91008461173Cancer Causes ControlCancer Causes ControlCancer causes & control : CCC0957-52431573-722523568533368442010.1007/s10552-013-0206-0NIHMS465597ArticleAnthropometric, behavioral, and female reproductive factors and risk of multiple myeloma: a pooled analysisWangSophia S.1+VoutsinasJenna1ChangEllen T.24ClarkeChristina A.34LuYani1MaHuiyan1WestDee34LaceyJames V.Jr.1BernsteinLeslie1Division of Cancer Etiology, Department of Population Sciences, Beckman Research Institute of the City of Hope, Duarte, California Health Sciences Practice, Exponent, Inc., Menlo Park, CA Cancer Prevention Institute of California, Fremont, CA Division of Epidemiology, Department of Health Research and Policy, Stanford University, School of Medicine, Stanford, CA To whom correspondence should be addressed: Sophia S. Wang, Ph.D., Division of Cancer Etiology, Department of Population Sciences, City of Hope and the Beckman Research Institute, 1500 East Duarte Road, Duarte, CA 91010, Phone: (626) 471-7316, FAX: (626) 471-7308, sowang@coh.org1652013094201372013017201424712791289Background

Risk of developing multiple myeloma (MM) rises with age and is greater among men and blacks than among women and whites, respectively, and possibly increased among obese persons. Other risk factors remain poorly understood. By pooling data from two complementary epidemiologic studies, we assessed whether obesity, smoking, or alcohol consumption alters MM risk and whether female reproductive history might explain the lower occurrence of MM in females than males.

Methods

The Los Angeles County MM Case-Control Study (1985-92) included 278 incident cases and 278 controls, matched on age, sex, race, and neighborhood of residence at case’s diagnosis. We estimated MM risk using conditional logistic regression to calculate odds ratios (OR) and 95% confidence intervals (CI). In the prospective California Teachers Study (CTS), 152 women were diagnosed with incident MM between 1995-2009; we calculated hazard ratios using Cox proportional hazards analysis. Data from the two studies were pooled using a stratified, nested case-control sampling scheme (10:1 match) for the CTS; conditional logistic regression among 430 cases and 1,798 matched controls was conducted.

Results

Obesity and smoking were not associated with MM risk in the individual or combined studies. Alcohol consumption was associated with decreased MM risk among whites only (pooled OR=0.66, 95% CI=0.49-0.90) for ever vs. never drinking). Higher gravidity and parity were associated with increased MM risk, with pooled ORs of 1.38 (95% CI=1.01-1.90) for ≥3 versus 1-2 pregnancies and 1.50 (95% CI=1.09-2.06) for ≥3 versus 1-2 live births.

Conclusions

Female reproductive history may modestly alter MM risk, but appears unlikely to explain the sex disparity in incidence. Further investigation in consortial efforts is warranted.

multiple myelomawomenreproductivemodifiablerisk factorsassociationpoolingcase-controlcohortepidemiology
Introduction

Multiple myeloma (MM) accounts for nearly one-fifth of hematologic malignancies (1). MM incidence rises with age and is higher in African Americans, men, and among those with a family history of hematologic malignancies (1). Identification of sex-specific factors may offer clues to explain the male predominance in MM, but female reproductive history has also been inconsistently associated with MM risk in prior studies (2-4). Among lifestyle factors, some, but not all, evidence supports modestly elevated MM risk among obese persons (5, 6); results have been largely inconsistent for associations with smoking behavior and alcohol consumption (1). Most studies of MM have been case-control studies, which are susceptible to differential biases in exposure ascertainment and the loss of cases through early death or substantial illness that could affect interpretation of results.

To shed additional light on potential risk factors, we evaluated data from two complementary epidemiologic studies of MM conducted in California: a case-control study of men and women in Los Angeles (LA) County, and the longitudinal California Teachers Study (CTS) cohort of women only. We evaluated the associations of obesity, smoking, and alcohol consumption with MM risk, updating and expanding upon previous assessments of associations with anthropometry and alcohol consumption in the CTS (7, 8). We hypothesized that female steroid hormones may alter the immune response in a manner that decreases MM risk. The comparison of case-control and cohort-based analyses permitted us to evaluate consistency of associations and thus address potential biases in retrospective exposure ascertainment and participation (e.g., survival and selection bias) in case-control studies.

Materials and MethodsStudy PopulationsThe Los Angeles County Multiple Myeloma Case-Control (LAMMCC) Study

The LAMMCC includes 278 cases (152 men, 126 women) diagnosed between 1985 and 1992 and 278 individually matched neighborhood controls from Los Angeles County. Incident MM cases were identified through rapid reporting in the Los Angeles County Cancer Surveillance Program, a population-based cancer registry; controls were recruited based on specific algorithms that placed their residence near the residence at diagnosis of their matched case, as previously reported (9). Controls were matched on date of birth within 5 years, sex, and race. All participants were interviewed in person. For evaluation of timing of exposures, a reference date was defined as the date of the case’s diagnosis for both the case and matched control. When compared to Surveillance Epidemiology and End Results (SEER) data for the study period in LA County, the race distribution of cases was similar among men, but as per study protocol, blacks were oversampled among women. Age distributions compared to SEER were also equivalent, with the exception of the LAMMCC’s exclusion of cases over 75 years. The study population was therefore largely representative of the general SEER population for LA County during the study period with the exception of study design attributes.

The CTS cohort

The CTS is a prospective cohort comprising 133,479 female public school professionals recruited through the California State Teachers Retirement System (10-12). In 1995-1996, participants completed a detailed, self-administered questionnaire that gathered information on demographics, anthropometric characteristics, lifestyle factors, and reproductive factors. Updates to menopausal status were obtained in follow-up questionnaires administered in 1997-98 and 2000-2001. For this analysis, we included 123,396 women after excluding those who, at baseline, had a previous diagnosis of MM or other hematologic cancers (n=536), had an unknown prior history of cancer (n=662), were not California residents (n= 8,867), or wished to participate only in breast cancer studies (n=18). Incident MM cases (ICD-O-3 9732-9733) were identified in the CTS through linkage with the California Cancer Registry. In the CTS, all cancers are identified prospectively by the Cancer Surveillance Program (>99% identification).

Follow-up time for each CTS participant began on the date the baseline questionnaire was completed and continued until the first of the following outcomes: a first diagnosis of any hematological malignancy, a move outside of California for >4 months, death, or the end of the follow-up period (December 31, 2009). The status of California residence was monitored through annual mailings, responses from participants and routine record linkages with multiple sources, including the US Postal Service National Change of Address database. The date and cause of death were ascertained through linkage with California and national mortality records (the Social Security Death Master File and the National Death Index).

Risk factors

We evaluated risk factors that were similarly queried in both studies. These characteristics included: cigarette smoking (never, current, or former; pack-years for current and former smokers), alcohol intake (number of glasses of beer, wine, and liquor consumed during a typical week prior to the reference date in the LAMMCC study and during the year prior to baseline in the CTS), body weight (pounds), height (inches), and corresponding body mass index (BMI) calculated as kg/m2: (i) at age 18 years old and (ii) as an adult, defined as one year prior to MM diagnosis/reference date for the LAMMCC study and the participant’s age at the date of the baseline questionnaire for CTS.

For analytic purposes, height, weight, and weight at 18 years old were evaluated as tertiles. The tertile values for the CTS were as follows for (a) height (inches): tertile 1 (<64), tertile 2 (64-65), tertile 3 (>65); (b) weight (pounds): tertile 1 (<131), tertile 2 (131-154), tertile 3 (>154); and (c) weight at 18 years (pounds): tertile 1 (<119), tertile 2 (119-130), tertile 3 (>130). The LAMMCC tertile values for females are as follows for (a) height (inches): tertile 1 (<62), tertile 2 (62-64), tertile 3 (>64); (b) weight (pounds): tertile 1 (<137), tertile 2 (137-164), tertile 3 (>164); and (c) weight at 18 years (pounds): tertile 1 (<111), tertile 2 (111-125), tertile 3 (>125). The LAMMCC tertile values for males are as follows for (a) height (inches): tertile 1 (<68), tertile 2 (68-70), tertile 3 (>70); (b) weight (pounds): tertile 1 (<168), tertile 2 (168-184), tertile 3 (>184); and (c) weight at 18 years (pounds): tertile 1 (<140), tertile 2 (140-160), tertile 3 (>160). Tertile values for pooled analyses (women only) are as follows for (a) height (inches): tertile 1 (<63), tertile 2 (63-65), tertile 3 (>65); (b) weight (pounds): tertile 1 (<131), tertile 2 (131-155), tertile 3 (>155); and (c) weight at 18 years (pounds): tertile 1 (<118), tertile 2 (118-129), tertile 3 (>129).

Female reproductive characteristics ascertained in both studies included number of pregnancies (combining live births, stillbirths, abortions, miscarriages, and ectopic pregnancies); number of live births; age at menarche; oral contraceptive use (years); and menopausal status (natural menopause, surgical menopause by bilateral oophorectomy or hysterectomy, premenopausal, or other/unknown). In the CTS, women under age 56 years who listed a simple hysterectomy as the cause of their last menstrual period were considered to have unknown menopausal status (13). In the LAMMCC study, menopausal status was ascertained at the referent date (date of diagnosis for the case) by asking if the women were still having menstrual periods. Those who indicated yes were identified as premenopausal, and those who indicated no were identified as menopausal. Surgical versus natural menopause was further delineated by their responses to whether they had a hysterectomy and/or an ovary removed before the referent date. In the CTS (but not the LAMMCC study), we were able to further discern between bilateral or unilateral oopherectomy with or without a hysterectomy.

Family history of hematologic malignancies was defined as having any first-degree relative with Hodgkin lymphoma, non-Hodgkin lymphoma, leukemia, or multiple myeloma (yes, no). Both studies ascertained family history of leukemia and other lymphomas, and the LAMMCC study also ascertained family history of multiple myeloma.

Statistical Analysis

To estimate relative risk for MM in the LAMMCC study, we used conditional logistic regression to compute odds ratios (ORs) and 95% confidence intervals (CIs) among men and women combined and by sex. In the CTS, we used multivariable Cox proportional hazards regression models to generate hazard rate ratios (HRs) and 95% CIs, with age in days from baseline until the end of follow-up as the time scale and models stratified by age in years at baseline and adjusted for race.

Both studies included additional multivariate adjustment for specific risk factors. BMI, smoking, and alcohol were adjusted for socioeconomic status (SES). Models for BMI and smoking were further adjusted for alcohol consumption. Numbers of pregnancies and live births were adjusted for age at menarche. SES in the LAMMCC study was based on education level (some high school or less; high school graduate; some college and above) because cases and controls were matched on neighborhood characteristics. The CTS data used residential neighborhood-level SES, based on address at baseline, as measured by a composite index that combined 1990 U.S. Census block-group-level data on education, income, and occupation, categorized into quartiles according to the California statewide distribution of the SES index (14). Different SES measures were used in the two studies because LAMMCC participants were matched on neighborhood, whereas nearly all CTS participants had at least a college degree. Additional adjustment for other risk factors, such as family history of hematopoietic malignancies, socioeconomic status, and further mutual adjustment for risk factors (e.g, BMI for reproductive factors, and vice versa) did not alter the risk estimates by ≥10%; we thus present the most parsimonious statistical models. Secondary analyses were conducted stratified by white or black race.

Pooled analysis

Cases and controls from the LAMMCC study were combined with a nested case-control sample of CTS participants. For each of the 152 MM cases in the CTS, we randomly selected 10 controls from risk sets matched on race, date of birth within 5 years, geographic region (San Francisco Bay Area, Southern California excluding Los Angeles, Los Angeles, Central/Southern California, Northern California excluding San Francisco, and the Central Valley). Our rationale for selecting 10 controls per MM case for the CTS was to maximize our power for detecting modest associations. Sensitivity analyses were also conducted whereby a 1:1 match was used to assure that our results were not weighted by the greater number of controls to CTS cases, compared to the LAMMCC study. The final pooled dataset thus comprised 430 cases and 1,798 controls.

We conducted conditional logistic regression to calculate ORs and 95% CIs for the pooled analyses. To retain matches, missing categories were created for each variable and covariates where applicable. The test for trend was obtained with logistic regression with each tertile coded as a 3-level ordinal variable for each tertile (0 = tertile 1, 1 = tertile 2, 2 = tertile 3). All statistical analyses were performed using SAS version 9.2 (SAS Institute Inc, Cary, NC).

Results

Table 1 shows selected demographic characteristics of participants by study, sex, and case status.

Anthropometric characteristics

Adult height was associated with increased MM risk among women but not men in the LAMMCC study (third versus first tertile OR=2.31, 95% CI=1.04-5.13), whereas the association was weaker among women in the CTS (HR=1.40, 95% CI=0.94-2.09) (Table 2). In the pooled analysis among women, MM risk remained marginally positively associated with elevated height (third versus first tertile OR=1.39, 95% CI=0.95-2.02) (Table 3). Adult and young-adult weight and BMI were not associated with MM risk in either study. Results stratified by race are shown in Supplemental Tables 1 and 2.

Cigarette smoking and alcohol intake

Smoking status was not associated with MM risk in either study (Table 2). Further evaluation by pack-years similarly yielded no associations with MM (data not shown).

Relative risk estimates for MM among ever versus never drinkers were all below 1.0 but not statistically significant in either the individual studies or in the pooled analysis. Evaluation of ever versus never alcohol consumption in the pooled analysis by race yielded a statistically significant decreased MM risk among whites (OR=0.66, 95% CI=0.49-0.90), but not among blacks (Supplemental Tables 1 and 2). We observed no specificity in risk by type of alcohol consumed (wine, beer, or liquor) and no evident dose-response association by number of drinks per day (data not shown).

Female reproductive characteristics

Compared with women with 1-2 pregnancies, women with 3 or more pregnancies had statistically significantly elevated MM risk in the CTS (HR=1.57, 95% CI=1.08-2.29); results for number of live births were similar (Table 2). In the LAMMCC study, the ORs were modestly above 1.0 and not statistically significant. In the pooled analysis, MM risk was statistically significantly associated with 3 or more versus 1-2 pregnancies (OR=1.38, 95% CI=1.01-1.90). Similarly, we observed a 1.50-fold increased MM risk (95% CI=1.09-2.06) among women with 3 or more versus 1-2 live births. The association was statistically significant among whites (OR=1.52, 95% CI=1.05-2.12), but not among blacks (Supplemental Tables 1 and 2). Surgical menopause was statistically significantly associated with increased MM risk in the LAMMCC study (OR=1.85, 95% CI=1.06-3.25), but not in the CTS (Table 2). In the pooled analysis, the OR remained above 1.0 but was of borderline statistical significance (OR=1.33, 95% CI=0.97-1.82) (Table 3). The positive association with surgical menopause was observed among whites, but not blacks (Supplemental Tables 1 and 2).

Discussion

This evaluation of anthropometric, behavioral, and reproductive risk factors revealed statistically significant increased risks of MM among women with higher gravidity or parity. Overall, we did not observe statistically significant associations for anthropometric features, alcohol consumption, or smoking behavior. Among whites only, we observed a statistically significantly modest decrease in MM risk with ever versus never alcohol consumption, but this was not accompanied by an inverse dose-response trend.

These null associations with anthropometric measures are consistent with findings from a nationwide Swedish cohort study and the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort study (15, 16), but not U.S. cohort studies among men and women (Nurses’ Health Study and Cancer Prevention Study cohorts) (17, 18). However, recent meta-analyses comprising a combined 8,982 incident multiple myeloma cases reported a modest 20% increase in MM risk among obese persons in both cohort (6) and case-control studies (5, 19). This positive association is further supported by recent evidence that risk for the precursor condition for MM, monoclonal gammopathy of undetermined significance, is elevated two-fold among obese persons (20). It is thus possible that our study lacked sufficient power to detect such a weak association.

The literature on smoking and MM risk remains unresolved, with one review article indicating no difference in MM risk with smoking (21), but more recent studies indicating some elevation in risk (1, 22). Our overall null association for smoking is consistent with results from other previous case-control and cohort studies (15, 23) and a recent meta-analysis of previously published case-control and cohort studies indicating no association (24). The evidence for alcohol consumption and MM risk is similarly unresolved, with results split between increased and decreased risks and an overall null association (1, 21). This is in contrast to the now relatively consistent protective association observed for non-Hodgkin lymphoma (25). Our findings add to the evidence that cessation of smoking is unlikely to influence MM risk, but further evaluation of associations with alcohol by race may be warranted.

We hypothesized that female steroid hormones may alter the immune response in a manner that decreases risk for MM. However, the positive associations of more pregnancies (resulting in fewer menstrual periods) and surgical menopause with MM risk may point to differences in hormone levels as a potential factor of interest for MM risk. Our results are consistent with findings from a recent case-control study of hematopoietic malignancies in Europe that also reported increased risk for women who reported being ever pregnant (3). However, we note that few studies have evaluated reproductive characteristics and of those who have, similar associations have not been reported (2). The modest positive association observed for increasing number of pregnancies may also suggest that repeated immunological changes, such as immune suppression that occurs during pregnancy (26, 27), promotes an environment that could contribute to MM development. Although we sought to identify reproductive characteristics that might explain the higher male to female ratio in MM incidence, our results suggest that higher female steroid hormones probably do not explain the lower MM incidence among female.

Differences in MM tumor characteristics by sex have previously been reported; female patients have more IgH translocations and inferior survival, whereas male patients have more frequent hyperdiploidy (28). Some epidemiologic associations have also varied by sex; for example, in the European Prospective Investigation on Cancer study, height was associated with increased MM risk among women but not men, and higher weight was associated with increased MM risk among men only (16). These differences could be due to chance, but biochemical studies have reported lower estrogen levels and a decreased estrogen-to-testosterone ratio among women with MM than women without MM (29, 30). These results point to potentially real biological differences in MM risk factors and tumor characteristics by sex.

The strengths of our analyses include our ability to compare results and pool from two complementary epidemiologic studies – a case-control study and a prospective cohort study conducted in the same geographic area – to identify consistent associations. Behavioral characteristics are particularly prone to recall bias in retrospective studies, and the parallel evaluation of the same characteristics in a prospective cohort is thus advantageous. Study limitations include the modest sample size for this rare cancer, limiting our ability to detect modest associations. Although our analysis included a sizable black population, stratified analyses by race were not robust due to small sample size. Thus, we cannot say with certainty if associations observed among whites (e.g., alcohol consumption) are definitely null among blacks. To increase validity, we limited our pooled analysis to variables that were asked in a consistent manner between the two studies. While differences in study design, exposure classification, and exposure prevalence, along with chance, may explain some of the lack of consistency in results between the two studies, they also make consistently observed associations particularly interesting. Consistency of associations between cohort data and case-control data was also beneficial due to potential biases among controls who participate. Specifically, as noted previously (31), selection bias among controls participants include higher education/SES and differences in reproductive factors, including lower birth order and parity than expected. These biases are known to result in muted risk estimates among case-control studies; while we cannot confirm whether ascertainment bias among controls occurred in the LAMMCC study, the lower risk estimate observed with the reproductive risk factors compared CTS would be consistent with this potential explanation.

In conclusion, our results suggest a possible role for female reproductive characteristics in MM etiology, albeit not in the direction that we hypothesized. We cannot discount the possibility that these findings are due to chance, but these consistent results from two complementary studies warrants further evaluation to confirm and better understand the roles of hormones and reproductive characteristics in MM development.

Supplementary MaterialAcknowledgements

This research was supported in part by federal funds from the NCI, NIH, under R01CA036388, R01CA077398, and K05CA136967 and by the City of Hope. We are indebted to the late Ronald K. Ross, M.D. who was the original principal investigator for the Los Angeles County Multiple Myeloma Case-Control Study and a former principal investigator for the California Teachers Study. We are grateful to the participants in the case-control study, to the women of the California Teachers Study for their continued participation and enthusiasm for the study, and to the California Teachers Study Steering Committee for their scientific guidance and insight. The collection of cancer incidence data used in this study was supported by the California Department of Public Health as part of the statewide cancer reporting program mandated by California Health and Safety Code Section 103885; the National Cancer Institute’s Surveillance, Epidemiology and End Results Program under contract HHSN261201000140C awarded to the Cancer Prevention Institute of California, contract HHSN261201000035C awarded to the University of Southern California, and contract HHSN261201000034C awarded to the Public Health Institute; and the Centers for Disease Control and Prevention’s National Program of Cancer Registries, under agreement U58DP003862-01 awarded to the California Department of Public Health. The ideas and opinions expressed herein are those of the author(s) and endorsement by the State of California, Department of Public Health the National Cancer Institute, and the Centers for Disease Control and Prevention or their Contractors and Subcontractors is not intended nor should be inferred.

ReferencesBeckerNEpidemiology of multiple myelomaRecent Results Cancer Res2011183253521509679TavaniAPregnolatoALa VecchiaCA case-control study of reproductive factors and risk of lymphomas and myelomasLeuk Res199721988589393604CostasLCasabonneDBenaventeYReproductive factors and lymphoid neoplasms in Europe: findings from the EpiLymph case-control studyCancer Causes Control201223119520622116538MortonLMWangSSRichessonDAReproductive factors, exogenous hormone use and risk of lymphoid neoplasms among women in the National Institutes of Health-AARP Diet and Health Study CohortInt J Cancer20091241127374319253366LarssonSCWolkABody mass index and risk of multiple myeloma: a meta-analysisInt J Cancer2007121112512617680557WallinALarssonSCBody mass index and risk of multiple myeloma: a meta-analysis of prospective studiesEur J Cancer2011471116061521354783LuYSullivan-HalleyJHendersonKDAnthropometric characteristics and multiple myeloma riskEpidemiology2010212272320160567ChangETClarkeCACancholaAJAlcohol consumption over time and risk of lymphoid malignancies in the California Teachers Study cohortAm J Epidemiol20101721213738320952595NelsonRALevineAMBernsteinLReproductive factors and risk of intermediate- or high-grade B-Cell non-Hodgkin’s lymphoma in womenJ Clin Oncol20011951381711230482BernsteinLAllenMAnton-CulverHHigh breast cancer incidence rates among California teachers: results from the California Teachers Study (United States)Cancer Causes Control20021376253512296510ChangETCancholaAJClarkeCADietary phytocompounds and risk of lymphoid malignancies in the California Teachers Study cohortCancer Causes Control20112222374921107674PrescottJLuYChangETReproductive factors and non-Hodgkin lymphoma risk in the California Teachers StudyPLoS One2009412e813519956586MarshallSFClarkeCADeapenDRecent breast cancer incidence trends according to hormone therapy use: the California Teachers Study cohortBreast Cancer Res2010121R420064209ReynoldsPHurleySGoldbergDERegional variations in breast cancer among california teachersEpidemiology20041567465415475725FernbergPOdenbroABelloccoRTobacco use, body mass index, and the risk of leukemia and multiple myeloma: a nationwide cohort study in SwedenCancer Res200767125983617575169BrittonJAKhanAERohrmannSAnthropometric characteristics and non-Hodgkin’s lymphoma and multiple myeloma risk in the European Prospective Investigation into Cancer and Nutrition (EPIC)Haematologica2008931116667718835833BirmannBMGiovannucciERosnerBBody mass index, physical activity, and risk of multiple myelomaCancer Epidemiol Biomarkers Prev20071671474817627013CalleEERodriguezCWalker-ThurmondKOverweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adultsN Engl J Med20033481716253812711737LichtmanMAObesity and the risk for a hematological malignancy: leukemia, lymphoma, or myelomaOncologist20101510108310120930095LandgrenORajkumarSVPfeifferRMObesity is associated with an increased risk of monoclonal gammopathy of undetermined significance among black and white womenBlood201011671056920421448AlexanderDDMinkPJAdamiHOMultiple myeloma: a review of the epidemiologic literatureInt J Cancer2007120Suppl 12406117405120NietersADeegEBeckerNTobacco and alcohol consumption and risk of lymphoma: results of a population-based case-control study in GermanyInt J Cancer200611824223016080191MonnereauAOrsiLTroussardXCigarette smoking, alcohol drinking, and risk of lymphoid neoplasms: results of a French case-control studyCancer Causes Control2008191011476018781390PsaltopoulouTSergentanisTNKanelliasNTobacco smoking and risk of multiple myeloma: A meta-analysis of 40 observational studiesInt J Cancer2012MortonLMZhengTHolfordTRAlcohol consumption and risk of non-Hodgkin lymphoma: a pooled analysisLancet Oncol2005674697615992695MorGCardenasIThe immune system in pregnancy: a unique complexityAm J Reprod Immunol20106364253320367629PriddyKDImmunologic adaptations during pregnancyJ Obstet Gynecol Neonatal Nurs199726438894BoydKDRossFMChiecchioLGender disparities in the tumor genetics and clinical outcome of multiple myelomaCancer Epidemiol Biomarkers Prev20112081703721680536EverausHHeinMZilmerKPossible imbalance of the immuno-hormonal axis in multiple myelomaLeuk Lymphoma1993115-645388124218EverausHLehtmaaJSex hormones and immune dysregulation in multiple myelomaCancer Immunol Immunother199336320598439982GlaserSLClarkeCAKeeganTHAttenuation of social class and reproductive risk factor associations for Hodgkin lymphoma due to selection bias in controlsCancer Causes Control2004157731915280631

Selected demographic characteristics of participants in the Los Angeles Multiple Myeloma case-control (LAMMCC) study (1985-1992) and the California Teachers Study (CTS) cohort (1995-2009).

LAMMCC StudyCTS
MenWomenWomen
Demographic characteristicsMM cases(n=152)Controls(n=152)MM cases(n=126)Controls(n=126)MMcases(n=152)Cohort(n=123,396)
n%n%n%n%n%n%
Race
White112(74)114(75)75(60)75(60)125(82)106,765(87)
Black21(14)21(14)39(31)39(31)12(8)3,288(3)
Other/Unknown19(13)17(11)12(10)12(10)15(10)13,343(10)
Age at diagnosis (years)
 < 4510(7)6(5)1(1)
 45-5422(14)22(17)11(7)
 55-6458(38)40(32)24(16)
 65-7460(39)58(46)45(30)
 75+2(1)0(0)71(47)
Family history of hematopoietic malignancies (first deg ree re lative s) α
 No145(97)149(98)122(98)121(98)149(98)116 ,447(97)
 Yes5(3)3(2)3(2)2(2)3(2)2,958(3)
Education
 Some high school or less26(17)26(17)26(21)34(27)
 High school grad28(19)32(21)47(38)37(29)
 Some college and above97(64)94(52)52(42)55(44)
Socioeconomic status (SES)
 Low11(7)5,455(4)
 Med-Low20(13)21,301(17)
 Med-High57(38)39,935(32)
 High63(41)55,123(45)

LAMMCC includes family history of multiple myeloma, leukemia, Hodgkin Disease or other lymphoma; CTS includes family history of leukemia, Hodgkin Disease or other lymphoma.

Evaluation of anthropometric, behavioral, and reproductive characteristics and multiple myeloma risk in the Los Angeles Multiple Myeloma case-control (LAMMCC) study and the California Teachers Study (CTS) cohort.

LAMMCC StudyCTS
AllΔMenΔWomenΔWomen0
cases/controlsOR95% CIcases/controlsOR95% CIcases/controlsOR95%CINPersonYearsCasesHR95%CI
Anthropometric characteristics
Height at interview/Q1(inches)
 Tertile 159/751.00-41/411.00-18/341.00-40483501987491.00-
 Tertile 2119/98 1.61 (1.01-2.55)59/491.24(0.68-2.27)60/49 2. 6 6 (1.23 - 5.77) 35827451149471.29(0.86-1.94)
 Tertile 397/1051.21(0.77-1.90)51/620.85(0.48-1.50)46/43 2. 3 1 (1.04 - 5.13) 46505589703551.40(0.94-2.09)
Weight at interview/Q1(pounds)
 Tertile 191/911.00-44/501.00-47/411.00-40914513633521.00-
 Tertile 284/910.92(0.62-1.38)51/481.16(0.68-1.99)33/430.69(0.37-1.26)37910479414470.89(0.60-1.32)
 Tertile 399/961.02(0.68-1.54)55/541.15(0.66-1.99)44/420.89(0.48-1.65)39324496648460.83(0.55-1.24)
BMI at interview/Q1* (kg/m2)
 <25116/1071.00-58/551.00-58/521.00-72079909860891.00-
 25-29.998/1190.75(0.51-1.10)65/740.85(0.52-1.39)33/450.62(0.34-1.17)29459370557390.84(0.57-1.23)
 30-34.943/400.98(0.59-1.62)19/190.96(0.46-2.01)24/210.92(0.45-1.88)10723135230120.74(0.40-1.36)
 35+21/121.86(0.84-4.14)8/41.80(0.51-6.30)11/81. 5 6(0.55-4.40)58437365050.69(0.28-1.72)
Weight at 18 (pounds)
 Tertile 185/921.00-44/501.00-41/431.00-40379509068491.00-
 Tertile 293/991.04(0.67-1.60)51/481.16(0.68-1.99)39/470.90(0.48-1.69)40693513470551.24(0.84-1.84)
 Tertile 392/831.22(0.79-1.88)55/541.15(0.66-1.99)42/341.33(0.69-2.58)36535461122421.17(0.77-1.78)
BMI at 18*α (kg/m2)
 <2080/871.00-31/331.00-49/541.00-39509498005541.00-
 20-21.984/821.07(0.67-1.72)45/401.21(0.61-2.41)39/420.92(0.48-1.77)39797503198410.75(0.50-1.13)
 22+106/1051.14(0.72-1.79)72/771.02(0.55-1.87)34/281.35(0.67-2.73)37661474682491.03(0.70-1.51)
Smoking*
 Never113/1111.00-52/481.00-61/631 .00-811971027688921.00-
 Ever165/1671.01(0.71-1.43)100/1040.88(0.54-1.46)65/631.10(0.67-1.80)41 42 7511790581.10(0.78-1.54)
  Current40/540.79(0.48-1.32)25/320.72(0.35-1.48)15/220.73(0.35-1.54)35201437341471.50(0.80-2.83)
  Former125/1131.09(0.75-1.59)75/720.93(0.55-1.56)50/411.31(0.75-2.29)622674449111.03(0.72-1.448)
Alcohol*
 Ever130/1211.00-54/461.00-76/751.00-39102483209561.00-
 Never148/1560.86(0.59-1.24)98/1050.81(0.49-1.34)50/510.93(0.54-1.60)76579972509820.81(0.57-1.15)
Female reproductive characteristics
Number of pregnancie
 08/120.64(0.22-1.84)25089312069241 .27(0.77-2.09)
 1-241/411.00-49631624835431.00-
 3+77/731.11(0.64-1.93)4619858252179 1. 5 7 (1.08 - 2.29)
Number of live births
 0N/AN/A15/170.96(0.43-2.16)32672407453311.23(0.79-1.91)
 1-248/531.00-59198749179541.00-
 3+63/561.29(0.76-2.18)2910036345361 1. 5 8 (1.09 - 2.31)
Menopausal status
 Natural menopause47/611 .00-47928633322151.00-
 Surgical menopause(hysterectomy oroopherectomy)67/48 1. 8 5 (1-06-3.25)39059469463721.12(0.76-1.64)
 Premenopausal11/160.64(0.19-2.19)20177242422420.80(0.33-1.94)

LA MM case-control study estimates matched on sex, date of birth +/− 5 years, and race

CTS estimates adjusted for age and race Additional adjustments for individual models include: *SES (CTS) or education (case–control)

alcohol

age at menarche

Multiple myeloma risk for select anthropometric, behavioral, and reproductive characteristics from pooled analyses of the Los Angeles Multiple Myeloma case–control (LAMMCC) study and the California Teachers Study (CTS) cohort with matched controls (10:1).

AllΔWomenΔ
cases/controlsOR95% CIcases/controlsOR95% CIP-heterogeneity
Anthropometric characteristics
Height at interview/Q1 (inches)
 Tertile 1115/4821.00-74/4411.00-
 Tertile 2174/7341.20(0.89-1.62)115/6851.21(0.85-1.72)
 Tertile 3137/5691.18(0.86-1.61)86/5071.39(0.95-2.02)
p-trend=0.33 p- trend=0.1 0 0.52
Weight at interview/baseline (pounds)
 Tertile 1133/5601.00-89/5101.00
 Tertile 2134/5630.95(0.71-1.28)83/5150.87(0.61-1.23)
 Tertile 3152/5730.95(0.71-1.27)97/5190.88(0.62-1.24)
p-trend=0.65 p-trend=0.34 0.93
BMI at interview/baseline* (kg/m2)
 <25205/9061.00-147/8511.00-
 25-29.9137/5260.79(0.60-1.04)72/4520.79(0.57-1.11)
 30-34.955/1950.85(0.58-1.24)36/1760.83(0.53-1.30)
 35+22/691.05(0.59-1.88)14/650.91(0.46-1.80)
p-trend=0.49 p-trend=0.38 0.63
Weight at 18 years (pounds)
 Tertile 1140/5521.00-96/5031.00
 Tertile 2138/5551.15(0.85-1.56)84/5031.14(0.80-1.63)
 Tertile 3138/5781.17(0.86-1.59)88/5291.18(0.83-1.70)
p- trend=0.3 0 p- trend=0.3 3 0.28
BMI at 18* (kg/m2)
 <20134/5531.00-103/5201.00
 20-21.9125/5800.87(0.64-1.19)80/5400.78(0.54-1.11)
 22+155/5391.02(0.75-1.39)83/4621.07(0.75-1.53)0.29
Smoking*
 Never205/10731.00-153/10251.00
 Ever223/7191.07(0.84-1.37)123/6151.09(0.83-1.44)0.15
  Current51/1170.97(0.64-1.47)26/851.11(0.67-1.87)
  Former172/6021.10(0.84-1.42)97/5301.09(0.81-1.46)
Alcohol*
 Never186/6281.00-132/5821.00-
 Ever230/11420.79(0.61-1.02)132/10370.78(0.58-1.06)0.74
Female reproductive characteristics
Number of pregnancies
 032/2681.04(0.66-1.66)
 1-284/6021.00-
 3+156/736 1.38 (1.01 - 1.90) 0.35
Number of live births
 046/3341.14(0.77-1.70)
 1-2N/A102/7391.00-
 3+124/534 1.50 (1.09- 2.06) 0.61
Menopausal status
 Natural menopause119/8081.000.01
 Surgical menopause (hysterectomy oroopherectomy)109/4381.33(0.97-1.83)
 Premenopausal26/2100.60(0.30-1.21)
 Other/unknown24/1901.25(0.75-2.07)

Matched on sex, date of birth +/− 5 years, and race Additional adjustments for individual models include:

education level

age at menarche

Malcohol