Sex differences in outcomes in multiple myeloma

Despite being perhaps the most fundamental distinction among human beings, the impact of sex on outcomes in multiple myeloma (MM) is unknown. Given that men are more likely to develop MM compared to women, and an analysis of all cancers using the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) Program suggested possible worse MM survival for men than women,¹ a comprehensive evaluation of sex differences in survival in MM is warranted. In an analysis of 1 960 MM patients enrolled in a phase III clinical trial, sex was associated with overall survival (OS) on univariate analysis (median OS 44·8 months female vs 49·9 months male, P = 0·02) but was not significant in multivariate analysis. Similarly, an analysis of the Myeloma XI trial found no sex differences in progression-free survival (PFS) or OS.³

In this report, we describe baseline characteristics and outcomes of patients with MM by sex using two large population-based datasets: the SEER dataset and the Multiple Myeloma Research Foundation (MMRF) CoMMpass dataset with clinical annotation.

We used MM data for the period 2000–2017 from SEER 18 that covers about 27·8% of the United States population. We identified MM patients (n = 78,351) using the third edition of the International Classification of Diseases for Oncology (ICD-O-3) code 9732. We also obtained data from the MMRF CoMMpass study (NCT01454297), a prospective observational study of newly diagnosed MM patients initiated in 2011. Data relevant to treatment outcomes of 1 143 MM patients were collected, including pre-treatment demographics, International Staging System (ISS), baseline MM parameters, cytogenetics, induction regimen, autologous stem cell transplant (ASCT) and maintenance therapy use, PFS and OS.

We defined OS as time from diagnosis until death from any cause. PFS was not available in SEER. In the MMRF data, PFS was defined as the time from the date of diagnosis until progression, death or the date of last follow-up. Survival curves were constructed using the Kaplan–Meier method and compared with the log-rank test. To examine the association between sex and outcomes, we first used Cox proportional hazards models to calculate age-adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) for PFS and OS. In all models, males served as the reference. Age at diagnosis as a continuous variable and categorical variable (<55, 55–69, 70+) generated similar findings. For SEER, basic multivariate models included age, race and diagnosis year (Table SI). For MMRF, basic multivariate models included age and established clinical prognostic factors including ISS, front-line ASCT and Eastern Cooperative Oncology Group (ECOG) performance status. High-risk cytogenetic abnormalities [defined as t(4;14), t(14;16), t(14;20), 1q gain and/or 17p deletion], gene expression profiling (UAMS70), estimated glomerular filtration rate (eGFR), treatment and race, were examined as potential covariates (Table SI), but were not found to change risk estimates by> 10%. The reported P values are two-sided. Statistical analyses were conducted in R v4·0.2 using the Survival package v3·1-12 (R Foundation for Statistical Computing, Vienna, Austria; https://cran.r-project.org).

There were more males than females in both datasets: 55·5% males vs 44·5% females in SEER and 60·4% males vs 39·6% females in MMRF data. Although there were no significant differences between males and females with respect to baseline MM clinical characteristics (age, race, performance status, ISS, presence of high-risk cytogenetic abnormalities and induction treatment), females received front-line ASCT at a higher frequency compared to males (54% vs 46%, P = 0·009). OS and PFS were significantly better for females than males (Fig 1). The OS benefit persisted for females (SEER: HR = 0·92, 95% CI 0·90–0·94, P < 0·0001; MMRF: HR = 0·66, 95% CI 0·51–0·85, P < 0·0001) but was attenuated for PFS (MMRF: HR = 0·85, 95% CI 0·72–1·00, P = 0·06) after controlling for age, ISS, performance status and ASCT (Table I and Table SI). Point estimates for OS by sex in the SEER cohort were similar for patients diagnosed between 2000–2010 and 2011–2017, the time period during which the MMRF cohort was enrolled.

Fig 1.

Kaplan–Meier survival curves for MM by sex. (A) Overall survival of SEER patients. (B) Overall survival of MMRF patients. (C) Progression-free survival of MMRF patients. SEER: National Cancer Institute’s Surveillance, Epidemiology, and End Results program. MMRF: Multiple Myeloma Research Foundation (MMRF) CoMMpass study.

Table I.

Hazard ratios (HRs) and 95% confidence intervals (95% CI) for overall survival and progression-free survival by sex.

	SEER*						MMRF*
	OS (HR, 95% CI)			OS (HR, 95% CI)			PFS (HR, 95% CI)
	Males	Females	P value	Males	Females	P value	Males	Females	P value
Model 1**	1·0 (ref)	0·93 (0·91–0·94)	<0·0001	1·0 (ref)	0·61 (0·47–0·78)	<0·0001	1·0 (ref)	0·82 (0·70–0·97)	0·02
Model 2†	1·0 (ref)	0·92 (0·90–0·94)	<0·0001	1·0 (ref)	0·66 (0·51–0·85)	<0·0001	1·0 (ref)	0·85 (0·72–1·00)	0·06
Model 3‡	1·0 (ref)	0·92 (0·90–0·93)	<0·0001	1·0 (ref)	0·64 (0·50–0·83)	<0·0001	1·0 (ref)	0·84 (0·71–0·99)	0·03
Model 4§	–	–	–	1·0 (ref)	0·66 (0·51–0·85)	<0·0001	1·0 (ref)	0·85 (0·72–1·01)	0·06

Abbreviations: OS, overall survival; HR, hazard ratio; CI, confidence interval; ASCT, autologous stem cell transplant; ECOG, Eastern Cooperative Oncology Group; eGFR, estimated glomerular filtration rate; HRCA, high-risk cytogenetic abnormality [t(4;14), t(14;16), t(14;20), 1q gain and/or 17p deletion]; IMiD, immunomodulatory imide drug; ISS, International Staging System; PI, proteasome inhibitor; UAMS70, 70-gene expression profile.

^*SEER: National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) Program. MMRF: Multiple Myeloma Research Foundation (MMRF) CoMMpass study.

^**SEER and MMRF: adjusted for age (<55, 55–69, 70+).

^†SEER: adjusted for age (<55, 55–69, 70+), race (whites/blacks/others) and diagnosis year. MMRF: adjusted for age (<55, 55–69, 70+), ISS (stage 1, 2 and 3), front-line ASCT (no vs yes) and ECOG performance status (0, 1, 2–4).

^‡SEER: adjusted for variables in Model 2 plus county-level household income. MMRF: adjusted for variables in model 2 plus front-line treatment (doublet, PI + IMid triplet, alkylator triplet, other), eGFR (>60, <=60) and UAMS70 (low risk, high risk).

^§MMRF: adjusted for variables in model 3 plus race and HRCA (0, 1, 2+).

These findings stand in contrast to previously published data from clinical trials that have suggested no difference in survival outcomes by sex. Though the afore-mentioned studies benefitted from more uniform treatment of patients under clinical trial conditions, they likely represent a subset of highly motivated and physically fit MM patients that could participate in such a trial.

That biologic sex plays a role in MM outcomes poses a vexing problem and requires further investigation as to the reasons why. Factors related to long-term disease control may explain better OS and PFS for women; chiefly, that they may have healthier attitudes and behaviours and are more likely to engage in health-promoting behaviours after intensive treatment.⁴ It is also possible that male patients have more comorbidities at the time of MM diagnosis than females, and these comorbid conditions may contribute to poorer MM survival.

There may be biologic underpinnings for sex differences in MM outcomes as well. Hormone-related pharmacokinetic variation, for example, has been demonstrated in lymphoma studies, where higher rituximab serum concentrations were associated with female sex and associated with remission quality of PFS.⁵

The cornerstones of anti-MM therapies involve immune-modifying drugs such as proteasome inhibitors, immunomodulatory imide drugs and corticosteroids. Thus, it is important to point out that there are clearly described sex differences in both innate and adaptive immune responses, including that males have higher frequencies of regulatory T cells (Treg) than their female counterparts.⁶ Increased Treg are associated with adverse clinical features and have an impact on MM progression.⁷ Depletion of Treg in the peri-ASCT setting of MM patients was shown to delay Treg recovery after ASCT and led to favourable outcomes in a pilot study of MM patients.⁸ Moreover, multiparametric flow cytometry of bone marrow samples from MM patients revealed that bone marrow Treg were lower in patients with long-term disease control than in symptomatic MM.⁹ This suggests that higher Treg frequencies in males may be responsible in part for inferior MM outcomes, though this requires further investigation.

Our findings have implications with respect to the clinical care of MM patients. Specifically, clinical trials in MM must ensure adequate representation of both sexes. If frailty and/or lack of social-support mechanisms are in fact part of the underlying factors for sex differences, then pretreatment and on-treatment comprehensive geriatric assessment can be undertaken to maximize patient function which may allow for higher-intensity treatment and improved MM survival.¹⁰ For a distinction among human beings as old as time, much remains to be uncovered regarding the impact of sex on MM outcomes.

Supplementary Material

Supplementary Table 1

Table S1. Characteristics of SEER and MMRF patients by sex.