Environmental and occupational risk factors associated with multiple myeloma: a multicenter, hospital-based, matched case-control study

Abstract

Introduction

Multiple myeloma (MM), a hematologic malignancy driven by neoplastic plasma cell proliferation, remains insufficiently characterized with respect to occupational and environmental risk factors and their impact on patients’ quality of life (QoL). This study explores modifiable exposures in the West Bank, Palestine, and evaluates their associations with MM risk and disease-specific QoL outcomes.

Methods

A multicenter, hospital-based case-control study was conducted between 2018 and 2025, including 227 MM patients and 176 matched controls. Matching was based on age, sex, hospital setting, and admission type. Occupational/environmental exposures including ionizing radiation, cosmetics-related agents, pesticides, organic solvents, and farming were assessed via structured interviews and chart reviews. MM diagnosis adhered to International Myeloma Working Group criteria. QoL was evaluated using the validated EORTC QLQ-MY20 instrument. Multivariable logistic and linear regression analyses were performed, adjusting for clinical confounders using LASSO selection.

Results

Cosmetics-related chemical exposure was independently associated with higher odds of MM (OR = 2.85; 95% CI: 1.56–5.21) and a mixed QoL profile. Specifically, it predicted increased disease symptoms (Coeff = 11.55; 95% CI: 2.82–20.28; p = 0.010), lower treatment side-effects scores (Coeff = -2.17; 95% CI: -8.57 to -0.23; p = 0.049), and a marked decline in future perspective (Coeff = -13.73; 95% CI: -22.88 to -4.58; p = 0.003). Pesticide exposure was significantly linked to lower disease symptom burden (Coeff = -3.77; 95% CI: -12.61 to -2.06; p = 0.041) and better future outlook (Coeff = 10.05; 95% CI: 0.77–19.34; p = 0.034). Meanwhile, organic solvent exposure (carcinogenic-organic compounds) was associated with a decline in future perspective (Coeff = -3.96; 95% CI: -5.70 to -2.62; p = 0.042).

Conclusion

This study highlights cosmetics-related agents, pesticides, and organic solvents as key modifiable risk factors for both MM development and QoL deterioration. Their significant physical and psychological impacts underscore the urgency of integrating preventive occupational health strategies with holistic myeloma care that addresses symptom burden and future outlook.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-24366-9.

Introduction

Multiple myeloma (MM) is a hematological lymphoid malignancy presenting as neoplastic plasma cell clonal proliferation in the bone marrow and is usually characterized by the overproduction of monoclonal immunoglobulin (M-protein), composed of heavy and/or light chains [1]. The excess monoclonal protein production occurs at the expense of normal polyclonal immunoglobulins, resulting in immunopanels and increased susceptibility to infections [2]. Disease may be asymptomatic but with laboratory abnormalities, fatigue, and weight loss [3], or more specific features may be present, which include those in the CRAB (hypercalcemia, renal insufficiency, anemia, and bone lesions) criteria [4, 5]. In Palestine, MM is relatively uncommon but increasing. The Palestinian National Cancer Registry reported 314 cases in 2022, accounting for 2% of all malignancies. In the Gaza Strip specifically, MM ranked as the 7th most reported cancer (4.5%) [8]. On the other hand, GLOBOCAN data estimated 96 new MM cases and 81 deaths in Palestine in 2022, with a 5-year prevalence of 4.7 per 100,000 population [6].

Despite the rising numbers, the etiology of MM remains poorly understood, with no definitive environmental or occupational risk factors established to date [7]. The exact cause remains unclear, and no specific lifestyle, occupational, or environmental risk factors have been firmly established. While several possible contributors have been suggested, research findings so far have been inconsistent [7, 8]. Few studies have explored multiple myeloma in Palestine [9]here is limited data on the disease’s epidemiological factors and management. This lack of understanding makes it difficult to fully comprehend how MM affects individuals, and it limits the knowledge available to physicians and policymakers working to prevent and manage the disease.

Many risk factors have been proposed to increase the risk of multiple myeloma, which can be broadly categorized into non-modifiable, occupational, and environmental factors. Non-modifiable risk factors such as increasing age [10, 11], male sex [12, 13], and Black ethnicity [14] have been consistently associated with higher disease risk [15]. In contrast, findings related to occupational and environmental exposures remain inconsistent. Pesticide exposure, for example, has shown variable associations with multiple myeloma risk [16–18]Similar uncertainty exists regarding agricultural occupations [19, 20]. Organic solvents have been more consistently linked to an elevated risk [21–23], while studies examining ionizing radiation exposure have yielded mixed results [24]. Likewise, the use of cosmetic agents such as hair dyes has been explored, but the evidence remains inconclusive [25, 26].

Large-scale epidemiologic studies and consortia have attempted to clarify these exposure–disease relationships in MM, with mixed results. For example, a pooled analysis by the International Multiple Myeloma Consortium (IMMC) found no significant association between cigarette smoking and myeloma risk [27]. Meanwhile, multi-center studies have observed only modest risk increases with certain exposures – notably farming occupations with prolonged pesticide use (odds ratios ~ 1.6–1.8) – and no clear excess risk from other suspected agents like organic solvents [22]. The International Agency for Research on Cancer (IARC) has classified chemicals such as benzene as carcinogenic to humans with a possible link to myeloma [28], and high-dose ionizing radiation is an established risk factor based on studies of exposed populations [29].

Palestine, in particular, presents a unique exposure profile: informal farming is common and pesticide use is largely unregulated [30]. Some communities are additionally exposed to industrial pollutants – for example, residents near certain industrial zones in the West Bank (such as in Tulkarm) experience unusually high rates of cancer and respiratory illnesses [31]. Combined with chronic political instability and limited public health resources, these factors contribute to underreporting of cancer cases and a lack of organized screening or prevention programs [32]. To address this gap,.In this multi-center, hospital-based case-control study, will investigate a spectrum of environmental and occupational exposures, including pesticide use, agricultural activities, ionizing radiation, organic solvent contact, and cosmetic agents, and study their potential associations with the development of MM. Additionally, we will explore how these risk factors will impact MM patients’ quality of life.

Methodology

Study design, setting, and MM diagnosis criteria

This is a retrospective 1:1 case-control study set in hospital clinics using data from Jenin, Al-Watani, Beit Jala, and Dura Governmental, An-Najah National University Hospital, and the Palestine Medical Complex, from January 2018 to December 2025. Multiple myeloma in patients was diagnosed according to the revised International Myeloma Working Group (IMWG) diagnostic criteria, which require histopathologic criteria to be met and 1 or more myeloma-defining events. Criteria for histopathology include either bone marrow biopsy presenting with ≥ 10% clonal bone marrow plasma cells or biopsy-proven bony or extramedullary plasmacytoma. Myeloma defining events include [1] end-organ damage due to the underlying plasma cell proliferative disorder, also known as the CRAB criteria (hypercalcemia, renal insufficiency, anemia and bone lesions) and [2] biomarkers of malignancy such as a percentage of clonal bone marrow plasma cells of ≥ 60%, serum free light chain involved to uninvolved ratio of ≥ 100 with the involved free light chain quantity being ≥ 100 mg/L, and, on MRI studies, > 1 focal lesions with each measuring 5 mm or more in size [33].

Study population

We conducted our study on 607 patients who were initially screened for eligibility, including 431 potential MM cases and 176 non-neoplastic controls. Among the MM cases, 150 deceased patients, 30 who declined participation, and 24 critically ill individuals were excluded, resulting in 227 eligible MM patients. Of these, 202 completed the quality-of-life assessment using the EORTC QLQ-MY20, after excluding an additional 25 patients (20 refused, 5 were critically ill). Controls were selected from the outpatient clinics of the same six hospitals and included patients receiving care for non-neoplastic conditions, including nephrological, cardiovascular, neurological, respiratory, rheumatological, orthopedic, gastrointestinal, and endocrine diseases. See Fig. 1.

Fig. 1.

Study flow diagram for case-control design and quality of life assessment

A total of 176 controls were included and matched 1:1 with 176 MM patients based on age (± 5 years), gender, inpatient/outpatient status, and hospital of treatment. Matching was performed manually during data collection by trained medical students and verified at the analysis stage to ensure comparability. Baseline characteristics of matching variables are presented in Table 1 to demonstrate post-matching balance. Matching variables were also included as covariates in subsequent multivariable regression models to control residual confounding. All participants were fully informed about the aim of the study. See Fig. 1 for the flowchart of participant selection.

Table 1.

Baseline sociodemographic, clinical, laboratory, and therapeutic characteristics of study participants (Cases and Controls)

Variable		Category	Frequency (%)
Demographic variables
Age, years (mean ± SD)			61.3226 ± 11.93328
Gender		Male	219(54.34%)
		Female	184(45.66%)
Type of residency		Village	213 (52.9%)
		City	177 (43.9%)
		Camp	13 (3.2%)
Weight (mean ± SD)			77.2491 ± 16.10066
Hight (mean ± SD)			1.6765 ± 0.09227
BMI (mean ± SD)			27.5207 ± 5.56040
Body habitat		Underweight	5 (1.2%)
		Normal weight	90 (22.3%)
		Overweight	181 (44.9%)
Smoking			78 (19.4%)
‘Incidental disease discovering			49 (12.2%)
Hospital sitting		Inpatient	15 (3.7%)
		Outpatient	388 (96.3%)
ABO Blood group		A+	62 (15.4%)
		A-	5 (1.2%)
		B+	16 (4.0%)
		B-	3 (0.7%)
		AB+	10 (2.5%)
		O+	49 (12.2%)
		O-	5 (1.2%)
History and Co-morbidities
Family history of MM			32 (7.9%)
Prior history of cancers			27 (6.7%)
COVID − 19			160 (39.7%)
Diabetes Mellitus			129 (32.0%)
HTN			167 (41.4%)
Osteoporosis			91 (22.6%)
Hepatomegaly			9 (2.2%)
History of Gastritis			49 (12.2%)
Anemia			103 (25.6%)
Hypothyroidism			10 (2.5%)
Hyperthyroidism			8 (2.0%)
Arthritis			50 (12.4%)
Acute pain			148 (36.7%)
Unspecific pain			125 (31.0%)
Neurological deficit			47 (11.7%)
Deformity			15 (3.7%)
Osteolytic lesion on x-ray			75 (18.6%)
Skin rash			29 (7.2%)
Autoimmune diseases			13 (3.2%)
Hepatitis C virus chronic infection			2 (0.5%)
AIDS			0 (0%)
Arrhythmia			13 (3.2%)
IHD			68 (16.9%)
CHF			12 (3.0%)
Cardiac arrest			4 (1.0%)
TIA			11 (2.7%)
Cerebrovascular disease			17 (4.2%)
Chronic lung disease			28 (6.9%)
Acute renal failure			26 (6.5%)
Peptic ulcer			30 (7.4%)
Neurological disease			12 (3.0%)
Peripheral vascular disease			9 (2.2%)
Gout			32 (7.9%)
Chronic kidney disease			39 (9.7%)
Dementia			15 (3.7%)
Lab value, (mean ± SD)
Hemoglobin			11.8746 ± 2.61628
Hematocrit			37.0836 ± 20.76559
WBC			7.9798 ± 3.96260
RBC			4.2959 ± 1.37300
Platelet count			237.4453 ± 95.71627
Plasma cells percentage in the bone marrow			45(10–75)
Kappa (KAP)			300.5(95.18- 775.07)
Lambda (LAM)			53(14.9-152.93)
Creatinine			1.3135 ± 1.62166
Beta2-Microglobluin			6.7667 ± 7.92043
ESR			68.0995 ± 50.51474
Albumin level			3.7530 ± 0.75372
Total protein			8.7112 ± 9.34810
Calcium			9.2958 ± 1.30462
Uric acid			6.3231 ± 2.27756
Therapeutic Option
Bortezomab			42 (10.4%)
Cyclophosphamide			11 (2.7%)
Dexamethasone			92 (22.8%)
VCD protocol			21 (5.2%)
Undergo homologues bone marrow transplant			70 (17.4%)
Symptom		Category	Frequency (%)
Weight loss			146 (36.2%)
Abdominal pain			138 (34.2%)
Headache and dizziness			147 (36.5%)
Joint pain			212 (52.6%)
Chest pain			145 (36.0%)
Bone pain			177 (43.9%)
Lower back pain			229 (56.8%)
Lower limb edema			153 (38.0%)
Constitutional symptoms			183 (45.4%)
Pathological fracture			71 (17.6%)
Change in bowel habits		None	294 (73.0%)
		Diarrhea	93 (23.1%)
		Constipation	16 (4.0%)

Variables

For both multiple myeloma cases and their matched controls, study variables are presented in terms of demographic, clinical, behavioral, occupational, environmental, laboratory, and treatment-related domains. Data were collected either from patients’ medical records or directly through patient interviews. Demographic variables included age, height, weight, body mass index (BMI), gender, and address of residence via electronic medical records, with residence categorized as village, camp, or city. Clinical variables covered a wide range of symptoms, signs, diagnoses, and comorbidities. We recorded symptoms such as weight loss, abdominal or epigastric pain, headache, dizziness, chest pain, bone pain, lower back pain, acute pain, joint pain (arthralgia), constitutional symptoms (e.g., fever, fatigue, cough), and bowel disturbances (constipation or diarrhea), all based on patient-reported history. Concurrent signs and diagnoses, including hepatomegaly, gastritis, anemia, hypothyroidism, hyperthyroidism, arthritis, swelling, deformity, neurological deficits, pathological fractures, and skin rash, were extracted from medical records and confirmed by history when applicable. Medical history and comorbidities included autoimmune diseases, hepatitis C virus infection, AIDS, and other chronic conditions such as diabetes mellitus, hypertension, chronic kidney disease, dementia, cerebrovascular disease, chronic lung disease, peptic ulcer disease, gout, acute renal failure (requiring one episode of dialysis), arrhythmia, ischemic heart disease (IHD), congestive heart failure (CHF), cardiac arrest, transient ischemic attack (TIA), early-stage heart failure (non-congestive), and post-surgical acute myocardial infarction (MI) and family history of multiple myeloma. All were confirmed through medical records and interviews. Occupational and environmental variables were considered positive when direct handling of an agent or when the job or residence of the patient had a clear presumption of exposure. These included pesticide exposure, ionizing radiation exposure, and organic solvents handling detailed definitions of these exposures are available in Supplement 1.

Laboratory variables included blood group, Hemoglobin A1c, hematocrit, white blood cell count, red blood cell count, platelet count, erythrocyte sedimentation rate, albumin, uric acid, total protein, and serum calcium. For cases, these values were obtained at the time of diagnosis; for controls, the most recent results were used if available. Additionally, we collected disease-specific markers, including beta-2 microglobulin, creatinine, bone marrow cellularity (plasma cell percentage), and kappa and lambda light chains.

Variables related to treatment and management included the chemotherapy regimen known as the VCD protocol, including Bortezomib (marketed as Velcade), Cyclophosphamide, and Dexamethasone, and history of autologous stem cell transplantation (ASCT).In the control group, the primary diagnosis of the underlying disease was recorded and categorized by system involvement (e.g., renal, cardiovascular, gastrointestinal, neurological, musculoskeletal).

Study outcomes

The primary outcome was the presence of MM concerning pre-defined occupational and environmental exposures, specifically: ionizing radiation, pesticide use, agricultural occupation, carcinogenic organic solvent exposure, and cosmetics-related chemical exposure (e.g., hair dye or salon product use). These exposures were assessed using structured interviews and medical records and were analyzed as independent risk factors for MM through multivariable logistic regression models. Confounding variables were identified a priori and adjusted for using a Least Absolute Shrinkage and Selection Operator (LASSO)-based variable selection method to ensure robustness and minimize overfitting. The secondary outcome focused on the health-related quality of life (QoL) among MM patients, measured via the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire for Multiple Myeloma (EORTC QLQ-MY20). The instrument captures four validated subdomains: Disease Symptoms (DS), Side Effects of Treatment (SE), Body Image (BI), and Future Perspective (FP). Multivariable linear regression models were constructed for each QoL domain to examine the independent associations of each exposure with patient-reported outcomes. These models were carefully adjusted for relevant clinical covariates and symptomatic predictors specific to each subdomain, thereby providing a nuanced assessment of how different exposures not only contribute to disease risk but also shape the lived experience of MM patients.

Data collection procedure

A dual-method approach was implemented by combining electronic medical record (EMR) review and structured patient interviews to ensure comprehensive data collection. Hospital EMR systems from six Palestinian medical centers were reviewed for cases identified between 2018 and 2025, with multiple myeloma diagnoses confirmed using International Myeloma Working Group (IMWG) standards.

Instruments

Interview assessment of risk factors

A structured face-to-face interview was employed as a primary data collection method to assess exposure to selected environmental and occupational risk factors associated with multiple myeloma patients and their matched controls. Interviews were conducted by trained medical students under the direct supervision of a hematologist, using standardized and validated questions adapted from prior peer-reviewed studies. The process was performed in participants’ native language to ensure clarity and reliability of responses, with both verbal and written informed consent obtained before participation. The interview covered five major risk domains: exposure to ionizing radiation, cosmetics and hair products, organic solvents, pesticides, and agricultural occupations. see the supplementary file titled “Interview Questions” to see the questions in Arabic and English.

EORTC QLQ-MY20 questionnaire

The European Organization for Research and Treatment of Cancer Quality of Life Questionnaire–Multiple Myeloma Module (EORTC QLQ-MY20) is a disease-specific instrument developed to assess health-related quality of life (HRQoL) in patients with multiple myeloma. The QLQ-MY20 focuses on issues pertinent to multiple myeloma patients, including disease symptoms, side effects of treatment, body image, and future perspective. The QLQ-MY20 comprises 20 items divided into four domains: Disease Symptoms (DS; 6 items), Side Effects of Treatment (SE; 10 items), Future Perspective (FP; 3 items), and Body Image (BI; 1 item). All items are scored using a 4-point Likert scale ranging from 1 (“Not at all”) to 4 (“Very much”), based on the patient’s experience during the past week. Scores are linearly transformed to a scale ranging from 0 to 100. Higher scores in the DS and SE domains indicate a greater symptom burden, while higher scores in the FP and BI domains reflect better functioning.

The official Arabic (Lebanon) version of the EORTC QLQ-MY20 was used in this study [2]. It has been culturally adapted and validated in previous studies, showing good reliability with Cronbach’s alpha values ranging from 0.74 to 0.90 [3]. The questionnaire results were further analyzed about environmental and occupational exposures to evaluate their differential impact on the four quality of life domains: Disease Symptoms (DS), Side Effects of Treatment (SE), Body Image (BI), and Future Perspective (FP).

Scoring method

Each of the 20 items in the questionnaire is scored using a 4-point Likert scale ranging from 1 (“Not at all”) to 4 (“Very much”), based on the patient’s experiences during the past week. The QLQ-MY20 includes three multi-item scales and one single-item scale, grouped as: Disease Symptoms (DS), which includes 6 items (items 31–36); Side Effects of Treatment (SE), comprising 10 items (items 37–46); Future Perspective (FP), covering 3 items (items 48–50); and Body Image (BI), assessed by a single item (item 47).

Rationale for exposure selection

We identified key occupational and environmental exposures including ionizing radiation, organic solvents (e.g., benzene), pesticides, and hair dye chemicals as candidate risk factors for multiple myeloma based on mechanistic and epidemiologic evidence. Ionizing radiation is an established carcinogen (Group 1, IARC) with known DNA-damaging effects, and high-dose exposures as well as prolonged low-dose exposures have been linked to increased multiple myeloma incidence [34]. Likewise, benzene and other organic solvents, which exhibit genotoxic and immunotoxic properties in bone marrow – are recognized by IARC as carcinogenic to humans (Group 1) with a positive association to multiple myeloma [35]. Meta-analyses of occupational cohorts provide further support, showing significantly elevated myeloma risks among benzene-exposed workers [36]. Chronic agricultural pesticide exposure has also been implicated: farmers and applicators experience higher myeloma incidence (pooled odds ratio ~ 1.4–1.5) [20], and a 2-fold excess of monoclonal gammopathy of undetermined significance (a premalignant precursor of myeloma) has been observed in pesticide applicators [37, 38]. Finally, hair dyes and related cosmetic chemicals particularly in occupational settings – have been examined given the mutagenic aromatic amines in some formulations [39]. A meta-analysis reported that hairdressers (with repeated dye and chemical exposures) have a significantly elevated risk of multiple myeloma (~ 1.5-fold higher than unexposed) [38]. Accordingly, IARC classifies occupational exposure as a hairdresser as “probably carcinogenic” to humans [39]. Although studies of personal hair-dye use in the general population have shown mixed results, some have noted increased myeloma (and related lymphoma) risk with long-term use of permanent dyes [39]. In light of these data, we included the above exposures as risk factors in our analysis, prioritizing recent high-quality epidemiological studies (cohort and case–control) and authoritative evaluations (e.g. IARC monographs) to ensure robust evidence. These exposures are also relevant in our region, given the historical use of agricultural pesticides, organic solvents, and cosmetic products in Palestine and the Middle East.

Exposure data were collected through structured, interviewer-administered questionnaires conducted by trained medical students. Participants were asked about lifetime exposure to occupational and environmental agents, including ionizing radiation, pesticides, organic solvents, and cosmetic agents. Exposures were classified dichotomously (Yes/No) based on self-reported contact or use, without quantification of intensity or duration. No environmental measurements or occupational records were available for validation, and exposure classification relied entirely on participants’ recall, which may introduce misclassification and recall bias. While efforts were made to standardize interviews and reduce information bias, the lack of biomarker or registry confirmation is a recognized limitation of the study.

Ethical considerations

Ethical approval was obtained from the Institutional Review Board (IRB) before conducting this study. Informed consent was obtained from all patients, ensuring each was given the detailed information necessary to make an informed decision on their participation, with aims, procedures, potential risks, and benefits being known and voluntary participation with the right to withdraw at any time assured. Confidentiality and patient privacy were of utmost concern, with access to data restricted to the research team.

Sample size calculation

The sample size for this hospital-based matched case–control study was estimated using standard formulas for detecting associations between exposure and disease in unmatched studies, assuming an equal 1:1 case-to-control ratio. Based on previous literature and population data, the expected proportion of exposure to the selected environmental or occupational risk factor (e.g., pesticide use) among controls (P₂) was estimated at 20% (0.20) [40]. To detect an odds ratio (OR) of 2.0, a moderate and clinically meaningful effect size, with a two-sided alpha of 0.05 and power of 80%, the corresponding proportion of exposure among cases (P₁) was calculated using the formula:

Solving this equation yielded an expected P₁ of 33.3% (0.333). Based on these parameters and the Fleiss formula with continuity correction [41], the required sample size was 172 cases and 172 controls. This estimation followed the standard formula for unmatched case-control studies and was further validated using OpenEpi (version 3.01). The final sample included 227 cases and 176 controls, ensuring sufficient power for subgroup and adjusted analyses. Our study included 227 cases and 176 matched controls, exceeding this threshold and providing adequate statistical power to detect meaningful associations between the specified exposures and the risk of developing multiple myeloma. See Supplement 1 file for further details.

Statical analysis

We conducted a matched case–control analysis to evaluate environmental and occupational risk factors for multiple myeloma (MM). Fisher’s exact test was used in place of the chi-square when any expected cell count was < 5, to ensure valid p-values. These univariate tests were chosen for their simplicity in assessing group differences: the independent t-test compares mean values under the assumption of independent samples and normality, and the chi-square test (or Fisher’s exact) assesses differences in proportions or frequencies between two independent groups. All significance tests were two-tailed, and a nominal p < 0.05 was considered statistically significant.

We employed multiple strategies to control for confounding at both the design and analysis stages. First, we identified a priori a set of clinically significant covariates based on prior literature and expert knowledge [30, 42, 43]. Subsequently, we screened variables through univariate analyses, retaining any variables exhibiting at least a marginal statistical association (p < 0.10). Moreover, we utilized individual matching at the study design stage, matching each case with a control on critical demographic and clinical factors (age ± 5 years, gender, hospital site, inpatient/outpatient status), thus inherently controlling for these variables. Finally, we applied Least Absolute Shrinkage and Selection Operator (LASSO) regression with 10-fold cross-validation, an advanced variable selection method designed to minimize multicollinearity, enhance predictive accuracy, and retain only the most informative covariates. The combination of these steps yielded a refined set of confounders integrated into the final multivariable logistic regression models. For the primary risk factor analysis, we developed five distinct multivariable logistic regression models, each focusing on one of the main occupational or environmental exposures hypothesized to be associated with the risk of developing multiple myeloma. Model development adhered to the Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) guidelines [44]. In each model, the exposure of interest was treated as the main independent variable, while controlling for a consistent set of confounders identified through a structured, multi-phase selection strategy. Specifically, Model 1 evaluated the association between ionizing radiation exposure and MM risk. Model 2 focused on exposure to cosmetics-related chemicals. Model 3 analyzed the effect of organic solvent exposure. Model 4 assessed the relationship between pesticide exposure and MM risk. Finally, Model 5 examined the impact of working in agricultural settings. All models were adjusted for relevant sociodemographic and clinical covariates, including age, gender, BMI, type of community, hospital setting, autoimmune disease history, hypertension, dexamethasone use, and either albumin level or diabetes mellitus, depending on the exposure. Variables such as MM Family and Prior history of cancers showed perfect prediction (i.e., present only among cases) and were therefore excluded from multivariable logistic regression models. Fisher’s exact test was used for group comparisons in these cases due to zero cell counts in controls. Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were computed to estimate the independent effect of each exposure, and Wald chi-square tests were used to evaluate statistical significance. To ensure model reliability, we conducted diagnostic checks, including assessment of multicollinearity using variance inflation factor (VIF), goodness-of-fit evaluation via the Hosmer–Lemeshow test, and discrimination performance through receiver operating characteristic (ROC) curve analysis.

Among the MM cases, we analyzed health-related quality of life outcomes as secondary endpoints. Quality of life was measured using the EORTC QLQ-MY20 questionnaire, which yields scores in four domains: Disease Symptoms (DS), Side Effects of Treatment (SE), Future Perspectives (FP), and Body Image (BI). We first summarized the QLQ-MY20 domain scores for the MM patients, reporting mean ± standard deviation (SD) for approximately normally distributed domains Given the adequate sample size (n = 202), and consistent with the Central Limit Theorem [45, 46], we assumed approximate normality of the QOL domain scores.

To compare quality-of-life across different subgroups of MM patients, we used appropriate univariate tests. For example, an independent t-test was used to compare mean domain scores between two groups (e.g., patients with a certain exposure vs. those without. These tests were chosen to detect any significant differences in quality-of-life domain scores across patient subgroups.

For multivariable analysis of quality of life, we treated each QLQ-MY20 domain score as a continuous outcome and fitted separate multivariable linear regression models to identify factors associated with poorer or better quality of life. Each linear regression model included the exposure variables of interest (e.g., specific occupational/environmental exposures) and relevant covariates (demographic or clinical factors) as independent variables, with the domain score as the dependent variable. Each QOL domain has five models, where each risk factor had one full model, then we reported an adjusted beta coefficients with 95% CIs, representing the mean difference in the QOL domain score associated with each predictor, adjusted for other variables. In this way, we assessed the independent impact of environmental and occupational factors on multiple myeloma patients’ quality of life across the four domains. All final linear models incorporated the same set of confounders identified through our multistep selection process (described above), ensuring consistency in adjustment. The models were constructed as follows: the five DS models were adjusted for incidentally diagnosed cases, arthritis, cardiac arrest, neurological disorders, weight loss, chest pain, bone pain, and lower back pain. The SE models were adjusted for osteoporosis, autoimmune disease, peptic ulcer, gout, and a range of pain and constitutional symptoms. The BI models included adjustments for cerebrovascular disease, CHF, serum calcium, and neurological and gastrointestinal symptoms, while the FP models were adjusted for comorbidities such as osteoporosis, arthritis, autoimmune disease, cardiac arrest, and chest or bone pain. Regression diagnostics were performed for each model, including checks for residual normality and homoscedasticity. When necessary, robust standard errors were used. A significance level was set at p < 0.05 for all analyses. All statistical analyses were conducted using Stata version 17 (StataCorp LLC, College Station, TX, USA).

Results

Data presented frequency (%) or SD (Standard Deviation), or median (Q1-Q3). BMI (Body Mass Index), ABO (Blood Group System), MM (Multiple Myeloma), COVID-19 (Coronavirus Disease 2019), HTN (Hypertension), AIDS (acquired immunodeficiency syndrome), IHD (ischemic heart disease), CHF (Congestive Heart Failure), Acute MI (Acute Myocardial Infarction), TIA (Transient Ischemic Attack), WBCs (White Blood Cells), RBCs (Red Blood Cells), ESR (Erythrocyte Sedimentation Rate), VCD drug protocol (Bortezomib, Cyclophosphamide, Dexamethasone)

Table 1 presents the comprehensive baseline characteristics of all study participants, including both patients diagnosed with multiple myeloma (cases) and their matched controls. Variables are categorized into sociodemographic data, comorbidities, laboratory profiles, treatment regimens, and symptomatology. The mean age of participants was 61.3 years (± 11.9), with a slight male predominance (54.3%). Residency distribution was primarily rural (52.9%) and urban (43.9%), with a small proportion from refugee camps (3.2%). Anthropometric measures revealed a mean BMI of 27.5 kg/m², with a predominance of overweight individuals (44.9%). Current smokers comprised 19.4% of the cohort. Regarding comorbidities, hypertension (41.4%), diabetes mellitus (32%), and osteoporosis (22.6%) were the most prevalent. Other notable conditions included anemia (25.6%), arthritis (12.4%), and COVID-19 history (39.7%). A family history of MM was reported in 7.9% of participants. Laboratory findings showed a mean hemoglobin level of 11.87 g/dL, with elevated inflammatory markers such as ESR (68.1 ± 50.5 mm/hr). Bone marrow plasma cell infiltration median was 45. Renal function, protein levels, and other biochemical markers are reported to provide clinical context. Therapeutically, a subset of cases received agents such as bortezomib (10.4%), cyclophosphamide (2.7%), and dexamethasone (22.8%), with 17.4% undergoing autologous bone marrow transplantation. Symptomatically, back pain (56.8%), joint pain (52.6%), and constitutional symptoms (45.4%) were frequently reported.

Table 2 presents the distribution of key occupational and environmental exposure variables among the entire study population. These factors including prior exposure to radiation, cosmetic agents (e.g., hair dyes), carcinogenic organic solvents, pesticides, and agricultural environments were selected based on their hypothesized association with multiple myeloma (MM) development. Each variable was captured through structured participant interviews and medical chart reviews. The frequency and percentage of exposure reflect their baseline prevalence in the study cohort before group comparisons. These exposures were later analyzed in relation to MM diagnosis (case vs. control) and subsequently evaluated for their potential association with disease-specific quality of life using the EORTC QLQ-MY20 instrument. This dual analytic approach aligns with the study’s central aim of identifying candidate risk factors and characterizing their impact on patient-reported outcomes in MM.

Table 2.

Prevalence of occupational and environmental exposure risk factors among study participants

Variable	Frequency (%)
Occupational and environmental risk factors
History of radiation exposure	118 (29.3%)
History of cosmetologist	76 (18.9%)
Carcinogenic-organic exposure	65 (16.1%)
Pesticide exposure	82 (20.3%)
Agriculture	114 (28.3%)

Figure 2 is a bar chart comparing the prevalence of five occupational and environmental exposures between patients diagnosed with multiple myeloma (MM) and their matched non-neoplastic controls. The exposures include agricultural work, cosmetologist-related chemical exposure, organic solvents, pesticide exposure, and prior radiation exposure. Notably, MM patients exhibited a higher prevalence of exposure across all categories, particularly for radiation (36.6% vs. 19.9%) and agricultural work (33.0% vs. 22.2%). Cosmetologist chemical exposure (24.2% vs. 11.9%) and organic solvents (21.6% vs. 9.1%) also showed substantial differences, indicating potential associations with MM development. These findings support the hypothesis that certain occupational and environmental exposures may serve as risk factors for MM and warrant further investigation in multivariable models.

Fig. 2.

Prevalence of occupational and environmental exposures among multiple myeloma cases and controls

Table 3 provides a detailed comparative analysis of sociodemographic profiles, clinical histories, laboratory parameters, exposure histories, and treatment characteristics between patients diagnosed with multiple myeloma (MM) and their matched non-neoplastic controls. Significant differences were observed in multiple domains, including body mass index (BMI) (29.0 ± 4.5 kg/m² cases vs. 26.5 ± 4.5 kg/m² controls; p < 0.001), and incidental disease discovery (21.6% vs. 0%; p < 0.001). Clinical comorbidities were significantly more frequent among MM cases, including osteoporosis (36.1% vs. 5.1%; p < 0.001), hypertension (37% vs. 21.6%; p = 0.040), anemia (37.9% vs. 9.7%; p < 0.001), arthritis (17.2% vs. 6.2%; p = 0.001), gastritis (15.4% vs. 8%; p = 0.023), and inflammatory bowel disease (2.6% vs. 0%; p = 0.030). Symptoms significantly associated with MM included lower back pain (75.3% vs. 33%; p < 0.001), bone pain (66.5% vs. 14.8%; p < 0.001), joint pain (68.7% vs. 31.8%; p < 0.001), and pathological fractures (29.5% vs. 2.3%; p < 0.001). Laboratory findings indicated significant differences in mean hemoglobin levels (10.8 vs. 13.2 g/dL; p < 0.001), erythrocyte sedimentation rate (ESR) (87.2 vs. 33.4 mm/hr; p < 0.001), and serum calcium (9.56 vs. 8.80 mg/dL; p < 0.001), reflecting characteristic laboratory abnormalities associated with MM. Therapeutically, exclusive use of targeted treatments such as bortezomib (18.5%), cyclophosphamide (4.8%), the VCD protocol (9.3%).

Table 3.

Comparative analysis of sociodemographic, clinical, laboratory, and exposure characteristics between multiple myeloma cases and matched controls

Variable	Subgroup	Group						P value
		Control(n = 176)			Disease(n = 227)
Demographic variables
Age (mean ± SD)		60.2557 ± 12.06778			62.1498 ± 11.78812			0.114a
Gender	Male	95 (54%)			124 (54.6%)			0.897b
	Female	81 (46%)			103 (45.4%)
Type of residency	Village	81 (46%)			132 (58.2%)			0.050c
	City	88 (50%)			89 (39.2%)
	Camp	7 (4%)			6 (2.6%)
BMI (mean ± SD)		26.540 ± 4.516			29.043 ± 4.458			< 0.001a
Body habitat	Underweight	2 (1.9%)			3 (1.8%)			< 0.001c
	Normal weight	21 (19.4%)			69 (41.1%)
	Overweight	85 (78.7%)			96 (57.1%)
	No	176 (100%)			149 (35.6%)
Incidental disease discovering	Yes	0 (0%)			49 (21.6%)			< 0.001b
	No	176 (100%)			178 (78.4%)
Hospital sitting	Inpatient	3 (98.3%)			12 (5.3%)			0.060b
	Outpatient	173 (1.7%)			215 (94.7%)
Blood group	A+	28 (46.7%)			34 (37.8%)			0.312c
	A-	1 (1.7%)			4 (4.4%)
	B+	9 (15%)			7 (7.8%)
	B-	2 (3.3%)			1 (1.1%)
	AB+	4 (6.7%)			6 (6.7%)
	O+	14 (23.3%)			35 (38.9%)
	O-	2 (3.3%)			3 (3.3%)
History and Co-morbidities
Family history of MM	Yes	0 (0%)			32 (14.1%)			< 0.001b
	No	176 (100%)			195 (85.9%)
Prior history of cancers	Yes	0 (0%)			27 (11.9%)			< 0.001b
	No	176 (100%)			200 (88.1%)
COVID − 19	Yes	60 (34.1%)			100 (44.1%)			0.043b
	No	116 (65.9%)			127 (55.9%)
Diabetes mellitus	Yes	60 (34.1%)			69 (30.4%)			0.430b
	No	116 (65.9%)			158 (69.6%)
HTN	Yes	83 (21.6%)			84 (37%)			0.040b
	No	93 (78.4%)			143 (63%)
Inflammatory bowel disease	Yes	0 (0%)			6 (2.6%)			0.030b
	No	176 (100%)			221 (97.4%)
Osteoporosis	Yes	9 (5.1%)			82 (36.1%)			< 0.001b
	No	167 (94.9%)			145 (63.9%)
Hepatomegaly	Yes	0 (0%)			9 (4%)			0.008b
	No	176 (100%)			218 (96%)
Gastritis	Yes	14 (8%)			35 (15.4%)			0.023b
	No	162 (92%)			192 (84.6%)
Anemia	Yes	17 (9.7%)			86 (37.9%)			< 0.001b
	No	159 (90.3%)			141 (62.1%)
Hypothyroidism	Yes	4 (2.3%)			6 (2.6%)			0.813b
	No	172 (97.7%)			221 (97.4%)
Hyperthyroidism	Yes	5 (2.8%)			3 (1.3%)			0.278b
	No	171 (97.2%)			224 (98.7%)
Arthritis	Yes	11 (6.2%)			39 (17.2%)			0.001b
	No	165 (93.8%)			188 (82.8%)
Acute pain	Yes	71 (40.3%)			77 (33.9%)			0.185b
	No	105 (59.7%)			150 (66.1%)
Unspecific pain	Yes	59 (33.5%)			66 (29.1%)			0.338b
	No	117 (66.5%)			161 (70.9%)
Neurological deficit	Yes	21 (11.9%)			26 (11.5%)			0.882
	No	155 (88.1%)			201 (88.5%)
Deformity	Yes	5 (2.8%)			10 (4.4%)			0.411b
	No	171 (97.2%)			217 (95.6%)
Osteolytic lesion on x-ray	Yes	0 (0%)			75 (33%)			< 0.001b
	No	176 (100%)			152 (67%)
Skin rash	Yes	15 (8.5%)			14 (6.2%)			0.364b
	No	161 (91.5%)			213 (93.8%)
Autoimmune diseases	Yes	6 (3.4%)			7 (3.1%)			0.855b
	No	170 (96.6%)			220 (96.9%)
Hepatitis C virus chronic infection	Yes	2 (1.1%)			0 (0%)			0.107b
	No	174 (98.9%)			227 (100%)
AIDS	Yes	0 (0%)			0 (0%)			N/A
	No	176 (100%)			227 (100%)
Arrhythmia	Yes	6 (3.4%)			7 (3.1%)			0.855b
	No	170 (96.6%)			220 (96.9%)
IHD	Yes	32 (18.2%)			36 (15.9%)			0.537b
	No	144 (81.8%)			191 (84.1%)
CHF	Yes	0 (0%)			12 (5.3%)			0.002b
	No	176 (100%)			215 (94.7%)
Cardiac arrest	Yes	0 (0%)			4 (1.8%)			0.077b
	No	176 (100%)			223 (98.2%)
Acute MI – Post Surgery	Yes	0 (0%)			5 (12.2%)			0.048b
	No	176 (100%)			222 (97.8%)
TIA	Yes	7 (4%)			4 (1.8%)			0.176b
	No	169 (96%)			223 (98.2%)
Cerebrovascular disease	Yes	10 (5.7%)			7 (3.1%)			0.198b
	No	166 (94.3%)			220 (96.9%)
Chronic lung disease	Yes	12 (6.8%)			16 (2.6%)			0.928b
	No	164 (93.2%)			211 (97.4%)
Acute renal failure	Yes	7 (4%)			19 (8.4%)			0.075b
	No	169 (96%)			208 (91.6%)
Peptic ulcer	Yes	9 (5.1%)			21 (9.3%)			0.117b
	No	167 (94.9%)			206 (90.7%)
Neurological disease	Yes	6 (3.4%)			6 (2.6%)			0.654b
	No	170 (96.6%)			221 (97.4%)
Peripheral vascular disease	Yes	2 (1.1%)			7 (3.1%)			0.189b
	No	174 (98.9%)			220 (96.9%)
Gout	Yes	9 (5.1%)			23 (10.1%)			0.065b
	No	167 (94.9%)			204 (89.9%)
Chronic kidney disease	Yes	17 (9.7%)			22 (9.7%)			0.991b
	No	159 (90.3%)			205 (90.3%)
Dementia	Yes	3 (1.7%)			12 (5.3%)			0.060b
	No	173 (98.3%)			215 (94.7%)
Symptoms of Multiple Myeloma
Weight loss	Yes	18 (10.2%)			128 (56.4%)			< 0.001b
	No	158 (89.8%)			99 (43.6%)
Abdominal pain	Yes	37 (21%)			101 (44.5%)			< 0.001b
	No	139 (79%)			126 (55.5%)
Headache and dizziness	Yes	45 (25.6%)			102 (44.9%)			< 0.001b
	No	131 (74.4%)			125 (55.1%)
Joint pain	Yes	56 (31.8%)			156 (68.7%)			< 0.001b
	No	120 (68.2%)			71 (31.3%)
Chest pain	Yes	52 (29.5%)			93 (41%)			0.018b
	No	124 (70.5%)			134 (59%)
Bone pain	Yes	26 (14.8%)			151 (66.5%)			< 0.001b
	No	150 (85.2%)			76 (33.5%)
Lower back pain	Yes	58 (33%)			171 (75.3%)			< 0.001b
	No	118 (67%)			56 (24.7%)
Lower limb edema	Yes	50 (28.4%)			103 (45.4%)			0.001b
	No	126 (71.6%)			124 (54.6%)
Constitutional symptoms	Yes	130 (73.9%)			90 (39.6%)			< 0.001b
	No	46 (26.1%)			137 (60.4%)
Pathological fracture	Yes	4 (2.3%)			67 (29.5%)			< 0.001b
	No	172 (97.7%)			160 (70.5%)
Change in bowel habits	NO	176 (100%)			118 (52%)		< 0.001c
	Diarrhea	0 (0%)			93 (41%)
	Constipation	0 (0%)			16 (7%)
Lab value, (mean ± SD)
Hg			13.1813 ± 2.22381		10.8161 ± 2.43014	< 0.001a
Hematocrit			39.8279 ± 6.59480		34.8430 ± 27.18120	0.043a
WBC			8.7393 ± 3.33485		7.3597 ± 4.32158	0.003a
RBC			4.7934 ± 1.19348		3.8929 ± 1.38009	< 0.001a
Platelet count			241.8085 ± 84.90417		233.8829 ± 103.84463	0.486a
Creatinine			1.1446 ± 1.03001		1.4460 ± 1.95842	0.122a
ESR			33.4194 ± 34.28312		87.1603 ± 47.8630	< 0.001a
Albumin level			3.7889 ± 0.80046		3.7355 ± 0.73203	0.621a
Total protein			6.6503 ± 0.83107		9.2087 ± 10.35289	0.147a
Calcium			8.8007 ± 0.86281		9.5626 ± 1.42189	< 0.001a
Uric acid			5.9397 ± 2.04564		6.5274 ± 2.37388	0.075a
Therapeutic Option
Bortezomab	Yes		0 (0%)		42 (18.5%)	< 0.001b
	No		176 (100%)		185 (81.5%)
Cyclophosphamide	Yes		0 (0%)		11 (4.8%)	0.003b
	No		176 (100%)		216 (95.2%)
Dexamethasone	Yes		45 (25.6%)		47 (20.7%)	0.249b
	No		131 (74.4%)		180 (79.3%)
VCD protocol	Yes		0 (0%)		21 (9.3%)	< 0.001b
	No		176 (100%)		206 (90.7%)
	No		176 (100%)		157 (69.2%)

^(a) is tested by T- test, ^(b) is tested by Chi-square test, ^(c) is tested by Fisher’s exact test. Odds ratio (Confidence Interval), N/A (Not Applicable), SD (Standard Deviation), BMI (Body Mass Index), MM (Multiple Myeloma), COVID-19 (Coronavirus Disease 2019), HTN (Hypertension), AIDS (acquired immunodeficiency syndrome), IHD (ischemic heart disease), CHF (Congestive Heart Failure), MI (Myocardial Infarction), TIA (Transient Ischemic Attack), Hg (Hemoglobin), WBCs (White Blood Cells), RBCs (Red Blood Cells), VCD drug protocol (Bortezomib Cyclophosphamide Dexamethasone)

Table 4 presents the unadjusted case-control analysis investigating the associations between specific occupational and environmental exposures and the risk of multiple myeloma (MM) development. Significant associations were observed across all evaluated exposures, highlighting radiation exposure (OR = 2.322; 95% CI, 1.468–3.672; p < 0.001), Cosmetics agents exposure (OR = 2.360; 95% CI, 1.365–4.081; p = 0.002), carcinogenic-organic solvents (OR = 2.753; 95% CI, 1.506–5.033; p = 0.001), pesticide use (OR = 1.765; 95% CI, 1.060–2.939; p = 0.028), and agricultural work (OR = 1.733; 95% CI, 1.105–2.720; p = 0.016). These findings identify potential modifiable risk factors linked to MM etiology and development, suggesting that occupational and environmental exposures may play a clinically meaningful role in disease pathogenesis. The identified associations warrant further validation through adjusted multivariate analyses to precisely quantify their independent contributions and evaluate implications for preventive and public health strategies.

Table 4.

Unadjusted odds ratios for occupational and environmental risk factors associated with multiple myeloma

Variable		Subgroup	Group			Odds ratio (CI)	P value
			Control(n = 176)		Disease(n = 227)
Occupational and environmental risk factors
History of Radiation Exposure		Yes	35 (19.9%)		83 (36.6%)	2.322 (1.468–3.672)	< 0.001b
		No	141 (80.1%)		144 (63.4%)
History of Cosmetics agents exposure		Yes	21 (11.9%)		55 (24.2%)	2.360 (1.365–4.081)	0.002a
		No	155 (88.1%)		172 (75.8%)
Carcinogenic- organic exposure		Yes	16 (9.1%)		49 (21.6%)	2.753 (1.506–5.033)	0.001a
		No	160 (90.9%)		178 (78.4%)
Pesticide exposure		Yes	27 (15.3%)		55 (24.2%)	1.765 (1.060–2.939)	0.028b
		No	149 (84.7%)		172 (75.8%)
Agriculture		Yes	39 (22.2%)		75 (33%)	1.733 (1.105–2.720)	0.016b
		No	137 (77.8%)		152 (67%)

(b) is tested by the Chi-square test. (a) is tested by the Fisher exact test

Table 5 presents the influence of occupational and environmental risk factors on various quality-of-life domains among patients with multiple myeloma. A statistically significant association was found between a history of cosmetologist procedures and multiple dimensions of quality of life. Patients who had undergone such procedures reported higher disease symptom scores (Mean = 37.9, SD = 27.8) compared to those without such history (Mean = 27.2, SD = 24.2; p = 0.014). They also exhibited poorer body image, with a mean score of 65.1 (SD = 37.0) versus 78.2 (SD = 27.9) in those without exposure (p = 0.012). Furthermore, their future perspective was significantly more pessimistic (Mean = 68.5, SD = 29.7) than that of unexposed patients (Mean = 83.5, SD = 25.4; p = 0.018). Additionally, a history of radiation exposure was positively associated with future perspective. Patients who had been exposed to radiation reported a significantly higher future perspective score (Mean = 77.5, SD = 28.7) compared to those without such exposure (Mean = 69.1, SD = 29.4; p = 0.048). Similarly, pesticide exposure was also linked to more favorable future expectations, as exposed individuals scored higher (Mean = 80.4, SD = 28.6) than non-exposed counterparts (Mean = 69.7, SD = 29.2; p = 0.024). Other factors such as carcinogenic-organic exposure and agricultural work showed no statistically significant differences across any of the evaluated domains.

Table 5.

Influence of occupational and environmental risk factors on Quality-of-Life domains among multiple myeloma patients

Variable	Subgroup	Disease symptoms scale		Side effects of the treatment scale		Body image scale		Future perspective scale
		(Mean, SD)	P-Value	(Mean, SD)	P-Value	(Mean, SD)	P-Value	(Mean, SD)	P-Value
History of radiation exposure	Yes	33.8 ± 26.7	0.558a	32.3 ± 19.3	0.0.936a	67.5 ± 34.6	0.752a	77.5 ± 28.7	0.048a
	No	36.1 ± 27.6		32.1 ± 19.9		69.1 ± 35.9		69.1 ± 29.4
History of the Cosmetologist procedure	Yes	37.9 ± 27.8	0.014a	29.5 ± 16.8	0.256a	65.1 ± 37	0.012a	68.5 ± 29.7	0.018a
	No	27.2 ± 24.2		33.1 ± 20.5		78.2 ± 27.9		83.5 ± 25.4
Carcinogenic-organic exposure	Yes	28.7 ± 26.3	0.069a	27.9 ± 21.4	0.095a	72.5 ± 36.7	0.386a	71.2 ± 29.4	0.276a
	No	37.1 ± 27.3		33.4 ± 19		67.3 ± 34.9		76.6 ± 29.1
Pesticide exposure	Yes	33.0 ± 28.4	0.515a	29.5 ± 19.7	0.275a	72.5 ± 35.7	0.343a	80.4 ± 28.6	0.024a
	No	35.9 ± 26.9		33.0 ± 19.6		67.1 ± 35.2		69.7 ± 29.2
Agriculture	Yes	31.7 ± 26.5	0.209a	30 ± 16.5	0.271a	70.6 ± 34.8	0.548a	74.3 ± 30.3	0.506a
	No	36.9 ± 27.5		33.2 ± 21.1		67.4 ± 35.6		71.4 ± 28.9

a is tested by T test

Table 6 provides multivariable-adjusted odds ratios (OR) assessing the independent association between occupational and environmental exposures and MM occurrence, after controlling for potential confounders (age, gender, BMI, community type, hospital setting, autoimmune diseases, hypertension, diabetes mellitus, dexamethasone use, and albumin level). Significant elevated risk was independently associated with a history of radiation exposure (OR = 2.322; 95% CI: 1.423–3.788; p = 0.001), Cosmetics agents exposure (OR = 2.850; 95% CI: 1.559–5.211; p = 0.001), carcinogenic-organic exposure (OR = 2.888; 95% CI: 1.493–5.588; p = 0.002), pesticide use (OR = 1.789; 95% CI: 1.033–3.096; p = 0.038), and agricultural work (OR = 1.673; 95% CI: 1.033–2.709; p = 0.036). These adjusted results substantiate the etiological relevance of these factors, highlighting their clinical importance and potential as targets for prevention strategies in MM development. See supplementary file for the full diagnostic models (from etable1 to etable5).

Table 6.

Multivariable adjusted associations of occupational and environmental exposures with risk of developing multiple myeloma

Variable	Adjusted Odds ratio (CI)	P value
History of radiation exposurea	2.322 (1.423–3.788)	0.001
History of cosmetics agents exposure	2.850 (1.559–5.211)	0.001
Carcinogenic- organic exposureb	2.888 (1.493–5.588)	0.002
Pesticide exposureb	1.789 (1.033–3.096)	0.038
Agriculture farmingb	1.673 (1.033–2.709)	0.036

^aAdjusted for age, gender, BMI, type of community, hospital setting, autoimmune diseases, hypertension (HTN), dexamethasone use, and albumin level. ^bAdjusted for age, gender, BMI, type of community, hospital sitting, autoimmune diseases, hypertension (HTN), dexamethasone use, and diabetes mellitus.CI: Confidence Interval

Figure 3 presents the receiver operating characteristic (ROC) curve analysis evaluating the discriminatory capacity of five adjusted occupational and environmental exposures—radiation, Cosmetics agents-related chemical exposure, organic solvent exposure, pesticide exposure, and agricultural work—in distinguishing patients with multiple myeloma (MM) from matched non-neoplastic controls. Each risk factor was assessed using multivariable logistic regression models, tailored to control for relevant demographic and clinical confounders. Specifically, the radiation exposure model was adjusted for age, gender, body mass index (BMI), type of residency, hospital setting (inpatient vs. outpatient), autoimmune diseases, hypertension, dexamethasone use, and albumin levels. The remaining four exposures— Cosmetics agents-related agents, organic solvents, pesticides, and agricultural activity—were all adjusted for age, gender, BMI, type of residency, hospital setting, autoimmune diseases, hypertension, dexamethasone use, and diabetes mellitus. The area under the curve (AUC) values ranged from 0.7093 for radiation to 0.7212 for organic solvents, indicating moderate yet consistent predictive accuracy. The composite ROC curve overlay, which combines all five exposures, reveals a substantial degree of overlap among the curves, implying a possible synergistic or cumulative risk effect rather than discrete, independent contributions. This analysis underscores the clinical relevance of incorporating detailed environmental and occupational exposure histories into multivariable risk prediction strategies for MM detection and early stratification, even when individual exposures do not yield high discriminatory power in isolation.

Fig. 3.

Discriminatory performance of adjusted occupational and environmental risk factors for multiple Myeloma

Table 7 presents the results of a multivariable linear regression analysis examining the impact of occupational and environmental risk factors on the quality-of-life domains of the EORTC QLQ-MY20 in patients with multiple myeloma. The analysis adjusted for relevant clinical and demographic confounders across each domain. A history of exposure to cosmetic agents emerged as a strong and consistent predictor across multiple domains. Patients with such exposure demonstrated significantly worse disease symptoms, with a positive regression coefficient of 11.55 (95% CI: 2.82 to 20.28, p = 0.010), indicating higher symptom burden. Additionally, they reported significantly worse side effects of treatment, with a coefficient of −2.169 (95% CI: −8.568 to −0.229, p = 0.049), and a markedly reduced future perspective, as evidenced by a regression coefficient of −13.728 (95% CI: −22.876 to −4.579, p = 0.003). Although the association with body image was negative (Coeff = −21.613, p = 0.088), it did not reach statistical significance. Pesticide exposure was also significantly associated with quality-of-life outcomes. It predicted greater disease symptoms, with a regression coefficient of −3.77 (95% CI: −12.61 to −2.063, p = 0.041), and a more optimistic future perspective, as shown by a coefficient of 10.054 (95% CI: 0.770 to 19.337, p = 0.034).

Table 7.

Multivariable linear regression analysis of occupational and environmental risk factors on EORTC QLQ-MY20 Quality-of-Life domains in multiple myeloma patients

Variable	DSa		SEb		BIc		FPd
	Coeff, (CI)	P value	Coeff, (CI)	P value	Coeff, (CI)	P value	Coeff, (CI)	P value
History of radiation exposure	−2.196(−10.09 5.704)	0.584	0.533)−5.183–6.250)	0.854	7.129 (−6.26–20.51)	0.294	8.689 (0.401–16.977)	0.474
History of cosmetics agents exposure	11.55 (20.28 − 2.82)	0.010	−2.169(−8.568– −0.229)	0.049	−21.613 (−35.90–0.33)	0.088	−13.728 (−22.876–−4.579)	0.003
Carcinogenic-organic exposure	−8.99 (−18.25–0.25)	0.057	−4.168(−10.78–2.44)	0.215	11.603 (−3.76–26.97)	0.138	−3.959 (− 5.700–−2.618)	0.0420
Pesticide exposure	−3.77 (−12.61–−2.063)	0.041	−3.00(−9.43–3.41)	0.357	5.547 (−10.21–21.30)	0.487	10.054 (0.770–19.337)	0.034
Agriculture farming	−6.413(−14.62–1.80)	0.125	−3.247(−9.141–2.645)	0.278	0.094 (−0.066–0.255)	0.249	4.753 (− 3.772–13.278)	0.273

^a Adjusted for Incidentally finding, Arthritis, Cardiac arrest, Neurological disease, Weight loss, Chest pain, Bone pain, Lower back pain. ^b Incidentally finding, Osteoporosis, Autoimmune diseases, Peptic ulcer, Gout, Weight loss, Abdominal pain, Headache/Dizziness, Chest pain, Bone pain, Lower back pain, Constitutional symptoms. ^c adjusted for Incidentally finding, Osteoporosis, CHF, Cerebrovascular disease, Calcium level, Peptic ulcer, Neurological disease, Gout, Abdominal pain, Headache/dizziness, Chest pain. ^d adjusted for Osteoporosis, Arthritis, Autoimmune diseases, Cardiac arrest, Gout, Chest pain, Bone pain. Abbreviations: DS: Disease Symptoms, SE: Side Effects of Treatment, BI: Body Image, FP: Future Perspective, Coeff: Coefficient, CI: Confidence Interval, SD: Standard Deviation

Carcinogenic-organic exposure was associated with a significantly lower future perspective (Coeff = −3.959, 95% CI: −5.700 to −2.618, p = 0.042). Other domains influenced by this exposure, including disease symptoms and body image, did not reach significance. In contrast, radiation exposure and agricultural farming did not show statistically significant associations across any of the quality-of-life domains.

Figure 4 illustrates the outcome–exposure relationships between key occupational/environmental risk factors and specific quality-of-life domains among multiple myeloma patients, using scatterplots with fitted regression lines. Each subplot visualizes how binary exposure variables (0 = No, 1 = Yes) relate to quantitative outcomes. In the top-left panel, a positive linear trend is evident between cosmetologist exposure and the Disease Symptoms Score, indicating that patients with a history of such exposure tend to report higher symptom burden. This visual observation is consistent with the regression results from Table 7, where cosmetologist exposure was significantly associated with increased disease symptoms.The top-right panel shows a negative association between cosmetologist exposure and Future Perspective Score. Patients with exposure exhibit a noticeably lower range of scores compared to those without exposure, confirming a decline in future outlook. This aligns with Table 7’s significant regression finding. In the bottom-left panel, pesticide exposure is negatively associated with the Disease Symptoms Score, with the fitted line slightly sloping downward consistent with Table 7 where pesticide exposure was significantly associated with reduced disease symptoms. Finally, the bottom-right panel reveals a clear decline in Future Perspective Score among patients with carcinogenic-organic exposure, as the fitted red line slopes downward. This supports the regression result in Table 7.

Fig. 4.

Visualizing the impact of key occupational exposures on quality-of-life domains in multiple myeloma patients

Discussion

This multi-center, hospital-based case-control study sought to explore the association between occupational and environmental exposures and the risk of developing multiple myeloma (MM) in the Palestinian population, while simultaneously examining the influence of these exposures on disease-related quality of life (QoL). Conducted in a setting where industrial and agricultural hazards are prevalent yet under-investigated, the study addresses a critical knowledge gap in regional MM etiology and survivorship. After controlling for key confounders, including demographic, clinical, and treatment-related variables, our multivariable logistic regression models revealed that radiation, pesticides, agricultural work, organic solvents, and especially cosmetics-related chemical exposure were all independently associated with significantly increased odds of MM development. Notably, the strongest associations were observed for Cosmetics agents and organic solvent exposures, each nearly tripling the likelihood of MM diagnosis compared to unexposed individuals. Extending beyond etiology, the study also evaluated the impact of these exposures on health-related QoL using the EORTC QLQ-MY20 tool. Multivariable linear regression models demonstrated that cosmetics-related exposure was significantly linked with worsened disease symptoms and reduced future perspective, suggesting a tangible psychosocial burden in addition to its etiological contribution. Organic solvent exposure similarly impaired future perspective, while pesticide exposure, although associated with lower disease symptom scores may reflect a unique exposure-health perception paradox or reporting bias.

Starting with pesticide exposure, this study found an association between pesticide exposure and the risk of developing multiple myeloma. Tual et al. [47] supported our findings, reporting that pesticide exposure was associated with an increased risk of multiple myeloma (HR = 1.73; 95% CI: 1.08–2.78; p for trend < 0.01). Pahwa et al. [18] found this association to be significant only for one specific pesticide class (carbamate insecticides) with an odds ratio of 1.81 (95% CI: 1.05–3.12). indicating that this association may be detected in specific types of pesticides only. Pesticides have frequently been reported to cause genetic and epigenetic alterations that contribute to carcinogenesis, including the development of multiple myeloma [48]. These mechanisms may explain the association observed in our study.

Our findings align with prior literature suggesting a link between pesticide exposure and the risk of MM. Notably, Blair et al. [49] emphasized that misclassification of pesticide exposure, especially when lacking biomarker validation, can substantially underestimate true associations in epidemiological research. This observation may explain the moderate magnitude of association observed in our study (adjusted OR = 1.79), as our exposure measurement relied on structured self-reported interviews without subclassifying pesticide types. The underestimation hypothesis aligns with studies by Tual et al. [50] and Pahwa et al., [51]who reported stronger associations when analyzing specific classes of pesticides, such as carbamate insecticides. Therefore, future studies incorporating more granular exposure data or biomarker-based validation may uncover even more robust relationships between pesticide exposure and multiple myeloma risk.

In addition, agricultural work was considered a significant risk factor for MM in our study. Perrotta et al. [52] also reported that working as a farmer was associated with an increased risk of multiple myeloma, with an odds ratio of 1.39 (95% CI: 1.18–1.65). Baris et al. [53] found a significant association between multiple myeloma and living or working on a sheep farm, reporting an odds ratio of 1.7 (95% CI: 1.0–2.7), However, no significant associations were found among residents or workers of cattle, beef, pig, or chicken farms. Nanni et al. [54] also reported that growing apples and plums specifically was associated with an increased risk of multiple myeloma (OR = 1.75; 95% CI: 1.05–2.91). However, the study found no significant association between agricultural work in general and multiple myeloma (OR = 1.31; 95% CI: 0.62–2.74). These variations across studies suggest that not all forms of agricultural exposure confer the same risk for developing multiple myeloma. The type of farming, nature of the crops or animals involved, may play critical roles in determining the level of risk.

This study also observed a significant association between carcinogenic-organic exposure and MM risk. This finding is supported by similar evidence, as a pooled study by De Roos et al. observed a statistically significant increased risk of MM for those in the highest quartile for exposure to benzene (OR = 1.42 (95% CI 1.08 to 1.86), p value = 0.01) [23]. This is also consistent with Onyije’s et al. meta-analysis, which found a significant association with petroleum industry work (estimated risk of 1.80; 95% CI: 1.28–2.55) [55]. A Swedish study showed a difference in association before 1985 and after. In this study, work on Swedish petroleum tankers for five or more years was significantly associated with MM (OR, 5.39; 95% CI, 1.11-26.1), while those whose first sea service was after 1985 had no significant association (OR, 2.70; 95% CI, 0.70–10.4). This is most likely attributable to regulations done to reduce such exposure later on [56]. In a study by Gold et al., a statistically significant association with MM risk was also found in those exposed to 1,1,1-trichloroethane (TCA) (OR = 1.8, 95% CI = 1.1–2.9) but not in perchloroethylene (PCE) and chloroform. This is true for trichloroethylene (TCE) and methylene chloride (DCM) as well, but these became significant when subjects were considered unexposed for occupations with low confidence scores (TCE: 1.7 (1.0 to 2.7); DCM: 2.0 (1.2 to 3.2)) [57]. Our study does not support the results of Vlaanderen’s et al. meta-analysis of 26 cohorts, which found an insignificant association with workers exposed to benzene (RR: 1.12; 95% CI: 0.98–1.27) [58]. Our findings are also unlike those of Costantini et al. who found no significant association between a range of solvents including benzene and MM (OR = 1.9, 95% CI = 0.9–3.9) [59], although the author states that this may be due to Italy’s regulations regarding benzene long before the initiation of the study.

Cosmetics agents’ exposure was an observed variable in our study, in which we also found an association. This is consistent with results from a meta-analysis done in 2009, which showed the pooled RR of occupational exposure as a hairdresser being 1.62 (95% CI 1.22–2.14) [60]. This is unlike many other studies which reported no significant association between hair dye use and multiple myeloma risk, such as Zhang et al. (n = 1807; hazard ratio = 1.00, 95% CI 0.91 to 1.10) for hematopoietic cancers overall [61], Koutros et al. (OR 0.8; 95% CI 0.5 to 1.1) [62] and Wong et al. [63]. A 2022 review including 7 different studies all showed insignificant association between personal hair dye use and increased MM risk [64] except for Grodstein’s et al. study which found a small protective effect (RR = 0.4; 95% CI = 0.2–0.9) [65]. The variety in study findings may be due to different components and regulations of hair products in each country, along with hygiene measures and ventilation of workplaces for cosmetologists.

Finally, our study found a significantly increased risk of multiple myeloma (MM) associated with a history of radiation exposure, consistent with previous literature. Hunter and Haylock [24] reported a statistically significant association between ionizing radiation and MM risk (p = 0.02), with an estimated excess relative risk per Sievert (ERR/Sv) of 2.63 (95% CI: 0.30–6.37, AIC = 2473.2). Similarly, the International Nuclear Workers Study (INWORKS) also identified a positive and statistically significant association, reporting an ERR per Gray (Gy) of 1.62 (90% CI: 0.06–3.64, n = 527) [66]. In contrast, Yiin et al. [67] observed only a weak association between MM risk and internal uranium dose estimated from urinalysis, with an odds ratio (OR) of 1.04 (95% CI: 1.00–1.09). A study of Czech uranium miners found no association between radon exposure and MM risk [68]. Similarly, Hatcher et al. [69] reported no significant link between diagnostic X-ray exposure and MM risk, with an OR of 0.9 (95% CI: 0.7–1.2) for overall exposure and 0.7 (95% CI: 0.4–1.3) for individuals reporting ten or more X-rays. Moreover, a pooled analysis of nine cohorts evaluating individuals first exposed to external radiation before age 21 also failed to show a significant association (ERR/Gy = 0.149; 95% CI: −0.513 to 1.063; p-trend > 0.4) [70]. These mixed findings underscore the complexity of radiation as a risk factor for MM, influenced by factors such as exposure source, dose, duration, age at exposure, and underlying susceptibility. Nonetheless, the robust association observed in our multivariable-adjusted analysis reinforces the etiologic relevance of ionizing radiation in MM pathogenesis.

After a thorough review of the literature, to the best of our knowledge, no studies have specifically examined the relationship between these risk factors and the QoL of patients with MM. In our study, as shown in Table 7, we examined this association, using the four domains of the EORTC QLQ-MY20 quality-of-life. Exposure to cosmetic agents demonstrated a distinct and somewhat paradoxical pattern across quality-of-life domains. It was independently associated with significantly higher disease symptom scores, indicating a greater reported symptom burden, while concurrently showing lower scores for treatment side effects and future perspective, reflecting reduced perceived treatment burden but a more pessimistic outlook [71]. This profile may reflect underlying psychological or behavioral traits; individuals with frequent exposure to Cosmetics agents-related products may be more vigilant about their symptoms, yet potentially more proactive in symptom management, thereby minimizing perceived treatment side effects. However, the reduced future perspective suggests deeper concerns about long-term health outcomes, possibly shaped by prior health experiences or perceived vulnerability [72].

Exposure to carcinogenic organic solvents demonstrated a significant negative association with the future perspective domain of quality of life, indicating a more pessimistic outlook among exposed multiple myeloma patients. This finding may reflect heightened psychological distress or diminished health-related optimism, potentially stemming from the well-established carcinogenicity and systemic toxicity of many industrial solvents. Such exposures may instill a heightened sense of vulnerability, contributing to greater anxiety about disease progression or long-term survival. This is consistent with prior evidence suggesting that individuals with significant chemical exposures particularly to organic solvents—often report psychological strain and reduced future orientation due to perceived long-term health risks [73]. These findings underscore the need for targeted psychosocial support for patients with known high-risk occupational histories, particularly those exposed to hazardous chemicals, as these exposures may shape not only biological risk but also emotional and cognitive dimensions of the illness experience.

Although radiation exposure did not show any independent connections for any of the QoL domains in this research, it did show signs of some associations that were consistent with weak effects that might prove evident in longitudinal observations or larger series. Participants’ exposure to pesticides was substantially linked to lower disease symptom scores, indicating that those exposed reported fewer symptoms. This could be because exposed people underreported, were more physically tolerant, or had different ethnic perspectives on symptoms, which are frequent among workers in rural areas. Despite the fact that we did not find any quality-of-life domain that was significantly correlated with agricultural occupation, larger research may reveal patterns that can be considered relevant [74].

While several associations in Table 7 reached statistical significance, we acknowledge that some were accompanied by wide confidence intervals, particularly those related to quality-of-life outcomes and environmental exposures. These wide intervals indicate reduced precision and may reflect variability due to sample size limitations or exposure measurement. As such, we interpret these findings with appropriate caution. To strengthen the reliability of these associations, future studies with larger sample sizes are essential. Such efforts would help narrow the confidence intervals, improve estimate stability, and enhance the overall accuracy and validity of the observed relationships.

Strengths and methodological innovations

This multi-center, hospital-based case–control study significantly advances the understanding of occupational and environmental exposures in relation to multiple myeloma (MM) within the unique context of the West Bank. A key strength of this research lies in its comprehensive approach, simultaneously examining both etiological factors and their nuanced impacts on quality of life (QoL). The matched design, employing stringent criteria such as age, gender, inpatient/outpatient status, and treatment facility, effectively minimized baseline differences and potential confounders.

The methodological rigor was further enhanced through advanced statistical modeling. Beyond conventional multivariate analyses, the implementation of Least Absolute Shrinkage and Selection Operator (LASSO) regression with 10-fold cross-validation provided robust control against model overfitting and allowed precise selection of influential confounders. This innovative approach ensured the retention of variables with true predictive contributions, thereby refining the accuracy of the association assessments between exposures and QoL outcomes.

Adjustments for confounding variables were specifically tailored to each QoL domain of the validated EORTC QLQ-MY20 instrument, ensuring precise and clinically relevant interpretation. For example, disease symptoms (DS) accounted for factors such as incidental diagnosis, arthritis, cardiac arrest, neurological conditions, and various pain indicators. Side effects of treatment (SE) adjustments incorporated osteoporosis, autoimmune disorders, gastrointestinal conditions, and constitutional symptoms. Body image (BI) analysis included adjustments for cardiac conditions, cerebrovascular disease, calcium levels, and neurological disorders, whereas future perspective (FP) models controlled for osteoporosis, arthritis, autoimmune conditions, cardiac events, and associated symptoms.

Validation and reliability of measures

The selection of occupational and environmental exposures evaluated—radiation, pesticides, organic solvents, Cosmetics agents-related chemicals, and agricultural work—was grounded in prior epidemiological evidence and biological plausibility, strengthening the study’s validity. Moreover, the use of the culturally adapted and internationally validated EORTC QLQ-MY20 questionnaire for assessing QoL outcomes enhances the reliability and comparability of results across different populations and clinical contexts.

Limitations

Despite rigorous methods, this study is not devoid of limitations. Its retrospective nature inherently restricts causal inference capabilities, and the potential for recall bias, especially regarding long-term occupational exposures, must be acknowledged. The hospital-based control group, though carefully selected, may not entirely represent the general population, potentially introducing selection bias. The cross-sectional nature of QoL assessments limits the capacity to capture temporal dynamics and changes over time. Moreover, the study did not assess the intensity, frequency, or duration of exposures, particularly for pesticide use, thereby limiting the evaluation of dose-response relationships. This limitation may obscure more nuanced exposure–outcome relationships and potentially lead to exposure misclassification, a known source of nondifferential bias that typically underestimates true associations Blair et al. [49]. Additionally, although multiple confounders were Additionally, although multiple confounders were adjusted for using LASSO and multivariate modeling, residual confounding and bias cannot be fully excluded. Some important variables—such as socioeconomic status (e.g., income and educational attainment) were not captured due to data unavailability. Moreover, an important limitation of our analysis is that certain clinically relevant variables, specifically family history of multiple myeloma and prior history of cancers, were excluded from the multivariable logistic regression models due to perfect prediction. In both cases, all individuals with a positive history belonged exclusively to the MM group, resulting in complete separation and preventing coefficient estimation. While this pattern may reflect a true and strong association, the lack of corresponding cases in the control group limited our ability to evaluate their independent effects statistically. We addressed this issue by using Fisher’s exact test in univariate comparisons and acknowledge the need for larger sample sizes or alternative modeling techniques (e.g., penalized likelihood methods) in future studies to properly assess these predictors. Finally, the limitations of relying on self-reported data without biomarker or environmental validation should be critically acknowledged, as such methods may introduce misclassification and reduce the accuracy of exposure assessment.

Policy implications

Our findings underscore the urgent need for public health policies targeting occupational exposures linked to multiple myeloma. The identification of cosmetics-related chemicals, pesticides, and organic solvents as significant risk factors aligns with global concerns about chemical carcinogenesis in occupational settings. Prior studies have established elevated MM risks among agricultural workers exposed to pesticides Perrotta et al. [20]. industrial workers exposed to organic solvents Lope et al. [75] and cosmetologists with prolonged chemical exposure IARC [76], Regulatory bodies should prioritize stricter control measures, including updated exposure thresholds, compulsory use of personal protective equipment, and regular surveillance in high-risk sectors such as agriculture, cosmetology, and manufacturing. These findings support international calls for integrated occupational cancer prevention frameworks Georgakopoulou et al., [38].

Clinical implications for healthcare providers

Beyond policy-level interventions, the medical community holds a pivotal responsibility in addressing the impact of occupational and environmental exposures on the risk and progression of multiple myeloma (MM). Clinicians, particularly hematologists, oncologists, and primary care physicians should incorporate detailed environmental and occupational histories into routine clinical assessments, especially in patients presenting with unexplained hematologic abnormalities. Identifying exposures to pesticides, solvents, or cosmetic agents can significantly heighten clinical suspicion for plasma cell dyscrasias such as MGUS or MM, even in the absence of overt symptoms. Notably, a landmark Swedish case–control study revealed that occupational chemical exposure, particularly to sensitizing agents, significantly increased MM risk, underscoring the importance of such history-taking in clinical practice Lope et al., [75] Furthermore, evidence suggests that exposure history may modify disease course; for instance, a U.S. study found that veterans with MGUS exposed to Agent Orange had an over 11-fold increased risk of progression to MM compared to unexposed individuals, suggesting exposure-based stratification may enhance surveillance and early intervention strategies [77]. Early detection is especially critical given that MM is often diagnosed at an advanced stage with irreversible organ damage; integrating exposure data into the diagnostic work-up could prompt clinicians to investigate subtle lab abnormalities or early symptoms more aggressively, leading to timely diagnosis and intervention. This is exemplified by emerging recommendations for proactive MGUS screening in high-risk, exposure-defined populations such as WTC-exposed firefighters, aiming to facilitate treatment at a stage that improves survival (Kazandjian et al.) [78],.Physicians should also educate patients about reducing continued exposure risks, advise on appropriate protective measures, and tailor follow-up intervals accordingly.

Implications and future directions

The findings from this study provide a robust foundation for future research initiatives, including prospective cohort studies designed to validate and expand upon these initial associations. Longitudinal investigations could elucidate the temporal relationships between exposures and QoL outcomes, enhancing causal understanding. Additionally, systematic reviews and meta-analyses could further consolidate evidence across diverse contexts, contributing to global knowledge on MM etiology and patient care. Ultimately, clinical trials targeting modifiable environmental and occupational exposures or focused on interventions to enhance patient QoL could translate these epidemiological insights into practical, patient-centered improvements in MM care and survivorship.

Conclusion

This study confirms significant associations between occupational and environmental exposures particularly cosmetics-related chemicals, organic solvents, ionizing radiation, pesticide use, and agricultural activities—and the risk of developing multiple myeloma. These exposures also demonstrated measurable impacts on patient-reported quality-of-life outcomes, especially disease symptoms and psychological well-being, with cosmetics-related exposures exhibiting the most pronounced effects. These findings underscore the necessity of integrating environmental and occupational history into routine clinical assessments and risk stratification models. Furthermore, they highlight the need for targeted preventive strategies, including workplace safety reforms and surveillance initiatives. Importantly, our results align with recent multicenter epidemiological research emphasizing regional variability in MM etiology and presentation, such as the Eastern European registry-based study by Ghiaur et al., [79] reinforcing the value of conducting localized investigations to capture unique exposure patterns and inform tailored interventions. Continued longitudinal research and preventive intervention trials are essential to deepen etiological understanding and optimize clinical outcomes for patients with multiple myeloma.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ABO

Blood Group System

AIDS

acquired immunodeficiency syndrome

AUC

Area Under the Curve

Beta2-Microglobulin

Serum marker for tumor burden in multiple myeloma

BI

Body Image

BMI

Body Mass Index

CHF

Congestive Heart Failure

CI

Confidence Interval

COVID-19

Coronavirus Disease 2019

Coeff

Coefficient

DS

Disease Symptoms (Quality-of-Life domain)

EORTC QLQ-MY20

European Organization for Research and Treatment of Cancer Quality of Life Questionnaire for Multiple Myeloma (20-item)

EMR

Electronic Medical Record

ERR/Gy

Excess Relative Risk per Gray

ERR/Sv

Excess Relative Risk per Sievert

FP

Future Perspective (Quality-of-Life domain)

HR

Hazard Ratio

HRQoL

Health-Related Quality of Life

HTN

Hypertension

IHD

Ischemic Heart Disease

INWORKS

International Nuclear Workers Study

IRB

Institutional Review Board

KAP

Kappa (light chains, immunoglobulin)

LAM

Lambda (light chains, immunoglobulin)

LASSO

Least Absolute Shrinkage and Selection Operator

MI

Myocardial Infarction

MM

Multiple Myeloma

OR

Odds Ratio

PCE

Perchloroethylene

QoL

Quality of Life

RBCs

Red Blood Cells

ROC

Receiver Operating Characteristic

RR

Relative Risk

SD

Standard Deviation

SE

Side Effects (or Side Effects of Treatment, Quality of- Life domain)

TIA

Transient Ischemic Attack

TCA

1,1,1-Trichloroethane

TCE

Trichloroethylene

VCD protocol

Combination chemotherapy with Bortezomib, Cyclophosphamide, and Dexamethasone

VIF

Variance Inflation Factor

WBCs

White Blood Cells

Gy

Gray (unit of absorbed radiation dose)

Author contributions

M.A. and N.A. contributed equally to this work and share first authorship. They conceptualized the study, performed statistical analysis, and co-wrote the manuscript. H.A. served as corresponding author, supervised the project, and critically revised the manuscript. I.A., A.Z., S.Z., B.B., F.H., O.E., D.N., and A.D. contributed to data collection and field coordination. R.J., M.Q., M.S., A.A., and S.A. assisted with data entry, validation, and database management. L.S., E.O., A.K., F.I., N.S., M.A., and A.B. participated in literature review and supported manuscript preparation. S.A., H.A., M.M., O.D., M.N., and O.I. contributed to statistical modeling and results interpretation. Y.F., O.B., M.E., M.Q., and Y.Q. critically revised the manuscript for intellectual content.All authors read and approved the final manuscript.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding authors upon request.

Declarations

Ethics approval and consent to participate

The study protocol was reviewed and approved by the Institutional Review Board of Najah National University in Palestine. Written informed consent was obtained from all participants prior to data collection.The study was conducted in accordance with the ethical standards outlined in the Declaration of Helsinki.

Consent for publication

Not applicable. The manuscript does not include any individual personal data in any form (e.g., images, videos, personal details).

Competing interests

The authors declare no competing interests.