African Americans With Acute Pancreatitis Present With Worsened Kidney Injury and Have Inadequate Access to Care

To the Editor:

Acute pancreatitis (AP) is a health care challenge¹ and a major cause for gastrointestinal hospitalization. Severe AP^1,2 is associated with up to 30% mortality and current management is limited to supportive care.³ Few studies identified alarming trends in the care and outcomes in AP of minority patients including increased rates of organ failure⁴ and mortality⁵ and lower rates of patient transfers⁶ in African Americans (AA) and longer Emergency Department wait times in Hispanics.⁷ These investigations established a link between health care disparities and clinical outcomes in AP⁸ including increased inpatient mortality in AA compared to Whites (odds ratio, 1.18, P = 0.025)⁵, lower transfer rates of AAs to institutions with a higher level of care⁶ and higher rates of organ failure among AAs with AP compared to the general population.⁴ However, they did not provide a detailed assessment of socio-demographic, clinical, and management metrics to determine whether specific differences in these variables underlie this disparity thus limiting our ability to address it. In this study, we assessed whether significant differences in socioeconomic status, clinical presentation, or management of AP subjects contribute to AP disparities.

African Americans and Whites 18 years of age or older who were admitted to University of Illinois at Chicago with AP between January 1, 2010 and December 31, 2015 were identified through the International Classification of Disease (ICD)-9 code for AP (577.0). Socio-demographic, clinical, laboratory and management data were collected and analyzed using Statistical Analysis Software version 9.4 (SAS, Cary, NC). T-test or Mann-Whitney U test for continuous variables and Fisher’s exact test or Analysis of Variance for categorical data were used with two-sided tests with a priori significance level of alpha = 0.05. A multivariate regression analysis was completed to determine independent predictors of creatinine by adjusting for race, body mass index (BMI), age, and comorbidities.

There were 253 subjects analyzed, of which 131 were AAs and 122 Whites (Table 1); 135 were female. There were no significant differences among AAs and Whites (Table 1) for many variables including in age, BMI, or sex (P = 0.42, P = 0.14, P = 0.08, respectively); alcohol, tobacco and illicit drug consumption (P = 0.37, P = 0.09 and P = 1.0, respectively); non-steroidal anti-inflammatory, opioid or statin use (P = 0.69, P = 0.48, P = 0.89, respectively); similar rates of diabetes mellitus, chronic kidney disease and hypertension; nor were there differences in family history of AP and underlying etiologies (P = 0.37 and P = 0.19, respectively). However, compared to Whites, AAs were more likely to have coronary artery disease (P = 0.05). In addition, AAs were more likely to live in underserved neighborhoods of Chicago (P = 0.02); had significantly lower predicted mean annual income (P < 0.0001); and the percentage of subjects transferred to UIC for higher level of care was significantly lower in AAs compared to Whites (20.6% versus 36.9%, P = 0.005) (Table).

TABLE 1.

Demographic and Clinical Data Among African Americans (AAs) and Whites With Index Acute Pancreatitis

	AAs (n = 131)	Whites (n = 122)	P
Age, mean (SD), y	48.9 (15.8)	50.8 (17.6)	0.42
Sex, female, n (%)	77 (58.8)	58 (47.5)	0.08
Body mass index, mean (SD), kg/m2	29.8 (15.8)	29.7 (8.5)	0.14
Location of residence, n (%)			0.02
Southern suburbs	74 (56.5)	58 (47.5)
Northern suburbs	43 (32.8)	36 (29.5)
Western suburbs	14 (10.7)	28 (23)
Predicted mean income, mean (SD), $	37,924 (16,111)	58,502 (24,978)	<0.0001
Transferred to UIC for higher level of care, n (%)	27 (20.6)	45 (36.9)	0.005
Past medical history, n (%)
Diabetes mellitus	22 (16.8)	27 (22.1)	0.34
Coronary artery disease	21 (16)	9 (7.4)	0.05
Chronic kidney disease	10 (7.6)	9 (7.4)	1.0
Hypertension	57 (43.5)	39 (32)	0.07
Family history of pancreatitis, n (%)	4 (3.1)	1 (0.8)	0.37
No. alcoholic drinks per week, mean (SD)	11 (20)	13 (28)	0.37
Duration of tobacco exposure, mean (SD), y	6 (10.4)	4.7 (10.9)	0.09
Etiologies, n (%)			0.19
Gallstones	43 (33.1)	51 (41.8)
Alcoholic	33 (25.4)	22 (18.0)
Idiopathic	40 (30.8)	29 (23.8)
Hypertriglyceridemia-induced	2 (1.5)	7 (5.7)
Post-ERCP	2 (1.5)	3 (2.5)
Other	10 (7.7)	10 (8.2)
SBP at presentation, mean (SD), mmHg	141 (23)	132 (24) (n = 121)	0.003
DBP at presentation, mean (SD), mmHg	81 (15)	74 (12) (n = 121)	<0.001
Creatinine level, mean (SD), mg/dL	1.6 (2.2)	1.2 (1.2)	0.02
Disease Severity, n (%)*			0.66
Mild	90 (68.7)	70.5 (86)
Moderate	33 (25.2)	21.3 (26)
Severe	8 (6.1)	8.2 (10)
Mortality, n (%)	5 (3.8)	3 (2.5)	0.15
Intensive care unit (ICU) admission rate, n (%)	21 (16.0)	14.8 (18)	0.86
ICU length of stay, mean (SD), d	1 (3.8)	1.1 (3.5)	0.87
Hospital length of stay, mean (SD), d	6 (5.9)	6.8 (7.6)	0.21
30-day readmission rate, n (%)	25 (19)	15 (12.3)	0.14

Bold value are statistically significant.

^*Disease severity based on the Revised Atlanta Classification.

SD, standard deviation; UIC, University of Illinois at Chicago; ERCP, endoscopic retrograde cholangiopancreatography, SBP, systolic blood pressure; DBP, diastolic blood pressure

Despite having higher mean systolic and diastolic blood pressures at presentation (P = 0.003, P < 0.001, respectively), AAs had significantly higher creatinine level at admission compared to Whites (1.6 mg/dL vs 1.2 mg/dL, P = 0.02). In a multivariate regression analysis, only race was found to be an independent predictor of creatinine (P = 0.037) when additional adjustments were made for age, BMI, and comorbidities.

African Americans and Whites received similar amounts of intravenous fluids within 8 hours (1.3 [standard deviation {SD}, 0.6] versus 1.4 [SD, 0.8] liters, P = 0.77) and 24 hours of admission (5.3 [SD, 2] versus 5.0 [SD, 2.1] liters, P = 0.19). Disease severity was similar between two groups (P = 0.66). Mortality rate was numerically higher in AAs compared to Whites (3.8% versus 1.6%, P = 0.29) and there were no significant differences in intensive care unit (ICU) admission rate, ICU length of stay and hospital length of stay (P = 0.86, P = 0.87, P = 0.21, respectively) between two groups. However, there was a positive correlation between creatinine levels and ICU admissions (P = 0.002) and AAs had higher 30-day re-admission rate compared to Whites (18.9% versus 12.3%, P = 0.14).

We identified significantly higher creatinine levels on admission in AAs compared to Whites. Race remained an independent predictor of creatinine in the multivariate regression analysis when adjusted for covariables. Therefore, this finding truly reflects the presence of end-organ damage in an acute setting. There were no significant differences in total intravenous fluids administration and disease severity between two groups. African Americans had higher mortality and 30-day re-admission rates compared to Whites but these did not reach statistical significance, most likely due to small study size.

African Americans with AP were more likely to live in underserved neighborhoods and have significantly lower predicted income levels. In addition, percentage of transfer among subjects who received care at our institution for AP was significantly lower in AAs compared to Whites and this is consistent with previous reports.⁶ This finding along with limited access to care in disadvantaged, lower income populations can explain part of existing disparities.

Our study has several limitations. It is i) a retrospective study which prevented longitudinal collection of key variables, ii) utilized ICD codes which may have resulted in under detection of AP cases,⁹ iii) focused on AAs and Whites, which decreased the study size. A larger study with an increased power may allow more in depth investigation of actionable factors which can be utilized to develop preventive strategies, improve access to care and optimize delivery of the care especially at high-risk groups.