MERS transmission and risk factors: a systematic review.

Background Since Middle East respiratory syndrome (MERS) infection was first reported in 2012, many studies have analysed its transmissibility and severity. However, the methodology and results of these studies have varied, and there has been no systematic review of MERS. This study reviews the characteristics and associated risk factors of MERS. Method We searched international (PubMed, ScienceDirect, Cochrane) and Korean databases (DBpia, KISS) for English- or Korean-language articles using the terms “MERS” and “Middle East respiratory syndrome”. Only human studies with > 20 participants were analysed to exclude studies with low representation. Epidemiologic studies with information on transmissibility and severity of MERS as well as studies containing MERS risk factors were included. Result A total of 59 studies were included. Most studies from Saudi Arabia reported higher mortality (22–69.2%) than those from South Korea (20.4%). While the R0 value in Saudi Arabia was < 1 in all but one study, in South Korea, the R0 value was 2.5–8.09 in the early stage and decreased to < 1 in the later stage. The incubation period was 4.5–5.2 days in Saudi Arabia and 6–7.8 days in South Korea. Duration from onset was 4–10 days to confirmation, 2.9–5.3 days to hospitalization, 11–17 days to death, and 14–20 days to discharge. Older age and concomitant disease were the most common factors related to MERS infection, severity, and mortality. Conclusion The transmissibility and severity of MERS differed by outbreak region and patient characteristics. Further studies assessing the risk of MERS should consider these factors.


Background
Middle East respiratory syndrome (MERS) was first reported in 2012 in Saudi Arabia [1]. Although most patients are linked to the Arabian Peninsula geographically, MERS has been detected in many other parts of the world [2]. A large MERS cluster was also observed in 2015 in South Korea [3].
MERS causes sporadic infection and intrafamilial and healthcare-associated infection. Its symptoms can vary from asymptomatic infection to death. Despite the infection's association with high mortality, specified antiviral therapy is lacking, especially for patients with concomitant diseases [2].
Many previous studies have assessed the risks of MERS, such as factors dictating severity or an infection risk, yet the indices they present vary. For example, the case fatality rate was found to be 25.9% in the Middle East area, but 20.4% in South Korea [4]. The incubation period was reported to be 6.83-7 days in South Korea [4,5], but 5.5 in a study using data from multiple areas [6] and 5.2 in Saudi Arabia [7]. Accurate assessment of the risk of MERS is essential for predicting and preventing infection.
A systematic review of the risk of MERS, as covered in previous studies, is potentially helpful for predicting this spread, and its future impact. This study aimed at reviewing the risk of MERS, focusing on indices related to infectivity and severity.

Methods
We searched international (PubMed, ScienceDirect, Cochrane) and Korean databases (DBpia, KISS) using the term "MERS" or "Middle East respiratory syndrome", encompassing articles published after 2000. The search process was conducted in October 2017. We also manually searched the reference lists of the included studies.
Human studies were included, while animal studies and reviews were excluded. Only articles in English or Korean were included. Even if a study collected data on humans, such as collecting specimens from religious pilgrims, it was excluded if there were no MERS patients in the study sample. Additionally, case studies including fewer than 20 MERS patients were excluded as they were considered as having insufficient MERS patient numbers and representative information.
The included studies were classified as epidemiologic studies and those covering risk factors of MERS. In the epidemiologic category, indices related to the risk of MERS were divided into two categories; related to infectivity and related to severity. The index related to infectivity included the reproduction number (R), attack rate, incubation period, serial interval, and days from onset to confirmation. The index related to severity included the case fatality rate (CFR), days from onset to hospitalization, days from onset to discharge, days from onset to death, and days from hospitalization to death.
In the risk factor category, factors related to infection, transmission, severity, and mortality of MERS were analysed. Even if the included studies investigated factors that were related to mortality, when they did not analyse risk factors of severity or mortality using appropriate statistical methods (e.g., regression analysis, Cox proportional hazards model) or only compared prevalence factors, we excluded them from the risk factor category. In all categories, we extracted the study period, number of participants, and geographical region where the data were collected using a data extraction form confirmed after pilot assessment.

Results
A total of 3009 studies were searched, and 2717 were reviewed, excluding 292 duplicate studies. After the title and abstract review, a further 1804 and 663 were excluded, respectively. Another four studies were included via a manual search, which left a total of 58 studies for analysis ( Fig. 1).

Epidemiologic studies
The 38 of total 58 included studies were classified as epidemiologic studies (Table 1).

Attack rate
A total of eight studies reported the attack rate. Four reported the overall or secondary attack rate, and the other four reported the attack rate of specific participant groups. Two studies conducted in Saudi Arabia showed 0.42% [21] and 4% [22] secondary attack rates. Studies in South Korea showed secondary attack rates of 3.7% in one study [23] and 14.3-15.8% in another [24].
Two studies reported the attack rate among healthcare workers (HCWs). One study in South Korea reported a MERS incidence of 1.5% among HCWs [20], and another study using multiple area data reported a 13.4-13.5% infection rate among HCWs [8]. The attack rate among hospital patients was 4% in one study [5] and 22% in the early and 1% in the later period in another [16].

Incubation period
The incubation period is the period between infection and appearance of signs of a disease. A total of 12 studies reported the incubation period of MERS. Nine used data from South Korea and showed a 6-7.8 day incubation period [3][4][5][23][24][25][26][27][28]. One study using data from Saudi Arabia reported a 5.2 day incubation period [7], and another using data from multiple areas reported a 5.5 day incubation period [6]. Sha et al. compared the incubation periods between the Middle East area and South Korea and reported 4.5-5 and 6 days, respectively [29].

Serial interval
The serial interval of an infectious disease represents the duration between symptom onset of a primary case and of its secondary cases. Two studies used South Korea data, reporting serial intervals of MERS of 12.6 and 14. 6 days, respectively [24,26].

Days from onset to confirmation
Among five studies reporting days from onset to confirmation, three studies used data from South Korea. One study analysing all South Korea cases reported 5 days from onset to confirmation [3]. Park et al. reported 6.5 days for all cases, 9 for second generation and 4 for third generation [28]. One study from Taiwan reported 6 days for HCWs and 10 for non-HCWs [30]. A study from Saudi Arabia reported 4 days from onset to confirmation [31]. Sha et al. compared the data from Middle East and South Korea areas and reported 4-8 and 4-5 days, respectively [29].

Days from onset to hospitalization
Two studies from Saudi Arabia reported days from onset to hospitalization. One reported 2.9-5 days [32], and the other reported 5.3 days [33].

Days from onset to discharge
Three studies reported days from MERS onset to discharge. Sha et al. reported 14 days in the Middle East area and 17 in South Korea [29]. One study from Saudi Arabia reported 17 days [36], and another in South Korea reported 20 [3].

Days from onset to death
Two Korean studies reported similar periods of 11-13 days from onset to death: 11-12.5 in Park et al. [24] and 13 in Ki et al. [3]. Although one study from Saudi Arabia reported longer than 17 days from onset to death [36], Sha et al., comparing data between the Middle East and South Korea, reported similar periods of 11.5 and 11 days, respectively [29]. One Taiwanese study also reported a similar period of 12-13 days [30].

Days from hospitalization to death
Two studies reported a similar length of hospitalization: 15 [33] and 15.2 days [19].

Risk factors related to mortality
Of the 20 studies included in the risk factor category, four were duplicates of studies in the epidemiologic category as they had information regarding the epidemiologic index and risk factors ( Table 2).

Factors related to MERS infection
Two studies reported on the risk factors of MERS infection. Alraddadi et al. [42] analysed the effect of nonhuman contact, including travel history, animal-related exposure, food exposure, health condition, and behaviour and reported direct dromedary exposure, diabetes or heart disease, and smoking as risk factors of MERS infection. Another study reported older age, outbreak week, and nationality as risk factors [43].

Factors related to MERS transmission
Three studies analysed factors associated with spreaders. Non-isolated in-hospital days, hospitalization or emergency room visits before isolation, deceased patients, and clinical symptoms, including fever, chest X-ray abnormality in more than three lung zones, and the cycle threshold value, were related to spreaders [34,44,45].

Factors related to MERS severity
Four studies reported risk factors of MERS severity. The included studies showed that the PRNT 50 and CD4 T cell response [46] as well as a high MERS virus load [47] were associated with the severity of MERS. Additionally, male sex; older age; concomitant disease, including hypertension; and symptoms, including fever, thrombocytopenia, lymphopenia, and low albumin concentration, were related to MERS severity or secondary disease [47][48][49].

Epidemiological index of MERS between the Middle East area and South Korea
The characteristics of MERS differ between South Korea and the Middle East area. The R value of MERS was reported to be below 1 in the Middle East area, except in one study [15], but was from 2.5-8.1 in South Korea [15][16][17][18][19]. Although studies using data from the Middle East area reported 0.42-4% secondary attack rates, studies in South Korea reported 4-6% secondary attack rates for patients or hospital visitors [5], and 3.7-15.8% for the overall attack rate [23,24]. The MERS incubation period was reported to be 4.5-5.2 days in the Middle East area [7,29], but this period was found to be slightly longer in South Korea [3][4][5][23][24][25][26][27][28].
The severity of MERS also differed between the Middle East area and South Korea. Mortality of MERS patients was found to be 20.4% in South Korea based on a report including all cases [27], but most studies from Saudi Arabia reported higher rates, from 22 to 69.2% [7,22,33,[37][38][39]. Days from onset to confirmation were similar, 4-8 days in the Middle East area [29,31] and 4-6.5 days in South Korea [3,28,29]. Days from onset to discharge were slightly longer in South Korea, 14-17 days in the Middle East area [29,36] and 17-20 days in South Korea [3,29] (Table 3).

Discussion
The transmissibility and severity of MERS were different by outbreak countries, especially between the Middle East area and South Korea. The virus, host, and environmental factors may be the causes of the MERS outbreakrelated differences between the two regions. From the standpoint of viral factors, there was a mutation of the MERS coronavirus (MERS-CoV) in the South Korea outbreak. Kim et al. [60] reported a point mutation in the receptor-binding domain of the viral spike protein in MERS-CoV, and another study showed that MERS-CoV in South Korea had higher genetic variability and mutation rates [61]. Individual characteristics can also affect MERS transmission. As previous studies showed, there   is an association between older age and MERS infection [43], severity [48], and mortality [4,50], and the population structure may be related to transmission and severity. Additionally, individuals aware of MERS were found to be more likely to practice preventive behaviour [62], which differed by demographic characteristics [63,64]. The transmission environment may also contribute to the difference. While many MERS cases were contracted through exposure to camels in Saudi Arabia [42], the South Korea outbreak involved multiple generations of secondary infections caused by intra-hospital and hospital-tohospital transmission [3,65]. Strategies considering various factors are therefore needed to assess the impact of MERS and to better control its spread.
Although several studies have reported the overall R value [9,10,14,19], others have shown that this value this can be variable based on the generation or a control intervention [11,16,19]. Especially in the South Korea epidemic, the R value was particularly high in the early stage or first generation, at 4.42-5.4, though it later decreased to 0.14-0.39 [16,19]. Further studies should consider and analyse the variation of the R value depending on the period or control intervention.
While earlier studies on infectious diseases assumed a homogeneous infection ability of a population, recent studies have shown the existence of so-called super spreaders, individuals with a high potential to infect others in many infectious diseases, including Ebola and severe acute respiratory syndrome (SARS) [66]. The role of the super spreader is also important in the spread of MERS. In South Korea, 83.2% of MERS patients were associated with five super-spreading events [27]. Stein et al. [67] asserted that super spreaders were related with the host, pathogen, and environmental factors, and Wong et al. [66] reported that individual behaviours could also contribute to disease spread.
There are variations in the mortality and attack rates among studies using South Korea data. For example, Park et al. [24] reported a 47.8% MERS mortality, while reports from the Korean Ministry of Health and Welfare showed 20.4% MERS mortality. This disparity may, in part, be due to small sample sizes. Park et al. [24] included only 23 patients because the study was conducted in an early phase of a MERS outbreak. We excluded studies that included cases with < 20 subjects, which were mostly case series, to reduce those types of biases.
The present review found that older age and concomitant disease were risk factors of MERS infection and mortality. These results are consistent with a recent systematic review that reported older age, male, and an underlying medical condition as predictors of death related to MERS [68]; therefore, these factors should be prioritized in protection and treatment procedures.
One limitation of this study was the possibility of subject duplication. Especially in South Korea, the Korean government publishes MERS reports that include all patients. The epidemiologic index in other studies might be biased since they included partial Korean patients and were analysed in the middle of an outbreak. However, we included those studies because they showed the characteristics of MERS in different situations and different stages.
We did not conduct a meta-analysis because of the small number of studies for each index, which might be another limitation of this study. Although this study reviewed the risk factors of MERS and their impact, assessing the effect size of each risk factor is important. More studies investigating the effect of risk factors on MERS need to be constantly conducted.

Conclusion
Most studies on the transmissibility and severity of MERS have originated from Saudi Arabia and South Korea. Even though the R 0 value in South Korea was higher than that in Saudi Arabia, mortality was higher in Saudi Arabia. The most common factors behind MERS infection and mortality were older age and concomitant disease. Future studies should consider the risk of MERS based on the outbreak region and patient characteristics.
The results of the present study are valuable for informing further studies and health policy in preparation for MERS outbreaks.