Asymptomatic SARS-CoV-2 Infection in Havana, Cuba, March–June 2020: Epidemiological Implications

INTRODUCTION The percentage of asymptomatic COVID-19 cases worldwide is estimated at 18–50%; 53% in Cuba speciﬁ cally, and 58% in Havana, the Cuban capital and the 2020 epicenter of the country’s COVID-19 epidemic. These ﬁ gures, however, do not represent the transmission capacity or behavior of asymptomatic cases. Understanding asymptomatic transmission’s contribution to SARS-CoV-2 spread is of great importance to disease control and prevention. OBJECTIVE Identify the epidemiological implications of asymptomatic SARS-CoV-2 infection in Havana, Cuba, during the ﬁ rst wave of the epidemic in 2020. METHODS We carried out a cross-sectional study of all conﬁ rmed COVID-19 cases diagnosed in Havana, Cuba, from March 16 through June 30, 2020. The information was obtained through review of the standardized form for investigation of suspected and conﬁ rmed cases. Examined variables included age, sex, occupation, case type and source of infection. Cases were divided into asymptomatic and symptomatic groups, and transmission was characterized through the creation of a contact matrix. Analysis was carried out in Epidat and R. RESULTS We studied 1287 conﬁ rmed cases, of which 57.7% (743) were asymptomatic, and 42.3% (544) were symptomatic. Symptomatic presentation was the most common for both imported and introduced cases, while asymptomatic presentation was more common in autochthonic cases and infections from an undetermined source. Asymptomatic infection was more common in groups aged <20 and 20–59 years, while symptomatic infection was more common in those aged >60 years. In the contact matrix, 34.6% of cases (445/1287) were not tied to other cases, and 65.4% (842/1287) were infectious–infected dyads, with symptomatic–symptomatic being the most common combination. The majority of primary cases (78.5%; 1002/1276) did not generate secondary cases, and 85.6% (658/743) of asymptomatic cases did not lead to other cases (although one asymptomatic superspreader led to 90 cases in a single event). However, 63.2% (344/544) of symptomatic primary cases generated secondary cases, and 11 symptomatic superspreaders spawned 100 secondary cases in diﬀ erent events. CONCLUSIONS infection was the most common form of COVID-19 in Havana during the study period, but its capacity for contagion was lower than that of symptomatic individuals. Superspreader events under speciﬁ c conditions played an important role in sustaining the epidemic.


INTRODUCTION
Research on COVID-19, the disease caused by the SARS-CoV-2 virus, is evolving, addressing questions like diff erent transmission routes, the infectious dose (the amount of virus required for transmission), the characteristics of those most susceptible to infection, situations that facilitate contagion events, the proportion of individuals who remain asymptomatic throughout the course of infection, the specifi c factors that drive asymptomatic and presymptomatic transmission, as well as the proportion of infections transmitted by asymptomatic and pre-symptomatic individuals. [1] The overall percentage of COVID-19 cases that are presymptomatic (infected individuals who are currently asymptomatic but will present with symptoms during a later stage of infection) or asymptomatic (individuals who will never present symptoms) is estimated at 18%-50% worldwide, although this fi gure has been much higher in certain contexts. [1][2][3] The data available in Cuba for the fi rst three months of the epidemic (whose fi rst case was offi cially reported on March 11, 2020) place the proportion at about 53% of all confi rmed cases nationwide, [4] and about 58% of all cases diagnosed in Havana. [5] The aforementioned data, however, represent the total proportion of asymptomatic or pre-symptomatic COVID-19 cases, and not the role of asymptomatic patients in SARS-CoV-2 transmission. According to a study by the University of Padua and the Imperial College of London, asymptomatic COVID-19 patients are about as infectious as symptomatic patients, [6] but there are still no reliable estimates of the contagiousness of asymptomatic individuals as compared with symptomatic individuals.
Due to its clinical importance, an understanding of the magnitude of the role asymptomatic transmission plays in the spread of SARS-CoV-2 is much needed. Additionally, it cannot be assumed that a lack of symptoms means there is no harm being done IMPORTANCE Identifi cation of asymptomatic SARS-CoV-2 infection and its role in COVID-19 spread and transmission contributed to the implementation of eff ective disease control measures in Havana, Cuba during the pandemic's fi rst wave in 2020.

Original Research
Peer Reviewed to the asymptomatic individual. [7] A thorough understanding of asymptomatic transmission is also important from an epidemiological point of view, as the diagnosis, confi rmation and subsequent isolation of symptomatic cases (as established in Cuban national action protocols for case and contact management)[8]-if not accompanied by similar measures for asymptomatic cases-may have a limited eff ect in reducing overall community transmission.
The basic reproduction number R 0 (the average number of secondary infections caused by a primary case in an epidemiologically naïve population) for SARS-CoV-2 person-toperson transmission was estimated at two to three cases, for the original wild-type virus circulating when this study was done. Fewer secondary cases have been reported in some countries than would have been expected with this R 0 , suggesting that not all primary cases cause secondary transmission. [9,10] Some literature suggests that 10%-20% of infected individuals are responsible for 80%-90% of all transmitted cases. [10,11] All of this suggests the need for epidemiological studies estimating not only R 0 for asymptomatic SARS-CoV-2-positive individuals (estimated as 1%; 95% CI: 0%−2%), but also other important parameters such as the k dispersion parameter, a measure of person-to-person transmission variation that is especially important in calculating superspreader events in which a single person infects tens or hundreds of others. [11,12] Given the importance of this information for pandemic control, this article explores asymptomatic and presymptomatic transmission in the Havana pandemic epicenter during the fi rst half of 2020, although the authors recognize that these epidemiological parameters have changed both in Cuba and worldwide with the appearance of new SARS-CoV-2 variants.

Study design and participants
We carried out a descriptive cross-sectional study that included all confi rmed SARS-CoV-2positive cases in Havana, Cuba, from March 16 (following the declaration of the fi rst cases in Cuba and the beginning of the Cuban epidemic) through June 30, 2020 (the end of the fi rst wave of the Cuban epidemic).
We also included the following clinical variables: We selected confi rmed COVID-19 cases in Havana from the country's database and established two study groups: symptomatic and asymptomatic. We recorded the symptom onset date for symptomatic cases, and the confi rmation date for each patient, regardless of whether they were symptomatic or not.
We constructed a contact matrix or transmission tree for all cases to characterize both the groups and their transmission routes. To do this, we identifi ed isolated cases, or primary-secondary contact pairs (infective-infected) in which both cases were symptomatic, both were asymptomatic, or only one patient was symptomatic. Contact matrices (epidemiological networks) were constructed with the index, primary and secondary cases forming part of either outbreaks (defi ned as ≤9 secondary cases) or events (≥10 generated cases); a few of which included superspreader cases.
We performed a percentage analysis of asymptomatic and symptomatic cases, according to age, sex, occupation and infection source, and bivariate analyses through calculation of prevalence ratios with 95% confi dence intervals. The dispersion threshold was calculated using the 99 th percentile of the Poisson distribution (λ = R 0 ) where Pr (Z ≤ Z(99)|Z ~ Poisson(R 0 ) = 0.01; where R 0 is the baseline reproductive number and Z is the number of secondary cases derived from an infected person in a susceptible population. [12] Given that the value of R 0 for COVID-19 (wild-type) has been estimated at 2-3, [7,9,10] the superspreader threshold was established at Z = 6 secondary cases. From this value, all subjects who generated six or more secondary cases were considered superspreaders.
For transmission chains, the eff ective reproduction number (R) was estimated from the negative binomial distribution's

Original Research
Peer Reviewed mean adjusted to the distribution of observed secondary cases. [12] This was completed for all pairs within the matrix clusters and for purposes of comparison, was additionally completed for all chains started by symptomatic and asymptomatic cases.   Confi dence intervals for the prevalence ratios (PR) for sex and occupation in symptomatic cases contained the value 1, which is not the case for individuals ≥60 years of age, which have a remarkably higher likelihood (2.13) of being symptomatic than individuals aged ≤20 years. Asymptomatic infection was most common in individuals <60, while adults ≥60 were more likely to be symptomatic ( Table 2).

As
The contact matrix allowed us to identify 445 isolated cases (34.6% of the total) and 842 interrelated cases (65.4%), for a total of 644 infector-infected dyads, among which symptomaticsymptomatic pairings were the most common (144/644; 22.4%).
During the study period, 177 transmission chains or clusters began in Havana, 49 of which had an asymptomatic individual as the fi rst case, while 128 were initiated by a symptomatic case ( Figure 3). Clusters with asymptomatic primary cases generated a smaller number of secondary cases (193), of which 80.3% were  Among asymptomatic cases, a single individual was identifi ed as a superspreader who generated 90 secondary cases in a single event (closed internment institution). Among symptomatic cases, 11 primary cases acted as superspreaders, leading to 100 secondary cases, related to either high social mobility, close contact between individuals in closed spaces with limited ventilation, or places with high concentrations of people (Table 3).

DISCUSSION
Claims that asymptomatic and presymptomatic transmission represent the main source of the pandemic's spread [13,14] are debatable, particularly given the emergence of new virus variants and new strategies to address them. What is not in doubt is the high proportion of asymptomatic individuals that are carriers of the virus in diff erent contexts, [15][16][17] so research is essential to determine the true role of asymptomatic carriers in SARS-CoV-2 transmission.
The upward linear trend in asymptomatic case notifi cation in Havana during the epidemic's fi rst wave was similar to that reported by other national [4,5,18] and international studies; [6,19,20] but diff ered from reports from researchers in Korea [21] and China. [22] By the time this study was done, Cuba's national strategy, contained in the country's national action protocols for all phases and stages of the pandemic,[8] included extensive research at the primary care level: testing everyone who had either direct or indirect contact with a confi rmed case; isolation of all suspected or confi rmed cases in hospitals, regardless of the presence or absence of symptoms; quarantine of contacts of confi rmed cases in isolation centers; epidemiological surveillance of travelers; and other control measures in accordance with WHO guidelines. [23] Most of these protocols, with the exception of strict isolation measures, are still in eff ect.
A shift in transmission from a predominance of symptomatic to asymptomatic cases during the fi rst wave was likely due to

Original Research
Peer Reviewed increased diagnostic capacity and contact tracing, in which more asymptomatic and pre-symptomatic cases were counted that would have gone undetected at the very beginning of the pandemic. At the same time, there was wide community transmission and defi ned diagnostic testing criteria, in an asymptomatic population that was mostly young and healthy and associated with more benign clinical progression.
This situation continued until the end of 2020, when borders closed due to the pandemic began to open, and cases in foreign visitors and Cubans visiting from abroad introduced new SARS-CoV-2 variants, which are associated with more severe forms of COVID-19 that are more likely to be symptomatic, [24] a situation that has continued through 2021.
The fi rst Cuban travel restrictions were implemented on March 21, 2020, and borders were defi nitively closed on April 7, 2020.
[25] Most cases were either imported or introduced during this period. From then on, outbreaks in closed communities or institutions, where confi rmed cases were not linked to travelers from aff ected areas, suggest that autochthonous transmission was a fundamental aspect of the pandemic, confi rmed by this and other national studies. [4,5,18,26,27] The observation that imported and introduced cases mainly presented as symptomatic before the borders were closed could be related to the fact that Cuba's epidemiological surveillance system during the very beginning of the pandemic defi ned a 'suspected case' of COVID-19 as an individual who presented with symptoms compatible with COVID-19 who came from countries or territories with documented transmission. After studies began of all contacts, identifi cation and diagnosis of asymptomatic cases increased, which led to a change in the defi nition of 'suspected case.' SARS-CoV-2 can aff ect people of all ages. In this instance, both globally and within groups, infection was more common in middle-aged adults, similar to results found in additional studies in Cuba, [5,18] and other countries, [2,28] with the exception of those carried out in Santiago de Cuba [26] and Camagüey [27] provinces, in which older adults were more susceptible to infection. These results could be due to the fact that middle-aged adults comprise the group with the highest levels of social interaction, as they continue to work and carry on family activities, while children remained at home after schools were closed and older adults self-isolated, among other posible causes. [2] In our study, asymptomatic infection was most common in individuals <60 years of age, while older adults were more likely to manifest the disease symptomatically. This result is consistent with other literature on the same period of the pandemic, in which children were generally less susceptible to infection and manifested milder symptoms than adults, suggesting however that they posed a high risk of transmitting the disease through asymptomatic presentation. In adults, advanced age is directly correlated with greater risk of infection, as well as more severe clinical presentations, and death. [29][30][31][32][33] In some studies, the highest rates of infection were reported among men, [5,33,34] while others report the opposite. [26] In this study, both sexes showed the same probability of becoming infected and presenting with symptomatic (or asymptomatic) forms.
According to offi cial information published in the Cuban media, [35,36] health workers (during the study period) were not among the risk groups with the highest SARS-CoV-2 infection rates, nor did theirs compare with such rates in other countries. Public health programming within the country led to increased staff awareness and compliance with biosafety protocols. The proportion of cases with occupations outside the healthcare sector was higher in this study; however, the overall proportion of symptomatic and asymptomatic cases was similar in both groups.
The degree to which SARS-CoV-2-positive individuals spread the virus depends to a great extent on environmental and seasonal conditions, and social behavior. [37] COVID-19 transmission dynamics are characterized by the fact that much of disease's spread is due to a small group of infected individuals. [38] The epidemiological networks and transmission chains of each group in this study led to the identifi cation of a lower R 0 among asymptomatic cases, with a lower percentage of transmission chains initiated by asymptomatic individuals, which translates directly into generation of a smaller number of secondary cases; a fi nding in direct contrast with one Chinese study, [11] and similar to that found by Ruiz Nápoles in Cuba's Holguín province and another study performed in China during the beginning of the epidemic. [39,40] We also found that superspreader events occurred more often among symptomatic patients, except for one asymptomatic superspreader case that occurred in a closed internment institution.
The results of this study affi rm that it is not possible to know in advance which persons have the ability to transmit the infection, so from an epidemiological point of view it is advisable to promote and respect measures to avoid specifi c socio-environmental conditions that favor transmission: not frequenting crowded public places or closed spaces with poor ventilation, maintaining physical distancing, masking, and frequent hand washing in all circumstances.
In agreement with Nishiura, [41] estimating the percentage of asymptomatic cases and their role in the spread of the epidemic has substantially expanded knowledge of the disease, and has improved understanding of SARS-CoV-2 transmission.
Ne w SARS-CoV-2 variants of concern Molecular surveillance carried out by the Pedro Kourí Tropical Medicine Institute in Havana has shown that COVID-19 transmission patterns in Cuba during the beginning of the epidemic were similar to those in Wuhan, China. During almost all of 2020, the mutation in position 614 (D614G) was the most common, coinciding with the period in which asymptomatic or pre-symptomatic infection was the most predominant form of infection. After borders re-opened in November 2020, the alpha and beta variants of concern (VOCs) were identifi ed and later isolated in all Cuban provinces. VOC beta was associated with signifi cant increases in case numbers and with case severity in certain regions, including Havana and Jagüey Grande, Matanzas province. This situation worsened after the introduction of VOC delta, fi rst identifi ed in a traveler at the end of April 2021 and detected in local transmission cases by May. Within a short period of time, this variant spread throughout most provinces, especially Havana, Matanzas, Ciego de Ávila, Cienfuegos, Holguín and Guantánamo. [42,43] Delta's rapid

Original Research
extension-associated with characteristics that facilitate greater transmissibility-led to its predominance over other variants, including beta, also the case in the Americas. [44] The change caused by the new variants in the proportion of symptomatic and asymptomatic cases at the time of diagnosis is remarkable. In 2020, asymptomatic patients exceeded 70% of all patients during certain periods, while by the end of August 2021, asymptomatic cases represented only 5% of daily confi rmed cases. [42] The increase in the proportion of symptomatic cases, together with a shortening incubation period, an increase in viral load and lengthened transmissibility periods suggest that the contribution of symptomatic cases could continue to be an important factor in community transmission of SARS-CoV-2 compared with asymptomatic transmission, as was observed in 2020.

Study limitations
As we used information from the Cuban database of confi rmed cases, pre-symptomatic cases may have been included in the asymptomatic case group, as there was no clinical follow-up as to whether these cases became symptomatic. It was also not possible to calculate disease attack rates, as it would have been necessary to know the exposed population in each epidemiological network.

CONCLUSIONS
Asymptomatic SARS-CoV-2 infection was the most common form of COVID-19 in Havana during the study period, and such individuals were less likely to spread the infection than those who were symptomatic. Superspreader events, most likely under specifi c conditions, played an important role in sustaining the epidemic.