Background: Parasite prevalence is a key metric used to quantify the burden of malaria and assess the impact of control strategies. Most published estimates of parasite prevalence are based on microscopy and likely underestimate true prevalence.

Methods: Thick smear microscopy was performed in cohorts of children (aged 6 month to 10 years) and adults every 90 days over 2 years, at three sites of varying transmission intensity in Uganda. Microscopy-negative samples were tested for sub-microscopic parasitaemia using loop-mediated isothermal amplification (LAMP). Generalized estimating equation models were used to evaluate associations between age and parasitaemia, factors associated with sub-microscopic infection and associations between parasitaemia and haemoglobin.

Results: A total of 9260 samples were collected from 1245 participants. Parasite prevalence among children across the three sites was 7.4, 9.4 and 28.8 % by microscopy and 21.3, 31.8 and 69.0 % by microscopy plus LAMP. Parasite prevalence among adults across the three sites was 3.1, 3.0 and 5.2 % by microscopy and 18.8, 24.2 and 53.5 % by microscopy plus LAMP. Among those with parasitaemia, adults and persons recently treated with anti-malarial therapy had the highest prevalence of sub-microscopic infection. Children with sub-microscopic or microscopic parasitaemia had lower mean haemoglobin levels compared to children with no detectable parasites.

Conclusions: Across a range of transmission intensities in Uganda, microscopy vastly underestimated parasite prevalence, especially among adults.