experience with a fully paperless, national, electronic medical record in the US Veterans Healthcare System may offer important insights(47, 63–66). Electronic health record (EHR) clinical reminders may help overcome documented failures of one-time universal screening, and risk based and indicator condition testing.

When effectively implemented and maintained, universal screening facilitates more timely diagnoses of HIV infection. In August 2009, The US Veterans Health Administration (VA) revised its HIV testing policies to promote voluntary routine one-time testing of all adults regardless of age and to streamline testing procedures through a clinical reminder. Streamlining eventually included a transition from requiring written informed consent to verbal consent. These changes tripled the lifetime HIV testing prevalence within the national VA(36).

A multimodal HIV testing intervention was also launched with site-specific study teams consisting of an infectious disease specialist, a primary care team leader, and other stakeholders. The intervention included an electronic clinical reminder, a multifaceted provider activation program, social marketing to providers and patients, regular informal conversations with providers, and quarterly feedback on rates of testing.(47) The proportion of newly diagnosed persons ≥60-years-old increased from 7.5% to 15.3% (p=0.10) and the proportion of patients with CD4 counts <200 cells/μL (well below the more commonly used “late presenter” threshold of <350 cells/ μL) decreased from 43% to 29% (p=0.04). A facility that implemented only the electronic reminder linked to a test order achieved the same improvement in testing as the facility with the full multimodal intervention suggesting that this was the element most critical to success(67). Similarly, clinical prompts could also improve adherence to risk based and indicator condition testing.

For resource limited settings, innovative approaches using solar power(68), cloud-based systems(69), and mobile phone applications(70) for data entry have been developed to support EHRs in the context of intermittent, or non-existent, electricity. These have been successfully applied in Kenya(71), India(72), and other low to middle-income countries(73). They have already demonstrated effectiveness at improving the timing of ART in Kenya(69).

Discussion of pros/cons of PrEP among older adults at risk for HIV

Among those at substantial risk of HIV infection, a frank discussion of the pros/cons of pre-exposure prophylaxis (PrEP), tailored to this age group, is indicated. Importantly, based on studies focused on HIV and non-HIV medications, older individuals are more capable of achieving excellent medication adherence than younger individuals(74). On the other hand, addition of two antiretrovirals (a fixed dose, single-tablet combination of tenofovir (300 mg) and emtricitabine (200 mg)) to a medication regimen that may already cross the line into polypharmacy (≥5 chronic medications) (75) and increased risks of hospitalization and mortality(75) has its downside. Polypharmacy is a growing problem among older individuals(76) and the long term safety of these medications in individuals 65 years of age or older is largely unknown(77).

Tenofovir is associated with nephrotoxicity and is contraindicated for those with a creatinine clearance less than 60 mL/min(78, 79). Tenofovir is also associated with bone loss and may contribute to osteoporosis (78, 79), a particular concern among older individuals, especially women. A careful consideration of what other medications the individual is taking and whether these toxicities might exacerbate those of the other medications is indicated(77).

Further, before initiating PrEP, patients must be tested for HIV since PrEP is not an effective treatment for HIV infection and can lead to viral resistance. While receiving PrEP, patients should be monitored every 3 months for declining renal function, sexually transmitted infections, and HIV infection. All this may seem like too much additional effort to patients who may only have sexual intercourse or use injection drugs intermittently(80).

Momentum is building for “on-demand” PrEP(79, 81, 82). The IPERGAY (Intervention Preventive de l’Exposition aux Risques avec et pour les Gays) randomized MSM to receive pericoital PrEP—two pills between 2 and 24 hours before anal intercourse and one pill daily for two days following sex but not more than 7 pills in a single week. This might substantially curtail concerns about toxicity. While this may be an appealing solution, further work is clearly needed.

Conclusion

Although older individuals more commonly present for HIV care late and have more contact with the healthcare system, few studies have focused on factors associated with late presentation specifically among older individuals. This is important because older age is independently associated with risk for indicator conditions possibly rendering them less informative for detection of undiagnosed HIV infection. As the population of older adults with HIV continues to grow, in-depth studies are needed to inform guidelines for HIV testing and determine how best to implement more wide-spread testing and earlier diagnosis and treatment in this growing age group.

Key Messages.

• Late presentation for HIV care is a major impediment to prevention and effective treatment of HIV infection.

• A growing proportion of adults presenting for HIV care are ≥50-years-old and nearly half of them have delayed presentation.

• In many regions of the world, the age associated gap in CD4 count at presentation is widening as the average CD4 count at presentation rises for younger adults.

• Few studies have focused on factors associated with late presentation specifically for older individuals.

• Early diagnosis and treatment of HIV for older individuals is particularly challenging because early signs and symptoms may be attributed to diseases of aging and because neither these individuals nor their care providers perceive them to be at risk for HIV.

• If the widening age associated CD4 gap is to be addressed, interventions will need to be explicitly targeted to older individuals.