> 10 studies were eligible for our systematic review. The weighted average risk reduction against reinfection was 90.4% with a standard deviation of 7.7%. Prote

> 10 studies were eligible for our systematic review. The weighted average risk reduction against reinfection was 90.4% with a standard deviation of 7.7%. Protection against SARS-CoV-2 reinfection was observed for up to 10 months. Before people dismiss these results for the typical reasons (eg it's a pre-print, etc): 1) Basically every peer-reviewed, large scale, and long-term serological study has demonstrated that immunity acquired through natural infection has been at least as effective as vaccination in preventing reinfection (eg take a look at the citations in OP). Mysteriously, it appears that none of the CDC's publications acknowledge, cite, or discuss discrepancies with any of this reputable literature. 2) The absolute risks of vaccination causing hospitalization and death are extremely low. For a vast majority of the population, the absolute risks of SARS-CoV-2 infection causing hospitalization and death are also extremely low. Folks here love to do armchair calculations of risks - or cite odds ratios from one study - to defend their view. But if you aren't stratifying by age and health then you will misrepresenting the true absolute risks for the majority of the population. 3) After infection your antibody levels will eventually drop below detectable levels, but thanks to long-lived bone marrow plasma cells and memory B cells and T cells, you are likely to retain effective protection against severe outcomes. 4) Variants of concern like delta will continue to accumulate mutations that confer a certain degree of immune escape, which will undermine protection in both vaccinated and naturally immune individuals. Comparing the effectiveness between both types of immunity (and blends of the two) is an ongoing area of research. Preliminary data from Israel suggests that natural infection may confer a more robust immune response that offers better protection when challenged by such variants (see Gazit et al., 2021, as cited in OP).