key: cord-0835216-ehyuesvf
authors: Méndez-Arriaga, Fabiola
title: The temperature and regional climate effects on communitarian COVID-19 contagion in Mexico throughout phase 1
date: 2020-05-19
journal: Sci Total Environ
DOI: 10.1016/j.scitotenv.2020.139560
sha: cd7451890932d6733c09ae3be36c1d2285e01a71
doc_id: 835216
cord_uid: ehyuesvf

Abstract Due to the close relationship between the incidence of infectious diseases by epidemics and environmental conditions, this research explores the temperature, evaporation, precipitation and regional climate effects on the local transmission of coronavirus SARS-CoV-2 inside 31 states and capital of Mexico since February 29 (national onset) to March 31, 2020. Statistical analysis was conducted to explore the association between the daily local COVID-19 confirmed positive cases (LCPC) and both climate characteristics and the daily weather reported by the regional meteorological stations. In this work, the local transmission ratio (LTR) was calculated with the regional LCPC divided by the number of the effective contagion days since regional onset in each state. The results showed a negative association between temperature (mean, max and min) and climate classification with both LCPC and LTR variables. The precipitation associated positively with LCPC and LTR. The associations between the climate classification with LCPC and LTR are statistically significant. The tropical climate (mean temperature around 25.95 °C and mean precipitation around 8.74 mm) delayed the regional onset. However, the regional onset in dry climates emerged earlier as consequence of the lower temperatures and higher precipitations (20.57 °C and 20.87 mm respectively) than the observed in the tropical climate. The fastest regional onsets were observed in tempered climates in states where the lowest temperatures and lowest precipitations were registered (19.65 °C and 8.48 mm respectively). Meteorological factors influenced the trend on the regional outbreaks in Mexican's states likely by the host predisposition and susceptibility during the cold winter season. In Mexico, the climate characteristics played a crucial role on the local infection during the phase 1 being the tempered regions (as Michoacán, Jalisco, Puebla, etc.) more vulnerable than the dry (as Chihuahua, Durango or Zacatecas, etc.) or tropical areas (as Colima, Campeche, Morelos etc.).

humans had emerged from the Huanan Seafood Market in Wuhan, China. The new coronavirus was termed SARS-CoV-2 and COVID-19 as its respiratory infectious disease. Since that, reports of positive confirmed COVID-19 cases have been informed in more than 205 countries worldwide (https://coronavirus.jhu.edu/map.html, May 14-2020). In Mexico, the official announcement on the first infected patient was on Feb 29, 2020. To date, there were informed 42,595 cumulated confirmed cases, 26,746 suspected cases, 10,057 confirmed active cases, and 4,477 deceases alongside the 31 provinces belonging to Mexico and its capital (https://www.gob.mx/salud, May 14-2020). For previous historical outbreaks of SARS, MERS or influenza viruses some epidemiological studies have suggested that environmental conditions altered the spread pattern. For instance, in the case of influenza infections, Tamerius and cols., assured that in temperate regions the virus exhibited a strong occurrence according to the seasonal cycle: May to September (annual summer) in the southern hemisphere and November to March (annual winter with the minimal levels of specific humidity and temperature) in the northern hemisphere (Tamerius et al., 2013) . They concluded that the virus is highly sensitive to high temperature. Lowen and cols., have showed in experiments with guinea pigs as host model at laboratory scale, that both cold (low-temperature) and dry (low-humidity) conditions favour transmission of influenza increasing its infectivity (Lowen et al., 2007) . In contrast, if temperature and relative humidity increased, the influenza transmission was entirely blocked. Regarding to SARS-CoV in studies developed at laboratory scale, Chan and cols., founded that such virus was active for at least 5 days at 22-25℃ within a relative humidity between 40 and 50%. In opposite, once the temperature reached 38℃ with a relative humidity of 95%, the virus lost its activity. At 56°C, virus survival lasted only 15 min (Chan et al., 2011) . Casanova and cols., employed the TGEV (transmissible gastroenteritis virus) and MHV (mouse hepatitis virus) as potential surrogates of SARS-CoV to determine the air temperature and relative humidity effects on the viruses survival once deposited in metallic surfaces (Casanova et al., 2013) . At 4°C, viruses persisted as long as 28 days. The lowest level of inactivation occurred at and cols., observed that the virus activity persisted for long time in low-temperature and low-humidity stable and controlled conditions at laboratory scale (Doremalen et al., 2013) . Concerning the new COVID-19 disease, recent statistical researches have demonstrated a solid relationship between the total number of positive cases and the cold-dry environment (between -4 ºC and 15,58 ºC and low humidity) registered in several countries during January and February, 2020 (Sajadi et al., 2020; Wang et al., 2020; Keshavarzi, 2020; Nazari et al., 2020) . According to the above-mentioned authors, spread of the new virus was associated with the latitude and longitude as geographical indicators and seasonal dynamics (Wang et al., 2020; Keshavarzi, 2020) . Additionally, Araujo and cols., assured that tropical countries have been marginally affected by SARS-CoV-2 virus and that a high COVID-19 transmission has been observed in countries with cold-dry weather in temperate climates (Araujo et al., 2020) . Similarly, Bukary and cols., estimated that the low number of positive confirmed cases observed in tropical countries might be explained due to their usual warm-humid environmental conditions. They also assessed that countries experiencing high absolute humidity above 10 g/m 3 could see in a short time a slowdown in the COVID-19 transmissions (Bukhari et al., 2020) . Ma and cols., assured that temperature and humidity were affecting factors related to the number of COVID-19 mortality in Wuhan, China (Ma et al., 2020) . Table 1 recapitulates the above-mentioned researches regarding to the effect of ambient factors on the survival, the transmission and/or the infectivity of influenza, SARS-CoV, MERS and the novel SARS-CoV-2 viruses. As possible to see in Table 1 , most of the reports rarely have considered neither the well-differentiated stages of the epidemic evolution (phases 1, 2 or 3) nor the origin type of affected patients (imported or local). Therefore, the present study shows the effects of the air temperature, precipitation and regional climate on the daily evolution of the local confirmed COVID-19 positive cases (LCPC) documented in 31 provinces and capital of Mexico since February 29, 2020 to March 31, 2020, i.e. during the phase 1 of the epidemic. This study contributes to elucidate on eventual SARS-CoV-2 reoccurrence considering weather conditions, environmental local characteristic and seasonal cycles in each regional climate existing in Mexico.

Onset data and the cumulative number of positive confirmed COVID-19 cases, alongside with age, sex, preceding type (local or imported), confirmation date, number of day with symptoms and deceases per state were daily collected from the Mexican Health Commission´s website (https://www.gob.mx/salud). 1,215 COVID-19 positive total confirmed cases J o u r n a l P r e -p r o o f statistical data (daily highest, lowest and mean temperature, monthly mean atmospheric evaporation and precipitation) were obtained from the historical climatology database of CONAGUA (https://smn.conagua.gob.mx/). The daily average air temperature at the height of 2 meters (T 2m ) and specific humidity at 1000 hPa (SH 1000hPa ) was provided by the Copernicus Climate Change Service database (https://climate.copernicus.eu/). Development of a dynamic database on Excel spreadsheet was designed in order to compile and handle the last information. The statistical association and qualitative correlation of both the regional climate and daily weather conditions with LCPC for each Mexican state were calculated by the Spearman's non-parametric test (r s ). r s determines the strength or weakness of monotonic association between variables through of positive or negative correlation (between -1 and +1) with a concomitant statistical significance (p). All the statistical analyses were two-sided at a 5% level of significance using the R software, version 3.5.3.

Mexico is geographically part of North America but culturally associated with Latin America. It has around 1,972,550 km 2 and 125,959,205 million people (2018). Mexico has almost all the climates according to the Köppen-Geiger climate classification, excepting polar and continental microthermal types. Table 2 shows the TCPC, ICPC, LCPC, gender, average age, days lasting till imported and local onset, the total number of deceases, LTR, monthly average temperature (high, low and mean), monthly mean evaporation and precipitation as well as the predominant climate for each Mexican state and its concomitant number of inhabitants. Moreover, daily cumulative evolution of confirmed positive cases (local, imported and total) since regional onset and daily maximum and minimum temperature recorded for each state alongside detailed climate map and geocoordinate information are depicted in Table 3 . Table 2 lists the data on ascendant LTR values and Table 3 lists the state´s information by climate clusters as tropical, dry or tempered classification.

The evolution of ICPC, LCPC and TCPC in each of the 31 Mexico´s states and its capital were used to determine whether regional climate and daily weather condition could be a factor in the spread of the COVID-19. This descriptive analysis have considered 3 main thoughts further explained: a) to focus on phase 1 of the epidemic, b) to focus on the communitarian confirmed positive cases (LCPC) observed in each state, and c) to enclose the local transmission rate (LTR) based on the effective regional spread since imported onset (I onset ) case until March 31, 2020. In Mexico and internationally, the COVID-19 disruption has raised high attention to governments and institutions in order to design sanitary emergency actions. The Mexican Health Commission announced the China´s epidemic situation in January 3, 2020. Special attention was paid on the international airports where was estimated to detect the earliest infected people arriving to Mexican territory via airplane. The first ICPC was announced on February 28, 2020 in Mexico City arrived from Italy. Since that, Mexican authorities declared phase 1 and the contention strategy emphasized the isolation of all detected imported cases J o u r n a l P r e -p r o o f and tracked down their primary near contacts. Few days later, several contagion cases emerged (both imported and local types) in most of Mexico´s states. In March 23, 2020 phase 2 was implemented with many major governmental intervention measures (for instance the massive quarantine and massive mobility reduction) in order to deaccelerate the communal infection and to avoid the hospital-system collapse. Since that, the massive quarantine stopped progressively the people´s daily activities through the social distance and the confinement. In March 31, 2020, the people´s quarantine and the mobility reduction up to 50% decrease were clearly manifested in all of Mexico´s states. In this study, the period since February 29, 2020 until March 31, 2020 was of analytic importance from an epidemic-evolution point of view. In phase 1, the human-to-human spread of the virus in both symptomatic and asymptomatic cases was developed predominantly from people with previous stay abroad under usual environmental conditions and unrestricted contention control of the transmission. Neither imported cases nor deceases account were considered as responses related to environmental effects, because the initial imported seed of infection is randomly distributed without reasonable association neither to climate nor weather conditions. On the other hand, during the pandemics events, the peak number of deceases normally take place during phase 2 or 3. Deceases are strongly related to previous individual health condition, comorbidity and/or hospital facilities availability rather than climate and weather conditions. Therefore, LCPC factor was chosen as reasonable variable to analyze the association with weather and regional climate instead of ICPC or TCPC. Moreover, due to the regional onset varied in each state, the Local Transmission Ratio (LTR) was defined as the number of local COVID-19 confirmed cases divided by the effective lapse of dissemination into the cumulated days since regional imported onset (I onset ) until March 31, 2020.

As possible to observe in Table 2 , 48.23% and 51.77% were classified as ICPC and LCPC respectively. On average, there were approximately 19.65 LCPC per day. Initial imported cases came from Europe (Spain, Italy, France and Germany), the US and Asia (Singapore). 51.77% and 48.23% were male and female respectively with an average age of 43.32 and 42.84 years old correspondingly. 87.5% of the states reported less than 50 ICPC, 9% between 51 and 100 ICPC and 3% higher than 100 ICPC. 94% of states shown 66% LCPC and only 2% (2 locations) represented 33% of local contagion including both Mexico City (CDMX) and Estado de Mexico (EdoMex) urban megacities as conjoined metropolitan area with more than 20 millions of inhabitants. As possible to observe, Colima is the state with the minimum number of ICPC (2) and did not report LCPC in the studied period. Since February 29, 2020 to March 31, 2020, temperatures in Mexico ranged from 0.87 °C to 38.55 °C with a mean temperature of 22.22°C. The maximum temperature reached between 30.71 to 38.55 °C with an average of 34.61°C and the minimum between 0.87 to 17.79 °C with an average of 9.97 °C, respectively.

Evaporation ranged from 4.5 to 9.2 mm and a mean of 6.14 mm. The average of total precipitation during this period was J o u r n a l P r e -p r o o f 6 13.99 mm. The T 2m and SH 1000hPa values in Mexico within the studied period were 19.82°C (13.5°C < T 2m < 24.9 °C) and 5.92 g/kg (3.1 g/kg < SH 1000hPa < 8.6 g/kg) respectively. Table 4 shows the correlation parameter r S resulted from association between TCPC, ICPC, LCPC and LTR with both meteorological measures (i.e., monthly average temperaturemax, min and mean-, evaporation and precipitation data) and the main climate regional characterization. Qualitative climate factor was associated to numerical equivalences as 1, 2 and 3 for tropical, dry and tempered climates accordingly. The association between of TCPC, ICPC, LCPC and LTR were in all cases negatively associated with temperature (-0.007< r S < −0.182) and evaporation (-0.221 < r S < -0.317) without statistical significance. The association between ICPC and monthly average T max was positive. Association between of TCPC, ICPC, LCPC and LTR were positively associated with precipitation (0.176 < r S < 0.0282) without statistical significance as similar with the TCPC positive association to SH 1000hPa (r S = 0.1677). In contrast, a negative association was observed between each TCPC, LCPC and LTR with climate factor with resulted r S values equal to 0.39041, 0.39031 and 0.46377 respectively. All of last associations resulted statistically significant with p values of 0.02716, 0.02721 and 0.00751 concomitantly. ICPC was also negative associated with the climate variable without statistical significance. TCPC and T 2m association resulted positive and statistically significant (r S = 0.866 and p = 0): as increasing T 2m , increasing TCPC accordingly. Similarly, SH 1000hPa associated positively to TCPC (r S = 0.167) without statistics significance.

On the other hand, Figure 1 shows the monthly average temperature (green bars) and monthly mean evaporation (black bars) per state alongside to the regional LCPC (white bars) in ascending order. As possible to observe, in general both monthly average temperature and monthly mean evaporation decrease as LCPC increases. Figure 2 shows the monthly average max and min temperatures (red and blue boxes respectively) and LCPC values clustered in tropical (pink bars), dry (brown bars) and tempered (green bars) climates. Figure 3 shows the mean temperature (boxes) and LTR (bars) also classified by climate clusters (tropical in pink color, dry in brown color and temperate in green color). As possible to observe in Figure 2 and 3, climate clusters pointed highlight out the role of environmental factors on LCPC. In the dry cluster, the average evaporation value was 6.65 mm similar to evaporation observed in tropical and temperate climates, 5.5 mm and 6.11 mm respectively. However, precipitation in the dry cluster (20.87 mm) was around two folds higher, than the observed in tropical and tempered clusters, 8.74 mm and 8.84 mm respectively. Tropical cluster showed the highest mean temperature of 25.95 °C in comparison to 20.57°C and 19.65 °C for dry and temperate clusters respectively. Similar behaviour was observed for the max and min temperatures with higher values in tropical areas (32.87 °C and 19.03 °C) than in dry (28.39 °C and 12.75 °C) or temperate (28.97 °C and 10.33 °C) regions. Thus in tropical areas infectivity of SARS-CoV-2 disfavour transmission under stable high temperature conditions. In temperate regions, both low temperature and low evaporation with poor precipitation favoured the local infectivity observed in cities such as Tlaxcala, Guanajuato or Michoacan. LTR is statistically significant correlated to climate factor as previously observed in Table 4 . LTR describes the dynamics of the epidemics into the effective elapsed time since regional onset. Average onset data occurs later in tropical climates but earlier in the dry and temperate climates (25.09, 22.35 and 19.57 days respectively) as observed in Figure 3 . Thus, the negative association between LCPC and local onset is considered statistically significant with r s = -0.65207 and p= 5E-05. Moreover, the number of days between the I onset and the L onset was 8.54, 8.35 and 5.28 days respectively in tropical, dry and tempered climates respectively. According with the last results, high LCPC values would be not expected in tropical climates such as Yucatán or Quintana Roo states. However, as possible to observe in Fig. 2 and 3, both Yucatán and Quintana Roo belonging to the tropical cluster, were states with high LCPC values. The last likely resulted due to the high touristic character of both states where the ICPC were the highest values reported (24 each) for the whole tropical cluster. In temperate climate, the infectivity and virus transmission were facilitated in local community due to the low-temperature and low-humidity reported compared with temperature in the dry and tropical clusters. Mexico city and EdoMex, classified as temperate states, reported the highest national local transmission not only justified to the environmental factors but also by the demographic aspects such as the population density (see Table 2 ). Therefore, in Mexico, the climate characteristics played a crucial role on the initial local infection during the phase 1 being the tempered regions (as Michoacán, Jalisco, Puebla, CDMX, among others) more vulnerable than the dry (as Chihuahua, Durango or Zacatecas among others) or tropical areas (as Colima, Campeche, Morelos among others). The last assumption agreed with previous evidences observed in tropical areas of Malaysia, Indonesia, and Thailand during the SARS-CoV outbreak in 2003 (Chan, el al., 2011) .

The most accepted scientific consensus agrees that the low temperature and the low humidity favor the outbreak, the virus survival and the human-to-human transmission as previously observed in historical epidemic events. However, currently there is not consensus of what climates or weather conditions constitute the lower or higher limit to be considered as keylevel factors for pandemics control or spread prediction. In the literature, is common to find several parameters that could be not equivalent for a precise comparison. For instance 1) the chosen response of the epidemics to be monitored i.e. total confirmed cases, suspected cases, mortality, hospital collapse, etc., 2) the chosen time period for evaluation (whole epidemic evolution, weekly, monthly or any arbitrary lapse), and 3) the reference of environmental factors for analysis as indoor or outdoor contexts. The vast majority of the reports are based on either laboratory results -through in situ viral activity observations-or meteorological statistical evidences and mathematical adjusting models. At laboratory conditions, J o u r n a l P r e -p r o o f there exist specific protocols at extreme well-defined experimental environments (incubation time, precise room temperature, controlled room humidity, host specimen specification, etc.). Such studies evaluated the effect of "specific controlled environmental conditions" on general infectivity of targeted viruses. Results of such studies normally pointed out precise values of room temperature or humidity as valid limits for virus survival or inactivation (Lowen et al., 2007; Chan, et al., 2010; Casanova et al., 2010; Dormalen et al., 2013; Dormalen, et al., 2020) . However, such experimental information cannot fully explain why epidemics are also reached in localities where temperature of humidity are out of the range experimentally observed at laboratory conditions. On the other hand, most of the statistical methods and mathematical models have employed data from different lapses or periods of pandemics. Such information is acquired with more or less detailed and in-depth data according its availability (Wang et al., 2020; Sajadi et al., 2020; Keshavarzi, 2020; Bukary et al., 2020; Ma, et al., 2020; Araujo, et al., 2020) . The vast majority of such studies disregard both the origin of the transmission (local or imported) -assuming the total infection cases as the suitable variable to monitor-and the welldifferentiated phases of the epidemic -assuming the whole pandemic lapsed as the most valid range for evaluation-. Such studies are not clearly conclusive respect to the environmental effects on epidemic´s overall outcomes. Inappropriate focus on the epidemic lapse could release contradictory conclusions as pointed out by Cai and cols. (Cai et al., 2007) . For instance, in this study, the association of temperature with confirmed cases resulted positive with perfect statistical significance (p=0): as increases temperature, increases the number of infected people as well (see Table 2 ). However, this expected result comes from the ordinary seasonal cycles that cross from winter to spring season with a concomitant temperature increase. Therefore, the positive association between temperature and TCPC with statistical significance could be mistakenly interpreted. In contrast, in the present analysis was possible to evaluate the correct negative association between environmental factors and LCPC similar than the high correlation coefficient statistically significant between LPCP and TCPC with the climatic character. Moreover, the LTR, which takes into account the effective number of days when the epidemics is regionally active, showed the highest correlation coefficients in association with the climate variable (see Table 4 ). The last results were observed taking in consideration that the principal effect of environmental factors on virus transmission occurred only in the early stage of the epidemics before any imposed contention strategy. Therefore, both the initial lapse of the epidemic and the contagion origin-type are of the most significant highlights here founded useful to assure that the regional climate character in each states did influence the initial transmission of the COVID-19 disease in Mexico. Finally, it is important underline that experimental results under controlled conditions at laboratory scale offer extremely useful information to be employed during confinement phases 2, 3 and beyond. As shown in Table 1 , SARS-CoV-2 is stable at low temperatures (between 5 and 11°C or generally below 20°C) and low humidity (generally dry J o u r n a l P r e -p r o o f condition) (Doremalen et al., 2020) . Commonly, in tropical and warm-dry climates the indoor environment is artificially adjusted by air-condition apparatus that reduces the temperature and the humidity for comfortable spaces. In the dry cluster analyzed in this study, the LCPC was high in states such BC and BCS likely linked to the intensive use of air-conditioning instrument in agreement with observations previously reported in Singapore and Hong Kong by Chan and cols. (Chan et al,, 2011) . The use of air-condition apparatus promotes the most appropriate conditions for survival of the SARS-CoV-2.

Therefore it is important avoid the low temperature and low humidity in both indoor hospitals and room-home level spaces during phase 2 and 3 under implemented strict contention strategies. In Mexico, the quarantine obligation in phase 2 and 3 resulted in reduction until up 70% of the mobility in metropolitan areas with more than 20 million of inhabitants. People stayed at home in a massive extent. Taking into account that the virus infectivity decreases as temperature and humidity increases some public health recommendations could generate massive non-contagion indoor areas. The employment of commercial humidifiers and heaters or even better, natural solar warm and fresh air circulation when possible, could be effective massive sanitary strategies in order to reduce the virus transmission as well as to decelerate the hospital system collapse.

The occurrence of confirmed local cases of COVID-19 in Mexico has been analyzed taking into account the weather and climate of each of 31 states and the capital during the phase 1 of the epidemic progress. The evolution of COVID-19 was analyzed through in the number of LCPC rather than TCPC or ICPC. Neither ICPC nor deceases described the environmental effect on transmission in phase 1. COVID-19 local spread was strongly associated with both the climate and daily environmental conditions due to during phase 1 the transmission occurred freely without contention strategy influences. Thus, the regional climate in each Mexico´s state has a central role in the initial local SARS-CoV-2 transmission. TCPC (i.e. ICPC+LCPC) associates positively with daily average air temperature. High temperature and high evaporation in tropical climates are the best-combined predictors as a disadvantaged condition for SARS-CoV-2 survival.

Daily incidence of COVID-19 increased slowly in states belonging to tropical cluster. Template condition promotes the highest LCPC observed. Temperature and evaporation were negatively associated with the LCPC without statistical significance. Regional climate was statistically significantly associated with LCPC and LTR. Dry and temperate cold climates are the more suitable environmental conditions for transmission. The tropical climate (mean temperature around 25.95 °C and mean precipitation around 8.74 mm) delayed the regional onset. However, the regional onset in dry climates emerged earlier as consequence of the lower temperatures and higher precipitations (20.57°C and 20.87 mm respectively) than the observed in the tropical climate. The fastest regional onsets were observed in tempered climates with the lowest J o u r n a l P r e -p r o o f temperatures and precipitations registered (19.65°C and 8.48 mm respectively). Therefore, the regional climate was strongly associated with initial COVID-19 local contagion during the phase 1 being the tempered regions (as Tlaxcala, Guanajuato, Jalisco, among others) more vulnerable than the dry (as Aguascalientes, Sinaloa, Queretaro, among others) or tropical areas (as Veracruz, Guerrero, Chiapas, among others). Any attempt related to evaluate the effect of outdoor environmental factors on the spread, transmission or virus survival should take into account both the phase of the epidemic evolution and the effective active period of the virus since regional onset on communitarian transmission.

The data used during the current study are available from the websites. 

The author declare that she has no conflict of interest. Regional onset in dry climate came early as the effect of the lower temperatures and higher precipitations (20.57°C and 20.87 mm respectively). The fastest regional onsets were in tempered climates with the lowest temperatures and precipitations registered (19.65°C and 8.48 mm respectively).

The tropical climate (mean temperature around 25.95 °C and mean precipitation around 8.74 mm) delayed the regional onset.

Enclosed specific phase of the epidemic and specificity on the contagion types (local o imported) are important factors for weather and climate association studies.

J o u r n a l P r e -p r o o f Table 3 . Cumulative evolution of confirmed positive cases (local, imported and total) since regional onset and daily maximum and minimum temperature recorded with the concomitant detailed climate map according to Köppen-Geiger classification adapted by INEGI 1 alongside the geocoordinate data and number of inhabitants per state. Colima Long 114°45'36"W 103°29'9.6"W, Lat 18°20'20.04"N 19°30'45 

Spread of SARS-CoV-2 Coronavirus likely to be constrained by climate

Will Coronavirus Pandemic Diminish by Summer?

Influence of meteorological factors and air pollution on the outbreak of severe acute respiratory syndrome

Effects of air temperature and relative humidity on coronavirus survival on surfaces

To National Council of Science and Technology through Catedras Program. In memorial to the COVID-19´s deceases worldwide.