It has been observed that lockdown restrictions during COVID-19 pandemic may have had a negative impact on HIV epidemic goals with disruption in care. We aim to analyse the trends in non-viral suppression for PLWH during and after the lockdown for COVID-19 pandemic in Italy compared to 2019.
We included all participants in the ICONA cohort for whom there was ?1 viral load (VL) in the window Nov 2019-Jan 2020 and with most recent VL?50 copies/mL (exposed to lockdown), and over Nov 2018-Jan 2019 (not exposed). New enrolments in the study period were excluded. At population level and separately by year, we calculated proportion with VL?50 copies/mL at each month over March-September and we performed an intermittent time series (ARIMA) model centred in March. In addition, we defined an individual outcome using the first VL over May-September (>50 vs. ?50 copies/mL), comparing proportion with VL>50 copies/mL between exposed and not exposed by means of logistic regression models. PLWH with missing VL in the outcome window were excluded from the analysis. We also performed an alternative analysis in which censoring bias was minimised using inverse probability of weighting. Sensitivity analyses were performed after restricting to clinical sites with electronic linkage with laboratory data and to the subset of PLWH under follow-up in both years.
A total of 3,684 PLWH were included (2019=2,948; 2020=736). PLWH exposed to lockdown were significantly older, less frequently MSM, non-Italian, had a higher CD4+ count and more frequently resident in north of Italy. The mean proportion of VL<50 copies/mL was 97% at March 2020 (ref.), 99% before March 2020, 82% at April 2020 (ARIMA estimates -21% 95% CI:-28%;-14%; P=0.01) and 97% after April 2020. In the 2019, the same proportions were 100%, 98%, 95%, and 97% with evidence for a lower drop in April (-6%, 95% CI:-8%;-3%, p=0.02). The results of the logistic regression model are reported in Table 1. When restricting to sites with electronic VL linkage and to those followed-up in both years the IPW OR of 2020 vs. 2019 were 1.23 (0.69-2.18) and 1.03 (0.48-2.19), respectively.
We found little evidence for a difference in the proportion of PLWH with a VL>50 copies/mL, following stable suppression, in the period post lockdown due to COVID-19 as compared to the previous year. Although selection bias was minimized, reasons for a missing VL should be further investigated.
