1. Introduction

Non-Hodgkin lymphoma (NHL) is the seventh most common cancer and the sixth leading cause of cancer death in the United states (US) [1]. Forecasts for 2024 in the US alone predicted 80,620 new cases and 20,140 patient deaths, with half being men [2]. NHL exhibits high heterogeneity, which includes a wide range of origins, symptoms, treatment modalities, and results. The two most prevalent subtypes, diffuse large B-cell lymphoma (DLBCL) and follicular lymphoma, collectively comprise approximately 35% and 20% of all lymphomas, respectively [3,4].

NHL has been associated with immunosuppression which may stem from inherited conditions, infectious diseases, radiation therapy or organ transplantation [5–9] Individuals with congenital immunodeficiency syndromes face a significantly higher risk of NHL—28–49 times greater—than the general population, [5] while those with HIV have a risk approximately 100 times greater [10,11]

While NHL incidence rates nearly doubled between 1970 and 1990, advances in treatment have helped to stabilize the general population since the late 1990s [12] The prevalence of NHL underscores the need to better understand its impact on patient outcomes. However, trends in NHL-associated mortality remain inadequately explored. Recent work by Tan et al. (2024) examined NHL mortality in the US using Centers or Disease Control and Prevention’s Wide-ranging Online Data for Epidemiologic Research (CDC WONDER) data, focusing on underlying causes of death across sex, race, and geography [13]. Our study extends this by incorporating multiple cause of death data, which include both underlying and contributing causes, yielding a more accurate estimate of the true mortality burden. Reliance on underlying cause of death alone may underestimate cases where NHL contributed to death but was not the primary cause. We further stratify by place of death, providing new insight into care settings and palliative needs. Together, these additions enhance understanding of NHL mortality patterns and underscore the value of a dedicated analysis. This review aims to assess trends in NHL-associated mortality in the US from 1999 to 2020 and evaluate differences by sex, race, ethnicity, geographic region, and urbanization, using data from CDC WONDER. Our analysis aimed to investigate demographic and regional changes in NHL-related mortality among US adults aged 15 and older from 1999 to 2020 to identify at-risk individuals.

2. Methodology

3. Results

A total of 560,469 NHL-related deaths occurred between 1999 and 2020 among adolescents and adults (≥15 years old) (Supplementary Figure 1, Supplementary Table 1 in S1 File). Information on the location of death was available for 559,079 deaths. Of these, most deaths occurred within medical facilities (44.58%), followed by the decedent’s home (30.27%), followed by nursing home/long term care (14.41%), followed by hospice facilities (6.09%), and others (0.25%). Annual trends in NHL-related mortality stratified by place of death are shown in Supplementary Table 2 in S1 File.

3.1. Annual trends for NHL-related AAMR

The AAMR for NHL-related deaths in adults was 12.86 in 1999 and 8.26 in 2020. (Table 1) The overall AAMR decreased steeply between 1999 and 2004 (APC: −3.22 [95% CI, −4.11 to −2.79]), followed by a relatively less steep decrease from 2004 till 2018 (APC: −2.15 [95% CI, −2.28 to −2.02]). From 2018 to 2020, an increase in AAMR was noted (APC: 0.89 [95% CI, −0.55 to 1.64]). Visual trends are illustrated in Fig 1 with comprehensive data present in Supplementary Table 3 and Supplementary Table 4 in S1 File.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Frequency and age adjusted rates per 100,000 in adults with NHL stratified by sex, race, census region and location of death.

https://doi.org/10.1371/journal.pone.0327809.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Non-Hodgkin Lymphoma related AAMRs per 100,000 stratified by sex in the United States from 1999 to 2020.

https://doi.org/10.1371/journal.pone.0327809.g001

3.3 NHL-Related AAMR Stratified by Race/Ethnicity

Throughout the study period, NH White Americans had the highest overall AAMR among all ethnicities (overall AAMR: 10.34 [95% CI, 10.32 to 10.37]). They were followed by Hispanics or Latinos (7.53 [95% CI, 7.44–7.61]), NH Black Americans (7.03 [95% CI, 6.95 to 7.10]), NH American Indian or Alaska Native (6.54 [95% CI, 6.25 to 6.84]) and NH Asian or Pacific Islander populations (6.04 [95% CI, 5.94 to 6.14]).

From 1999 to 2018, all ethnicities experienced a decline in AAMR values. The AAMR for NH American Indian or Alaska Native continued to decrease till 2020 (APC: −1.44 [95% CI, −2.32 to −0.46]) as well as the AAMR for NH Asian or Pacific Islander (APC: −1.72 [95% CI, −1.96 to −1.44]). The AAMR for NH Black Americans decreased steeply until 2005 (APC: −3.77 [95% CI, −7.81 to −2.30]), followed by a lesser decrease till 2020 (APC: −1.40 [95% CI, −1.77 to −0.21]). AAMR for NH White Americans decreased steeply till 2004 (APC:-3.06 [95% CI,-3.98 to −2.64]) followed by a lesser decrease till 2018 (APC:-2.12 [95% CI, −2.24 to −1.97]) with Hispanic or Latinos following a similar trend (APC: −1.77 [95% CI, −3.24 to −1.03]). The AAMR for NH Whites increased from 2018 to 2020 (APC: 1.33 [95% CI, −0.17 to 2.16]). A similar trend was noted for Hispanic or Latinos with a much steeper increase from 2018 to 2020 (APC: 3.11 [95% CI, −1.61 to 5.79]).

Visual trends are illustrated in Fig 2 with comprehensive data present in Supplementary Table 3 and Supplementary Table 5 in S1 File.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Non-Hodgkin Lymphoma related AAMRs per 100,000 stratified by race in the United States from 1999 to 2020.

https://doi.org/10.1371/journal.pone.0327809.g002

4. Discussion

Our study, examining NHL-related mortality trends from 1999 to 2020, is the first to comprehensively analyze CDC WONDER data by gender, race/ethnicity, region, and place of death. In this 22-year analysis of mortality data, we report several key findings. First, we observed a sharp decline in mortality rates from 1999 to 2018, followed by an increase until 2020. Second, among all racial groups, NH White Americans had the highest NHL-related AAMR, which experienced an upward trajectory after 2018. Third, we observed variations in AAMR geographically, with states in the upper 90th percentile (Minnesota, Nebraska, Iowa, Ohio and Vermont) displaying nearly 1.4 times the AAMR compared to those in the lower 10th percentile (Georgia, Hawaii, District of Columbia, New Mexico and Nevada). Notably, from 1999 to 2000, metropolitan areas exhibited higher NHL-related AAMRs compared to nonmetropolitan areas, a trend that subsequently reversed, with nonmetropolitan areas surpassing their metropolitan counterparts until 2020. These findings hold substantive implications for public health policy formulation and underscore the necessity of addressing the nuanced temporal and geographical dynamics observed in NHL-related mortality trends. (Fig 3)

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Central Illustration.

https://doi.org/10.1371/journal.pone.0327809.g003

Our study offers a broader and more nuanced view of NHL-related mortality by using multiple cause of death data, rather than limiting the analysis to underlying cause of death as done in the study by Tan et al.[13] This distinction is important, as multiple cause of death also captures deaths where NHL contributed to mortality but was not the primary cause—particularly relevant during the COVID-19 pandemic. Notably, reliance on underlying cause of death data alone may underestimate the true mortality burden of chronic illnesses like NHL. This concern has been previously highlighted by Minhas et al., who demonstrated that underlying cause of death-based analyses can underrepresent disease impact by excluding cases where a condition is a contributing factor rather than the primary cause [18]. By incorporating multiple cause of death data and performing additional stratifications,such as by place of death, and sociodemographic subgroups,our study offers a more nuanced understanding of NHL-related mortality trends and disparities in the US. Lastly, unlike previous studies, ours includes stratification by both place of death and state, offering a deeper understanding of regional differences in mortality trends.

From 2018 to 2020, a modest increase in AAMR was noted, marking the first uptick in nearly two decades. The AAMR for NHL listed as the Multiple Cause of Death rose from 8.00 in 2019 to 8.26 in 2020. Notably, this increase coincides with the onset of the COVID-19 pandemic. During this period, the number of deaths with NHL as the underlying cause of death remained relatively stable. In fact, the AAMR for NHL as the underlying cause of death, slightly declined from 6.38 in 2019 to 6.21 in 2020. This pattern is consistent with findings from previous studies, suggesting that the overall rise in NHL-related mortality is largely attributable to COVID-19–related deaths in patients with NHL, where NHL was a contributing factor rather than the primary cause of death [13]. This trend appears to be linked to the heightened susceptibility of NHL patients to COVID-19 infection, resulting in deaths from COVID-19 where NHL is listed as a contributing factor [19,20].

Race-wise stratification of NHL-related mortality trends revealed an overall decrease in all races until 2018, followed by an increase in AAMR for the White and Hispanic or Latino populations from 2018 to 2020. Nevertheless, Black individuals and Hispanics have some of the lowest AAMRs, which may be attributable to their low incidence rates. Although AAMR remains a valuable summary measure, it should be interpreted cautiously, as it conflates disease incidence with case-fatality and thus cannot reveal whether observed mortality differences reflect fewer new diagnoses or differences in survival after diagnosis. Studies have highlighted that NH White Americans not only have the highest incidence of NHL but also experienced a 65.9% increase between 2014 and 2018, whereas the rise among Black adults was only 7.8% [21]. Once diagnosed, however, prognosis is poorest for Black patients, with five-year relative survival of 69.5% compared with 77.6% in White patients [21]. Advanced-stage presentation is more common in Black individuals, and socioeconomic barriers, including lower income and limited insurance, further delay treatment [22,23]. It is also likely that access to high-quality oncology centers and persistent under-representation in trials also widen survival gaps [24]. A review of phase I oncology trials found that White/Caucasians made up 84.2% of U.S. participants, while Black individuals, Asians, and Hispanics represented only 7.3%, 3.4%, and 2.8%, respectively [25].

Gender-wise stratification of NHL-related mortality trends revealed an overall decrease in both male and female populations. Nevertheless, AAMRs for men remained notably higher compared to women, consistent with global NHL trends [26–28]. While large prospective and epidemiologic studies consistently indicate male sex as an indicator for worse survival, the underlying reasons remain inadequately understood. The elevated risk among men may be attributed to a higher prevalence of specific risk factors, including HIV and chemical exposures [29]. Additionally, certain NHL subtypes, such as Burkitt lymphoma, diffuse large B cell lymphoma (DLBCL), Sezary syndrome, and various T cell lymphomas, exhibit a pronounced male/female incidence ratio [30,31].

Differences in mortality trends may also be elucidated by gender-associated responses to therapy. A study demonstrated that women exhibit a superior response to R-CHOP therapy [32]. It was also found that the addition of rituximab to therapy showed no response difference between young males and females. However, among older adults, women benefited more, attributed to the lower clearance rate of rituximab in older females than males [33]. A UK-based phase II study suggested a significant correlation between sex and response to lenalidomide in patients with relapsed/refractory mantle cell lymphoma, indicating that female patients are more sensitive to lenalidomide than their male counterparts [34].

Pregnancy emerges as a potentially protective factor in NHL. Studies have established an inverse relationship between pregnancy and the risk of B-cell NHL [35]. A decreased risk across different NHL subtypes with oral contraceptive use has also been observed, particularly among women who initiated use at an earlier age (<25 years) [36,37]. However, the molecular mechanisms underlying the influence of sex hormones on NHL development remain incompletely understood. Potential mechanisms include the impact of estrogen on the host immune response. 17β-estradiol decreases interleukin 6 (IL6) production by mononuclear cells, resulting in lower serum IL6 levels [38]. Elevated serum IL6 levels correlate with reduced complete response and diminished overall survival in DLBCL patients [39]. This suggests a protective effect of estrogen by downregulating IL6 levels. Another proposed mechanism involves the direct effect of estrogen on lymphocytes expressing estrogen receptors [40]. There is evidence for the anti-proliferative effect of estrogen on lymphoid cells through estrogen receptor β (ERβ) signaling, inhibiting the proliferation of T cell lymphoma and Burkitt lymphoma cells in vitro [41].

The reasons behind the higher mortality rates observed in non-metropolitan areas are multifaceted. Tao et al. has shown that a lack of insurance is associated with increased mortality from non-Hodgkin lymphoma. Given that non-metropolitan regions tend to have lower rates of insurance coverage compared to their metropolitan counterparts, this finding supports our observation of elevated mortality in non-metropolitan areas. Additionally, socioeconomic status has been identified as a significant factor influencing mortality rates [42]. It’s conceivable that non-metropolitan areas in the U.S. face lower socioeconomic status levels compared to metropolitan areas, a trend observed in other countries, which could contribute to the higher mortality rates in non-metropolitan regions [43]. Lastly, the prevalence of obesity, a known risk factor for NHL, tends to be higher in non-metropolitan areas, further exacerbating mortality rates in these regions [28,44].

Over the past two decades, there has been a decline in NHL incidence and mortality, attributed to advancements in treatment and diagnosis. Mortality from AIDS related NHL, which makes up 4.6% of all NHL cases, has decreased in recent era due to improvements in HAART regimens and increased utilization of HAART in the population [45]. Rituximab, which was introduced in 1998 and is used in combination with chemotherapy, has greatly improved response rates, length of remission, and survival in patients with follicular lymphoma and diffuse large B-cell lymphoma [46–53]. Key studies, such as the US intergroup trial, highlight the advantages of rituximab-based regimens. For instance, the 3-year failure-free survival rate was significantly higher at 53% for patients treated with R-CHOP compared to 46% for those treated with CHOP alone [53]. However, the exact mechanisms of rituximab’s action are still under investigation, and its limited efficacy in B-cell NHL results in poor outcomes in certain NHL subtypes [54,55].

Supporting information