Ending Zero-COVID is challenging, particularly when vaccine coverage is low. Considering Wallis and Futuna, a French Zero-COVID territory affected by reluctance to vaccination, low immunity and high levels of comorbidities, we investigate how targeted use of nirmatrelvir/ritonavir (brand name Paxlovid) can complement vaccination and non-pharmaceutical interventions (NPIs) and mitigate the epidemic rebound expected when Zero-COVID ends. We developed a discrete age-stratified compartmental model describing SARS-CoV-2 spread and healthcare impact once Wallis and Futuna reopens. It accounts for comorbidity risk groups (CRG), vaccine coverage (2 doses, 3 doses), the effectiveness of vaccines (recent or old injection), treatments and NPIs. In our baseline scenario, cases aged 65+ in intermediate/high CRG and 40+ in high CRG are eligible for treatment. Findings The epidemic is expected to start 13-20 days after reopening with a doubling time of 1·6-3·7 days. For medium transmission intensity (R0=5), 134 (115-156) hospital admissions are expected within 3 months, with no pharmaceutical measures. In our baseline scenario, admissions are reduced by 11%-21% if 50% of the target group receive treatment, with maximum impact when combined with NPIs and vaccination. The number of hospitalisations averted (HA) per patient treated (PT) is maximum when 65+ in high CRG are targeted (0·124 HA/PT), quickly followed by 65+ in intermediate/high CRG (0·097 HA/PT), and any 65+ (0·093 HA/PT). Expanding the target group increases both PT and HA, but marginal gains diminish. Modelling suggests that test and treat may contribute to the mitigation of epidemic rebounds at the end of Zero-COVID, particularly in populations with low immunity and high levels of comorbidities.