Performance data of all athletes with results at age 100 years and older were collected from the publicly available “World Master Rankings” data base (www.mastersrankings.com). As the data base reached back to the year 2006, and since restrictions due to the COVID-19 pandemic led to the absence of data in the year 2020, results of both sexes from 2006 to 2019 were considered. For all athletes who had a result at an age of 100 years and older, the data base was searched for additional results from previous ages going back to as early as the age of 90 years to obtain a longitudinal trajectory of the individual.

The new data were compared to the data of all 80+-year-old athletes in a data set from Sweden that was published previously (Ganse et al., 2020b). Briefly, this data set comprises all data from the Swedish master athletics rankings from the years 1901 to 2019 and is the largest longitudinal master athletics data set published to date.

Changes in weights of javelins, discusses, and shots with age did not affect the results in the present analysis, as implement weights in the throwing disciplines stay constant for athletes 80 years and older. The following implements are used by athletes 80 years and older: shot put: men 3 kg, women 2 kg; discus throw: men 1 kg, women 750 g; javelin throw: men and women 400 g.

All statistical tests were executed with IBM SPSS Statistics version 25. Regression statistics were employed to compare performance decline rates between disciplines and age groups. Regression coefficients and their corresponding p-values were calculated (significance level 0.05). Disciplines were included in the study if the following inclusion criteria were met: at least three athletes with at least two results in the data set and at least one result at an age of 100 years or more. Linear regression decline rates were computed for the younger (the data from Sweden) and the older (the data of the athletes who still compete at 100 years) data set separately to allow direct comparison of declines, even though other regression types would have delivered higher R²-values. Two separate analyses were conducted: linear regression on all pooled data points of each group, and linear regression on the individual regression lines of each athlete, combined with t-tests. We chose linear regression, as the optimal type of regression function differed between disciplines, and to be able to compare decline slopes between the centenarians and the non-centenarian old athletes.

We collected 138 results from 42 athletes (5 women and 37 men) of whom results at the age of 100 years or more up to the age of 105 were present in the data base (Table 1). The most data in this age group were available from discus throw and shot put, followed by javelin throw, 100-m sprint and long jump. Table 1 also shows the numbers of athletes and results in the disciplines in the other longitudinal data set of athletes between 80 and 96 years of age with 1,134 results from 374 athletes.

Figures 1A–E and Table 2 show that in all five disciplines the average decline rate was higher in the older compared to the younger athletes. The steepest increase in performance decline was found in the 100-m sprint (t-test, p < 0.05), where the decline rate was 8x as steep in the older compared to the younger group (Table 2). The least steep increase in performance decline was found in the javelin throw. The increases in decline were not significant in long jump and in the throws (t-test, Table 2). On average, the decline rate of centenarians was 2.53x as steep as compared to that between 80 and 96 years. Figure 1F shows a direct comparison of the disciplines.

Only 13.5% (5 out of 42 athletes) among the centenarians were women, which is in line with the 13.9% (52 out of 374 athletes) in the Swedish data set. Three women had results in the 100-m sprint, one in the shot put and one in the javelin throw. Due to the low number of women in the data-set, an analysis of sex differences in decline rates was not possible.

It has been reported that the rate of performance decline was more than three times as fast in athletes older than 80 years compared to 30- to 69-year-old athletes (Ganse et al., 2020a,b). Accelerated decline rates after the age of 70 have also been shown in other disciplines and sports (Young and Starkes, 2005; Berthelot et al., 2012; Rubin et al., 2013; Dahl et al., 2020). The “plateau” in mortality rate in the oldest old (Barbi et al., 2018) may suggest a stabilization of the age-related rate of decrease in physical decline. Our data, however, do not show evidence for such a plateau in physical performance, but rather that the rate of decline is accelerated even further in 90+-year-old athletes beyond that seen in 80+-year-old athletes. Ongoing increases in decline rates were also previously reported from master world records (Baker and Tang, 2010; Akkari et al., 2015). At the same time, our data suggest that there is variation in the decline rate between individuals, that was also shown in the Swedish data set (Hoog Antink et al., 2021). Understanding the underlying factors for these differences could help to develop and improve interventions to slow down performance decline rates.

Differences in decline rates between track and field disciplines may be explained by specific skills required in each discipline, such as power, speed, agility, coordination or endurance, that may show different rates of age-related decline. For instance, the fastest performance decline was observed in the 100-m sprint that requires mainly speed and power (Dahl et al., 2019), while the javelin throw showed the least steep decline rate and depends besides power, on agility and coordination (Ganse and Degens, 2018). This needs to be studied in more detail, however, as previously we have observed that there was no significant difference in the rate of decline in aerobic and anaerobic power in master athletes (Bagley et al., 2019). It is clear that different physiological systems contribute to the success in different disciplines, where for instance power events are particularly dependent on muscle power, and endurance events are most limited by the cardiovascular system. It is therefore of interest to assess to what extent performance declines in different disciplines are attributable to proportional decrements in the muscle, respiratory (Degens et al., 2013), or cardiovascular function (Ganse and Degens, 2021).

Data from the oldest-old athletes were only available in 1 of the 10 individual running events, 1 of the 4 jumping events, but 3 of the 5 throwing events contested at the International Masters Games¹. Interestingly, the event selection of athletes changes in the oldest old compared to earlier ages. Perhaps these disciplines were chosen, as the oldest-old athletes are physically unable to compete in other events. Such a limitation may already occur before the age of 100 years, as in another data set none of the 80–94-year-old athletes participated in hurdles or in other jumps (Ganse et al., 2020a). Possible factors that may contribute to the absence of oldest-old competitors in theses disciplines may be lack of sufficient muscle power and/or fear of injury.

Despite the higher life expectancy of women², only 13.5% of the centenarian athletes in the present study were women. In comparison, 22% of the athletes in the largest longitudinal master athletics data set were women (Ganse et al., 2020b). Among the athletes with results at age 80 and older from the same data set, however, only 13.9% were women. The low proportion of oldest-old women may be related to a lower participation of women in sports in the past, where changes in society over the decades have stimulated a growing interest in sports and fitness in women, explaining a larger proportion of competing women at younger ages (O'Brien and Robertson, 2010).

The main strength of the present study is that it is the first to present longitudinal data on changes in performance after the age of 95 of athletes still competing at the age of 100 or over. The main weakness is that a low amount of data from women led to the inability to test for sex differences. This is particularly interesting, as differences in decline rates between men and women have been shown in previous studies (Ganse et al., 2018; Gava et al., 2020). Larger data sets may show that the increases in decline between the younger and the older group are actually significant.

We presented the first data set of performance decline rates after the age of 90 of athletes still competing at the age of 100 or over in five track and field disciplines. In this age group, the age-related performance decline was even faster than that previously reported in 80+-year-old athletes.