Chapter 9. Atherosclerosis and malignant tumours

Chapter 9. Atherosclerosis and malignant tumours N. H. STERNBY Aortic and coronary atherosclerosis were studied in subjects with a malignant disease and compared with those in the three atherosclerosis reference groups. In general, subjects with a malignant disease had little atherosclerosis, except for men with lung or prostatic cancer; in particular, men with lung cancer tended to have more extensive aortic atherosclerosis. Atherosclerosis was less extensive in subjects with tumours specific to females. The ratios between the different types of lesions was preserved and the change was therefore quantitative rather than qualitative. No proof of a real negative influence on atherosclerosis by malignant diseases wasfound but wasting could be one factor influencing the development ofatherosclerosis in tumour subjects. Such subjects, except those with lung andprostatic cancer, could have been included in the low atherosclerosis group which was an index of the mean basic level of atherosclerosis in the populations studied. It is well known that atherosclerosis is, in general, not extensive in subjects dying from malignant dis- ease. A negative relationship between malignancy and atherosclerosis has therefore frequently been reported. The question has been discussed for several decades, and the literature on it has been extensively reviewed by McGill (7) and Sternby (11). The results of many studies have, however, been conflicting regarding both malignant diseases in general and specific tumours in relation to atherosclerosis. Ac- cordingly, it was thought that the use of extensive population-related autopsy material, collected and evaluated in a standardized way, might help to clarify this question. MATERIAL AND METHODS During a 5-year period the aorta and the coronary arteries were collected from almost all subjects dying in 5 European towns, by means of the methods described in Chapter 1. Specimens were collected from all age groups above 10 years, but both the number of tumour cases and the autopsy frequency were highest in ages between 40 and 79 years. Therefore the analyses to elucidate a possible relationship between atherosclerosis and malignant tumours concerned mostly subjects in this age range. The number of subjects aged 40-79 years with malignant tumours, by sex and town, is shown in Table 27, which also gives figures for some specific tumours. The tumour frequency figures are in re- spect of non-violent deaths only. The subjects indi- cated are those in whom the malignant disease was considered to be the principal disease. The figures relate to the number of subjects and not to the number of tumours, although according to some investigations multiple tumours occur in about 10% of all tumour cases. Also, the duration of the malig- nant disease and the extent of spread of the tumour found at autopsy were not taken into account. The tumour cases were compared for atherosclerosis with one or all of the three reference atherosclerosis groups. RESULTS Prevalence of atherosclerosis The prevalence of atherosclerotic lesions in the abdominal aorta and the left anterior descending coronary artery was compared in the tumour sub- jects and the low atherosclerosis group. For aortic atherosclerosis, subjects with any malignant tumour had a higher prevalence of complicated lesions (espe- cially in men) and calcified lesions, while there were no differences as regards fibrous plaques and raised lesions. Consequently, milder forms of athero- sclerosis were somewhat less frequent among tumour subjects than among those belonging to the low atherosclerosis group. In the coronary arteries the findings were similar for complicated lesions, while for other types of lesion the differences were rather small and in some age decades non-existent. The prevalence of stenosis in the left anterior descending BULL. WORLD HEALTH ORGAN., Vol. 53, 19763472 -555 N. H. STERNBY cooCD 0) oo 04 CO N 00 N 04 CD CD* L) 00 CC) CN co co 0 CC) CY) CD LO -f 0 N 0) CC) b 00 C1 o) 0) t 00 0 CD CD CD o CD N CC) Nt 0) CX) co 0 N N CY) N4 N4 N- N- N CN CD4 0) 0V) co 0) 0 N 0 v- aN 0 CN N CC) CC) 0Y) co CY) NO N CD 0 CC) 0 Co6 0) N CC) NO CD4 0) LO a) CD 0 CN CC)O CO NO CD 0) CD CD cD 0) CD N r0 NO N N CO N cD N- cD CC CC Nq 0) co 6 0 CC) N co CN CC) CC) CD4 N C- ) b CD C C - 0N N N0 NCD N CD N X CC CC) C 0 m N N 000)U) N CC)N N CD N CD CD CD N CD ID N N ID N cN LO II Lo o~~~~~c I I O It II co CD 0 >° C Cu -6 .>i 9 - r 0- > -I..D. -o c" 556 a 0 E 0 C: 0 CN LO) N1 N1 0) CC) N CD4 CY) 04 r LCD CD LC) 0 CD4 N- N- N- WI 6 CLZF 6 > z DC E Z 6:ze C Z co Cu Z CuZ co N 1= co X 6 X0 Z E 0: c o 00 3 z 0 -j a) oN LO) CDco CD CO CO It coo co Nt CD CY) LO CD4 co CN CD4 0) 0 CD- CD LO It) cn E a O -J C/) co cn co to I'* co 0) CD N LO 0 CD CC) 0 N Lo m CD cC,I) N CN N r-- 0 'D N 0 N 0 CC Nl CC) 00 N N N N 0- 0- LO 0 C x CD n 0 4U co0 cn 0 E c cu E .C cn CD 0) (D CUE -3 0) CU Cl) CD co -a a .2 ao cn 4- 0 C.) a) ,C ATHEROSCLEROSIS AND MALIGNANT TUMOURS coronary artery in men with a tumour was consider- ably higher than in the corresponding low athero- sclerosis group; in women no such difference was observed. The prevalence of atherosclerotic lesions was also studied for some specific tumour sites and the figures compared with those for the low atherosclerosis group and the group of all malignant tumours. In men with lung cancer, both complicated and calci- fied aortic lesions were more prevalent than in men with tumours in general and also more prevalent than in subjects of the low atherosclerosis group. Consequently, milder forms of aortic atherosclerosis were less common in subjects with lung cancer. Coronary complicated and calcified lesions, on the other hand, did not differ very much in prevalence among these three groups. But coronary stenosis was considerably more prevalent in lung cancer subjects. Subjects with prostatic cancer generally showed lower prevalence rates for complicated and calcified lesions than those with lung cancer and were mostly similar to those with any kind of malignant tumour or to the low atherosclerosis group. This pattern was also found for coronary stenosis. Subjects with tumours at other sites showed a somewhat different pattern. Men with stomach cancer generally had low prevalence rates of com- plicated and calcified aortic lesions and complicated coronary lesions in comparison with all subjects with malignant tumours, and a lower prevalence of compli- cated coronary lesions than the low atherosclerosis group. They also showed a similar prevalence of mild aortic and coronary atherosclerosis (i.e., fatty streak and fibrous plaque only) in both the aorta and the coronary arteries than the low atherosclerosis group. Women with stomach cancer did not differ significantly from the low atherosclerosis group in any respect. The prevalence of the different types of lesion was as low in subjects with ovarian carcinoma as in the low atherosclerosis group, and the pre- valence of coronary stenosis even lower. Extent of atherosclerosis The extent of different types of atherosclerotic lesions was studied both in age-grouped material (by decade) and in a pooled 40-79-year age group. Tumour subjects showed more extensive aortic le- sions of all types (except fatty streaks and calcified lesions in women) than did the low atherosclerosis group, the differences being mostly significant in the age-standardized average group and for some decades. The extent of coronary lesions showed a similar pattern, although the differences were signifi- cant only for men. In the case of men with lung cancer the differences from the low atherosclerosis group for both aortic and coronary lesions were much more evident and significant in most age groups. Women with lung cancer also had more extensive lesions than the low atherosclerosis group, although the differences were only occasion- ally significant. Subjects with stomach carcinoma showed no significant differences in the extent of aortic or coronary lesions in comparison with the low atherosclerosis group but they tended to have more extensive lesions. Subjects with tumours speci- fic to females did not generally differ from the low atherosclerosis group. Calcified aortic and coronary lesions were, however, significantly less extensive in the cancer than in the low atherosclerosis group. The extent of atherosclerotic lesions was also studied in the pooled subjects aged 40-79 years from all towns. Raised and calcified lesions in the average aorta and average coronary artery were used as indices and the figures were compared with those for the three atherosclerosis reference groups. The tumour subjects as a whole (both men and women) had significantly less extensive raised aortic and coronary lesions than the standardized average atherosclerosis group but significantly more exten- sive lesions than the low atherosclerosis group. The extent of calcified aortic lesions was low in women but not in men, while coronary calcified lesions were significantly lower than in the average athero- sclerosis group in both men and women. Men with lung cancer formed the only tumour group with significantly more extensive aortic raised lesions than the standardized average atherosclerosis group. Subjects with other tumours had similar or significantly less extensive raised aortic lesions than the corresponding average atherosclerosis groups. The extent of calcified aortic lesions was high in men with lung cancer compared with the average athero- sclerosis group, while other tumour subjects gener- ally showed low figures. In female tumour subjects calcified aortic lesions were never very extensive. In tumour subjects (both men and women) raised coronary lesions were less extensive than in the average atherosclerosis groups and these differences were mostly statistically significant. Raised coronary lesions, unlike raised aortic lesions, were not more extensive in men with lung cancer. Calcified coron- ary lesions were in general significantly less exten- sive than in the standardized atherosclerosis group, in both men and women, irrespective of tumour site. 557 N. H. STERNBY Table 28. Extent (% of surface) and prevalence (%) of some atherosclerotic lesions in subjects aged 40-79 with malignant tumours, as related to the thickness of their subcutaneous fat Average aorta Average coronary artery Raised lesions Calcified lesions Raised lesions Calcified lesions arteryastenosis (extent) (extent) (extent) (extent) (prevalence) Men Women Men Women Men Women Men Women Highest third a 46.8 38.0 2.6 3.2 45.1 28.3 4.2 1.2 32.2 11.5(0.95) (0.84) (0.86) (0.78) (0.98) (0.79) (0.99) (0.42) (0.90) (0.50) Lowest third a 46.1 38.5 3.3 3.3 38.1 26.0 3.3 1.7 23.4 10.5(0.96) (0.87) (1.16) (0.84) (0.85) (0.76) (0.85) (0.64) (0.68) (0.47) a Subjects with cancer as the principal disease were ranked in order of the thickness of their subcutaneous fat. The highest and lowest third were compared and the results are shown above. The figures in brackets denote the ratio to the value for the standardized average athero- sclerosis group. Since subjects with hypertension and diabetes were by definition excluded from the low athero- sclerosis group, it was thought reasonable to exclude such cases also from the cancer group for a further comparison, which showed that even though the differences between the cancer and the low athero- sclerosis groups became smaller, they still remained. Atherosclerosis was thus less extensive on the whole in tumour subjects, except that subjects with lung cancer tended to have more extensive aortic atherosclerosis, as did those with prostatic carcinoma, especially in the younger age groups. Subjects with the specific female tumours generally showed very little atherosclerosis. In view of these findings, a separate analysis was carried out of all tumour subjects excluding those with lung and prostatic carcinoma and the specific female tumours. The differences between subjects with any malignant tumour and the corresponding low atherosclerosis groups almost disappeared when lung and prostatic carcinoma subjects were removed from the tumour subjects and subjects with the specific female cancers showed almost the same figures as the low athero- sclerosis group. The prevalence of coronary stenosis in tumour subjects did not differ significantly from that in the low atherosclerosis group in both men and women. In men with lung cancer, this prevalence was inter- mediate between those of the low and average atherosclerosis groups. The prevalence in women with lung cancer was higher than in women in the low atherosclerosis group, but not significantly so. The prevalence of fresh myocardial infarction was very low in tumour subjects and did not vary greatly with sex or tumour site. Large myocardial scars were more prevalent in men than in women with tumours, but did not vary much with tumour site in either sex. Thus, lung cancer subjects did not show increased prevalence of myocardial lesions compared with other tumour subjects. To study the possible influence of wasting in subjects with cancer as the principal disease, such subjects were ranked in order of the thickness of their subcutaneous fat. Those belonging to the highest and lowest third, respectively, were then compared (Table 28). The analysis showed that raised aortic lesions in both men and women were of similar extent in these two groups, as were calcified aortic lesions, but there was a tendency for those in the lowest third to have more than those in the highest. Coronary raised lesions were more extensive in those belonging to the highest third, especially in men, while the extent of coronary calcified lesions showed no certain trend. Coronary stenosis was significantly more common in men of the highest than in men of the lowest third but in women no difference was observed between these two groups. The frequency of fresh myocardial infarction was too low in both groups to permit any conclusion. Large myocardial scar was significantly more preva- lent in the highest third in men but not in women. When comparisons were made in cancer subjects with and without hypertension, the same trends were apparent. Inter-town differences Raised aortic and coronary lesions differed among the towns in their mean extent but when they were compared with those in the standardized average groups the ratios were very similar. The same was 558 ATHEROSCLEROSIS AND MALIGNANT TUMOURS true of the extent of calcified coronary and aortic lesions and the prevalence of coronary stenosis and fresh and old myocardial infarction. These ratios generally varied more than those for the extent of atherosclerotic lesions. When differences between towns occurred they did so haphazardly and no certain trends emerged. Individual cancer sites gen- erally followed this pattern. In this respect, subjects with lung cancer were especially studied. In all towns except Tallin they showed more extensive raised aortic lesions than the corresponding standardized average atherosclerosis groups; the differences were significant for Malm6 and Prague. But raised coron- ary lesions were less extensive than in the standard- ized average group in all towns. DISCUSSION The overall frequency of malignant tumours in subjects aged 40-79 who had died from natural causes was 30% in men and 33% in women. There were, however, considerable inter-town differences. Malmo was highest for men and women (both 40 %) and Tallin was lowest (23 and 28%, respectively). For men, Prague occupied an intermediate position between Malmo and Yalta/Ryazan, while for wom- en Ryazan was second. The lung was the most common tumour site in men in Malmo, Prague (especially), and Yalta, while the stomach predomi- nated in Ryazan and Tallin. The prostate was very rarely reported to be the primary site in the towns of the USSR. In women stomach cancer was reported more often from the towns of the USSR (especially Ryazan) than from Malmo and Prague, while mam- mary carcinoma was much more common in these latter towns than in the USSR. The observed differences do indicate that the pattern of malignant disease varies among towns and, generally, between Malmo and Prague on one hand and the towns of the USSR on the other. However, the autopsy rate for ages above 70 was rather low in the towns of the USSR compared with the other two towns. Subjects in the older age groups suffering from multiple diseases including a clinically unknown cancer may thus after autopsy have been assigned to the cancer death group in Malmo and Prague. This may explain some of, but probably not all, the differences. The general finding was that subjects with malig- nant tumours had less aortic and coronary athero- sclerosis than the standardized average atherosclero- sis group but the same amount or more than the low atherosclerosis group. They also had a low preva- lence of coronary stenosis, lower than that for the standardized average group but higher than that for the low atherosclerosis group. Myocardial lesions were also much less frequent among tumour subjects than among the reference groups. These findings were also true for tumours at most sites, exceptions being lung cancer and, to some extent, prostatic cancer. Men with a lung tumour thus had signifi- cantly more thoracic and abdominal aortic athero- sclerosis than the standardized average group and almost as much as the high atherosclerosis group. Coronary atherosclerosis, however, was not more extensive than in tumour subjects in general, but coronary stenosis showed a slightly increased preva- lence in lung cancer subjects in comparison with other tumour subjects. Women with lung cancer followed the general pattern, i.e., did not show any increased aortic atherosclerosis or extensive coron- ary atherosclerosis or stenosis. When subjects with lung and prostatic cancer were removed from the tumour group, it was very similar to the low athero- sclerosis group. The literature on the relationship between malig- nant diseases and atherosclerosis is vast and has been reviewed elsewhere (7, 11). In general, all studies have shown that subjects with malignant tumours have less atherosclerosis than subjects dying from non-malignant diseases. The conclusions drawn from these findings have, however, been different. Some have stated that a true negative relationship exists (2, 3, 4, 5, 6, 10), while others have found no such relationship (1, 7, 8, 9, 11). The results of this study do not support the assumption that there is a true negative relationship between atherosclerosis and malignant diseases. The results appear to depend upon the material with which the cancer cases are compared. Except for women with the specific female cancers, cancer subjects almost never had less atherosclerosis than the low athero- sclerosis group and only rarely such a low level. In this respect the tumour subjects occupied a position intermediate between those of the low and the standardized average groups. A possible negative relation between malignant tumours and atherosclerosis might operate through one of two mechanisms: a " biological " effect on the development of atherosclerosis or a retarded devel- opment or regression related to the wasting often present in cancer subjects. The rate of development with age of atherosclerosis, as well as the proportion between different types of lesions, was similar in, for 559 N. H. STERNBY Table 29. Ratios between the extent of raised aortic and coronary lesions in tumour subjects and in the standardized average atherosclerosis group, in rank order Average aorta Average coronary Men Women Men Women Lung 1.07 Kidney 0.99 Prostate 0.96 Liver 0.97 Bladder 1.01 Liver 0.87 Liver 0.93 Bladder 0.96 Prostate 0.99 Breast 0.86 Lung Lung 0.86 All f0.91 Pancreas Colon 0.83 Liver 0.95 All 0.5 Brain 0.85 Kidney 0.89 Pancreas 0.81 Leukaemia 0.93 Bladder All 0.88 Brain 0.78 Kidney Uterus 0.82 0.92 Colon 0.87 Kidney Pancreas Pancreas 0.81 0.77 Stomach 0.84 Stomach Colon 0.90 Ovary 0.80 Bladder 0.83 All 0.76 Stomach 0.86 Leukaemia Breast 0.74 Brain 0.85 0.81 Brain ) Leukaemia 0.68 Uterus 0.66 Ovary 0.60 instance, subjects who had died from stomach carci- noma or myocardial infarction, the difference being merely quantitative, not qualitative. The analy- sis also showed that aortic atherosclerosis in sub- jects dying from cancer had no relation to the amount of subcutaneous fat but that coronary atherosclerosis and coronary stenosis in men varied with this factor. The influence of body build thus seemed to be more marked than a possible influence of cancer. Except for lung and possibly prostatic carcinoma, there was no indication that certain tumours behave differently from tumours in general. Men with lung cancer, however, showed significantly more aortic but not more coronary atherosclerosis than tumour subjects in general, and subjects with prostatic carci- noma seemed to have slightly more atherosclerosis than tumour subjects in general. In women with lung cancer such a difference was not apparent. The reason for this feature in connexion with lung cancer is not known but theories have been expressed regarding the possible influence of smoking on the development of both atherosclerosis and lung cancer. The finding that it is the extent of aortic but not of coronary atherosclerosis that is increased in lung cancer patients does not, however, support this assumption. It is unlikely that this observation is the result of methodological differences, since the fre- quency of coronary stenosis and myocardial lesions was not significantly higher in men with lung cancer than in other cancer subjects and since the myocar- dial lesions were estimated without knowledge of the grading results. A reason for increased aortic atherosclerosis in lung cancer subjects could be that the thoracic aorta became relatively more severely involved than the abdominal part because of improved blood or lymph flow, treatment, etc. There was an indication, but nothing more, that this could be the case. The ratio between the extent of raised lesions in tumour subjects and the corresponding standardized average atherosclerosis group was used in an attempt to rank the different tumour sites with regard to severity of involvement. It is evident from the re- sults, given in Table 29, that even though athero- sclerosis is not extensive in cancer subjects there are variations even within this group. These may be mainly the result of the well-known individual varia- tion with regard to atherosclerosis, but probably in part the result of environmental (lung cancer), hor- monal (breast, ovarian, and/or other uterine cancer), or other factors. 560 ATHEROSCLEROSIS AND MALIGNANT TUMOURS 561 RESUME CHAPITRE 9. ATHEROSCLE'ROSE ET TUMEURS MALIGNES L'ath6rosclerose aortique et coronarienne a ete etudiee chez des sujets atteints d'une affection maligne et com- paree a celle des personnes se rangeant dans les trois groupes de r6ference de I'ath6rosclerose. En general, les sujets atteints d'une affection maligne ont peu d'athero- sclerose, & 1'exception des hommes porteurs d'un cancer du poumon ou de la prostate; en particulier, le cancer du poumon s'accompagne volontiers d'une extension supe- rieure de I'ath6rosclerose aortique, alors que le contraire s'observe chez des sujets pr6sentant des tumeurs propres au sexe feminin. Les proportions entre les diff6rents types de l6sions sont conservees, et la modification est donc quantitative plutot que qualitative. Aucune preuve d'une influence negative reelle des tumeurs malignes sur l'atheroscl6rose n'a ete trouvee, mais la denutrition pourrait 8tre l'un des facteurs influengant le developpe- ment de F'atherosclerose chez les porteurs de tumeurs. Si l'on excepte les malades atteints de cancer du poumon ou de la prostate, de tels sujets pourraient 8tre ranges dans le groupe 'a faible degre d'atherosclerose qui est un indicateur du degre fondamental moyen d'athero- sclerose dans les populations etudiees. REFERENCES 1. CREED, D. L. ET AL. Am. J. med. Sci., 230: 385 (1955). 2. ELKELES, A. Br. J. Cancer, 10: 247 (1956). 3. ELKELES, A. Br. J. Cancer, 13: 403 (1959). 4. GIERTSEN, J. C. Acta pathol. microbiol. scand., 66: 341 (1966). 5. JuHL, S. Acta pathol. microbiol. scand., 37: 167 (1955). 6. JUHL, S. Acta pathol. microbiol. scand., 41: 99 (1957). 7. McGILL, H. C., Jr, ed. The geographic pathology of atherosclerosis. Baltimore, Williams & Wilkins, 1968. 8. MrrCHELL, J. R. A. & SCHWARTZ, C. J. Arterial disease. Oxford, Blackwell, 1965. 9. ROBERTSON, W. B. J. Atheroscler. Res., 2: 79 (1962). 10. SJOVALL, J. & WIHMAN, G. Acta pathol. microbiol. scand., Suppi. 20 (1934). 11. STERNBY, N. H. Acta pathol. microbiol. scand., SuppI. 194 (1968).