11. The centenarians
To live one hundred years has no particular significance to the organism as it is simply a fortuitous product of the speed of the earth on its orbit around the sun, and the way man began to use his fingers to count things. Yet, already in remote times, he became intrigued with the possibility of living ten times ten cycles of the sun, undoubtedly because it was perceived as being at or close to the maximum possible. The length of the solar year was for both agricultural and religious purposes painstakingly measured with great accuracy by civilizations which did not learn it from each other. Even today, to live to the age of 100 years carries with it a distinction and is the source of such prestige and pride that where strict records have not been kept, most claims to such age are false.
The recent increase in the number of centenarians has given rise to a debate on whether true centenarians are a new phenomenon or whether some of the historical claims to such age can be credited. VAUPEL and JEUNE (1994) argued that centenarians must have been very rare in most countries before the modern era, and demonstrated that the growth in their number is mostly due to improved chances of surviving from age 80 to 100. This finding is supported by our data in Chapter 17 and since the substantial improvement of these chances is a very recent development, the probability of reaching age 100 must have been very much smaller in the recent past. Basing on the view that the tail end of the survival distribution has been moving to the right, JEUNE (1994a) made the rather provocative assertion that no person lived to age 100 before 1800, nor to 110 before 1950. In sharp contrast to this, is the calculation of WILMOTH that the first centenarians appeared when the world population reached 100 million and life expectancy at age 50 was 14 years; according to him, the two conditions were met around 2500 BC (Wilmoth, 1995). Most of the sources he cites, however, would indicate 12 years as a more likely value for e(50) during the early agricultural era of mankind but if, as most historical demographers are inclined to believe, there were prolonged periods of higher and lower mortality, 14 years may have been characteristic of some better period.
How rare then is a centenarian today? The answer is: still quite rare but less and less so. In low-mortality countries, there is today one centenarian per about 20,000 people while thirty years ago, there was only one per 200,000. In the rest of the world, despite claims here and there to the contrary, a true centenarian is much more difficult to find. This author estimated that there were in 1985 about 30,000 centenarians in the world (Kannisto, 1990).
The number and proportion of centenarians in the countries of the Odense database are given in Table 19. Female centenarians are much more numerous than male (Table 20) outnumbering them at age 100 by 4:1 and at still higher ages even more.
A centenarian life table is presented as Table 21 for an aggregate of fourteen countries by adding to the usual thirteen Belgium for which the requisite data are not available for earlier decades. The life table is relatively accurate for men up to age 107, for women to 109 - ages where the numbers of observations are not less than 100. Beyond these ages, all rates become so unsteady as to give nothing more than rough indications at best. It can be expected that the above-mentioned limits (107 and 109) will be pushed ahead by one or two years in each decade as long as the mortality decline continues at present speed.
Table 22 presents a Nordic centenarian life table in which the five countries are separated from the main group of 14 on the grounds that the accuracy of age information is in them more rigorous because the age of every deceased is routinely verified against a permanent population register where the date of birth is based on a birth certificate. The drawback is that in this group, the numbers are much smaller, about one-tenth of the larger group.
When the qx values of both life tables are shown together in Figure 22, the female rates agree very well up to age 104 and the difference after that is not very important. The shape of the male curve for the larger group may, however, raise doubts about its accuracy because of its low initial slope which contrasts with the curve for Nordic males and for both female groups. The existence of some weakness in the basic data can therefore not be excluded.
Survival after age 100 is expressed adequately by life expectancy at 100 because significantly different mortality patterns at these ages have so far not been recorded. Even this summary measure is often volatile because of small numbers or due to short term fluctuations. In Table 23, the indicator is given for countries where it is considered meaningful. For the aggregate of thirteen best-documented countries we note in the post-war period a steady increase, slow for men, moderate for women and an acceleration in the last decade. The differences between countries are small and not very consistent from one period to the next. One could perhaps very tentatively single out Japan and Norway as countries with possibly above-average survival of centenarians in case future years confirm the tendency.
Regarding pre-war development, our data indicate since 1930 slow improvement for bith sexes in Norway and Sweden as well as among English women. For English men the data show a fall in life expectancy which, however, may be an artefact by improving accuracy.
The data which VINCENT (1951) and DEPOID (1973) assembled from France, the Netherlands, Sweden and Switzerland also indicated improvement in life expectancy at age 100:
Male Female Vincent, pre-1948 1.30 1.42 Depoid, pre-war 1.26 1.47 post-war 1.37 1.66 Our own data based on fully extinct cohorts in thirteen countries confirm this trend but, lagging behind period data, do not yet show the recent acceleration.
Male Female Cohorts 1850-59 1.60 1.76 1860-69 1.64 1.84 1870-79 1.67 1.92 Figure 23 presents the survival curve after age 100 according to two life tables, 20 years apart. The difference in each pair of curves seems narrow but corresponds at highest ages to a ratio of nearly 1:2. In complete life tables, the two survival curves would reach age 100 at different levels but their form after that point would be equal to the one shown in the graph. In this actual case, the proportion of live-born surviving to age 100 might have grown by a factor of 4 and close to age 110 the growth factor might therefore be 7 or 8. The proportional increase in the number of survivors to extreme ages is therefore quite remarkable but the increase in the highest age reached would probably be modest.
If, as is sometimes assumed, mortality reaches a plateau at some advanced age such as 110 years, the survival curve in log. scale would from that point on not bend down any more but fall in straight line. A look at Figure 23 shows that the effect of the plateau on human survival would not be important in practice.
There are by now sufficient observations to trace the probability of a centenarian to survive ten more years and to become what James VAUPEL has termed a supercentenarian. Combining the data for thirteen countries we obtain the probabilties given Table 24.
Though unsteady, particularly on the male side, they bear witness to improvement of the chances. The female survival ratios, based on much larger numbers, show that the relative improvement has been considerably greater in the 105-110 interval. The chances of surviving this interval are now for women one-third of the chances of making it from 100 to 105, and two out of every 1000 female centenarians live to be 110.
Updated by V. Castanova, 1 November 1999