Malaria is a mosquito-borne infectious disease of humans and other animals caused by a parasite of the genus Plasmodium [1, 2]. Malaria in humans is caused primarily by four species of Plasmodium: Plasmodium falciparum, Plasmodium vivax, Plasmodium ovale, and Plasmodium malariae; in addition, it can be caused by Plasmodium knowlesi, which was recently discovered in Malaysia [1, 2]. P. vivax malaria has lower fatality compared with P. falciparum malaria, but its occurrence is widespread across tropical, subtropical, and temperate regions [3].

In Korea, P. vivax malaria was an endemic disease before the 1970s, and its elimination was declared in 1979. It re-emerged, however, in 1993. In the initial stages of re-emergence, it occurred mostly in regions near the Demilitarized Zone (DMZ), such as Paju-si and Yeoncheon-gun, but it rapidly spread to the east and west. Recently, the incidence of malaria has increased notably in the Northwestern region of Gyeonggi-do, such as Gimpo-si, Paju-si, and Goyang-si, as well as in Incheon (Fig. 1) [3-5].

The incubation period of malaria appears to vary according to the type of parasites. In P. vivax malaria, strains from tropical regions show an onset of symptoms 2-3 weeks after infection, whereas strains from temperate regions show both a short incubation period (early primary attack, onset within 1 month after infection) and a long incubation period (late primary attack, onset 8-12 months after infection) [3]. This is attributable to the fact that, after invading liver cells, some strains of P. vivax malaria become a hypnozoites rather than schizonts; in such cases, relapse or long incubation periods (late primary attack) can occur after a certain period even when the parasites have completely disappeared from the blood [6]. It is presumed that, even with strains of the same P. vivax malaria, onset can differ according to varying proportions of sporozoites morphing into either hypnozoites or schizonts, which grow and multiply right away [7].

The spread of P. vivax malaria to other regions of Korea is mainly attributable to residents from P. vivax malaria non-risk regions visiting risk regions for military service, travel (visiting, business), or other purposes and being infected with latent hypnozoites, causing late primary attack [3]. As a result, cases occur even at times when vector mosquitoes are not active, causing difficulties in extermination, case management, and treatment of P. vivax malaria in Korea.

In Korea, the proportion of P. vivax malaria cases with long incubation periods is estimated at 50%. Therefore, it is necessary to analyze the onset of such cases for effective preventive measures and case management [8]. Previous research on the incubation period of P. vivax malaria has mostly focused on estimating the mean incubation period using a small number of cases over a short period [9-14], due to the difficulty of accurately identifying the time at which the incubation period begins following the bite of a malaria-infected mosquito.

This study used long-term epidemiological data regarding P. vivax malaria cases on a national scale to epidemiological investigate the influence on incubation period of various factors: demographic factors, clinical symptoms, time of infection and symptom onset, and environmental factors. This study aims to provide basic information that will help in establishing management plans for P. vivax malaria, which is one of the most important public health problems among vector-borne diseases.

This study used epidemiological data regarding P. vivax malaria cases in Korea from 2001 to 2010 to classify cases infected while visiting risk regions into long and short incubation periods. Then, we analyzed the correlation between incubation period and demographic characteristics, time of infection and symptom onset, and environmental factors such as latitude and meteorological factors. The results showed that the mean incubation period for short incubation period cases, with an onset of within 1 month, was 25.5 days. The mean incubation period for long incubation period cases was 329.4 days. These results are similar to those of precedent studies [9, 10].

The 10-year mean proportion of long incubation periods was approximately 36%, which is lower than the long incubation period proportion of 75% reported in a P. vivax malaria study that investigated a North Korean strain in Moscow from 1953 to 1972 [9]. However, it is similar to the results of recent studies of P. vivax malaria cases in Korea [10, 11].

In this study, incubation period appeared to be associated with the region of infection, with infections presumed to have occurred in Gangwon-do having a higher proportion of long incubation periods compared with infections presumed to have occurred in Gyeonggi-do or Incheon. According to precedent studies, since 2005, cases have been concentrated in the western P. vivax malaria risk regions of Ganghwa-gun, Seo-gu, Jung-gu and Dong-gu in Incheon and Goyang-si, Gimpo-si, and Paju-si in Gyeonggi-do; meanwhile, there have been very few cases in Gangwon-do other than in Cheorwon-gun [3]. Our study also showed that the number of cases infected in Gangwon-do gradually decreased while those infected in Incheon increased, and the proportion of short incubation periods also increased. It is expected that this trend will continue in the future and, presumably, have an effect on the increase of short incubation period cases.

Differences in incubation period were observed according to the time of infection and symptom onset. A precedent study that investigated aspect of symptom onset among soldiers who entered after November that short incubation cases occurred from early June [4]. In our study, however, although there were a small number of cases, 123 cases were presumed to be infected from March to May, and 103 cases (83.7%) showed a short incubation period. Infections during this period showed symptom onset from May to August, or 2 to 4 months later, rather than within 1 month of infection.

In precedent studies, cases with symptom onset before May were all classified as having long incubation periods [8], but our results included 26 cases with symptom onset before May who were presumed have short incubation periods. The KCDC's vector mosquito (Anopheles sinensis) survey is conducted at 27 stations in the P. vivax malaria risk regions (Gangwon-do 6, Gyeonggi-do 12, Incheon 9) from April to October every year. In the vector mosquito (Anopheles sinensis) survey results from 2007-2011, the mean first collection time was late May, while the appearance of vector mosquitos infected with parasites started from mid-July [15]. However, the vector mosquitoes of some regions had been collected as early as the first week of April [16], and in a military investigation, mosquitoes infected with the P. vivax malaria parasite were reported to appear in May, 2 months earlier than in the investigation of civilians [15].

The risk of being infected by P. vivax malaria is low from March to May, but there were a considerable number of cases who responded that they had visited the P. vivax malaria risk regions in those months, without any other risk factors for exposure to the vector mosquito of P. vivax malaria outside this time period. If they were not infected by P. vivax malaria during this period, it is possible that the P. vivax malaria non-risk regions in which they reside have, in fact, become risk regions. However, in the 'Surveillance of vector mosquitoes and investigation of P. vivax malaria parasites,' conducted in non-risk regions near the P. vivax malaria risk regions by the KCDC in 2012, P. vivax malaria vector mosquitoes were found to exist in non-risk regions, but no mosquitoes infected with the P. vivax malaria parasite were discovered [17]. Further research is needed, but it is currently assumed that there is a low risk of infection in regions near the P. vivax malaria risk regions.

As a limiting factor, subjects in our research who travel to P. vivax malaria risk regions as the cause of their infections, which were not linked to mosquito survey data. Hence, there is a need for additional research studies, such as vector mosquito monitoring and epidemiologic investigations, to examine whether these results are due to errors in memory or whether, as precedent studies have suggested, vector mosquitoes have begun to appear earlier and with longer periods of activity due to climate change, creating conditions favorable for P. vivax malaria transmission [18, 19].

In addition, among cases infected in June, only about 20% were long incubation period cases, whereas about 40% of infections occurring in July or August had long incubation periods despite the fact that the temperature and precipitation conditions were sufficient for the onset of P. vivax malaria. Hence, there is a need for further research.

This study analyzed the correlation between the mean proportion of long incubation period cases and monthly meteorological factors for each infected region from 2001 to 2010, with the following results. The proportion of long incubation-period cases showed a tendency to increase with increasing precipitation in October, with lowering mean temperature in September, and with lowering minimum temperatures in August and September. From 2005-2006, the highest 10-days occurrence of P. vivax malaria extended to early September; this increase of patients in September and October is considered to result from an increase in early-primary-attack cases due to short incubation period cases [5]. If the temperature in September increases in the future for reasons such as climate change, the malaria transmission period will be extended; in this case, it is expected that infections occurring in late August to early September will show short rather than long incubation periods, with the onset of symptoms developing by late October.

Among the strains of Plasmodium, only P. vivax malaria has a short incubation period in low-latitude tropical regions and both long and short incubation periods in subtropical regions. It is known that the proportion of short and long incubation periods depends on latitude [20]. In addition, P. vivax hibernans, the strain found in high-latitude regions such as Russia, is known to have only a long incubation period [20]. White [21] predicted that as the latitude increases, malaria cases with short incubation periods decline given that the mosquito breeding seasons shorten. Our results showed a clear positive correlation between the 10-year mean proportion of long incubation periods and the latitude of the relevant P. vivax malaria risk region in Korea. The difference in latitude between such P. vivax malaria risk regions in Korea is not big, but high-latitude regions such as Goseong-gun, Inje-gun, and Cheorwon-gun in Gangwon-do and Yeoncheon-gun in Gyeonggi-do are considered to have considerably different meteorological conditions, such as temperature in winter, compared with the low-latitude regions of Goyang-si in Gyeonggi-do and Incheon. Using the annual climatological report issued by the Korea Meteorological Administration, the mean and minimum temperatures of P. vivax malaria risk regions with weather stations were compared. Comparing Cheorwon-gun in Gangwon-do and Incheon revealed that the mean annual temperature over 10 years (2001-2010) was 2.5℃ lower in Cheorwon-gun compared with Incheon, while the mean annual minimum temperature was about 4.6℃ lower in Cheorwon-gun compared with Incheon. The difference in mean monthly temperature was clearer, and the difference in both mean and minimum temperature was biggest from January to March. In particular, the 10-year mean minimum temperature in January showed a difference of 7.5℃ [22]. Various explanations have been proposed for the two different incubation periods of P. vivax malaria [23], but the cause is likely not a single factor but rather a complex interaction of environmental factors, such as the meteorological conditions, latitude of risk regions, the principal vector's biological characteristics, adaptation to the environment, and characteristics of the infected cases.

The limitations of this study are as follows. First, the most important element in any incubation period study is having an accurate estimation of the time of infection, but this epidemiological investigation had to rely on the memory of individuals; thus, there could be an error in the travel histories. Second, as this research only investigated P. vivax malaria cases who had been infected during short-term visits, there were a small number of cases; this factor may have limited our ability to explain the incubation period of all P. vivax malaria cases in the country. However, as the purpose of the research was to identify the factors affecting incubation period, it was not necessary to examine all cases. The last limitation is that the cause-and-effect relationship cannot be confirmed using only cross-tabulation and correlation analysis, leaving open the possibility that the results happened by chance.

This study is significant in that it is the first study to comprehensively consider environmental and epidemiological factors influencing the incubation period of P. vivax malaria cases in Korea. In terms of factors that affected long and short incubation periods, results showed that there were significant differences according to the region of infection, time of infection, and time of symptom onset. In addition, there was a high positive correlation with the latitude of the region of infection, and there was a correlation with temperature in August and September. It is anticipated that P. vivax malaria cases with short incubation periods will increase in the future. Increasing proportion of patients with short incubation periods may be convenient for case management and treatment perspective, but there will be a need to improve vector mosquito control as the transmission period of P. vivax malaria extends. In addition, there is a need for additional research to comprehensively examine the vector mosquito (Anopheles sinensis), parasite strains, and epidemiologic investigation period in order to identify differences in the incubation period of P. vivax malaria in Korea.