To the Editor Dr Matta and colleagues1 importantly recommend that those with persistent COVID-19 symptoms be adequately evaluated for conditions that may mimic post-acute sequelae of COVID-19. In arriving at this conclusion, they present results that replicate ours2: the only long-term symptom correlated with positive SARS-CoV-2 serology is anosmia. However, the evidence suggests that anosmia may be more helpful in assessing one’s likelihood of positive serology results than one’s history of COVID-19 infection. We have previously shown that those reporting persistent COVID-19 who are untested or who have tested negative have symptom time courses—with exception of change in smell and/or taste—that overlap with those who have had a positive result on reverse transcriptase–polymerase chain reaction (RT-PCR), antigen, or antibody test. We interpret this to mean that the clinical syndrome in people reporting persistent COVID-19 with positive, negative, or untested status, is the same.2 The authors1 make a key error in assuming that serology has good sensitivity for history of SARS-CoV-2 infection in all populations. Indeed, male sex and hospitalization were found to be predictors of greater antibody titers.3 Additionally, though participants were asked to indicate whether they thought they had been infected since March 2020,1 the serology results were obtained between May and November of 2020, suggesting that the time from the RT-PCR positive finding was significantly greater than the mean of 44 days in the study assessing sensitivity of the antibody assay.4 As mentioned by the authors, antibodies wane with time and are less likely to be found in those with persistent symptoms.1 There is additionally no mention of statistical methods used to account for extremely large differences in between-group number of patients in the cohort, which can invalidate the results of the logistic regression analysis. Crucially, nonseroconversion after SARS-CoV-2 infection (that is, failure to test positive for antibodies at any time after a positive RT-PCR test result) occurs in an estimated 24% of cases.5 Clearly, a negative serology finding is insufficient to rule out prior COVID-19 illness. Rigor must be applied to competing differential diagnoses, particularly those that are impossible to objectively disprove, such as some psychiatric conditions. In the absence of sufficient objective measures of prior infection, the patient’s clinical presentation should be the deciding factor when determining likelihood of prior infection. Further, a more important clinical question is why those with clinical syndromes consistent with persistent COVID-19 appear less likely to have sustained antibody responses.