Journal Article

OC8 Did the COVID-19 pandemic shift the landscape of late HIV diagnosis?

Mazzotta, V, Cozzi-Lepri, A, A d’Arminio Monforte, Lucchetti, I, Girardi, E, Mondi, A, Taramasso, L, Roen, A, Sarmati, L, Mussini, C, Micheli, G, Gori, A, Lapadula, G, Antinori, A, Marchetti, G

2025

Overview

BackgroundThe COVID-19 pandemic profoundly reshaped healthcare services, affecting the management of chronic pathologies, including HIV. This study aimed to evaluate the interplay between COVID-19 pandemic and late HIV diagnosis (LD) in Italy.Material and MethodsAll people with HIV (PWH) enrolled in the Icona cohort between January2016 and December2023, diagnosed with HIV within 3 months before enrolment, were included. The incidence of LD (defined as CD4 count<350 cell/mm3 at diagnosis or AIDS-defining event [ADE] regardless of CD4 within 3 months from diagnosis) and AIDS presentation (AIDS-P, ADE within 3 months from diagnosis) were compared between pre-COVID (2016–2019) and post-COVID (2021–2023) period. The difference in prevalence of LD by period and by participants’ socio-demographics was compared using Poisson regression models with an interaction term. Survival analysis was used to assess the effect of LD and AIDS-P on all-cause mortality risk and its interaction with the calendar period.ResultsOverall, 5,081 newly diagnosed PWH were enrolled in the Icona cohort over the study period, of whom 2,928 (58%) were LD and 612 (12%) were AIDS-P. The proportions of LD and AIDS-P remained stable across the pre- and post-COVID period: 1,817/3,172 (57.3%; 95%CI:55.5–59%) and 1,111/1,909 (58.2%, 95CI% :5.9–60.4%) for LD and 372/3,172 (11.7%; 95%CI:10.6–12.9%) and 240/1,909 (12.6% 95%CI:11.1–14.1%) for AIDS-P, respectively [figure 1]. LD were more likely to be women, elderly, migrants, and heterosexuals (compared to MSM), and to have lower educational level. Risk patterns for LD remained consistent pre- and post-COVID, although there was weak evidence that the risk of LD in heterosexuals vs. MSM weakened in post-COVID (interaction p-value=0.38, table 1). Over the follow-up, a total of 138 deaths occurred: 114 in pre-COVID and 24 in post-COVID period (41 and 73 for ADEs; 16 and 8 for not-AIDS related events). At survival analysis, LD/AIDS-P had a significantly higher probability of all-cause mortality both in the pre-COVID (3-year risk [95%CI]: non-LD (nLD) 0.94% [0.38–1.49%], LD 4.3%[3.3–5.3%], AIDS-P 9.6% [6.8–13%],p<0.001) and post-COVID (3-year risk [95%CI]: nLD 0.18%[0.02–0.52%], LD 2.7%[1.6–3.9%], AIDS-P 7.5%[3.4–11.5%], p<0.001). After controlling for confounding, LD and particularly AIDS-P were confirmed to be at higher risk of death compared to nLD, with some modest evidence that this excess in risk was exacerbated in the post-COVID period (aHR LD vs. nLD and AIDS-P vs. nLD: 13 and 49.9 in the post-COVID vs 3.56 and 7.07 in the pre-COVID [figure 2, interaction p>10%]).Abstract OC8 Figure 1Proportion of late HIV diagnosis and AIDS presentation by period[Figure omitted. See PDF]Abstract OC8 Table 1Prevalence risk ratio of late HIV diagnosis by socio-demographic characteristics by fitting univariate Poisson regression modelAbstract OC8 Figure 2Hazard ratio of all-cause mortality stratified by late diagnosis and period by univariable and multivariable Cox regression models[Figure omitted. See PDF]ConclusionsOur data shows that the frequency of LD remained stable before and after COVID-19, regardless of participants’ socio-demographics. LD was confirmed to be associated with an increased mortality risk, with evidence suggesting that this risk may have been more pronounced in the post-pandemic period. Longer follow-up is needed to further validate these findings.

• Publisher: BMJ Publishing Group LTD
• Subject: Acquired immune deficiency syndrome / AIDS / COVID-19 / Demographics / Health risks / Health services / Heterosexuality / HIV / Human immunodeficiency virus / Medical diagnosis / Mortality / Pandemics / Regression analysis / Sociodemographics / Survival analysis