From October 1, 2020 to April 15, 2021, 1016 cases tested positive for SARS-CoV-2 through UT’s voluntary testing program. Four cases were hospitalized and excluded from the study. All cases were attempted to be investigated; 895 (88%) were successfully investigated and of those, 656 cases reported contacts. There were 287 index cases with at least one contact who: (1) tested 3–14 days following the exposure; (2) had no other known exposures to COVID-19 in the 14 days prior to their test; and (3) had no known history of SARS-CoV-2 infection. Twenty-five cases were excluded because either case or contact in a pair had received at least one dose of a COVID-19 vaccine by the time of exposure. Following manual inspection of case-contact pairs, 50 cases were excluded due to uncertainty about the direction of transmission, resulting in 212 index cases for analysis.
Index cases were predominantly students with a median age of 21 years (Table 1). Approximately, 56% of cases were female, 14% had a comorbidity, and 79% had presented with symptoms by the time of investigation. Among cases with symptoms, 64% reported congestion or runny nose, 43% reported cough and 89% tested on or after the day of symptom onset. Cycle threshold values of the three genes used to detect SARS-CoV-2 were highly correlated (Fig. 1). Viral loads were estimated using the ORF1ab gene given that it was consistently detected among all test-positives and less likely to vary among variants containing mutations on the spike protein. The median viral load among index cases was 5.6 log10 copies of RNA per mL of saliva (range = 1.8–10.4). A small proportion of cases (4%) had viral loads > 8 log10 RNA copies per mL. Only one case (< 1%) had a viral load > 9 log10 copies per mL. Cases reporting one or more comorbidities also had higher viral loads compared to those with no comorbidities (5.83 log10 copies per mL vs. 5.31 log10 copies per mL, t = 2.05, p = 0.04, Additional file 1: Figure S1). Cases presenting with cough on the day of exposure had significantly higher viral loads compared to cases with no cough (5.79 log10 copies per mL vs. 5.21 log10 copies per mL respectively, t = 2.5, p = 0.01, Additional file 1: Figure S2). No significant differences in viral load were found between cases with and without congestion or runny nose (5.59 log10 copies per mL vs. 5.42 log10 copies per mL, t = 2.03, p = 0.05, Additional file 1: Figure S3).
There were 365 close contacts that met inclusion criteria. Most cases (62%) were associated with only one close contact who met the inclusion criteria (Fig. 2). Contacts ranged in age from 16 to 92 years (median = 21 years) and were predominantly female (54%); 15% reported comorbidities and 89% were students, staff or faculty at UT.
Just over half of all contacts were exposed to an asymptomatic or pre-symptomatic case (Table 2). The majority of exposures were indoors (90%), lasted greater than 30 min (90%), involved no mask use by the case and contact (83%) and non-physical contact (87%). Approximately 41% of contacts were household contacts.
Out of 365 contacts, 343 (94%) tested for SARS-CoV-2 using a nucleic acid amplification test, 12 (3%) tested using an antigen test and 10 (3%) had an unknown test type but were able to provide the date and location of their test. Among all contacts that tested, 70 (19%) tested positive for the virus; 36 (51%) were exposed to an asymptomatic or pre-symptomatic case and 54 (77%) were exposed to a case with viral load ≥ 5 log10 copies per mL. The proportion of contacts that tested positive increased monotonically with viral load of the index case at the time of test (12%, 23% and 25% corresponding to < 5, 5–8 and > 8 log10 copies per mL, respectively; X2 = 7.18, df = 2, p = 0.03; Fig. 3). The proportion of contacts that tested positive was highest after exposure to cases with cough compared to cases with no report of cough (35% vs. 16%, respectively; X2 = 9.85, df = 1, p = 0.002). The difference in test positivity of contacts by congestion or runny nose of index case was not statistically significant (Additional file 1: Table S1). When exposures were restricted to those outside the home (N = 190), test positivity significantly increased with viral load (11%, 27% and 33% corresponding to < 5, 5–8 and > 8 log10 copies per mL, X2 = 7.04, df = 2, p = 0.03, Additional file 1: Table S2). In this restricted sample, test positivity was higher among those exposed to cases with cough compared to cases without cough (44% vs 19%, df = 1, X2 = 5.08, p = 0.02). The majority of exposures occurred within 2 days of the case’s test (Additional file 1: Figure S4). When exposures were restricted to those on the day of the index case’s test (N = 75), test positivity was higher among contacts exposed to cases with viral loads 5–8 log10 copies per mL compared to contacts exposed to cases with viral loads < 5 log10 copies per mL (28% vs. 3%, respectively) and among contacts exposed to cases with cough compared to cases without cough (43% vs 8%, df = 1, X2 = 8.34, p = 0.004).
The risk of transmission to a close contact was significantly associated with viral load (adjusted RR = 1.27 per log10 increase in viral load, 95% CI 1.22–1.32, Table 3). Risk of transmission was also associated with the number of days between test and exposure (adjusted RR = 1.14, 95% CI 1.11–1.17), the index case presenting with cough on the day of exposure (adjusted RR = 1.41, 95% CI 1.04–1.92) and physical contact (adjusted RR = 1.78, 95% CI 1.26–2.53). These associations remained statistically significant when viral load was categorized as < 5, 5–8 and > 8 log10 copies per mL (Additional file 1: Table S3). Little to no correlation was found between risk factors in the final model (Variance Inflation Factor range = 1.1–1.8). Factors not associated with transmission in this population included age, presentation of any symptom, mask use, duration of exposure, location of exposure and household contact.