Temporal analysis of sickle cell disease mortality in adults (1999–2020): insights from the CDC WONDER database

Abstract

Objectives

This study aimed to evaluate temporal trends in adult sickle cell disease (SCD) mortality in the United States from 1999 to 2020, stratified by age, sex, race, and geography, to identify disparities and guide targeted interventions.

Methods

A retrospective observational study was conducted using national mortality data from the Centers for Disease Control and Prevention Wide-Ranging Online Data for Epidemiological Research (CDC WONDER) database. The sample included adults aged ≥15 years who died from SCD between 1999 and 2020 (n = 17,443). Age-adjusted mortality rates (AAMRs) were calculated and stratified by demographic and geographic variables. Temporal trends were assessed using Mann-Kendall trend tests, and t-tests were applied to compare continuous variables across subgroups. Statistical significance was defined as P < 0.05.

Results

The AAMR for adult SCD increased by 132% over the study period (P = 0.014). The greatest rise in mortality was observed among adults aged ≥65 years (P = 0.008) and women (P = 0.015). Black individuals accounted for 97.5% of SCD-related deaths, underscoring severe racial disparities. Geographically, the Southern region exhibited the highest AAMR and was the only region with a statistically significant increase in mortality over time (P = 0.001).

Conclusions

Adult SCD mortality in the United States has risen significantly from 1999 to 2020, with disproportionate increases among older adults, women, and individuals in the Southern region. The findings highlight urgent needs for targeted national interventions, development of age-specific care models, and implementation of equity-focused health policies to address persistent racial and regional disparities in SCD outcomes.

Sickle cell disease (SCD) is a hereditary hemoglobinopathy characterized by chronic hemolytic anemia and recurrent vaso-occlusive crises, resulting in significant morbidity and reduced life expectancy.^1,2 While advances in neonatal screening, hydroxyurea therapy, and supportive pediatric care have led to declines in childhood mortality, outcomes for adults with SCD remain disproportionately poor.^1–3 Despite early survival improvements, the transition from pediatric to adult care frequently occurs in fragmented healthcare systems, contributing to elevated rates of acute care utilization and premature mortality in adulthood.⁴ There has also been a documented shift in primary causes of death—from acute complications such as infection and stroke in childhood to chronic multi-organ failure in adulthood.⁵

SCD-related mortality shows pronounced disparities across geographic, racial, and age groups. The highest burden is observed in the Southern United States, where limited access to comprehensive SCD care may contribute to increased mortality.⁶ Black individuals, who account for much of the US SCD population, experience significantly higher age-adjusted mortality rates (AAMRs) compared to their White counterparts.^7–10 Recent analyses also suggest a rise in mortality among older adults with SCD, potentially reflecting both increased survival into adulthood and insufficient management of comorbidities in aging populations.^1,10 While prior studies have highlighted pediatric survival trends or have drawn on single-center or state-level data, there remains a paucity of national, longitudinal analyses of adult mortality patterns in SCD. In particular, comprehensive evaluations of how AAMRs have evolved over the past two decades—stratified by age, sex, race, and geographic region—are limited.^2,3,5,6

To address this gap, we conducted a national, population-based analysis of adult SCD mortality trends in the United States from 1999 to 2020 using the Centers for Disease Control and Prevention Wide-ranging Online Data for Epidemiological Research (CDC WONDER) database. Our objective was to assess temporal changes in AAMRs and to characterize demographic and geographic disparities by age, sex, race, and region to better understand the evolving burden of SCD in adults.

METHODS

Study design and data sources

This retrospective analysis followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE)¹¹ guidelines for observational studies to investigate trends in mortality from SCD in adults using data from the CDC WONDER database.¹² CDC WONDER is a publicly accessible online system that reports de-identified mortality data. The mortality data is sourced from the National Vital Statistics System, which compiles information reported on the official death certificates filled across all 50 states and the District of Columbia. The data include demographic information, geographic location, and cause of death as determined by the certifying physician or coroner.

For this study, we used the Multiple Cause of Death files for the period of 1999 to 2020. ICD-10 codes D57.0, D57.1, D57.2, D57.3, and D57.8 were used to identify cases of SCD in death certificates that listed SCD as either an underlying cause of death or a contributing cause. CDC WONDER uses ICD-10 codes throughout the period; data prior to the transition from ICD-9 to ICD-10 have been officially recoded using comparability ratios by the National Center of Health Statistics (NCHS), ensuring that the ICD-9 to ICD-10 shift did not affect this analysis. As the data were publicly available and anonymized, the study was exempt from institutional review board approval.

Study population and variables

AAMRs were calculated per 100,000 of the population by the system (based on the formulas set by the NCHS). NCHS uses the “direct method” to weigh an age-specific death rate against a standard age distribution to calculate the AAMRs. The weights used in this calculation are the age-group proportions of the 2000 US standard population (Supplemental Table 1).^12,13 The uncertainty in AAMRs is reported as 95% confidence intervals (CI). CDC WONDER calculations account for the propagation of variance by using a weighted sum of the variance of the individual age-specific rates. The final variance is used to calculate the CIs by using a normal distribution for death counts higher than 100 and a Poisson-based distribution for counts fewer than 100.¹³

In accordance with CDC WONDER’s confidentiality protections, the mortality data were “suppressed” for subgroups with death counts between 1 and 9 and deemed “unreliable” for death counts less than 20. This is a standard protocol across datasets to prevent potential identification of individuals.¹³ The data for these individuals were available for calculations in higher level aggregates (national and regional levels).

The AAMRs were analyzed across four demographic dimensions: age, sex, race, and geography. The study included individuals aged 15 years and older whose cause of death was attributed to SCD. CDC WONDER reports AAMR using standard 10-year age subgroups only, and the final pediatric subgroup is “5 to 14” in the database. Age 15 was considered the optimal starting point to exclude pediatric subgroups and properly use the system’s AAMR calculation. Further, age was stratified into two age groups (15–64 and >64) and 20-year age bands (the minimum set by the database for calculation of AAMRs). Geographic analysis was conducted at the level of the four US census regions (South, Midwest, West, and Northeast). State-level analysis was performed to further show overall geographic heterogeneity over the period. Suppression constraints in the database limited year-to-year analysis for states. While the database reports data for four racial groups (White, Black or African American, Asian or Pacific Islander, and American Indian or Alaska Native), this analysis focused on Black and White racial groups. Due to the limited number of deaths among White patients, a significant portion of the data was either deemed unreliable or suppressed to maintain privacy. These limitations were more pronounced in non-Black, non-White groups, where year-to-year AAMRs for these populations were not available, necessitating their exclusion from analysis.

Outcomes

The primary outcome that was analyzed was the trend in AAMR from SCD between 1999 to 2020. Secondary outcomes included stratification of the AAMR trend based on sex, age, race, and census regions.

Statistical analysis

The Mann-Kendall test, a nonparametric method, was employed to identify monotonic trends over the entire 22-year time period (1999 to 2020). This test evaluates the full series data by evaluating rank correlations between annual AAMR over time (rather than comparing the start and end points only). The Kendall tau coefficient was used for calculation, and statistical significance was determined at a threshold of P < 0.05. Prior to formal analysis, the AAMRs were plotted over time and visually inspected. As the time series demonstrated generally consistent directional trends, the Mann-Kendall test was deemed appropriate to assess monotonicity. To compare continuous variables between demographic subgroups, an independent sample t-test was employed. The data were analyzed using the IBM SPSS Statistics (Version 27) software.

RESULTS

Between 1999 and 2020, 18,047 adults died due to SCD in the United States. The overall AAMR for all patients was 0.334 per 100,000 people (95% CI: 0.329–0.339) and a significant increase of 132% was seen between 1999 and 2020 (AAMR: 0.334 vs 0.442 per 100,000 people; 𝜏=0.382, P = 0.014) (Figure 1, Table 1).

Figure 1.

Overall age-adjusted mortality rates per 100,000. SCD adults saw an increasing trend in AAMR between 1999 and 2020. 𝜏 refers to Mann-Kendall tau and P refers to its significance level.

Table 1.

Age-adjusted mortality rates according to demographic variables among the US population, 1999 to 2020

Year	Age-adjusted rate [95% CI]	Total number of deaths
Total: 1999–2020	0.334 [0.329–0.339]	18,047
Sex
Male	0.334 [0.327–0.341]	8894
Female	0.335 [0.309–0.322]	9153
Race
Black or African American	2.624 [2.585–2.664]	17,006
White	0	391
Race-sex
Black or African American males	2.710 [2.634–2.751]	8834
White males	0	227
Black or African American females	2.477 [2.43–2.536]	8862
White females	0	164
Age group
Young	0.369 [0.364–0.375]	1,6383
Old	0.152 [0.145–0.16]	1664
Age-sex
Young males	0.369 [0.361–0.377]	8218
Young females	0.345 [0.337–0.352]	8165
Older males	–	667
Older females	0.165 [0.154–0.175]	988
Census region
Northeast	0.314 [0.302–0.325]	3026
Midwest	0.276 [0.206–0.287]	3167
South	0.507 [0.491–0.511]	10,006
West	0.145 [0.142–0.156]	1848
Census region-sex
Northeast males	0.297 [0.281–0.312]	1464
Northeast females	0.314 [0.289–0.33]	1562
Midwest males	0.296 [0.282–0.311]	1581
Midwest females	0.259 [0.246–0.271]	1586
South males	0.507 [0.492–0.521]	4906
South females	0.518 [0.503–0.532]	5100
West males	0.149 [0.139–0.159]	943
West females	0.149 [0.139–0.159]	905

Stratified by sex

A total of 9153 women died from SCD during the period. The AAMR significantly increased (𝜏=0.387, P = 0.012) by 139% between 1999 and 2020 (0.305 vs 0.425, respectively). The total AAMR for women was 0.335 (95% CI: 0.309–0.322).

A total of 8894 men died during this period. Adult men saw no statistically significant change in AAMR during the period (𝜏=0.213, P = 0.167). The overall AAMR for men was 0.334 (95% CI: 0.327–0.341) (Figure 2, Table 1). The difference between men and women was not found to be significant (t-test, P = 0.766).

Figure 2.

Age-adjusted mortality rates per 100,000 people for sex. Female adult patients with sickle cell disease saw a rising trend in their AAMR while male rates remained stable between 1999 and 2020. 𝜏 refers to Mann-Kendall tau and P refers to its significance level.

Stratification by geographic regions

For census regions, the South had the highest AAMR of 0.507 (95% CI: 0.491–0.511) followed by the Northeast with an AAMR of 0.314 (95% CI: 0.302–0.325), Midwest with 0.276 (95% CI: 0.206–0.287) and West with 0.145 (95% CI: 0.142–0.156). When looking for the trend in AAMR, only the South (𝜏=0.34, P = 0.028) showed an increase in AAMR over the years (P = 0.429 for Northeast; P = 0.134 for Midwest; P = 0.296 for West) (Figures 3 and 4, Table 1). Females in the South had the highest overall AAMR followed by males in the South. The lowest AAMR for both males and females was in the West (Figure 5, Table 1).

Figure 3.

Age-adjusted mortality rates (AAMR) per 100,000 people of different census regions. An increasing trend in AAMR of sickle cell disease patients in the South was observed. 𝜏 refers to Mann-Kendall tau and P refers to its significance level.

Figure 4.

(a) Census region map and (b) statewide map of the overall age-adjusted mortality rate from sickle cell disease in adult patients from 1999 to 2020. (Source: https://wonder.cdc.gov/.)

Figure 5.

Age-adjusted mortality of census regions based on sex: (a) Northeast, (b) Midwest, (c) South, and (d) West.

For states, the District of Columbia (AAMR: 1.487), South Carolina (AAMR: 1.026), and Mississippi (AAMR: 0.964) were the top three with the highest AAMR. West Virginia (AAMR: 0.098), Iowa (AAMR: 0.071) and Oregon (AAMR: 0.048) were the three states with the lowest AAMR. For several states (Alaska, Hawaii, Idaho, Maine, Montana, New Hampshire, North and South Dakota, Vermont, Wyoming) data were suppressed by the CDC (Figure 4, Supplemental Figure 1).

Stratification by race and sex

Between 1999 and 2020, 17,006 Black individuals and 391 White people died from SCD. Black people had an overall AAMR of 2.624 in 100,000 people (95% CI: 2.585–2.664) and this was significantly different compared to White people (P < 0.001). Both Black and White people showed no significant variation in AAMR during this period (P = 0.446 for Black people; P = 0.372 for White people) (Table 1).

After stratifying for sex, Black men showed a consistently higher AAMR than Black women (overall AAMR for Black men: 2.710; overall AAMR for Black women: 2.478), which was significant (t-test, P < 0.001) (Table 1). However, both Black men (P = 0.554) and Black women (P = 0.8) showed no significant change in AAMR through the period.

The number of White people who died was very small; the CDC deemed their overall AAMR too low (∼0).

Stratification by age and sex

During the study period, 16,383 deaths occurred in the 15- to 64-year group (the younger group) and 1664 deaths in the ≥65-year group (the older group). Younger people had a significantly higher overall AAMR than older people (0.369 vs. 0.152, t-test P < 0.001). However, compared to the younger group (P < 0.323), only older individuals saw an increase in AAMR over time (𝜏=0.689, P < 0.001) (Figure 6, Table 1).

Figure 6.

Age-adjusted mortality rates (AAMR) stratified according to age group.

When stratifying by sex, similar results were seen where younger individuals exhibited no change in AAMR over the years (P = 0.172 for young females; P = 0.691 for young males) while older individuals showed a significant rising trend in AAMR over the years (𝜏=0.63, P < 0.001 for older females; 𝜏=0.481, P = 0.032 for older males).

Young males had the highest AAMR at 0.369 (95% CI: 0.361–0.377) followed by young females with 0.345 (95% CI: 0.337–0.352) and older females with 0.165 (95% CI: 0.154–0.175). Both younger males and females had significantly different AAMRs than their older counterparts (t-test, P < 0.001). However, when younger males were compared to younger females and older males with older females, no difference was seen (t test, P = 0.262 and P = 0.68, respectively) (Figure 7, Table 1).

Figure 7.

Age-adjusted mortality of the younger population based on sex.

Older males’ data was deemed unreliable for many of the years in the study period.

When stratifying into 20-year age bands, age 35–54 had the highest AAMR (0.455; 95% CI: 0.445–0.465) and saw a stable trend over time (𝜏=.257, P = 0.33). This was followed by ages 15–34 with the second-highest AAMR (0.299; 95% CI: 0.291–0.307) but a significant decreasing trend in mortality (𝜏=-0.308, P = 0.03). Third was age 55–74 (AAMR: 0.257; 95% CI: 0.248–0.265), which saw an increasing trend in this time period (𝜏=0.707, P < 0.001). Due to suppression constraints, total AAMR for age 74 and above could not be computed or used for trend analysis (Supplemental Table 2). Supplemental Table 3 shows the Mann-Kendall tau values for all demographic subgroups.

DISCUSSION

Our national, population-based analysis of adult SCD mortality from 1999 to 2020 demonstrated an overall increase in age-adjusted mortality. Younger adults had the highest mortality burden throughout the study period, while increasing mortality over time was observed only among older adults. Although overall mortality did not differ significantly between men and women, a rising trend was observed exclusively among women. Geographic variation was evident, with the South experiencing both the highest mortality rates and the only significant regional increase over time (Central Illustration).

Our analysis reveals a complex and shifting disease burden in adult SCD patients, with divergent mortality trends across the lifespan. Our decreasing trend in ages 15–34 was reflected in prior studies that have documented substantial improvements in pediatric survival due to advances in neonatal screening, vaccination, and early intervention, resulting in more individuals living into adulthood.^1–3 Conversely, the stable and high mortality in 35- to 54-year-olds and increasing trend in the 55 to 74 age group demonstrate that as this population ages, the management of chronic SCD complications becomes increasingly complex, and mortality patterns begin to reflect the cumulative burden of end-organ damage, comorbidities, and gaps in adult-directed care.^4,5 While improved survival is an encouraging trend, our findings underscore the need for focused attention on the aging SCD population, including longitudinal surveillance, development of age-specific care models, and health system preparedness for this growing demographic.

Our analysis revealed substantial geographic variation in adult SCD mortality across US census regions. The Southern region had both the highest overall AAMR and the only significant increase over the study period. In contrast, the West had the lowest mortality and no evidence of a rising trend. At the state level, the highest AAMRs were observed in the District of Columbia, South Carolina, and Mississippi. These findings are consistent with the known geographic distribution of SCD prevalence and underscore the importance of regionally tailored public health surveillance and resource planning. Future studies are needed to evaluate how healthcare access, care delivery models, and local disease burden contribute to these geographic differences.

While our study found no significant difference in overall AAMRs between men and women with SCD, only women exhibited a significant increase in mortality over the study period. Historically, males with SCD have been reported to experience more severe disease and earlier mortality, with prior data showing a median age at death of 42 years in males compared to 48 years in females with sickle cell anemia.² This survival advantage in females has been partly attributed to higher fetal hemoglobin levels and decreased hemolysis of red blood cells, which provides protection against certain complications.¹⁴ Women with SCD face unique health risks, particularly during pregnancy, where maternal mortality has been reported to be markedly elevated compared to the general population.¹⁵ While our dataset did not allow for exploration of these underlying causes, the rising mortality trend among women underscores the need for further research to investigate sex-specific disease progression, healthcare utilization, and social determinants in adult SCD populations.

A notable sharp increase was seen in 2020, coinciding with the COVID-19 pandemic. Literature has reported a 2- to 7-fold increase in COVID-19–related hospitalizations and a 1.2- to 2.5-fold increase in COVID-19–related deaths in adult SCD patients (compared to their non-SCD peers).¹⁶ The potential for SARS-COV-2 to trigger vaso-occlusive crisis, clinical overlap between acute chest syndrome and COVID-19, and preexisting organ damage caused by SCD have been postulated to explain these findings.¹⁶ In addition, the lack of healthcare access likely played a part. Multiple states saw up to a 25% decrease in emergency department visits by SCD patients in 2020.¹⁷ This reduction in care seeking likely led to worse complications and greater mortality.

Collectively, these findings offer a comprehensive national assessment of adult SCD mortality trends over a 22-year period, revealing important demographic and regional disparities. The observed rise in mortality among older adults and women, alongside persistently elevated rates in the Southern United States, highlights the shifting epidemiologic profile of SCD and underscores ongoing inequities in care delivery. These results reinforce the need for expanded surveillance, age- and sex-responsive care models, and regionally targeted public health strategies. Advancing health equity in SCD will require coordinated efforts across clinical, research, and policy domains to address both medical complexity and systemic barriers to high-quality, longitudinal care (Central Illustration).

Limitations

This study has certain limitations due to the inherent nature of CDC WONDER. The data are divided into aggregate groups, which prevents calculations of vital metrics like mean age (which require individual-level patient data). To ensure patient confidentiality, the database has protocols to “suppress” data or deem the data “unreliable” in the case of few death counts. This precluded a comprehensive state-by-state comparison and limited the analysis of some subgroups (e.g., older males).

The data are also subject to unmeasured confounding. Vital clinical variables like SCD genotype, use of disease-modifying therapy, and socioeconomic data such as insurance status are not available. These factors likely impact mortality and prevented the authors from drawing causal conclusions from the observed data. The Multiple Cause of Death dataset, while robust, is dependent on proper coding by the death-certifying physician. Finally, the trend analysis assumes a monotonic trend. While this was validated with visual assessment, more complex patterns cannot be detected.

Conclusion

In our study we identified a rising AAMR over time, driven predominantly by increases among older adults and women. Racial and geographic disparities remain pronounced, with the Southern United States exhibiting the highest and only increasing regional mortality. These findings underscore the need for enhanced national surveillance, improved models of adult SCD care, and health policy efforts focused on addressing demographic and regional inequities. Future research should incorporate individual-level clinical and structural data to inform targeted interventions that reduce mortality and improve the quality of life for adults living with SCD.

Disclosure statement/Funding

The authors report no funding or conflicts of interest.