**2. Results**

#### *2.1. Reported COVID-19 Cases in Guangdong Province between 19 January and 1 July 2020*

Between 19 January and 1 July 2020, Guangdong recorded 1961 laboratory-confirmed SARS-CoV-2 infections, with 1642 symptomatic COVID-19 cases, over three waves of activity that were characterized by distinct infection sources (Figure 1). The first wave, from January 14 to February 29, with a total of 1350 COVID-19 cases, was largely seeded by cases originating from Wuhan or Hubei [9] and remained the largest epidemic wave in the province so far. Cases were reported in 20 of the 21 prefecture-level cities in Guangdong (Yunfu reported no case) and were also highest in cities that had high migration activities from Wuhan in the five days prior to the national lockdown (Table 1, Figure 2A). The second outbreak, from March 1 to April 1, involved 161 infections that were mainly imported internationally while the third outbreak from 2 April to 2 May, involved 310 imported and associated-local cases. By 1 July, 1020 (62%) of the COVID-19 cases were reported mainly in Guangzhou and Shenzhen, which were also the major international port-of-entries, while the surrounding cities in the PRD recorded 478 (29%) cases (Table 1). As minimal cases were reported in all other cities after the first wave, the transmission trend across the province during the sera sampling period largely remained the same as in March.

**Figure 1.** Cases of SARS-CoV-2 infections in Guangdong between 19 January and 1 July 2020. Case numbers represent the total of symptomatic and asymptomatic infections. (Local cases: Cases infected in Guangdong or imported from other provinces in China; Internationally imported cases: Individuals with SARS-CoV-2 infections returned from overseas; Internationally imported associated cases: Local cases identified as being associated with the internationally imported cases.).

#### *2.2. Seroprevalence in Guangdong*

Between 11 March and 24 June 2020, 14,629 sera were collected from 983 institutions across Guangdong. A total of 5264 (36%) and 9365 (64%) sera were collected from the low and high-risks cities, respectively. Large cities in high-risks area were generally better sampled while samples in the 10 to 19 years old group were generally under sampled, particularly in the-low risk cities (Table S2).

Out of 14,629 sera, we identified 21 (0.14%) samples positive for SARS-CoV-2 IgG by magnetic particle based chemiluminescent enzyme immunoassay (CLIA). We calculated that in Guangdong province overall, the estimated age and sex-weighted seroprevalence was 0.15% (95% CI 0.06% to 0.24%) (Figure 3A, Table S3). The weighted seroprevalence in high-risk cities was 0.19% (95% CI, 0.06% to 0.33%) (Figure 3B), approximately 2.7-fold higher than the weighted seroprevalence for low-risk cities of 0.07% (95% CI, 0% to 0.24%) (Figure 3C). In the whole of Guangdong, the lowest seroprevalence was detected in the youngest age-group ≤9 years old (0.07% (95%CI, 0.01% to 0.24%), while the seroprevalence estimates in the other age-groups were between 0.13% to 0.22% (Figure 3A). We noted apparent differences in the age-specific trends of seroprevalence estimates between the low and high risks region in Guangdong. In high-risk cities, age-specific seroprevalence was lowest in children ≤9 years of age, highest in adolescents, and lower in the three adult age groups (Figure 3B). In the low-risk cities, age-specific seroprevalence was higher in children 9 years old and older adults ≥40 years old, and lower in younger adults (Figure 3C).

**Table 1.** Demographic of the twenty-one prefecture-cities in Guangdong; number of confirmed COVID-19 cases between January 19 to March 3, representing the first COVID-19 wave in Guangdong; and between January 19 to July 1, representing the first 6-months post COVID-19 emergence in Guangdong.


<sup>a</sup> Does not include asymptomatic cases. <sup>b</sup> Source: Based on 2018 population data (Guangdong Provincial Bureau of Statistics). <sup>c</sup> 0 = low migration activity with Wuhan, 1 = high incoming migration from Wuhan, 2 = high outgoing migration to Wuhan. All data was for the period between 18 to 22 January 2020. \* Prefectural cities in the Pearl River Delta.

> In terms of geographical distribution, the seropositivity correlated with the number of reported COVID-19 cases in each city, with the notable exception of Guangzhou. Despite reporting the highest number of reported COVID-19 cases, it had the lowest seropositivity at 0.08% out of 2520 tested samples (Figure 2B, Table S2). This could be due to the high numbers of imported-associated cases that were detected during border screening and were subsequently quarantined at centralized facilities until determined to be PCR-negative, which effectively reduced the risk of virus spreading. In the low-risk region, seropositive samples were detected in Qingyuan (*n* = 1, 0.09%), Jiangmen (*n* = 1, 0.50%) and Shantou (*n* = 3, 0.60%). These three were amongst the low-risk cities in Guangdong that reported cases during the first wave and had high migrant connectivity with Wuhan (Table 1).

**Figure 2.** (**A**) Reported confirmed COVID-19 cases based on local official surveillance data in the different prefectural cities in Guangdong province. Asymptomatic infections were not available at a prefectural-city level. (**B**) Seropositivity of antibodies to SARS-CoV-2 as identified from the present study, in prefectural cities of the Guangdong province within the first six-months post COVID-19 emergence. Cities that had relatively higher connectivity with Wuhan prior to 23 January 2020, were underlined and highlighted in blue.

**Figure 3.** Estimates of age-specific SARS-CoV-2 seroprevalence in Guangdong province in (**A**) all cities or stratified by (**B**) high or (**C**) low risk of COVID-19 activities according to local official surveillance data. Seroprevalences in all ages were age- and sex-weighted according to the population structure of the included cities.

## *2.3. Proportion of Seropositive Samples with Neutralizing Titers*

We tested the SARS-CoV-2 IgG positive samples for neutralization titers with the pseudovirus neutralization (pN) assay using a construct expressing the Spike (S) protein. Of the 21 samples, 14 (67%) had detectable neutralization activity at titers >20 (Table S4). The seven samples that did not have detectable neutralization activity had signal to cut-off readout (S/CO) that ranged from 1.002 to 2.442. There was no significant correlation between IgG-titer (expressed by the S/CO readout) with the measured IC50 titer (Pearson's coefficient, r = 0.223, *p* = 0.33. Figure 4).

**Figure 4.** Correlation between the SARS-CoV-2 IgG titers and pseudovirus neutralization titers, expressed as 50% inhibitory concentration (IC50). The signal to cut-off (S/CO) readout on the X-axis were loge transformed to aid visualization of data points.

#### **3. Discussion**

In collaboration with a clinical testing laboratory, we were able to use residual serum samples that were submitted for clinical tests to conduct a cross-sectional serosurvey across the expanse of Guangdong province shortly after the emergence of COVID-19 in January 2020. Using a pseudovirus neutralization assay, we confirmed that 67% of the samples had neutralization titers, suggesting that most of the IgG-positive samples were true-positives. The remaining seven samples may still represent true positives as some infections may not have induced neutralizing antibodies, or their neutralizing antibodies may have waned to below the threshold of detection since being infected [10,11]. Studies have shown that the long-term antibody dynamics, particularly in those mild or asymptomatic COVID-19 cases can be variable [12,13]. The lack of correlation between the SAR-CoV-2 IgG-readout measured by CLIA with the pseudovirus neutralization titer could be due to the assay choices. The CLIA detects IgG against both S1 and N protein whereas our pseudovirus in the neutralization assay expressed only the S-protein and would therefore only account for neutralizing activity afforded by S-specific IgG.

Six months into the pandemic, the seroprevalence estimates based on residual sera collected from a clinical diagnostic laboratory network reported in our study were similar to other studies in the general community and lower than studies of high-risk individuals such as healthcare workers or hospital visitors in China (summarized in Table 2). Notably, two studies that included cohorts from Guangdong reported higher seroprevalence rates than ours. For example, by April 2020, Xu et al. found a seroprevalence of about 4% in healthcare workers or their exposed contacts in Wuhan, compared to approximately 1% in healthcare workers or factory workers in Guangzhou [14]. Separately, Liang et al. reported a seroprevalence of 2.1% and 0.6% amongst the 16,000 hospital patients and visitors in Wuhan and Guangzhou between January 25 to April 30 [15]. Collectively, these studies confirmed that SARS-CoV-2 virus transmission in other areas were minimal compared to

Wuhan, but they did not include any validation tests to confirm the antibody specificity and provided limited information with regard to age-specific seroprevalence.



Our seroprevalence estimate is in line with the results of a large-scale nucleic acid testing conducted by Guangdong Centers for Disease Control and Prevention, which reported low number of PCR-positivity [9,19]. Between January 30 to March 19, they reported 1388 PCR-positive cases out of 1.6 million samples tested (0.089% positivity) [9]. In a follow-up study for the period up to July 9, only 385 samples in over 3.2 million samples tested by third-party institutions, were found to be positive (0.012% positivity) [19]. Notably, they observed changing age trends amongst the positive cases over time. In contrast to the first COVID-19 wave, during which the elderly (≥60 years old) comprised a significant proportion of infected cases (22.2%), PCR-positivity rates in the elderly declined during the subsequent waves (1.9%) but increased in the younger demographic, particularly in those between 20 to 39 years old (59.4%, which increased from 34.4% during the first wave). This was consistent with our data as we did not observe a higher seroprevalence among adults ≥60 years old. This trend was in contrast to those observed in Hubei [16], where evidence of infection was more common in those ≥60 years old. This suggests that cases in Guangdong after the first wave were more likely amongst the young, mobile travelers, particularly in the high-risk cities.

Guangdong was the epicenter of the first SARS outbreak in 2004 and regularly experienced zoonotic avian influenza cases. Consequently, the province had established an efficient surveillance and response system to emerging pathogens. The rapid response in the province appeared to have been effective in controlling the spread and emergence of SARS-CoV-2 locally. However, our seroprevalence trends suggests that the younger and mobile population in the urban centers of Guangdong as well as smaller cities with high connectivity may be a transmission risk and should be monitored. In conclusion, our data suggests an extremely low seroprevalence across Guangdong.
