Document rBq2DDeOdnzb109394rR6mmVr

ISCIDENCE OF M C L T I P L E MYEL0.M.A - T U R E S S O S E T AL. 42 1 .: TOMPARISON OF TRENDS INTHE INCIDENCE OF MULTIPLE MYELOMA IN M A L M ~ , SWEDEN, AND OTHER COUNTRIES, 1950-1979 I INCEMAR TURESSONM, .D., OLLE ZETTERVALLM. .D., JACK CUZICKP,H.D..JAN G. WALDENSTROMM.D, ., AND RAMON \'ELEZ,M.D., M.Sc. Abstract Dramatic increases in mortality from multiple myeloma have been reported in the UnitedStates and the United Kingdom over the past three decades. To assess what fraction of this increase, if any, might be attributable to a change in the incidence of this disease. we examined the incidence of multiple myeloma during 1950 to 1979 in Malmi), Sweden, a city chosen because of its medi- cal community's longstanding interest in this disease. The average annual incidence rates per 100.000 pop- ulation were 4.9 for males and 3.7 for females (ad- justed to the European age-standardized population). These rates are among the highest in the world. Unlike secular trends for myeloma in the United States and the United Kingdom, where large increases in mortality rates have been reported, the rates in Malm6 have increased only slightly, and the increase was restricted to males, suggesting the possibi!ity of environmental causes. We suggest that the rates in Malmo may repre- sent the asymptote of myeloma's incidence that will be approached in other white populations as case ascertainment among them becomes complete. (N Engl J bled 1984; 310:421-4.) MULTIPLE myeloma, like other fatal diseases. is characterized by incidence and mortality rates that are similar. .A dramatic rise in multiple myeloma's incidence and mortality over the past three dec- *'des has been reported in studies from the United ~tltcs,1*E2ngland and Wales,3 and most other counFraumeni compared data from the United l~w8$huSecond and Third Cancer Surveys (1947 to and 1969 to 1971) and found the percentage in- Qcase in the age-adjusted incidence of multiple mye- k a to be greater than that of any other cancer, in- @ 'bg lung ~ a n c e r T. ~his apparent change in the 1. ,xnce of multiple myeloma has led some investiga'(Rto hypothesize a relation between this disease and iocnased exposures to environmental or occupational Qtcinogens.6.7 Arecent study conducted in Olmsted County, Min- nesota, identified incidence rates for multiple mye- loma in that community that were higher than those expetted from national mortality rates. However, under further analysis, no increase in incidence was detected during the period 1945 through 1977.8 This k h e r but stable incidence was attributed to more complete case ascertainment. The obvious etiologic importance of an increase in multiple myeloma prompted us to undertake an invesa p t i o n of its incidence in Malmo. h e d e n . where the medical community has had a longstanding interest in disease, Our objectives were to describe time t m d s for multiple myeloma's incidence in Malm6 1950 to 1979, to compare these findings with the data reported for Olmsted County. and to compare %*specific incidence rates over time in Malmo with national mortality rates in the United States and in and Wales. The results of the study indicate 'b n the D e p m e n t of Internal Medicine. Univenin of Lund. Malmo GenerH"Wlp(,M W . Sweden: the Cancer Epidcmioloe) and Clinical Tnals Unit. a%@" W V Of Oxford. RadcliffeInfirmary.Oxford: the Dcpmment of Medicine. Uavenity Medical Center. Durham. N.C.; and h e Department of Mcdi-on of General Medicine. Uruvenic of Teaas H d t h Science Center at b o . Address reprint rcqucsu to Dr Velez at dr Depuvnent of Medi- of Texas Health Science Center. S.n Ankmm. TX 78284. S.PPOIPCd in part by the Miibank Mcmonal Fund tDr. Vekzl. the Imperial ' -&arch Fund, and the T.and R. Soderkrg Fauwt.tlon that national mortality rates for multiple myeloma are not an accurate index ortrends in true incidence rates. .Although mortality rates have increased dramatically, much of the increase can be attributed to improved case ascertainment. However, in a stable study population under comprehensive medical surveillance, an increase in the incidence of multiple myeloma in males suggests increased or newly introduced environmental hazards as an etiologic factor. This study illustrates the difficulty of interpreting changes in the incidence and mortality of cancer over time.g.'OChanging rates may be real -e.g., cases are due to environmental exposures that affect underlying incidence of the disease -or they may be artifacts - cases are due to improved diagnostic techniques or case-ascertainment methods or both. Medical surveillance in Malmo provides nearly ideal conditions for accurately determining the rate of multiple myeloma. METHODS Population and Initial Case Identification Xlalmo is the third lnrsest citv in Sweden. T h e population increased from 192.668 in 1950 to 236.716 in 1979. T h e community's medical needs are s c n c d by three hospitals: Malmo General Hospital (the main hospital and the only one with an emergen? service), Vacrnhcms Hospital for chronic diseases. and the Malm6 Eastern Hospital for mental disorders. Records were obtained for all patients with a clinical diagnosis of multiple mveloma. plasmacytoma. or extramrdullan. plasmaqtoma from medical-records discharge files. p a t h o l o q files. and plasma protein-clcctrophoresis reports at these three hospitals from 1950 through 1979. For the period 1969 through 1979. a computerized diagnostic-registration system provided another data source. Autopsy records were checked for additional cases. T h e frequency of autopsies was over 95 per cent for hospitalized patients and and 60 per cent for all persons dvitlg in Y a l r n o since 1960. ensuring accurate estimates of cases not diagnosed during life. Inclusion and Exclusion Criterla Minimal criteria for a diagnosis of myeloma were a monoclonal immunoglobulin in serum or monoclonal light chains in urine (or both) plus at least one of the following: IO pcr cent or more plasma cells and atvpical plasma-crll m o r p h o l o q in the bone marrow; os- teolytic lesions found on x-ray examination that could bc explained onl! by increased plasma cells: or macroscopic ostrolytic lesions found at an autopsv in 1%hich microscop! revealed plasma-cell pro- $ *E r .; .-,'. t ..;d. ? 422 THE NE\\' EXGLAND JOURNAL OF MEDICINE liferation. .\ diagnosis was further confirmed by a clinical course compatible with m)eloma. Cases without an obscned M compo- nent in serum or urine %ereaccepted iftherc was histologic evidence oi J. solitary plasmacytoma or if the diagnosis was made first at autopss and no serum or urinary protein electrophoresis was performed. Patients with an .LI component whose conditions did not fulfill previouslv stated criteria were excluded. as were patients not living in Slalmo at the time of diagnosis. Statistical Procedures Sex. vear. and age at diaqnosis were recorded for each case. Cen- sus data for Stalmo's population were used to calculate age-specific and sex-specific incidence rates in IO-vear 3ge qroups. For comparison of rates over time 3nd with other sources. the direct method of age standardization by five-year aqc groups was used." The total U.S. population for 1950" and, alternatively. the European standard population were the standard populations.'' .Analyses for trends were performed bv means of the .\lantcl-Haenszel extension chisquare test.'' RESULTS In SIalmo: multiple myeloma was diagnosed in 329 residents from January 1, 1950,to December 31, 1979. Of these, 163 were male and 166 were female. More than 90 per cent of the patients were seen by one of us (I.T.,O.Z., 0rJ.G.W.) at the Department of Internal Medicine, Malmo General Hospital. The average, age-adjusted incidence rates for multiple myeloma in .Malm6 during three decades are shown in Table 1. There is a small and statistically significant trend toward increasing rates among males. No increase in the age-adjusted rates is evident among females. Figure I indicates that the increases in the incidence rates of myeloma among males were not confined to older age groups. These data may reflect a small but genuine increase in the incidence of the disease. The average. age-specific incidence rates per 100.000 population in hlalmo for 1970 throuqh 1979 are shown in Table 2 . These rates are comparable to rates reported from Olmsted County.8 Sex-specific and age-specific incidence rates for the periods 1950 through 1959. 1960 through 1969, and 1970 through 1979 are plotted on a log-log scale in Fiqure 1. I t is noteworthy that the incidence rates continue to increase with age. even in the older age Table 1 . Average Annual Age-Adjusted Incidence Rates of Multi- ple Myeloma per 100,000 Population (Malmo. Sweden, 1950 to 1979). According to Sex: hRlOD M.1ES CASES RArE ' FEMALLI C~SES u r E Toru CUES RATE 195&59 1W9 197679 Significvlcc of mend 32 2.9 (3.9) 38 2.6 (3.5) 50 3.5 (4.6) 69 3.5(4.5) 81 4.6 16.2) 59 2.2 (3.0) 5.52. *P = 0.02$ 1.10. tNS f 70 1.7 (3.7) 119 3.4 (4.6) I40 3.2(4.3) *0.81. ?IS f &-'Table 2. Average Annual Age-Speufc incidence Rates of ple Myeloma per i00.000 Population (Maimb, Sweden, 1970 1979), According lo Sex. FWrU C N U UTE C * p I UTE s39 4049 00 4 2.7 s59 60-59 14 8.4 25 18.1 7&79 a80 24 35 4 -I4 64.5 Torah 81 Adjusted rates * 4.6 16.2) 00 2 1.3 6 3.3 16 9.5 18 16.9 -17 36.6 39 2.2 (3.0) * ~ ~ ~ - d j u tao c19d50 U S i d ) poplhm md E m p u n s-t pmnrhm). 0 6 20 41 42 -31 140 0 2.0 3.1 13.4 24.0 45,s 3.2 14.3) popll.llOO (nhcl,. groups, and that the data are well fitted by a straight line on a log-log DISCUSSION Changes in the incidence of multiple myeloma time may provide clues to its cause.I7 Since most m- cers are thought to be associated with e n v i r o n m a d exposures, a change in the incidence of a panic& cancer should alert investigators to the possibie I&of environmental carcinogen^.'^ A classic urampic'd such a time trend is the steep increase in lung & -9.that first occurred among men earlier in this The major etiologic factor implicated in the inaeore'm lung cancer was cigarette ~ m o k i n g . 'R~etrospsetIve reasoning related this increase to differential exponue to the major etiologic agent - cigarette smoling. When investigating or interpreting an increase in any disease. however, it is necessary to ascertain first whether the increase is real or is a reflection of im- provements in recognition and clinical diagnoskM This first step is particularly important in the case of multiple myeloma, a disease in which diagnosis is highly dependent on the use of medical-care facilities and laboratory in~estigation.I~n addition, since m y - loma is a disease of old age. factors that increase the use of medical senices by the elderly are likelv tC produce concomitant increases in the frequency ofdi- agnosis and reporting of myeloma. -1he general structure of Swedish medical care" and the particular organization of medical services in Llalmo make this city especially suitable for popula- tion-based Improved medical care for the elderly has been a national priority in Sweden since 1950. In addition, Malmo's population has remained relatively stable since 1950 and has had unresviFlted access to the city's medical center, which is staffed by diagnosticians especially interested in multiple rnyt loma. Since there are few private doctors in Md6,a majority of outpatient visits (60 per cent) are made within the medical-center clinics. Thus, Malm6 p'p vides a stable study population for whom diagn-lJ INCIDENCE OF MULTIPLE MYELOMA - TCRESSOS ET AL. 423 100 trend in incidence rates for either males or females during this 33-year period. I t is instructive to compare age-incidence curves for 50 cases occurring in three decades in Malm0 with national mortality statistics for the Cnited Statesj4 and England and iL'ales3- countries that have had strik- ing increases in reported' mortality from m) eloma. 20 T h e age-specific incidcnce rates in 51almo increased progressively wi:h ace. continued to rise. and included the oldest age group I Fig. 1 \ . This age-incidence pattern has been relatively unchanqed durinq the past three decades. with the esception or age-specific rates for males. For both sexes combined. the age-specific rates for the group who were 80 years old or older increased from 37.9 per 100.000 for the decade 1950 through 1939 to 45.5 per 100.000 in the most recent decade. 1930 through 1979 - a 30 per cent increase In contrast. for white .Americans. the aqe-adjusted AGE (LCG SCALE1 . Annual Incidewm Rates Of Multiple Myekma per Wales, the age-adjusted rates per 100,OOO increased from 0.5 for males and 0.4 for females in 1950 to 3.3 fgr I < J "C --54 THE YEW ENGWND J O U R N A L OF MEDICISE In conclusion, we found the increase in the incidence of multiple myeloma in populations with comprehensive medical survedlance to be much smaller 7. Agu VU. Chnsunsen BL. Buffler PA. Geographic patrerns of muti*' myeloma: mid and indur~nalcorrelates. state of Texas. 1969-71. 1980: 65:735-8 a. Linos ,A. ~ y Rk.4. O'Fallon WM. Kurland LT. Incidence dnd V C U I ~ ~ than the increase found when national mortality statistics were analyzed. This suggests that a large portion of the increase in mortalitv from multiple mveloma can be explained bv improved case ascertain- ment, especially in the elderly. This conclusion is sup- of multiple myeloma in Olmsrcd County. Minnesota: lY65-77. JNCI 1981; 66.17-30. 9. Felmgtn AR. aimcal epidemiology. 11. The identification rater or-. & [num Med 1968. 691037-61. IO. Gmnberg ER. Colton T. Bagne C. Measurementof cancer i n c i d~ h United Staus: sources and uses of data. JNCI 1983: 68::43-50. 1 1 . a t a g e p. staustlCal methods in medical research. vcw y e : & ported by previous studies of other types ofcancer. We also found that incidence rates of multiple myeloma among males have increased in Malmij, Sweden. The Swedish study population was stable, had access to comprehensive medical care, was subject to uniform Wiley. 1971. 12. U.S. Dcpanmcnr of Hcdth. Education. d Welfare. Viral S ~ ~ U S I I CoJf& united States 19%. Vol 1. Washington. D.C.: Government Pnnting Oflicc, 195456. 13. Watchuse I . MUV C. Coma P. Powell 1. eds. Cancer inciderye in & cononens. Vol. 3. Lyon: lnicrnational Agency for Research on h, 1976. diagnostic criteria, and was part of a health-care-delivery system committed to excellent geriatric care. It is therefore unlikely that the increased incidence in 14. Manu1 N. C h i - j q w (csts with one d e g m of freedom: exunsiom d Q Mantel-Haenszcl pracedurr. J .Am Stat .ASSOC 1963: 58:6%%700. 15. Arnutagc P. Doll R. Stochasac models for cucinogenesis. In: Hey= J, d. Roccedings of h e Founh Berkeley Symposium on .Mathematical-S males is a n artifact due to improved case ascertainment. Rather. the data support the hypothesis that some occupationally related factor (or factors) that and F'robabilit). &rkclcy. Cdif.. University of Cdifornia h.196~:I~ *-38. 16. Cmk PI. Doll R. Fellmgham SA. A mathemaucal m d e l for the ogc buuon of cancer UI mm. Inc I C m e r 1969: 1~93-I1?. has increased or has been introduced into Malmo is a causative agent. Accumulating evidence, including a 17 .MuoCS. Timetremiss indicatorsotcrioloey. In: Magnus K. ed. T-i, cancer iocidcne: cauvs and pracucd implicailons. Washington. D.C.: Hcm~spherc.1982:S99-101 recent review article implicating radiation exposure as an environmental hazard of etiologic importance in some cases of myeloma,32suggests that some cases of 18. Wyndcr EL. Con GB. Contribution of the environment to cancer:-I an epidemolopc cxe~cisc.JNCI 1977. 58:825-32. - -.19. Doll R. Hill AB r\ snrdy ot the aetiologyof carcinomaof he lung. &- J 1952: 21171-86. .r the disease may be preventable. We are indebted to Prof. N. H. Sternbv (Department of Pathology)and Prof. C. 9. Laurel1 (Department of Clinical Chemistry) of v20. Saxen M .Trends: facts or fallacy. In: Magnus K. cd. Trcndr imlocrdcnce: causes and pncucal implications. Washington. D.C.; ~pherc.1982:5-16. 21. Anderson OW Health care: can there be quity?: The United Malm6 General Hospital for autopsv records and reports on protein electrophoresis; to Prof. Robert Silber (New York University Medi- 22. Sumy NH.A d ~ m s c l e ul Center) and Geoffrey Rose (London School of Hygiene and Tropical Medicine) for support, advice, and encouragement; and to Ms. Sondra Coonw and Ms.Sandra Rexrode for help in preparing the manuscript. hREwcEs fMH publicahon w . 882-2435), ,s;+. 25. Doll R. Peco R. Tbc causes of cancer quantitauve estimale of nslu of cancer UI rbe United Starcs today. JNCI 1981: 66:1191-=. 7. CE? 1rn1-I . Blamer WA. Blau A. Mason TI. Multiple myeloma in chc United Stam. 26. Lcukenua: towards conml. London: Office of Hcdth Econom~CS. --# 0- t t/a 19501975. Cancer 1981. 48:2547-W. 2 . Dcvesa SS. Silverman DT. C m e r incidence md monality mnds in rhc 47 27 Coun Brown %V.Doll R Leukaemia in childhad Jnd young adult life: 5t-. United States: 1935-74. JNCl 1978. 60.545-71 vends in m o d i t ) . in relation to aetiology Br Med j !%I.I 981-8. 3 Velez R. B e d V . Cuzick 1. Incrtasing trcnds of multiple myeloma mortal- 28 M c P h c h P. Heath CW Jr. Garcia IS Racial vanation3 in leukcmu mi- s;6- iry in England and Wales: 195079 arc Ihe changes na17 JNCI 1982; dence among the cldcrly JNCI 1970: 4 5 ~ 2 s - x . m- t ~ ~ 3 a 7 - 9 2 . ?9. Meadon GF Epidemiology of' leukemia. Public Healfh Rep 1956: 71:103- 4. Cuzick I. Velcz R. Doll R. lnternauonal vanauons a d temporal trends in a. momlity from mulople myeloma. Int J C m e r 1983: 3213-9 30 Linos A. Kyle R4.Elyeback LR. Kurland LT Lrukcmi~i n Olmsrcd Coub 5 . Fraumcni IF JI. Environmental and generic dercrminanls of cancer. J Envi- rv. Minnesota. 1965.1973 Mayo CIin PmL. 1978. S ? . : l l - X . mn PathOl Toxicol 1977: 1:19-30. 3 I Bnncker H. Populmon-based age- md ux-.;pecinc incidence rates in Ihe 4 6. Davis DL. Bndbrd K. Schneideman .M. Cancer prevention: messing main ? ~ osi !eukxrnia Scand J Hacmatol IYXZ: 19 111-9 causes. exposures. and recent wends in mortality for U.S.males IW-1978. 32. Cuzick J Radiauun-induced mvclomatosts. V Engl J Mcd ; % I ; 304:204- Tentogenesis Cycinog Mutagen 1982: 2: 105-35 IO.