Early transmission dynamics, spread, and genomic characterization of SARS-CoV-2 in Panama

Bioethics Permit

The protocol EC-CNBI-202-04-46 was approved by the National Committee on Bioethics of Research of Panama to use de-identified epidemiological data to analyze SARS-CoV-2 transmission and spread, as well as to sequence the complete genome of SARS-CoV-2 from Gorgas Memorial Institute of Health Studies (GMI)’s confirmed cases.

COVID-19 Surveillance system

The surveillance program for COVID-19 was implemented by the Panama Ministry of Health (MoH) on January 20th, 2020. Suspected cases were actively sought at international airports using the WHO/PAHO case definitions and recommendations¹⁶. In this first stage, suspected cases were defined as presenting Influenza Like Illness (ILI) and severe acute respiratory infections (SARI) symptoms and signs, as well as coming from China. Later, on March 4^th, the list of countries was expanded to include Italy, Iran, South Korea, and Japan. Travelers from countries with confirmed circulation of SARS-CoV-2 were isolated at home or hotel-quarantine for 14 days and monitored for onset of symptoms. Patients that met the criteria of suspected cases were then confirmed via laboratory diagnosis. For confirmed cases, contact tracing was performed by the National Epidemiology Department of the MoH, and cases and contacts were placed under quarantine. Clinical evaluation and temperature measurement for travelers and contacts were done every day; if symptoms developed, nasopharyngeal or oropharyngeal swabs were taken by the rapid response teams of MoH and sent to the GMI for laboratory diagnosis. Details of COVID-19 surveillance system are described in the Supplementary Information.

Laboratory diagnosis

Nasopharyngeal and oropharyngeal swabs samples were submitted to the GMI for laboratory confirmation since January 23rd and to additional laboratories from the National COVID-19 Laboratory network since March 16th. Molecular diagnosis was performed first using a generic coronavirus RT-PCR and then specific Real Time RT-PCR for SARS-CoV-2¹⁷ as described in the Supplementary Information.

Epidemiological Investigation

Epidemiological data on suspected cases and their contacts were recorded by physicians when nasopharyngeal and oropharyngeal swabs were taken, using a standardized epidemiological form. Data entry was independently undertaken by the National COVID-19 Laboratory network, and then checked by the National Department of Epidemiology of MoH to confirm the accuracy of the information. A dataset of daily incidence based on date of symptom onset was created. The first 61 days of COVID-19 epidemic in Panama from February 15 to April 16 were analyzed. More details are given in the Supplementary Information.

Epidemic parameter estimation

Early transmission dynamics of COVID-19 in Panama were evaluated estimating the daily growth rate, doubling time, basic and effective reproductive numbers. We estimated the basic reproductive number (R₀) using the time series of confirmed cases, an early R₀ with Likelihood-based estimation using a branching process: This estimation follows Poisson likelihood standards¹⁸. A serial interval mean of 4·7 days and standard deviation of 2·9 was used for the estimation⁶. The time-varying effective reproductive number (R_t) was estimated in a Bayesian framework since the cumulative number of cases reached 25 cases as previous decribed⁶, serial interval with 95 % credible intervals (95% Cl), using the EpiEstim packaged implemented in R¹⁸. Details are provided in Supplementary Information.

SARS-CoV-2 Sequencing

Reverse transcription and multiplex PCR were performed according to ARTIC Network protocol (https://artic.network/), with modifications. Amplicons generated were pooled and prepared for Illumina Sequencing with the Nextera XT library preparation protocol. All samples were sequenced with MiSeq V2 reagent kits for 500 cycles. Reads obtained were aligned with bowtie-2 and consensus genome was generated with Quasitools-Hydra. Details are provided in Supplementary Information.

Genome Analysis

To evaluate virus lineage distribution in Panama, lineages were classified using Pangolin¹⁴, and their distribution was assessed per geographic region of origin of the sample and epidemiological information. Time-trees were estimated using BEAST v1.10.4¹⁹ and BEAGLE v3²⁰ to enhance computational speed. Posterior estimates of most recent common ancestor (tMRCA) of each locally transmitted lineage was obtained as described in Supplementary Information.

Control measures in Latin America

Reports from the WHO, the Organization for Economic Co-operation and Development (OECD) and official communications from the ministries of health were used to obtain data on control and mitigation strategies at the country level to reconstruct a timeline representing the strategies adopted by Latin American countries. Only when no official information was found, press articles from local newspapers were consulted, as described in Supplementary Information.

Early epidemic dynamics in Panama

Epidemiological surveillance began on January 23 and, by April 16, a total of 18 559 suspected cases of COVID-19 had been investigated. Of these, 4 210 were confirmed as SARS-CoV-2 infections at that date. On March 7th, the National Surveillance system detected a fatal case in which the SARS-CoV-2 infection was confirmed on March 9th. This case was not associated with previous travel history outside the country. The first case detected by active surveillance at international airports, came from Spain, had mild disease and was confirmed on March 9th. Most imported cases were associated with travel from Europe, North America, and one from China.

Epidemiological investigation confirmed that the date of onset of symptoms of one local case in our study goes back as early as February 15 (Figure 1A). Most local cases had mild disease (Figure 1B). By April 16, a total of 341 hospitalized cases (77 were hospitalized at diagnosis confirmation) and 116 fatalities (31 were dead at the diagnosis confirmation) were reported in Panama (Figure 1C and 1D). The highest proportion of confirmed cases were observed in the age group 20-59 years (Supplementary Figure 1A). A higher proportion of females were tested while more males were positive (Supplementary Figure 1B). A rapid growth rate of 0·13 cases/per day (Figure 2A) and a short doubling time was also observed during the early stages of the epidemic, with an increasing doubling time over the analyzed interval (Figure 2B). We estimated an early epidemic potential of transmission of SARS-CoV-2 in Panama of R₀= 2·22 (95% CI, 2·08 - 2·37).

• Download figure
• Open in new tab

Figure 1. Epidemic Curve of confirmed SARS-CoV-2 infections.

Daily incidence of: A) imported and local confirmed infections, B) ambulatory cases (asymptomatic, pre-symptomatic, mild outpatient cases), C) hospitalized cases and D) fatal cases, with symptom onset from February 13^th to April 13^th. In all four charts, the y axis represents the daily incidence and the x axis represents the date of symptom onset for each reported case. The blue vertical dashed lines represent the first recorded onset of symptoms and vertical red dashed lines the date of first confirmed case by the surveillance system.

• Download figure
• Open in new tab

Figure 2. Epidemiologic parameters:

Number of COVID-19 daily confirmed infections (left axis and bars) overlaid with estimates of A) Fitted exponential growth in cases daily growth rate, B) Doubling time and C) Time-varying effective reproduction number R, for a time frame of 45 days (x axis). For C), green shaded areas show 95% credible intervals around the median estimated R. The threshold value R_t = 1 is indicated by the horizontal red dashed line. For both B) and C), the vertical blue dashed line represents the schools’ closure, the vertical lines red and green represent night curfews and restrictions of movement respectively, and the purple dashed line the 24-hour curfew.

Panama detected the first SARS-CoV-2 infection one month after Brazil¹⁵, being the 11^th country with confirmed cases in Latin America (Supplementary Figure 2). School closures were implemented the day after first case confirmation, social distance measures six days later, and stay-at-home 24-hour curfew 14 days later. Compared to other countries in the Latin American region, Panama was among those that rapidly implemented control strategies after its first case (Supplementary Figure 2). The effective reproduction number since the accumulated 25 cases was R_t= 3·01 (95% Crl: 2·54 - 3·52) and rapidly dropped over the course of 17 days to Rt= 1·08 (95% Crl: 1·00 −1·17) (Figure 2C, Supplementary Table 1). Until April 16, Panama was the country from Central America with the highest cumulative incidence of cases and fatality (Supplementary Figure 3).

SARS-CoV-2 lineage distribution in Panama

To characterize the distribution of genetic lineages in Panama, we generated 313 SARS-CoV-2 sequences, which represents 7.4 % of the total confirmed cases by April 15^th (Supplementary Figure 4A). Analyzed sequences were obtained from patients with symptom onset between February 15 and April 14. The genome coverage of the analyzed sequences increased as the real-time PCR cycle threshold (CT) value of the sample decreased, with the best coverage obtained at CT values lower than 25 (Supplementary Figure 4B). Lineage analysis showed the circulation of ten previous reported lineages (Figure 3A, Supplementary Figure 4). The most successfully distributed lineage was A.2 (71.2 %), followed by B.1 (16.7%), and A.1 (3.5%). Lineages A.3, B, B.1.5, corresponded to early detected cases, most of them from returning travelers (squared tips in Figure 3A and black outlined circles in Supplementary Figure 4). Lineage A.2 was found in subjects with an epidemiologic link to cases related with a school outbreak (1.8%) where the first local transmission was detected, and in police officers (6.3%). Additionally, a group of 170 sequences which had an C to T nucleotide substitution at position 12815 emerged as a sister clade from lineage of A.2 (Figure 4A). These sequences, indicated a substantial local transmission as they corresponded to 54·3% of the sequenced cases in the country. Estimates of the tMRCA of the Panamanian sequences place the median age of A.1, A.3, B.1 between two weeks range before the first reported COVID-19 case, which agrees well with the time of onset of symptoms observed in the data (Figure 1A, Figure 3A), however the A.2, had the oldest median tMRCA corresponding to early February (Figure 3A).

• Download figure
• Open in new tab

Figure 3. Time-tree resolution, geographic and epidemiological characteristics of the SARS-CoV-2 genomes obtained.

A) Bayesian maximum clade credibility tree of 1261 SARS-CoV-2 sequences, 133 from Panama (red), 91 from Latin America (blue), 904 from other location (grey), tips with box shapes represent samples that have a clear epidemiological travel link. Posterior density estimates of time of the most recent common ancestor of each lineage with local transmission are show in their branches. B) Distribution of the lineages among regions in the country. C) Distribution of the lineages by type of exposure detected by the surveillance system.

• Download figure
• Open in new tab

Figure 4. Mutation profile of the sequences obtained in the study.

A) Frequency of single nucleotide polymorphism in the genome (position according to MN908947) for all analyzed sequences (n=313) in the study. Distribution of S protein variants (D614 or G614) B) in different regions of the country, and C) among lineages.

There is no structured spatial distribution of the lineages across the country, however the locally transmitted A.2 linage was found in all regions. Additionally, Center Panama and West Panama (regions that correspond to Panama City and its suburbs), had a more diverse lineage distribution (Figure 3B). Using active contact tracing of positive cases at the beginning of the outbreak, the National Epidemiology Department of MoH defined 30 epidemiological clusters (supplementary table S2), linked to a school, police officers and local outbreaks. Most sequenced cases were not classified inside these epidemiological clusters (unknown category in Figure 3C) and had a predominance of lineage A.2.

The previously described spike glycoprotein variants D614 and G614^21,22 were co-circulating early in the epidemic among all the regions analyzed, but G614 variant was predominant in 18.8% of the sequences analyzed (Figure 4B). Both variants were also distributed among lineages, however D614 was highly predominant in A.1 and A.2. (Figure 4C).