Acute lymphocytic leukemia severity and mortality hospitalizations in the United States: A population-based study

Introduction

Acute lymphocytic leukemia (ALL) is multiplying lymphoid progenitor cells in the bone marrow, blood, and extramedullary sites. It is the most prevalent type of cancer in the pediatric population, it is also prevalent as a devastating disease in the adult population.[1] Factors such as radiation, exposure to benzene, infection with EBV, HTLV-1, down syndrome, male, sex, and Caucasian or White are risk factors for ALL.[2,3] Survival rates and effectiveness of therapeutics have been found to vary by age.[4-7] It is therefore imperative to understand the dynamic of presentation at the population level. Such an understanding would aid in developing effective strategies to address the potential inequity in access to care that may exist. These advances in treatment modalities often come at a cost and availability limited to large metropolitan-populated centers. The reason for these variations is multiprone including technical expertise and resource availability. Other important factors associated with better outcomes in these patient populations include age at diagnosis, prognosis, presence of the Philadelphia chromosome, and response to chemotherapy. Studies have shown improvement in health outcomes for these populations attributed to advancements in treatment modalities and targeted therapies,[6,7] but as with other medical conditions, factors influencing health-care access and equity directly or indirectly impact the care patients receive. This study explores the relationship between social and economic indicators of health and its association with hospitalization severity and mortality risk among ALL patients.

Methods

In this retrospective study, hospitalizations with primary and secondary ALL diagnosis were identified using International Classification of Diseases (ICD 10) codes (C91.00, C91.01, and C91.02) in the National Inpatient Sample (NIS) between 2016 and 2018. All cases with any missing variables were excluded from any analysis. The NIS dataset is part of a family of all payers’ inpatient datasets available through the Healthcare Cost and Utilization Project (HCUP) sponsored by the Agency for Healthcare Research and Quality (AHRQ). The population set covered in this dataset is estimated to cover about 97%of the US population (47 States). The dataset is de-identified to protect the privacy of patients, providers, and hospitals. The demographic information available in the dataset was analyzed for the patient population that meets the inclusion criteria. Outcomes of hospitalization such as LOS, mortality, severity and mortality risk at presentation, cost of care, Number of diagnosis (NDX), Number of procedures (NPR) were analyzed by race and ethnicity, household income, and patient location among patients with primary and secondary ALL diagnoses. The most common Major Diagnostic Category (MDC) procedures were also analyzed by race, household income, and patient location.

The study cohort consisted of 158090 patients, with ALL hospitalized between 2016 and 2017. A total of 84645 hospitalizations were identified as meeting the inclusion criteria without missing cases with a primary or secondary diagnosis of ALL using ICD-10 codes. Primary ALL diagnosis included ALL without remission (21855), ALL in remission (3080), ALL in relapse (5275) between 2016 and 2018. Of the 54435 (64%) secondary ALL diagnoses, the most common primary cause of hospitalizations included disease and disorders of blood (14%), infectious and parasitic infections (9%), diseases or disorders of the espiratory system (4%), circulatory (3%) and digestive system (3%), nervous system (3%), and hepatobiliary and pancreases (1%).

Logistic regression was performed to estimate the predictors of mortality in the patient population. The independent variables include number of diagnoses (NDX), number of procedures (NPRs), region, sex, patient location, payer, household income (quartiles), and length of stay (LOS). The mortality rate for each patient’s demographics was calculated over time (2016−2018).

Statistical Package for the Social Sciences (SPSS) version 23.0 (IBM Corp., Armonk, NY) was utilized to analyze data in this study. Descriptive statistics of the variables of interest were conducted, and records with missing data were not included in any further analysis. A Chi-squared test was utilized to evaluate the relationship between variables of interest and the study sample. All analyses were conducted based on recommended HCUP guidelines; discharge weight was applied to obtain a nationally representative sample. Power B.I. (2019, Microsoft Corporation, Redmond, WA) was used to graph and calculate the rate of presentation among the population of interest.

Results

A total of 158,090 hospitalizations were identified as meeting the inclusion criteria without missing cases with a primary or secondary diagnosis of ALL using ICD-10 codes. Primary ALL diagnosis included ALL without remission (21855), ALL in remission (3080), and ALL in relapse (5275) between 2016 and 2018. Of the 54435 (64%) secondary ALL diagnoses, the most common primary cause of hospitalizations included Disease and Disorders of blood (14%), infectious and parasitic infections (9%), diseases or disorders of the respiratory system (4%), circulatory (3%) and digestive system (3%), nervous system (3%), and hepatobiliary and pancreases (1%). The rate of the extreme likelihood of dying at presentation increased between 2016 and 2018 in all regions across the country, as shown in Figure 1 (P < 0.001). Severity risk at presentation varied from one area to the next, with the highest rate (per 10K) presentations in the New England region for both extreme likelihood (778) and extreme loss of function (2198) at presentation. The severe severity of disease at presentation was highest in large metropolitan (2112). The extreme likelihood of dying was highest in non-metropolitan or micropolitan (702) per 10000 hospitalizations, as shown in Table 1.

Close

Figure 1

Rate (per 10,000) of mortality risk by region, race/ethnicity, and year among patients with ALL in the United States

Export to PPT

Table 1 Descriptive statistics(weighted) of ALL hospitalizations by socioeconomic and demographic characteristics

Evaluation of mortality and severity risk by payer shows a consistent rate by year with a higher rate among Medicare (1394, 3142) and others (484, 1817). Mortality and severity among uninsured patients were the second highest (614, 2193) compared to other payers. Extreme mortality risk at presentation was higher among Black (711), White (648), and Native American (612) populations compared to other racial and ethnic groups, as shown in Table 1. The highest extreme severity risk at presentation was observed among Asian and Pacific islanders (2297), Native Americans (2143), White (2057), and Black (2040). The highest mortality rate among ALL presentations by household income was observed in the middle two quartiles. The highest severity rate was observed among household income of <71K (1713) and <54 (1495) between 2016 and 2017(P < 0.001).

In the logistic regression, females (OR: 1.27, P < 0.001), East South Central (OR: 1.62, P < 0.001), West South Central (OR: 1.28, P < 0.001), Black (OR: 1.82, P < 0.001), Hispanic (OR: 1.32, P 0.002), Asian/Pacific Islander (OR: 1.44, P -0.003), and median income quartile (OR: 1.32, P < 0.001) predicted mortality among patients with ALL, as shown in Table 2. The rate of patients with the emergency department (E.D.) as a point of admission continuously increased between 2016 and 2018. The severity and mortality risk among patients with E.D as the end of admission continued to increase.

Table 2 Weighted multivariable logistic regression

The highest mortality rate was observed among individuals aged 75 + (672) and 65−74 (358) per 10,000 patient presentations. The lowest mean mortality rate was observed among patients <18 (57), 18−24(173), as shown in Figure 2 (P < 0.001). The highest likelihood of dying and loss of function at presentation was also observed among the older population compared to younger ones. Moreover, the highest rate of presentation was observed among household income less than $71,000 annually, as shown in Figure 2.

Close

Figure 2

Rate (per 10,000) of severity risk by age, income quartiles and year among patients with ALL in the United States

Export to PPT

Table 2 shows that the predictors of mortality were significant (P < 0.05) in most regions, except for East North Central, West North Central, and South Atlantic regions. Mortality by patient location was also significant in large populated areas compared to less populated areas. Mortality was more likely among patients with Medicaid, Medicare, and self-pay than other payer groups. Higher odds of mortality were observed among the White, Black, and lower-income quartiles population.

Discussion

This study’s utilization of primary and secondary diagnosis of inpatient ALL presentations allows for a better understanding of presentations for a more comprehensive population range. It aids in understanding the prevalence of the disease by age group and other demographics characteristics, which could help tailor therapeutics aimed at outcome improvement. Evidence from multiple studies shows advances in treatment modalities for the younger population and better outcomes.[4,5,8,9] Other publications have also recommended improved focus on sex differences and rural–urban and adherence to up to date guidelines to improve care and preventive strategies.[10,11] It is therefore imperative to understand the indicators that impact outcome improvement among diverse populations with ALL. Even though the presentation rate decreased in 2018 for some age groups, older adults had poorer outcomes and severity at presentation than the younger population.[12-14] These findings further support other publications on the effectiveness of treatment modalities for older populations with ALL.[12,14]

As shown in this study, a high severity and mortality risk among these populations may be associated with other comorbidities, given the higher number of potential comorbidities that may complicate treatment approaches. Mortality and severity rate variance by population demographics shows a changing trend with higher rates observed in 2018 in micropolitan or less populated areas. These presentations may be associated with variance in access to cancer treatment in these areas. The difficulties associated with reaching large urban centers for care have been found in other studies focusing on the socioeconomic impact of other malignancies.[15,16] The observed higher mortality for the Black population is similar to other studies that have found high mortality for Blacks for other cancer diagnoses.[17,18] Further analysis shows higher severity and loss of function rate at a presentation by race for the lowest income quartiles among all racial groups – a finding supported by other studies outside the U.S on cancer care for poor and minority communities.[14,19,20] A study conducted in Sudan by Muddathir et al.[21] on acute myeloid leukemia (AML) found the mean age of presentation to be 32 compared to this ALL study with the mean age of 18. A multinational study conducted by Bond et al.[22] on adult ALL found varied treatment outcome based on specific implicated gene and therapeutic intervention. The findings of these studies further support the argument of more efforts by providers to develop strategies to provide tailored care to these populations, given the diversity of factors that impacts presentation or severity. Understanding presentation dynamics and the impact of modifiable factors that impact impatient presentation, severity, and mortality risk for these patient populations could promote the development of practical population-based treatment approaches to improve health outcomes.

Conclusion

This study shows the trends associated with severity and mortality rate associated with inpatient ALL presentations. The findings of this study suggest a relative decrease in presentation rate by year and higher mortality among specific group-based demographics indicators. This study confirms the impact of advanced therapeutics and improved severity and mortality among younger populations with ALL compared to the older population.

In addition, the findings of this study could aid in the development of population-based risk stratification of patients based on these socioeconomic and demographic indicators. Our observation highlights the need to access and develop treatment modalities to improve health outcomes among the more aging population with ALL. It also highlights the need to evaluate care approaches to ensure patients in less populated areas and micropolitan locations have an equitable distribution of care resources.

Authors’ Declaration Statements

Authors’ Contributions

Saanie Sulley: Conceptualization, data analysis, data visualization, methodology, writing, review, and editing. Abimbola Saka: Review, Editing, data validation, methodology and review. Memory Ndanga: Review, Editing, data validation, methodology and review.