Mortality in Relation to Weight

BRANDRETH SYMONDS, A. M., M.D February 1 1909

Mortality in Relation to Weight

BRANDRETH SYMONDS, A. M., M.D February 1 1909

UNTIL life insurance came into existence, the proper relation of height to weight and the effect of this relation upon longevity had no commercial significance. The aesthetic standards of physique ranged from the waddling obesities, who are admired by Arabs and adored by Hottentots, to the Greek ideals, as shown in the Apollo Belvedere and the Venus of Melos. Each people adopted a different standard of physical perfection, but no one knew whether that standard made for a long life. With the advent of life insurance this matter became important, but even then its real significance was not recognized. In fact, only of late years have we become fully alive to the fact that the physique of an individual is a fundamental element in his selection as a life insurance risk. If a proper relation of weight to height and age is not secured when selecting a given group of risks, the mortality in that group will be high in spite of the utmost care in excluding all other unfavorable elements.

In 1897 Dr. George R. Shepherd compiled for the Association of Life Insurance Medical Directors a table of height and weight for each quinquennium from 15 to 69. This was based upon the heights and weights of 74,162 accepted male applicants for life insurance in the United States and Canada. The weight included the clothing and the height of the shoes. In other words, the conditions were the same as those under which the applicants presented themselves to the medical examiner. At the extremes of age and of height, the number of individuals in any one class was small and the curve of weight showed abrupt changes which had to be equalized. The net results was a table of heights and weights varying according to the age which was adopted by the leading insurance companies as being the standard.

In 1900 a table of heights and weights varying according to age was compiled by a committee of the medical section of the National Fraternal Congress. This was based upon an analysis of 133,940 applicants of selected risks from the United States and Canada. The number of weights given in it amounted to 112, and 111 of these are identical with the similar numbers in Dr. Shepherd’s table. This coincidence is so thorough as to border on the marvelous, and one felt some scepticism as to whether this table was constructed quite independently of Dr. Shepherd’s. I therefore took the liberty of writing to the Chairman of the Committee who had charge of the compilation, and was assured by him in reply that the table was constructed de novo. This wonderful corroboration of Dr. Shepherd’s table shows that it is undoubtedly an exact standard for the United States and Canada.

We must remember that these heights and weights were taken when the parties were shod and clad in ordinary clothing. The shoe of a man will ordinarily raise him about 1 or 1 1-4 inches. According to Quetelet, we should allow in the case of a man one eighteenth of his total weight for clothing. If the weight of the man is 170 pounds, the clothing, therefore, should weigh 9.5 pounds. Actually, the weight of his clothing, including shoes, but excluding any form of overcoat, varies considerably according to the season. In the neighborhood of New York a man of 170 pounds will wear during the summer 6 or 7 pounds of clothing, while in winter its weight may rise as high as 12 or 14 pounds. This difference is undoubtedly a factor in the increase in weight during the winter which so many people believe in. In examinations for life insurance we do not attach any great importance to the differences due to clothing, for they are not large enough to modify materially our results.

It will pay to glance over this table of Dr. Shepherd’s for a moment. You will note that the weight rises steadily as you go down each vertical column. Dr. Oscar H. Rogers has formulated the rule from a study of this table that each added inch in height calls for an addition of 3 per cent. in the weight. This rule will apply, if liberally interpreted, to all but small men. The weight increases steadily with age in each horizontal line up to the year 45 among the small men, the 50 among the middle-sized men, and the year 55 and even 60 among the tall men. One is almost tempted to say that the taller the man, the longer it takes him to reach full maturity as shown by his weight. In the very tall this rule does not seem to apply, but the number of these was so small that a slight error may have crept in.

When I read my paper on this subject at the one hundred and forty-second meeting of the Medical Society of New Jersey, we had no standard table of heights and weights for women. We assumed in a rough way that they were about six to nine pounds lighter than men at the age of 25, and that this difference gradually diminished until it practically disappeared after the age or 45 or 50. At my instance, Dr. Faneuil S. Weisse prepared a standard table of height and weight for women (see page 99), which he presented at the Nineteenth Annual Meeting of the Association of Life Insurance Medical Directors.

The women on the measurements of whom this table is based were all healthy residents of the United States and Canada, who had been accepted for life insurance since 1895. As in the case of men, they were shod and supposed to be dressed in ordinary clothing. According to Quetelet we should allow one twenty-fourth of the total weight for the clothing of a woman. From my limited observations I am inclined to think that this is about correct, though naturally it should vary with the season. There is no doubt that a woman’s clothing as a rule is lighter than a man’s. The shoes of the average woman will raise her about 1 1-2 to 1 3-4 inches.

It is a great advantage to have a standard for women definitely settled. Fortunately the rough empirical method by which he had recently been working is so nearly accurate that this new standard will not materially modify our conclusions.

It will pay to study this table, especially in conjunction with Dr. Shepherd’s table for men. Women attain their maximum weight more uniformly than men, as practically all of them reach this point at from 50 to 59. After the age of 25, women increase in weight more rapidly than men. Thus a man of 5 ft. 6 in. gains 11 pounds between the ages of 25 and 50, while a woman of the same height gains 18 pounds during that time. This is well shown in the curves prepared by Dr. Weisse, which appear on next page.

The horizontal lines represent the weight, starting at 125 pounds and allowing 5 pounds to each line. The vertical lines represent the age at the mid-point of the quinquenniums used in the tables. Thus the quinquennium 25-29 is put down at 27.5.

The upper curve is that of men, the lower of women. The distance between them is due to the fact that the average woman is about 3 inches shorter than the average man. Note now that the average man at the age of 27.5 weighs 148.1 and at 57.5, when he has reached his maximum, he weighs 159.7, a gain of 11.6 pounds. The average woman at the age of 27.5 weighs 129.1 and at the age of 57.5 she weighs 145.7, a gain of 16.6 pounds, which is just 5 pounds more than that of the average man during that period. The result of this is that at heights 5 ft. 3, 4, 5, and 6 in., women will weigh the same as men when they reach the age of 50 or thereabouts. At other heights, women get within a pound or two of men, but no closer. These curves are constructed from the unadjusted weights and thus represent the actual facts as determined from the original figures. In the standard tables prepared by Dr. Shepherd and Dr. Weisse, these unadjusted figures have been slightly modified, either up or down, at some points in order to secure a perfectly uniform curve at all ages. Dr. Weisse says: “In preparing the adjusted table, it is interesting to note that the weights of over 80 per cent. of all these women needed practically no adjustment. The average weight of all the 58.855 women, after all adjusting, involved a difference of less than one-tenth, .056 of a pound.

We are now very comfortably fixed, for we have accurate standards of weight, according to height and age, for both women and men, at least for the United States and Canada. It is a curious fact that the lowest death rate does not coincide with the standard. In general terms it may be said that the lowest death rate is found in the class who are about 5 per cent. below the standard, but in ages below 30 the lowest rate is found among those who are 5 to 10 per cent. above standard. These differences are not great, and I wish to discuss in fuller detail the more marked cases of overweight and underweight. Before doing so, I will explain as briefly as possible the statistical method by which we determine whether a given class of insurance risks is furnishing a satisfactory mortality.

The duration of an individual life is most uncertain, but the average duration of 100,000 lives is very certain. Many mortality tables have been constructed showing the probabilities of dying at each age of life. The one most used for these statistical investigations is known as the Modified Healthy English, and is the one adopted by the Actuarial Society of America in their Specialized Mortality Investigation. According to this table, the probability of dying at the age of 30 is .00821, or, to put it in another way, out of 100,000 living persons 30 years old, 821 will die during that year. This only holds true, however, if these cases are not influenced by medical selection, for we find that the influence of this extends for at least five years and probably longer. We must, therefore, make an allowance for medical selection.

According to the Actuarial Society, the probability of dying at the age of 30, during the first year after medical selection, is only one-half. To put it in another way, only 410 would die instead of 821. In the second year of insurance, the percentage is 68 at this age. As our group of cases are now in the second year of insurance, they are 31 years old, and the mortality figure is .00828, or 828 out of 100,000. We therefore have to take 68 per cent. of 828, and our expected deaths will be 564 out of 100,000 living, instead of 828. I think that this will give you some idea of the method employed. It is tedious and the details are very intricate, but the principle, as you will see, is comparatively simple. By this means we calculate the deaths that are expected to occur in a given group of individuals. If this group shows 100 actual deaths and 200 expected deaths, we say that the mortality is 50 per cent. If the group shows 200 actual deaths and 100 expected, we say that the mortality is 200 per cent. That is the technical meaning of the term “mortality” as employed in this and similar statistical researches. In a rough way we consider that a mortality between 90 per cent. and 100 per cent. is fair; if between 80 per cent. and 90 per cent., it is good, if below 80 per cent., it is very good.

Let us first take up overweights. A case is not considered overweight unless it is more than 20 per cent. above the standard weight for the height and age. For example, at the age of 40 the standard weight of a man 5 ft. 6 in. tall is 150 pounds. We would not regard him as an overweight until he had passed 180 pounds, which is 20 per cent. above his standard. Even in the classes of smaller excess than this the mortality increases, but at this point it begins to be a serious matter. For the sake of convenience, we will call those overweights who are between 20 per cent. and 30 per cent. above the standard “moderate overweights.” Similarly, we will call those overweights who are more than 30 per cent. above the standard “excessive overweights.” Thus, men 40 years old who are 5 ft. 6 in. tall and weigh between 180 and 195 pounds would be termed moderate overweights, but if they exceed 195 pounds, they would be called excessive overweights.

The effect of overweight is influenced by two fundamental factors. These are (1) percentage of overweight; and (2) age of the individual. The following table shows this very closely. The first column contains the age periods, the second column the mortality of the moderate overweights, the third column the mortality of the excessive overweights.your eye follows down each of these columns, you will note that the mortality rises rapidly both among the moderate overweights and the excessive overweights. It is true that the old in both classes have a little better mortality than the elderly, but I think that the old overweights were selected with a little more care than the e'derly, and for that reason their mortality is a shade better.

As you compare the moderate overweights with the excessive overweights, you will note that the latter have a higher mortality at every age. The young overweights have a good mortality in both groups. The mature moderate overweights are bad, while the mature excessive overweights are very bad. Still worse than they come the elderly moderate overweights, and at the unenviable apex stand the elderly excessive overweights, with a mortality of 162 per cent.

It may be said, then, that an overweight in a person below 29 is not harmful even up to 30 per cent. or 35 per cent. above the standard, provided the person does not get actually heavier with advancing years. You will note that this remark refers to actual weight and not relative weight. Our standard increases with advancing age, so that an excess of 33 per cent. at age 22 is almost exactly equal to an excess of 20 per cent. at age 45. If a boy 22 years old and 5 ft. 0 in. tall weight 100 pounds, his weight is 33 per cent. above the standard of 150 pounds at that height and age. When that boy gets to be 45 years old, if his height and weight still remain the same, we find that his weight is only 20 per cent. in excess, for the standard at age 45 is 166 pounds. He has kept just at the edge of the danger zone, and people of his class will give an almost uniform mortality, slightly in excess of the normal, irrespective of age.

Beyond 30 years of age the mortality among overweights rises rapidly with the age and with the weight. This will happen even when the utmost care is used in examining and selecting these risks. A long-lived family history, one in which neither parent has died below the age of 70, will improve the mortality by 10 or 15 points. Such a gain as this would make the young overweights of both classes very good, and it would make the mature moderate overweights show a fair mortality, about 90 per cent. In all other classes, however, the mortality would still remain very bad.

If the family history is an average of a short-lived one, the mortality will be increased by 5 to 15 points. Under these circumstances we find that the young moderate overweights still retain a good mortality, but the young excessive overweights get up to about 100 per cent. The other classes, of course, are rendered just so much worse.

A tuberculous family history seems to have about the same effect as a short-lived family history. In the younger ages it certainly has no worse effect, for, as might be supposed, the overweight tends to counterbalance the tuberculous predisposition.

Increasing abdominal girth is a very serious matter for overweights. When this exceeds the expanded chest, the mortality is markedly increased by 15 or 20 points at least, and much more if the abdomen greatly exceeds the expanded chest. You will see that by combining these factors, an elderly excessive overweight with a large abdomen and short-lived family history may easily be exposed to a mortality of 200 per cent.

In fact, we can say that any other blemish, whether in the personal history or the physical condition, regularly increases the mortality of these overweights. Conversely, overweight adds greatly to the gravity of any other defect. For example, I have lately analyzed a class of cases which gave a history of renal colic or renal calculus, and which showed a mortality of 99 per cent. In this class was a small group of overweights. Now, those cases which combined a history of renal colic or renal calculus with the overweight showed a mortality of 150 per cent.

As regards foreigners, it may be urged that our standard table is based entirely upon selected lives in the United States and Canada, and that it would not apply to another race, like the Germans, who are usually stouter and heavily built. We have found, however, that overweight foreigners are, if anything, a little worse risks than overweight natives. In truth, human fat seems to be the same wherever found, and has the same effects upon the prospects of life, whether in England, Germany, Holland, Belgium, France, Italy, Mexico, or the United States.

We find that overweight women, measured by their own standard, show practically the same mortality as overweight men. They are fewer in number, for women do not often permit themselves to become fat.

Now, let us consider the effect of underweight. As long as the weight is not below 80 per cent. of the standard, that is, not more than 20 per cent. below the standard, the effect seems to be slight. The mortality rises slowly, but the increase is gradual and not alarming. Below this level, however, the mortality rises to a point where it is of consequence, especially in the younger ages. Among the young underweights, we can expect a mortality of 110 per cent. when the weight is from 80 to 75 per cent. of the standard. The older ages in this group show a uniform mortality of about 95 per cent. We should call this a fair mortality, not very bad, but on the other hand not very good.

In the next group, in which the weight ranges from 75 per cent. to 70 per cent. of the standard, that is, from 25 to 30 per cent. below the standard, the number of entrants in my company below age 20 numbered only 30 in the 30 years from 1870 to 1899. This number is too small to furnish any figures of consequence. Even in the decade 20-29 the total number of exposures only amounted to 299, an average of less than 30 for each year. These gave 2 deaths as against 1.9 expected. The numbers are too small to be of consequence. In the decade 30-39 the number exposed rises to 1,391, and these give a mortality of 100 per cent. After this the mortality is fairly satisfactory, ranging from 90 to 95 per cent.

For weights below 70 per cent. of the standard, that is, more than 30 per cent. below the standard, our experience is very limited and too small to divide into different age periods. The actual deaths amounted to 12 and the expected to 13.5, showing that our selection was reasonably good. The number of entrants below age 40 was too small to give any information. Above age 40 we can only say that when they are picked with care, these extreme underweights live a good while.

As regards the other factors which modify the influence of underweight, we have to deal with a problem quite different from that of overweight. The influence of age is reversed among underweights. The younger ages are the ones most affected, while the older ages are but slightly disturbed.

The mortality increases as the weight diminishes, but even among those who are more than 30 percent, below the standard, the mortality is not excessive.

The association of dyspepsia with underweight is a serious matter with the young, and will give us a mortality as high as 150 per cent. I have no doubt that the combination of dyspepsia and underweight in the young is often indicative of incipient tuberculosis, the extent of which is so small that it is not determined on physical examination.

The association of underweight and tuberculous family history has long been recognized as serious, especially in younger ages. Thus, we find this combination gives a mortality of 180 per cent. in the ages below 35. Above that age the influence of tuberculosis depends upon the number of cases in the family. If we have two or more cases occurring in the family of an underweight, the mortality is 107 per cent, for all ages above 35. In these older ages the underweight who has had only one case of consumption in his family runs little risk, perhaps for the reason that he takes better care of himself.

As regards women, we find that the lesser grades of underweight from 80 to 75 per cent. of the standard give a mortality of only 77 per cent., an excellent result and practically uniform for all ages. For weights below this, the mortality becomes bad, in fact over 100 per cent., but the cases are so few that no deductions can be made from them. There is no reason to suppose, however, that underweight women are any worse than underweight men, and I have no doubt that they will give as good a mortality if selected under the same circumstances.

In conclusion, I think that I can do no better than to quote from the Medical Record the summary that ended my previous paper on this subject before the Medical Society of New Jersey:

“Now, let us sum up in a general way the differences between overweights and underweights. The mortality among all those, irrespective of age, who are between 20 and 30 per cent. below the standard weight, is 96 per cent., while the mortality of all, irrespective of age, who are between 20 and 30 per cent. above the standard, is 113 per cent. These figures alone would show that overweight is a much more serious condition than underweight. On the other hand, we must take into account the fact that until recent times overweights were accepted more freely by insurance companies than underweights. To put it in another way, underweights were selected with more rigid care than overweights. The old idea that an overweight had a reserve fund to draw upon in case there was a run on his bodily bank was prevalent, although it was recognized that excessive fat might be harmful and should exclude the risk on the ground, perhaps, that it was a form of capital which was not active. Similarly, an underweight was considered to be under-capitalized, and if his bodily bank had to go through a panic like pneumonia, or hard times like organic heart disease, he would become insolvent and bankrupt.

“As a result of this method of thought, our underweight mortality is rather better and our overweight mortality rather worse than if both sets had been accepted under exactly the same conditions. But, even if we make full allowance for the difference in selection, I am convinced that the same percentage of overweight is a more serious matter than if it were underweight. The excessive weight, whether it be fat or muscle, is not a storehouse of reserve strength, but it is a burden that has to be nourished, if muscle, and that markedly interferes with nutrition and function, if fat. This does not apply to the young, those below 25 years of age. Here a moderate degree of overweight is much more favorable than underweight. In fact, up to age 25 an overweight not to exceed 110 per cent. of the standard is upon the whole good for the individual. It seems to indicate a certain hyper-nutrition and robustness of physique that is favorable to the subsequent life. Underweight among these young people, on the other hand, is unfavorable, and in some cases indicates commencing disease or the tendency thereto. But when we pass the age of 30 these conditions are reversed, and the difference between overweight and underweight in their influence upon vitality becomes more marked with each year of age.

“Of course, for the best interests of health, one should be as near standard weight as possible, and that is the sermon which you should preach to your patients. Impress upon them the advisability of their being within 10 per cent. of the standard, for within that range is found the lowest mortality and the greatest vitality.”