Male and female Ethiopian and Kenyan runners are the fastest and the youngest in both half and full marathon

In major marathon races such as the ‘World Marathon Majors’, female and male East African runners particularly from Ethiopia and Kenya are the fastest. However, whether this trend appears for female and male Ethiopians and Kenyans at recreational level runners (i.e. races at national level) and in shorter road races (e.g. in half-marathon races) has not been studied yet. Thus, the aim of the present study was to examine differences in the performance and the age of female and male runners from East Africa (i.e. Ethiopians and Kenyans) between half- and full marathons. Data from 508,108 athletes (125,894 female and 328,430 male half-marathoners and 10,205 female and 43,489 male marathoners) originating from 126 countries and competing between 1999 and 2014 in all road-based half-marathons and marathons held in one country (Switzerland) were analysed using Chi square (χ2) tests, mixed-effects regression analyses and one-way analyses of variance. In half-marathons, 48 women (0.038 %) and 63 men (0.019 %) were from Ethiopia and 80 women (0.063 %) and 134 men (0.040 %) from Kenya. In marathons, three women (0.029 %) and 15 men (0.034 %) were from Ethiopia and two women (0.019 %) and 33 men (0.075 %) from Kenya. There was no statistically significant association between the nationality of East Africans and the format of a race. In both women and men, the fastest race times in half-marathons and marathons were achieved by East African runners (p < 0.001). Ethiopian and Kenyan runners were the youngest in both sexes and formats of race (p < 0.001). In summary, women and men from Ethiopia and Kenya, despite they accounted for <0.1 % in half-marathons and marathons, achieved the fastest race times and were the youngest in both half-marathons and marathons. These findings confirmed in the case of half-marathon the trend previously observed in marathon races for a better performance and a younger age in East African runners from Ethiopia and Kenya.


Background
Marathon and half-marathon races are very popular running events held all over the world with an increasing number of both races and participants during the last decades. For instance, in the USA, there were more than 1200 marathons held in 2014 compared to about 300 marathons held in 2000 (www.runningusa.org/2015national-runner-survey). The number of successful marathon finishers increased from 25,000 in 1976 to the all-time high of 550,637 in 2014. Compared to marathons, however, most of the runners competed in the USA in half-marathons. The number of successful halfmarathoners increased from 303,000 in 1990 to the alltime high of 2,046,600 in 2014 (www.runningusa.org/ half-marathon-report-2015). In fact, 3.7 times more halfmarathoners than marathoners competed in the USA in 2014. In smaller countries such as Switzerland in Europe, a total of 226,754 half-marathoners and 86,419 marathoners competed between 2000 and 2010 (Anthony et al. 2014). In other terms, 2.6 times more half-marathoners competed than marathoners. In 2010, 8690 women and Open Access *Correspondence: beat.knechtle@hispeed.ch 1 Facharzt FMH für Allgemeinmedizin, Gesundheitszentrum St. Gallen, Vadianstrasse 26, St. Gallen 9001, Switzerland Full list of author information is available at the end of the article Knechtle et al. SpringerPlus (2016) 5:223 21,583 men finished a half-marathon in comparison to 2904 female and 9333 male finishers in 2000, respectively, corresponding to an increase of 299 % for women and of 231 % for men over 10 years. In contrast, the number of male and female full marathoners increased until 2005 only and decreased thereafter (Anthony et al. 2014).
The dominance of East-African women and men in marathon running is well known (Hamilton 2000;Onywera et al. 2006;Tucker et al. 2015;Wilber and Pitsiladis 2012). Athletes from both Ethiopia and Kenya dominate marathon running for a long time (www.iaaf. org). In the top list of the International Association of Athletics Federations (IAAF) for male marathoners, the first best 37 marathon race times were achieved by athletes from Ethiopia and Kenya (www.iaaf.org/records/ toplists/road-running/marathon/outdoor/men/senior). In women, however, the three fastest marathon race times were achieved by an athlete from Great Britain followed by two female marathoners from Kenya (www.iaaf. org/records/toplists/road-running/marathon/outdoor/ women/senior). In the 'World Marathon Majors' with the largest city marathons worldwide, female and male champions are exclusively from East African particularly from Ethiopia and Kenya (www.worldmarathonmajors. com/champions/current-champions).
The reasons for the dominance of East-African runners in long and middle distance running events such as marathons included environmental conditions such as a specific geographic background Scott et al. 2003;Tucker et al. 2015). The dominance of East-African distance runners is primarily a Kenyan phenomenon, with majority of the Kenyan runners originating from the Kalenjin tribe in general and the Nandi sub-tribe in particular Tucker et al. 2015). Similar to Kenyan runners, elite Ethiopian runners are also of a distinct environmental background where marathoners mainly originate from the altitudinous regions of Arsi and Shewa (Scott et al. 2003).
However, there is paucity of information with regards to basic characteristics such as age and trends in performance of East-African half-marathoners (Aschmann et al. 2013;Cribari et al. 2013). These studies investigated all African half-and full marathoners competing in one country (Switzerland) together without a separation of East-African runners in their nationalities (Aschmann et al. 2013) or investigated a limited sample of the best athletes . Indeed, East African runners particularly those from Ethiopia and Kenya account for the largest percentage of African runners in halfmarathon and marathon (Aschmann et al. 2013). A recent study showed different barriers across both sex and distance (Wegner et al. 2015); hence, these trends might vary between half-marathon and marathon. The knowledge of East African's basic characteristics such as age, participation and performance trends might help coaches, fitness trainers and sports scientists to improve their understanding of half-marathon's demands.
Therefore, the aim of this study was to investigate performance and age of Ethiopian and Kenyan half-and full marathoners who competed between 1999 and 2014 in races held within one country (Switzerland) in a sample of more than 500,000 successful finishers. We hypothesized that female and male runners from Ethiopia and Kenya would also be the fastest in half-marathon races.

Ethics
The study was approved by the Institutional Review Board of St. Gallen, Switzerland, with waiver of the requirement for informed consent given that the study involved the analysis of publicly available data.

Data collection and data analysis
All half-marathons and marathons held in Switzerland from 1999 to 2014 were identified by using 'Laufkalender Schweiz' (www.laufkalender.ch). Since 1999, all running races in Switzerland started with an electronic chip system and full race results (i.e. name, age, sex, nationality and race time of the finishers) were available since then on the website of the specific races. Of all races, only those half-marathons and marathons were considered which were held on a road, not on a trail. No mountain marathons were included; start and finish of the race had to be on the same altitude. Athletes with missing age and/ or missing nationality were excluded from data analysis. In order to avoid a selection bias due to a limitation to top runners, we considered all finishers from all countries. To investigate a trend in participation and performance, athletes from countries where at least one women and/or one man competed in at least 8 years (i.e. half of the investigated period of time) were considered.

Statistical analysis
Each set of data was tested for normal distribution (D' Agostino and Pearson omnibus normality test) and for homogeneity of variances (Levine's test) prior to statistical analyses. Trends in participation across calendar years were analysed using regression analysis with linear growth equation models. Differences in the participation of East African runners by nationality and sex to halfmarathons and marathon were examined by using Chi square (χ 2 ) test. To investigate changes in performance across calendar years, we used a mixed-effects regression model with running speed as the dependent variable. We analysed women and men separately for each country for both half-marathon and marathon and included calendar year, sex, centered age, and squared centered age as fixed variables. To investigate changes in age across calendar years, we used a mixed-effects regression model with age as the dependent variable. For the change in age over time, we combined women and men for each country and included sex and calendar year as fixed variables. Differences in age and performance between athletes from multiple countries were compared using one-way analysis of variance (ANOVA) with subsequent Tukey's multiple comparison tests with a single pooled variance. Statistical analyses were performed using IBM SPSS Statistics (Version 22, IBM SPSS, Chicago, IL, USA) and GraphPad Prism (Version 6.01, GraphPad Software, La Jolla, CA, USA). Significance was accepted at p < 0.05 (two-sided for t tests). Data in the text and tables are given as mean ± standard deviation (SD).

Participation
Data from a total of 508,108 (125,894 female and 328,430 male half-marathoners and 10,205 female and 43,489 male marathoners) athletes was considered. These runners originated from a total of 126 countries spread around the globe. Table 1 summarizes the athletes from the considered countries for data analysis across calendar years in half-marathons (35 countries) and marathons (15 countries).
In half-marathons, 48 women (0.038 %) and 63 men (0.019 %) originated from Ethiopia and 80 women (0.063 %) and 134 men (0.040 %) from Kenya. In marathons, three women (0.029 %) and 15 men (0.034 %) were from Ethiopia and two women (0.019 %) and 33 men (0.075 %) from Kenya. There was no statistically significant association between the nationality of East Africans and the format of the race [χ 2 (1) = 0.001, p = 0.978]; that was, both Ethiopians and Kenyans equally participated to half-marathons versus marathons. Also, there was no association between male East Africans and the format of the race [χ 2 (1) = 0.001, p = 0.922]; i.e. both male Ethiopians and Kenyans accounted equally to the two formats.
Most of the successful finishers originated from Switzerland, Germany and France in both half-marathons and marathons. In half-marathons, the number of women (r 2 = 0.98, p < 0.0001) and men (r 2 = 0.98, p < 0.0001) increased significantly. Similarly, the number of women (r 2 = 0.46, p = 0.0041) and men (r 2 = 0.51, p = 0.0019) increased significantly in marathons. Regarding the considered countries, the number of female half-marathoners from Canada (r 2 = 0.81, p = 0.002), Germany (r 2 = 0.97, p = 0.005), Switzerland (r 2 = 0.97, p = 0.005) and Belgium (r 2 = 0.72, p < 0.0001) increased significantly. For male half-marathoners, the number of participants from France (r 2 = 0.97, p = 0.018), Great Britain (r 2 = 0.88, p = 0.036), Principality of Liechtenstein (r 2 = 0.87, p < 0.0001), Poland (r 2 = 0.65, p < 0.0001), South Africa (r 2 = 0.63, p = 0.006) and Argentina (r 2 = 0.70, p < 0.0001) increased significantly. In marathoners, there was no significant increase in the number of men regarding the country. In women, however, participants from France (r 2 = 0.46, p = 0.0275) and Japan (r 2 = 0.47, p = 0.0039) increased significantly their numbers. Table 2 shows the running speed of the female and male half-marathoners. Running speed decreased significantly in women from France, Switzerland, and Australia, but increased in women from Norway and Portugal (Table 3). In men, running speed decreased in athletes from Germany (Table 4). Table 5 presents running speed of female and male marathoners. Running speed remained unchanged in female marathoners (Table 6) but increased in British men (Table 7). Table 8 presents the age of the female and male half-marathoners. Age increased significantly across calendar years in women from Austria and Norway and in men from Japan and Norway (Table 9). In marathoners (Table 10), age decreased significantly in men from Italy and Principality of Liechtenstein, but increased significantly in men from Poland (Table 11). Table 12 presents running speed and age of female and male half-marathoners and marathoners sorted from the fastest to the slowest and from the youngest to the oldest. In absolute values, women from Kenya and Ethiopia were running the fastest. Kenyan women were not faster than Ethiopian women (p > 0.05) but they were significantly faster than all other women (p < 0.001 to p < 0.0001). Ethiopian women were not faster than women from Kenya, Portugal, Principality of Liechtenstein and Hungary (p > 0.05), but significantly faster than all other women (p < 0.001 to p < 0.0001). For men, Kenyans and Ethiopians were running the fastest regarding in absolute terms. Kenyan men were not faster than Ethiopian men (p > 0.05), but significantly faster than all other men (p < 0.001 to p < 0.0001). Ethiopian men were not faster than men from Portugal, Principality of Liechtenstein, Italy, Switzerland and Hungary (p > 0.05), but significantly faster than all other men (p < 0.001 to p < 0.0001).

Performance of the fastest and age of the youngest
Considering age, women from Ethiopia and Kenya were the youngest in absolute terms. However, Ethiopian women were not younger than women from Russia, Czech Republic, Argentina, India, Slovenia, Ireland, USA, Great Britain, Poland, Canada, Greece, Denmark and Spain (p > 0.05). Considering athletes from the other countries, women from Ethiopia were In marathon, women from Ethiopia and Kenya were faster than women from all other countries (p < 0.001 to p < 0.0001). However, Ethiopian women were not faster than Kenyan women (p > 0.05). For men, the fastest running speeds were achieved by athletes from Kenya, Ethiopia and Principality of Liechtenstein. Kenyan men were faster than men from all other countries (p < 0.001 to p < 0.0001) with the exception of Ethiopian men (p > 0.05). Ethiopian men were, however, not faster than men from Liechtenstein, Switzerland, Belgium, Spain, Italy, France, Great Britain, Germany and USA (p > 0.05).
Women from Ethiopia and Kenya were the youngest in absolute terms. However, only women from Japan were significantly older than women from Ethiopia (p = 0.001) but not all other women (p > 0.05). Considering Kenyan women, no statistical significant differences were found between the countries (p > 0.05). For men, Ethiopians and Kenyans were the youngest in absolute terms. Ethiopian men were not younger than Kenyan men (p > 0.05), but significantly younger than men from all other countries (p < 0.001 to p < 0.0001). Men from Kenya were not younger than men from Liechtenstein, Great Britain, Poland and the USA, but significantly younger than men from all other countries (p < 0.001 to p < 0.0001).

Discussion
This study intended to investigate performance and age of female and male Ethiopian and Kenyan half-marathoners and marathoners competing in races held in one country. The most important findings for female and male half-marathons and marathoners from Ethiopia and Kenya were that, (1) they accounted for less than 0.1 %, (2) they were running the fastest and, (3) they were the youngest.

Low participation of East African runners
A first important finding was that runners from Kenya and Ethiopia accounted for less than 0.1 % in both halfmarathons and marathons. The small percentage of participants from these countries should be attributed partially to the distance between these countries and the place of race. Considering the nationality of participants, one might observe a very large number of local participants followed by participants from the neighbouring countries.
Although athletes from neighbouring countries such as Germany, France, Italy and Austria were very numerous, also athletes from very remote countries such as the United States, Japan and Australia competed more numerous than athletes from Ethiopia and Kenya. A very likely explanation could be the income of persons living in these countries since they need to spend money for the travel to and the stay in Switzerland. Costs of living are very high in Switzerland compared to other countries (www.numbeo.com/cost-of-living/country_result. jsp?country=Switzerland). When we compare the gross domestic product (GDP) per capita for persons living in East African countries such as Ethiopia (www. indexmundi.com/ethiopia/gdp_per_capita_%28ppp%29. html) and in Kenya (http://www.indexmundi.com/ kenya/gdp_per_capita_%28ppp%29.html) with $1300 and $1800, respectively, persons from the other countries such as the United States of America (www.indexmundi.com/united_states/gdp_per_capita_%28ppp%29. html), Japan (www.indexmundi.com/japan/gdp_per_ capita_%28ppp%29.html) and Australia (www.indexmundi.com/australia/gdp_per_capita_%28ppp%29.html) have a GDP of $52,800, $ 37,100, and $43,000, respectively. With these higher GDP, persons from the United States of America, Japan and Australia might easier travel to Switzerland for competing in a marathon than persons from Ethiopia and Kenya.
The finding that mainly local athletes compete in races followed by athletes from surrounding countries confirms recent findings for other races. For example, in longdistance triathletes competing in the 'Ironman Hawaii' , women and men from the United States of America dominated both participation and performance (Dähler et al. 2014). In solo swimmers crossing the 'English Channel'        ). However, the most likely explanation for the very low participation of East African runners in half-marathons and marathons held in Switzerland are economic reasons. For Kenyan runners, marathon running is a means of making money to help their families, parents and siblings Onywera 2009). Onywera (2009) described economic reasons for Kenyan athletes as one of the most important factors to compete in marathon running, which might be undercharged so far (Hamilton and Weston 2000). Prize money in Swiss half-marathons and marathons is very low compared to prize money offered in the 'World Marathon Majors' (www.worldmarathonmajors.com). For the winner in the 'Zurich Marathon' in Switzerland, the prize money is 10,000 Swiss Francs (www. zurichmarathon.ch) which is very low in contrast to the prize money offered in large city marathons. Indeed, overall prize money in races of the 'World Marathon Majors' is considerably higher (www.worldmarathonmajors.com). In the 'BMW Berlin Marathon', the 'Tokyo Marathon' , and the 'Virgin London Marathon' the prize money is $1,000,000,    php/prizes). The differences in prize money seem very similar in half-marathon compared to marathon. In a large half-marathon held in Switzerland such as the 'Hallwilerseelauf' , the prize money for both women and men for the top five is, however, only CHF 600, 400, 300, 200, and 100, respectively (www.hallwilerseelauf.ch). In an elite halfmarathon such as the 'IAAF/AL-Bank World Half Marathon Championships' , a total prize purse of US$245,000 will be paid by the IAAF for the men's and women's races (www.iaaf.org/news/news/prize-money).

East African runners were the fastest in half-marathons and marathons
A second finding was that female and male runners from Kenya and Ethiopia were the fastest in both half-marathons and marathons. The dominance of East African runners was evident for both marathon and half-marathon but differed from longer distances. For instance, it has been shown that male Japanese runners were the best in 100-km ultra-marathons (Cejka et al. 2014). The trend in performance across years should be explained by a model showing that human speed after having progressed fast in the past has now reached a plateau and further progression should be attributed to an enlarged  population of runners and improved training practices (Desgorces et al. 2012).

East African runners were the youngest in half-marathons and marathons
A third important finding was that women and men from Kenya and Ethiopia were the youngest in both halfmarathons and marathons. Their mean age is considerably lower as has been reported for elite and recreational marathoners. The age of elite marathoners is at around 29-30 years when the nationality was not considered (Hunter et al. 2011). In female and male marathoners competing between 1979 and 2014 in the 'Stockholm Marathon' , the age of the fastest marathon performance was even higher with 34.3 ± 2.6 years (Lehto 2015). In a study investigating runners competing in Swiss halfmarathons and marathons from 2000 to 2010 and considering the top five African and Non-African runners, the mean age of the male runners was significantly higher for Non-African runners than for African runners in both half-marathons (Non-African runners 31.1 ± 6.4 years, African runners 26.2 ± 4.9 years) and marathons (Non-African runners 33.0 ± 4.8 years, African runners 28.6 ± 3.8 years). In marathons, the top five female Non-African runners (31.6 ± 4.8 years) were ~4 years older than the top five female African runners (27.8 ± 5.3 years) (Aschmann et al. 2013). The difference in age between East Africans and Europeans found in the present study was not in agreement with a previous comparison between African and non-African runners of marathons and half-marathons  indicating that the younger age was a specific characteristic of East Africans and should not be generalized to all African runners.

Physiological interpretation
For the dominance of East African runners such as Kenyan runners, physiological aspects need to be considered (Larsen 2003;Larsen and Sheel 2015). It has been supported that running speed sustained over a prolonged time depends on the maximal sustainable VO 2 (oxygen uptake) and running economy (Millet et al. 2012). A comparison between European and Eritrean long-distance runners showed that Eritreans, despite having a lower VO 2 max (maximum oxygen uptake), had a better running economy at 19 km h −1 . A better running economy might explain the supremacy of East Africans in the marathon, and the delayed glycogen depletion and reduced thermal stress have been suggested to be associated with a better running economy (Millet et al. 2012). An exceptional biomechanical and metabolic economy, chronic exposition to altitude, sociocultural background and a strong psychological motivation were highlighted as other factors of this supremacy (Onywera 2009;Wilber and Pitsiladis 2012). Moreover, the impact of stereotypes has also been noticed because, independently from the possible existence of physiological advantages in East Africans, the belief that such differences exist can impact performance by creating a psychological atmosphere (Baker and Norton 2003). With regards to their nutritional habits, a research on the dietary intake of Ethiopian long distance Data for Non-African runners are sorted in order of the number of finishers of each country  runners has shown that they met most recommendations for endurance athletes (Beis et al. 2011). A study on the diet of Kenyan endurance runners revealed that it composed mostly by carbohydrates (~67 %) and less by protein (~15 %) or fat (~17 %) (Fudge et al. 2006). In addition to the abovementioned physiological factors, Eastern African runners might differ from runners of other origin with regards to other specific anthropometric characteristics (Kohn et al. 2007;Lucia et al. 2006;Prommer et al. 2010;Vernillo et al. 2013). For instance, compared to elite German 10-km runners, elite Kenyan runners had a similar VO 2 max (ml min −1 kg −1 ) but were lighter by more than 9 kg (Prommer et al. 2010). Xhosa 10-km runners had also similar VO 2 max (ml min −1 kg −1 ) as their Caucasian counterparts, but they were lighter and shorter (Kohn et al. 2007). Eritrean distance runners had a lower body mass index and a better running economy at 21 km h −1 than Spanish runners, whereas their VO 2 max was similar (Lucia et al. 2006). In top class Kenyan marathoners, ectomorphy is dominant, but endomorphy and mesomorphy is more than one-half unit lower (Vernillo et al. 2013).
A review of genetic and lifestyle factors of the performance of the East Africans distance runners concluded that the findings on candidate genes linked to performance of Caucasian populations were not confirmed in East Africans showing research methods' limitations and the polygenic nature of performance (Tucker et al. 2013). This was in agreement with another review showing that distance running success of East Africans was not based on a unique genetic profile (Wilber and Pitsiladis 2012). Another parameter that has not been studied previously as much as the abovementioned parameters might be the physical activity and inactivity levels when athletes did Data for Non-African runners are sorted in order of the number of finishers of each country Cage centered age, Cage 2 centered age squared  not practise their sport. Surprisingly, a study in marathon and half-marathon runners showed that these athletes trained for 6.5 h weekly, but they also spent much more time sitting (Whitfield et al. 2014). The aforementioned study found no relationship between sitting time and performance. However, potential differences in non-sport physical activities and inactivity levels between East Africans and Europeans should be examined in future studies.

Limitations
A limitation of this analysis is the fact that an athlete may have changed his/her nationality, where, for example, an athlete from an African country might have been naturalized in another country. As an example, the Swiss marathoner Tadesse Abraham was born in Eritrea but is now a Swiss citizen. He won three marathons and one halfmarathon in Switzerland (www.tadesse-abraham.ch).
On the other hand, the focus of the present study was on half-marathon runners' characteristics (i.e. age, participation and performance) with regards to marathon. Since there was no evidence that the above-mentioned concern about the nationality appeared differently to the two events (half-marathon vs. marathon), it might be supported that it did not affect the overall findings.

Conclusions
In summary, women and men from Kenya and Ethiopia, despite they accounted for less than 0.1 % in half-marathons and marathons, achieved the fastest race times and were the youngest in both half-marathons and marathon. These findings confirmed in the case of half-marathon the trend previously observed in marathon races for a better performance and a younger age in East African runners compared to Non-African runners.
Data for Non-African runners are sorted in order of the number of finishers of each country Cage centered age, Cage 2 centered age squared