556
CAN C E R
duo to accident or to extended practice. Impressions derived from such data have often proved to be fallacious, and no weight should be attached to them in themselves. The Registrar - General’s returns stand on a different footing, and merit very careful attention. On the face of them, the statistics show an enormous increase, and similar records in other countries tell the same tale. “ In England and Wales,” says Dr Roswell Park, Professor of Surgery in the University of Buffalo, “the cancer deathrate has risen from 1 out of 5646 of population in 1840, to 1 out of every 1306 of population in 1896; that is, in fifty years the proportion has increased nearly five times. All over this country the increase is quite as alarming.” In England and Wales the mean annual death-rates per million living for the ten years 1861-70 were—males, 242 ; females, 519 ; in 1891-97 they were— males, 571 ; females, 882, showing an increase of 136 per cent, and 70 per cent, respectively; and in 1899 they had further risen to—males, 672; females, 977. Again, the registered mortality for both sexes in the five years 1891-95 was 712 per million; in 1896 it had risen to 761 per million, and in 1899 to 829. At first sight this great and progressive increase looks conclusive, but there are sound reasons for doubting if it represents a real increase in the prevalence of cancer. In any case the Registrar-General’s figures must be largely discounted. In the first place, it has been pointed out that since cancer is a “ disease of degeneracy, the frequency of which increases as age advances ” (Paget), a rise of mortality from this cause would naturally follow on the greater longevity which has been attained in recent years, and the survival of more persons to cancer age. In particular, Dr J. F. Payne has pointed to “the greatly diminished mortality from phthisis and from infectious disease. Since both these causes of death act much earlier in life than cancer, in proportion as they carry off a small number of persons at earlier ages, a larger number will survive so as to be the victims of cancer at a later period of life.” That, no doubt, is so, but Dr Newsholme has shown, by a careful comparison of the figures, that survival can only account for a small proportion of the cancer increase. The same statistician points out a far more important source of error, which applies also to other causes of death that show an apparent increase. The Registrar-General’s returns are based on the death certificates signed by medical men, and only represent the certified, not the actual, causes of death. They are imperfect, inexact, and liable to a considerable amount of error. In the past the imperfection and margin of error was much larger than in later years. With improved knowledge and greater care the returns tend to become more accurate, and the large number of deaths certified from indefinite causes constantly diminishes. For instance, the number of deaths so certified in 1866-68 was 143,472, and in 1894-96 only 68,650. This diminution represents an addition of 74,822 deaths ascribed to various causes, and it is quite certain that a large proportion of such additional certifications go to the credit of cancer. In 1889, as the result of special inquiries made by the Registrar-General, 421 deaths were transferred to cancer, and in 1896 the transfers rose to 597, and in 1899 to 760. This factor makes the comparison between earlier and later periods quite untrustworthy. The influence of improved diagnosis in swelling the cancer returns is clearly shown by the fact that the apparent increase is far greater among males than among females, for which no reason can be assigned except that cancer in men is more often internal and difficult of diagnosis, and was therefore less frequently diagnosed in former years. This view is further confirmed by the carefully classified records kept for twenty-seven years at Frankfort-
on-Main. During that period cancer of “ accessiblen parts remained stationary, that of “ inaccessible ” parts showed a large increase. The difference can only be explained by improved diagnosis. Whether this is sufficient to account for the whole of the apparent increase, it is impossible to say at present. As the returns improve in accuracy the factor will drop out, and if it is responsible for the whole of the increase, that will tend to cease pari passu. This does in fact seem to be happening, at least in Great Britain, for the rate of increase is less year by year. In other countries the mortality returns are certainly neither more complete nor more accurate, and it is highly probable that the same calculation appliesequally to them. On the whole, it must be held that though there has probably been some increase, it has not been satisfactorily proved and is certainly less than is supposed. The rise of the parasitic theory of infectious disease has given a great impetus to the study of cancer from the point of view of . causation. It has „ ^ . been . shown . Causation. that there are m cancers certain microscopic bodies peculiar to cancer, and possibly of a parasitic nature. In some cancers they are present in enormous numbers. They have been isolated and cultivated outside the body. The cultures, when introduced into certain animals, have caused death with the production of tumours resembling cancer, from which again cultures have been made producing similar growths on inoculation. The existence of these bodies may be taken to be proved, but their nature and significance remain still undetermined. It may be that their presence is the exciting cause of that abnormal growth of the natural cells which constitutes cancer, or, on the other hand, it may be merely a stage in the process. In any case their origin is quite unknown, and if cancer be infectious at all, both experience in nature and experiment go to show that its infectivity is of a very low degree. It should not be forgotten that cancer is essentially a disease of old age, and that the incidence of all known diseases of a micro-parasitic nature affects rather childhood and early adult life. With regard to the influence of heredity, the trend of modern research is to minimize or deny its importance in cancer, as in phthisis, and to explain family histories by locality or other conditions. In view of the wide hold obtained nowadays by popular fallacies through the agency of elementary education and a cheap press, it may be not superfluous to add that there are no reasons for supposing that tomatoes have anything whatever to do with cancer. Several years ago Mr Haviland investigated the distribution of cancer in England and Wales, and disclosed the existence of what he called “cancer fields,” that is, districts in which the disease prevailed in a marked degree. Such districts were invariably associated with “ seasonally flooded areas traversed by, or in close propinquity to, fully formed rivers,” and were geologically characterized by alluvial and clayey soils. On the other hand, districts having the lowest mortality were found to be situated on elevated lands, well drained, not subject to floods and geologically characterized by the Carboniferous limestone formation. He has more recently returned to the question, and by comparing a later period with the former one, has shown that this distribution remains constant, though the mortality has everywhere increased. Incidentally it may be remarked that the fact of the increase being generally distributed, and not confined to cancer districts, is an additional reason for regarding the increase as only apparent. Mr Haviland concludes that flooded clays are always associated with the highest mortality, and limestones with the lowest. The Thames I valley, which is the greatest cancer field in England,
