Changes in malaria epidemiology in Germany, 2001–2016: a time series analysis

Background German surveillance data showed a sharp rise of malaria cases in 2014 and 2015 due to the increased arrival of refugees from malaria endemic countries. A time series analysis of data from 2001 to 2016 was performed in order to describe the epidemiology of imported malaria in Germany in general and of the recent increase in particular. Results In total, 11,678 malaria cases were notified between 2001 and 2016 (range 526–1063 cases/year). Newly arriving refugees averaged 10 cases/year (1.5%) in 2001–13 and 292.5 cases/year (28.3%) in 2014–15. Plasmodium (P.) falciparum was the most frequently reported species (range 57.2–85.8%), followed by P. vivax (range during 2001–2013: 7.6–18.1%; during 2014–2015, mean 31.3%). In 2014–15, 22.3% of all P. vivax cases were refugees from Eritrea and 3.3% from other countries of the Horn of Africa; in 2015 and 2016, 19.5% were refugees from Afghanistan and Pakistan. Five P. knowlesi malaria infections were reportedly acquired in Thailand between 2012 and 2016. Total numbers of malaria notifications among native Germans and residents with migration background showed an increasing trend since 2007. Chemoprophylaxis use was reported for 24.3% (1695/6984) of cases and showed a declining trend. Native German cases took significantly more frequently chemoprophylaxis than cases with migration background (32.6% vs. 17.9%; p < 0.001). Discussion/conclusions The steep rise in vivax malaria notifications in 2014 and 2015 was mainly due to newly arriving refugees from Eritrea but also from other countries of the Horn of Africa and South Asia. Clinicians should include malaria in their differential diagnosis in case of a febrile illness in the respective population and consider vivax malaria even if arrival to Germany dates back several months. Over the past 10 years, malaria notifications among native Germans and residents with migration background showed an increasing trend. Use of chemoprophylaxis was insufficient in both groups and deteriorating. New strategies need to be found to increase compliance to chemoprophylaxis recommendations. The surveillance provides valuable data for epidemiological assessment of imported malaria in Germany.

to chemoprophylaxis recommendations. Laboratory malaria diagnosis is notifiable to the national public health institute [Robert Koch-Institute (RKI)] according to the German Infection Protection Act [5]. Malaria notification data are an important source for monitoring trends and provide information about the risk of travellers to different destinations and about the impact of pretravel advice regarding chemoprophylaxis.
A sharp increase in malaria notifications occurred in 2014, coinciding with a large influx of newly arriving refugees in Germany. Roggelin et al. [6] reported an increase of vivax malaria in 2014-2015 in refugees from Eritrea admitted to the University Medical Centre in Hamburg [6,7], as also described in Sweden in 2014 [8,Sonden et al.,personal communication], and the Netherlands in 2017 [9].
In order to obtain a better understanding of malaria epidemiology in Germany in general and of the recent increase in malaria cases in particular, time series analysis using national notification data from the past 16 years was performed, aiming to describe demographic characteristics of cases, distribution of Plasmodium species, countries of infection, and geographic origin of cases over the time period between 2001 and 2016. Chemoprophylaxis use and disease outcomes were also analysed.

Methods
Direct or indirect detection of any type of Plasmodium in human blood is notifiable by laboratories directly to RKI, in an anonymous manner, regardless of the applied laboratory method. The RKI sends standardized data forms to laboratories for this purpose. Clinical data are added by an attending physician, filling in a second reporting form, which is forwarded to him or her by the laboratory. Both reporting forms are matched together at national level and entered into a database. The form asks for information on month and year of birth, gender, the three first digits of the patient's home post code, month and year of malaria diagnosis, Plasmodium species, geographic origin, if applicable date of immigration to Germany, travel history with dates, travel purpose and destinations, information on prophylaxis and treatment, hospitalization and death. An automated algorithm identifies potential double entries. These are manually screened and datasets are excluded if they prove to belong to the same patient. Laboratories or clinics are contacted by telephone or telefax in order to complete data.

Case definition
Case definition includes any person who had a direct or indirect proof of malaria parasite from their blood and whose country of residence was Germany at the time of diagnosis.

Data analysis
STATA SE14 and Microsoft Excel 2010 was used for performing time series analysis on notification data. Since 2001, when the German Infection Protection Act came into force, data was collected in a consistent manner, therefore data from 2001 to 2016 was included. The 'year of notification' was set as time variable. In order to be able to distinguish sub-groups among malaria cases according to their travel history and country of infection, four groups of different origin were defined: Group 1: 'native German cases' were cases with country of origin indicated as Germany. Group 2: 'residents with migration background' were cases whose country of origin was reported to be any other country than Germany and who were, not categorized as 'recent refugees' (group 3). Group 3: 'recent refugees' were cases fulfilling the following criteria: • Person reported as being in Germany as 'immigrant/ asylum seeker/refugee' or whose purpose of travel was indicated as 'immigrant/asylum seeker/refugee' or with no information of why being in Germany or purpose of travel AND who had arrived in Germany during 1 year prior to malaria diagnosis, OR. • In case of missing information on why being in Germany and purpose of travel and/or date of arrival in Germany, person whose country of origin was a country which was among the top 10 countries of the refugee statistics of the German Federal Office for Migration and Refugees (BAMF) for the respective year of malaria notification. As the date of arrival was often incomplete with missing month and day, only the year of arrival was used for calculating the 1-year time interval.
Group 4: cases not belonging to group 1-3, with incomplete or missing data concerning the criteria for group 1-3.
In order to calculate the percentage of cases who reported the use of chemoprophylaxis only cases with reported countries of infection, for which chemoprophylaxis was recommended for at least part of the country, based on the current malaria prophylaxis recommendations of the German society of tropical medicine and international health [10], were taken into account. Travel destinations for which in the past, but not any more in 2016, malaria prophylaxis was recommended were excluded from the calculations. The denominator included only cases with information on chemoprophylaxis use. Native German cases and cases with migration background were analysed separately. Recent refugees who acquired malaria in their home countries or on their travel route were not taken into account when looking at use of chemoprophylaxis.
If 'mixed infection' was indicated for Plasmodium species, the data collection form did not indicate the different species involved. Therefore, mixed infections were treated as 'information on species missing' and excluded when looking at the percentage of species and hospital admission rates according to different species.
Chi squared-test, t-test and Wilcoxon-Mann-Whitney-test were used in order to compare sub-groups. Differences were considered statistically significant at p < 0.05. Malaria incidence in refugees was calculated using data from BAMF [11] and malaria incidence in travellers for 2012-2016 using data on air travel departures to malaria-endemic countries, from the German Federal Statistical Office as denominators [12]. Recent refugees were excluded when calculating incidences of malaria based on air travel departure data.

Data completeness
Overall, 11,678 malaria cases were included in the analysis, which were notified in Germany over 16 years (2001-2016). 618 cases not fulfilling the reference definition and 804 datasets, which were double notifications, were excluded from the analysis. Since 2001, reported malaria case numbers steadily declined from 1049 cases in 2001 to 542 in 2007. Subsequently they remained relatively stable until 2013, and rose in 2014-2016 by 43% compared to the mean of the previous 10 years. The lowest annual number of malaria cases was reported in 2009 (n = 526) and the highest in 2015 (n = 1063).

Gender and age distribution and geographic origin of cases
During all years more male than female cases were observed. Between 2001 and 2013 male cases accounted for about two-thirds of cases (69.2%; 5878/8492). In 2014-2016, the male proportion rose to more than threequarters of cases (75.7%; 2277/3008).
Native German cases were the largest group in 2001 (45.5%; 477/1049) (Fig. 1) The percentage of children and adolescents (< 18 years) among all malaria cases notified in Germany was fairly stable during 2001-2013 with a mean of 9.7% (in average 64.3/665.2 cases/year were < 18 years old) and rose to 16.9% in 2014-16 (mean: 171/1010.3 cases/year). This increase was almost entirely due to the group of recent refugees, who were significantly younger than other cases (p < 0.001).

Plasmodium species
During the entire observation period, P. falciparum was the most frequently reported species (  The percentage of tertian malaria cases without differentiation was around 1.3% during 2001-2013 and rose to 4.2% in 2014/15. The percentage of Plasmodium malariae, Plasmodium ovale and mixed malaria cases remained fairly stable over 16 years at around 2.5 (292/11.678), 2.8 (331/11.678) and 3.2% (374/11.678), respectively. In absolute numbers, all types of malaria, except knowlesi malaria, showed an increase in 2014 and 2015.
Over the whole observation period, 6 P. knowlesi cases were notified between 2012 and 2015. Five of those were native German persons who acquired malaria in Thailand: four were tourists and one visited friends and relatives (VFR). For one knowlesi malaria case, neither travel destination nor reasons for travel were indicated.

Regions and countries in which infection was acquired
Malaria was acquired on the African continent by 7734 out of 8710 (88.8%) cases, with indication of country/ region of infection (Fig. 4); 95.4% (3803/3987) of cases Cases with countries of infection in Europe were making up 0.06% (5/8710) of all cases with information on place of infection. Two cases had falciparum malaria, 2 vivax malaria and 1 had a mixed infection. Two cases were reported with acquisition of malaria in France (without further specification, likely acquired in French overseas departments) and 3 in Germany. Nosocomial transmission was the probable cause for 2 cases with acquisition in Germany in 2007 and 2016. Both cases were women in their 30s, who were hospitalized at the same hospital ward as a patient with imported malaria caused by P. falciparum. Intense epidemiological investigations could not clarify the circumstances of transmission. For the remaining case information on the suspected mode of transmission was not provided.

Use of chemoprophylaxis and reasons for travel
For calculation of the percentage of cases who had reportedly taken chemoprophylaxis only cases with reported countries or regions of infection, for which chemoprophylaxis was recommended were taken into account.  (Fig. 7). Malaria incidence among air travellers to malaria-endemic countries has increased continuously from 0.52/1000 travellers in 2012 to 0.72/1000 travellers in 2016.

Hospital admissions and fatalities
The overall hospital admission rate was fairly stable over the observation period (annual mean: 65.1%; 7601/11,678) and equal among the different groups. Children and patients with falciparum malaria or knowlesi malaria were significantly more frequently hospitalized than adults and patients with other types of malaria (p = 0.002 and p < 0.001, respectively). Hospitalization was reported for 68.7% (5641/8217) of falciparum malaria, 59.3% (1352/2279) of tertian malaria, 47.3% (138/292) of quartan malaria cases and for 5 out of 6 knowlesi cases. The mean lag time between symptom onset and start of treatment was slightly longer in cases with migration background and refugees (6.1 days) than in native German cases (5.6 days; p = 0.121).
During the observation period, 51 deaths due to malaria were notified, making up a yearly mean of 3.2 deaths (0.4%, 51/11,678; range 1-8). The largest number of fatal cases occurred in the year 2001. Plasmodium falciparum was causative for 43/51 (84.3%) of those fatalities, 4 were due to mixed infections. In 4 fatalities, Plasmodium species was not specified. Forty-two fatal cases were of German origin (82.4%; 42/51), 7 were residents with migration background (13.7%, 7/51), 1 was a recent refugee, and for 1 no information about the origin was indicated. Among the fatal cases, infection was Seven of the fatal cases, 6 native German cases and 1 person with migration background, had reportedly taken chemoprophylaxis. Three had taken mefloquine, one a atovaquone and proguanil combination medication, for three information on the regimen was unavailable. The delay between symptom onset and start of treatment could be calculated for 7645 non fatal cases and 24 fatalities. The mean was in both groups 5.8 days with an interquartile rage of 0-213 and 2-8 days, respectively. For treatment 28/51 (38.9%) fatal cases had received quinine, 17/51 (23.6%) doxycycline, 4/51 (5.6%) artesunate, 3/51 (4.2%) artemether, and 2/51 (2.8%) each atovaquone and mefloquine. Twenty (39.2%; 20/51) had received more than one drug and for 16/51 cases the information on the administered treatment was missing.

Discussion
Time series analysis provides valuable data on trends of imported malaria cases in Germany. The decreasing trend in 2001-2009 did not continue. On the contrary, a sharp increase in malaria notifications was observed in 2014-2016 due to a high number of refugees from malaria-endemic countries arriving in Germany. At the same time, a mild increasing trend of malaria notifications over the past 10 years was observed among the German (non-refugee) population, along with changes in travel reasons and destinations and a decline in chemoprophylaxis use.
The largest part of the rise in P. vivax notifications was due to young male Eritrean refugees, as has been reported in 2014 from Sweden [8] and in 2016 from a German tropical medicine hospital [6]. Gender distribution in the German surveillance data showed a predominance of male cases for all malaria notifications during all years (men:women 2:1). This was attributed to differences in travel and prevention behaviour and to the composition of groups of people migrating to Germany [7]. The gender difference in 2014-2016 was more pronounced (men:women 3:1) and cases were on average significantly younger than in previous years. Why was there such a sudden increase of young men fleeing from Eritrea in 2014? According to the UN Human Rights Council [13] and Amnesty International [14] the main driver to leave Eritrea was the national conscription practice into military services which in practice amounts to forced labour [13,14]. Since the law of the national service came into force in 1995, Eritreans have been fleeing their country, with large numbers residing in Sudan and Ethiopia [15]. According to the United Nations High Commissioner for Refugees (UNHCR) an intensified recruitment drive in 2014 and increasingly bad living conditions in the camps in neighbouring countries were causing a nearly 3-fold increase of Eritrean refugees in Europe in 2014, mainly in Germany, Sweden, Switzerland and the Netherlands [15]. The German Federal Office of Migration and Refugees reported a more than 23-fold increase of newly arriving Eritreans in 2014 and 2015, compared to the years 2001-2012 [11]. However, Eritreans accounted for 2.5% of asylum applications registered in 2015 in Germany only [11].
Most of the Eritrean cases presented with vivax malaria, only 5% were notified with falciparum malaria. According to the World Malaria Report 2016 [1], the predominant Plasmodium species in Eritrea and along the travel route (Ethiopia and Sudan) is P. falciparum (Eritrea: 75%; P. vivax: 25% [1]). As most Eritrean refugees travelled overland via Ethiopia and Sudan to Libya and then by boat to Italy [15], their travel time to arrival in Germany is longer than the usual incubation period for falciparum malaria. Sonden et al. (pers. comm.) report that a considerable number of cases arriving in Germany and Sweden had sought malaria treatment during migration, notably in Italy where national health services and non-governmental organizations provide medical care [16]. The high incidence of P. vivax hints at an accumulation of cases due to an increased probability of malaria acquisition in people who may have spent the nights outside during several months of travel, inadequately taken medication for relapse prevention (i.e., primaquine), due to unavailability or poor compliance, in addition to ongoing increased P. vivax transmission or outbreaks along the migration route (Sonden et al. pers. comm.).
Alongside refugees from Eritrea, there was an increase of malaria infections in refugees arriving from other countries of the Horn of Africa, Sudan, Pakistan, and Afghanistan. When plotting species distribution against time after withdrawing the population of recent refugees and cases without information on their origin, the rise of P. vivax cases in 2014 and 2015 was no longer visible (Fig. 3b). The refugee population, comprising many teenagers and young adults, might be more vulnerable and potentially less self-competent in attending medical care than adults who are acquainted with the German health care system. Clinicians must be aware of the possibility of malaria infection in refugees from the aforementioned countries, even if their arrival to Germany dates back several months. Relapses may occur frequently in inadequately treated patients and can occur when patients have resided in Germany for some time. Timely diagnosis is especially important, particularly as severe clinical pictures in this population were reported from different European countries probably due to delayed diagnosis and a generally impaired health status, presumably partly caused by the long and strenuous migration (Sonden et al. pers. comm.). This is an interesting observation, given that people from malaria endemic countries are likely to have acquired semi-immunity and should thus be partly protected against severe disease. The surveillance data do not provide any clinical information that could help to better understand this contradiction.
Overall, from 2001 to 2016, by far the most malaria infections were acquired in Africa with little variation in the most frequently reported countries (Ghana, Nigeria, Cameroon). The percentage of Asian countries of infection was low. It was fairly stable over the 15-year period (around 7%) with the exception of a 10% increase in 2012, which was mainly due to cases recently immigrated to Germany from Pakistan [17], and a rise in 2015 due to cases from Afghanistan and Pakistan. At the same time, groups who acquired malaria in Asia and countries of infection within Asia have changed. During the first half of the observation period the majority of cases were native Germans who had travelled as tourists to Southeast Asia, mainly to Indonesia and Thailand. From 2009 onwards, most cases were refugees or had migration background and were reported to have acquired malaria in Pakistan, India and Afghanistan. Southeast Asian countries were four times more frequently named as country of infection between 2001 and 2008 than between 2009 and 2016. This change may reflect an altering pattern of malaria epidemiology in Southeast Asia, or changes of travel destinations, or both. The risk for malaria in most regions in Southeast Asia is categorized as low, very low or non-existent. No regular chemoprophylaxis is recommended and carrying of standby emergency treatment for only some destinations [10]. The risk is still considered high for some regions in Indonesia [10]. But a decrease of malaria incidence in European travellers to Indonesia over the past decade, as described by Johansson Århem et al. [18], was also reflected in surveillance data for Indonesia and neighbouring Papua New Guinea.
Five knowlesi malaria infections acquired in Thailand were reported, and one additional without indication of travel destination. All those cases were reported between 2012 and 2016, mainly in tourists. It is likely that other P. knowlesi infections might have been misclassified as other malaria species, as correct microscopy diagnosis is complicated due to morphological similarity of the parasite with P. falciparum and P. malariae [19]. In addition, the same drugs as used for other types of malaria are effective in P. knowlesi cases [19]. It remains to be seen whether there will be more reports of P. knowlesi infections in the future as awareness of this Plasmodium species among microbiologists and clinicians rises and diagnostic tests improve.
An overall moderately rising trend of malaria notifications over the past 10 years was present among native German cases and residents with migration background. According to the German Federal Statistical Office, the number of air travel departures to malaria-endemic countries has increased by 10% between 2012 and 2016 [12]. Since 2006, the number of cases with migration background has exceeded native German cases. VFR was the foremost reported travel reason for cases with migration background during all 16 years of malaria surveillance. The largest sub-group of malaria cases overall were adult male residents with migration background who acquired falciparum malaria in Africa when VFR, sub-Sahara Africa and in particular West Africa being the regions with the highest risk of malaria acquisition. Among native German cases the picture has changed over the years. Tourism used to be the most important motive for travel, but its predominance was becoming less and less pronounced, while VFR, and work, education and training were becoming equally important reasons for travel. In the context of globalization, an overall growing mobility and an increasing number of people who are working and living abroad, and thus receiving friends and relatives for visits, might partly explain the overall increase of cases.
This data, showing an increasing trend of imported malaria in Germany, underlines the importance of pretravel counselling, as malaria in travellers is at a large part preventable [20]. Schlagenhauf et al. [20] postulate that pre-travel consultations are associated with reduced morbidity of falciparum malaria in all traveller subgroups. German surveillance data and notably data on countries or regions of infection are used for updating the malaria prophylaxis recommendations of the German society of tropical medicine and international health.
Inherently, malaria cases show a low percentage of chemoprophylaxis use, as compliant prophylaxis users are less likely to develop malaria. Nevertheless the reported share of chemoprophylaxis use was low (24%) compared to reports from other European countries [21,22] and the USA [23], where it ranged from 28 to 40%. When calculating this percentage only cases with "yes" or "no" for chemoprophylaxis use were taken into account and not cases lacking information on this topic. Therefore, the percentage of chemoprophylaxis users was probably overestimated, as one can assume that cases with no information are more likely to have not taken chemoprophylaxis. Furthermore, the overall share of cases who did take chemoprophylaxis was continuously decreasing over the observation period. A similar trend is observed in other European countries [24] and the USA [23,24]. The observed decrease was most pronounced among native German cases. Reasons can only be speculated about. Growing mobility and perceived easiness and affordability of travel to exotic destinations may alter a person's attitude towards travel preparations and risk perception. Fear of side effects may also be a reason not to take chemoprophylaxis. An increasing trend of malaria incidence among air travellers to countries with known malaria transmission was observed between 2012 and 2016. Single incidence calculations might be quite inaccurate as neither overland (non-air) travel nor the varying malaria endemicity in different countries, nor seasonality were taken into account, but the overall trend is worrying [12].
The proportion of cases who took chemoprophylaxis was markedly lower among cases with migration background than among native German cases. This disparity is known [22,[24][25][26] and is mostly attributed to a lower risk perception of acquiring malaria and of suffering from severe malaria, and the lower uptake of pre-travel counselling by the population with migration background [25,26]. This is particularly worrying since people whose families originate from malaria-endemic countries tend to use less mosquito protection measures, travel for longer periods and often to more remote areas than tourists and business travellers. In addition, this group of people is reported to seek medical attention in case of a febrile illness later upon return to Europe [25,26]. Interestingly, despite residents with migration background taking less chemoprophylaxis, there were fewer fatal cases among them than among native German cases. This might be due to persisting partial immunity among cases originating from malaria-endemic countries. The surveillance data does not give information on attendance of pre-travel consultations. Awareness of the risk of malaria infection among travellers needs enhancement. Strategies aiming to increase uptake of and compliance to malaria chemoprophylaxis recommendations are needed, especially targeting those groups who were identified as taking only rarely chemoprophylaxis, e.g. residents with migration background. The present data reveal that malaria cases frequently report travel to highrisk areas, such as West Africa and therefore provide important travel medicine information for the training of physicians.
Notification data suffers from under-reporting. As a large number of cases was primarily diagnosed or transferred to large specialized centres for tropical medicine, which have well established notification mechanisms, a selection bias is unlikely to occur. Education of physicians to completely fill in the notification forms (e.g. country of infection) is needed to improve completeness of data. Misclassification of cases to different patient groups is possible. Recent refugees and residents with migration background might have less complete notification forms due to communication barriers. Therefore, more cases from those two groups might have been classified into the group with unknown origin than native German cases. This might have led to underestimation of a number of results, such as the problem of vivax malaria in refugee cases and the predominance of cases with migration background among all malaria cases. The notification form asks for the probable country of infection and does not include questions on exact travel destinations within the countries. When calculating percentages of chemoprophylaxis use all cases, who had travelled to countries with recommendation for chemoprophylaxis for at least part of the country were included. This might have led to an underestimation of chemoprophylaxis use. Changes over time in recommendations for prophylaxis could not be taken into account. Travel destinations for which in the past malaria prophylaxis was recommended, e.g. some countries in Southeast Asia, were thus excluded from the calculations. There is no reason to assume that those cases were systematically different from others.

Conclusions
The steep rise in vivax malaria notifications in 2014 and 2015 was mainly due to newly arriving refugees from Eritrea but also due to refugees from other countries of the Horn of Africa and South Asia, predominantly young men. Clinicians should include malaria in their differential diagnosis in case of a febrile illness in this population even if arrival to Germany dates back several months.
Over the past 10 years malaria notifications among native Germans and residents with migration background showed an increasing trend. Use of chemoprophylaxis was insufficient in both groups and deteriorating. New strategies need to be found in order to increase compliance to chemoprophylaxis recommendations. The surveillance provides valuable data for the epidemiological assessment of imported malaria in Germany.
Authors' contributions SVB conducted the data analysis, interpreted the results and wrote the manuscript; KS gave scientific support and revised the manuscript. Both authors read and approved the final manuscript.