To the Editor:
Racial/ethnic segregation (hereafter referred to as segregation), the inequitable allocation of resources and opportunities in racially/ethnically segregated neighborhoods, leads to numerous disadvantages for residents1 and associates with access to kidney transplantation (KT).2 Based on the psychosocial framework,3 segregation likely associates with post-KT outcomes,2 including dementia, a diagnosis that is substantially more common among older KT recipients than in community-dwelling older adults,4 and increases the risk of allograft loss and mortality.4 We studied the potential associations between segregation and dementia, death-censored allograft loss, and all-cause mortality among older KT recipients.
Our cohort study included 23,044 older (aged ≥55 years) KT recipients (January 2007 – December 2019) identified using the United States Renal Data System/Medicare claims datasets, from which we obtained recipient characteristics and post-KT outcomes, including dementia diagnoses (Table S1).5 Recipient-specific residential segregation scores were calculated based on Iceland’s Multigroup Entropy Index using population counts from the 2020 American Community Survey (ACS) 5-year estimates and were linked to recipients’ 5-digit ZIP code (Item S1).6
Based on our conceptual framework (Fig S1), we used adjusted Cox proportional hazards models (mortality) and Fine and Gray’s subdistribution hazards models (death-censored allograft loss and dementia) to quantify the association between segregation levels (low- vs medium- vs high-segregation) and post-KT outcomes. To quantify the association between segregation and racial disparities in these outcomes, we included interactions in the fully adjusted models. These interactions were examined using the Wald test with robust variance to determine statistical significance (Item S1). Specifically, we analyzed 2 types of interactions: (1) the association between segregation levels and race/ethnicity (non-Hispanic [NH] White [reference], NH Black, and NH Other [including Hispanic and Asian]); and (2) the association between racial composition of high-segregation neighborhoods and race/ethnicity.
Recipients residing in high-segregation neighborhoods were more likely to be younger, Black, and less educated (Table S2). After adjustment, recipients from high-segregation neighborhoods had an increased risk of dementia (all-type dementia [adjusted subhazard ratio (aSHR) = 1.19, 95% CI: 1.04–1.37]; vascular dementia [aSHR = 1.57, 95% CI: 1.06–2.31], and other/mixed-type dementia [aSHR = 1.18, 95% CI: 1.01–1.38]) compared to those from low-segregation neighborhoods (Table 1, Fig S2, Fig S3). Black recipients living in high-segregation neighborhoods had a higher vascular dementia risk (aSHR = 1.67, 95% CI: 1.05–2.65) compared to White recipients living in low-segregation neighborhoods(Table 2).
Table 1.
Association Between Residential Segregation and Post-KT Dementia, Death-Censored Allograft Loss, and Mortality Among Older (Aged ≥55 Years) KT Recipients Between 2007 and 2019 (n = 23,044)
| Segregation Levels | Crude Model | Minimally Adjusted Modela | Fully Adjusted Modelb |
|---|---|---|---|
| sHRc (95% CI) | Adjusted sHRc (95% CI) | Adjusted sHRc (95% CI) | |
| All-type dementia (n = 1,288) | |||
| Low | Reference | Reference | Reference |
| Medium | 1.08 (0.94, 1.23) | 1.12 (0.98, 1.29) | 1.07 (0.93, 1.22) |
| P = 0.3 | P = 0.2 | P = 0.3 | |
| High | 1.17 (1.02, 1.34)d | 1.26 (1.10, 1.44)d | 1.19 (1.04, 1.37)d |
| P = 0.02d | P = 0.003d | P = 0.01d | |
| Alzheimer disease (n = 135) | |||
| Low | Reference | Reference | Reference |
| Medium | 0.80 (0.53, 1.21) | 0.85 (0.56, 1.30) | 0.83 (0.55, 1.25) |
| P = 0.3 | P = 0.3 | P = 0.4 | |
| High | 0.90 (0.60, 1.35) | 1.00 (0.66, 1.51) | 0.93 (0.62, 1.41) |
| P = 0.6 | P = 0.9 | P = 0.7 | |
| Vascular dementia (n = 165) | |||
| Low | Reference | Reference | Reference |
| Medium | 1.43 (0.96, 2.12) | 1.45 (0.98, 2.16) | 1.37 (0.93, 2.04) |
| P = 0.08 | P = 0.07 | P = 0.1 | |
| High | 1.59 (1.08, 2.35)d | 1.66 (1.12, 2.46)d | 1.56 (1.06, 2.31)d |
| P = 0.02d | P = 0.01d | P = 0.02d | |
| Other/mixed dementia (n = 988) | |||
| Low | Reference | Reference | Reference |
| Medium | 1.04 (0.99, 1.22) | 1.08 (0.92, 1.27) | 1.03 (0.88, 1.21) |
| P = 0.6 | P = 0.5 | P = 0.7 | |
| High | 1.14 (0.97, 1.34) | 1.23 (1.05, 1.45)d | 1.18 (1.01, 1.38)d |
| P = 0.1 | P = 0.03d | P = 0.04d | |
| Death-censored allograft loss (n = 3,209) | |||
| Low | Reference | Reference | Reference |
| Medium | 1.06 (0.97, 1.16) | 1.06 (0.97, 1.16) | 1.04 (0.95, 1.13) |
| P = 0.2 | P = 0.2 | P = 0.4 | |
| High | 1.27 (1.16, 1.38)d | 1.26 (1.15, 1.37)d | 1.19 (1.10, 1.30)d |
| P< 0.001d | P< 0.001d | P< 0.001d | |
| HRe (95% CI) | Adjusted HRe (95% CI) | Adjusted HRe (95% CI) | |
| All-cause mortality (n = 4,939) | |||
| Low | Reference | Reference | Reference |
| Medium | 0.99 (0.93, 1.05) | 1.00 (0.94, 1.06) | 0.97 (0.91, 1.03) |
| P = 0.7 | P = 0.9 | P = 0.3 | |
| High | 0.97 (0.92, 1.03) | 1.00 (0.95, 1.07) | 0.96 (0.91, 1.02) |
| P = 0.4 | P = 0.9 | P = 0.2 | |
Abbreviations: KT, kidney transplant; sHR, subhazard ratio; HR, hazard ratio.
Minimally adjusted model adjusted for recipient age at transplant and recipient sex.
Fully adjusted model adjusted for recipient age at transplant, recipient sex, education, cause of end-stage kidney disease (glomerulonephritis, diabetes, hypertension, and other), years on dialysis, tobacco use, cerebrovascular disease, body mass index, hypertension, coronary artery disease, heart failure, immunosuppression use, and donor type.
Fine-Gray subdistribution hazard model and subhazard ratios (sHR) were used for dementia and death-censored allograft loss.
Statistically significant results with P values <0.05 denoted.
Cox proportional hazard model and hazard ratios were used for examining all-cause mortality.
Table 2.
Association Between Residential Segregation and Post-KT Dementia and Death-Censored Allograft Loss by Recipient Race Among Older (Aged ≥55 Years) KT Recipients Between 2007 and 2019 (n = 23,044)
| Recipient Race | Adjusted sHR (95% CI), P Value | ||
|---|---|---|---|
| Low Segregation | Medium Segregation | High Segregation | |
| All-type dementia | |||
| White | Reference | 1.01 (0.86, 1.20), P = 0.9 - |
1.20 (1.01, 1.44), P = 0.04a - |
| Black | 1.01 (0.79, 1.29), P = 0.9 | 1.11 (0.91, 1.36), P = 0.3 | 1.13 (0.94, 1.35), P = 0.2 |
| Pinteraction | - | 0.4 | 0.5 |
| Other | 0.52 (0.35, 0.77), P = 0.001a | 0.71 (0.49, 1.03), P = 0.07 | 0.71 (0.45, 1.12), P = 0.1 |
| Pinteraction | - | 0.07 | 0.03a |
| Vascular dementia | |||
| White | Reference | 1.07 (0.66, 1.73), P = 0.8 - |
1.04 (0.60, 1.81), P = 0.9 - |
| Black | 0.56 (0.23, 1.33), P = 0.2 | 1.37 (0.80, 2.35), P = 0.3 | 1.65 (1.04, 2.63), P = 0.03a |
| Pinteraction | - | 0.4 | 0.09 |
| Other | 0.34 (0.08, 1.42), P = 0.1 | 0.89 (0.35, 2.31), P = 0.8 | 0.28 (0.04, 2.13), P = 0.2 |
| Pinteraction | - | 0.7 | 0.2 |
| Other/mixed dementia | |||
| White | Reference | 0.95 (0.78, 1.16), P = 0.6 | 1.21 (0.98, 1.50), P = 0.08 |
| Black | 1.13 (0.85, 1.49), P = 0.4 | 1.21 (0.96, 1.54), P = 0.1 | 1.13 (0.91, 1.40), P = 0.3 |
| Pinteraction | - | 0.04a | 0.5 |
| Other | 0.51 (0.32, 0.83), P = 0.006a | 0.65 (0.41, 1.03), P = 0.07 | 0.82 (0.49, 1.37), P = 0.4 |
| Pinteraction | - | 0.1 | 0.1 |
| Death-censored allograft loss | |||
| White | Reference | 0.97 (0.86, 1.09), P = 0.6 - |
1.16 (1.02, 1.31), P = 0.02a - |
| Black | 1.32 (1.14, 1.52), P < 0.001a | 1.42 (1.25, 1.60), P < 0.001a | 1.40 (1.25, 1.56), P < 0.001a |
| Pinteraction | - | <0.001a | 0.003a |
| Other | 0.85 (0.68, 1.07), P = 0.2 | 0.85 (0.67, 1.07), P = 0.2 | 1.05 (0.80, 1.37), P = 0.4 |
| Pinteraction | - | 0.3 | 0.5 |
Model adjusted for recipient age at transplant, recipient sex, education, cause of end-stage kidney disease (glomerulonephritis, diabetes, hypertension, and other), years on dialysis, tobacco use, cerebrovascular disease, body mass index, hypertension, coronary artery disease, heart failure, immunosuppression use, and donor type. Abbreviations: KT, kidney transplant; sHR, subhazard ratio.
Statistically significant results with P-values <0.05 denoted.
Segregation was not associated with post-KT mortality (Table 1). However, recipients living in high-segregation neighborhoods had an increased death-censored allograft loss risk (aSHR = 1.19, 95% CI: 1.10–1.30) compared to those in low-segregation neighborhoods (Table 1, Fig S4). Specifically, the association between segregation and death-censored allograft loss differed by race with Black recipients having higher risk than White recipients (Table 2). Minoritized recipients living in minority-predominant high-segregation neighborhoods had a differential risk of death-censored allograft loss (Pinteraction = 0.008) (Table S3). Sensitivity analyses showed similar results (Item S2, Tables S4–S7).
Among older KT recipients, residing in a segregated neighborhood was associated with a higher risk of dementia and allograft loss, but not mortality. Segregation, a known driver of health disparities in the United States, reduces the distributions of resources and opportunities, thereby increasing health risk.1 Our finding that segregation is associated with dementia aligns with previous studies of community-dwelling older adults7 and may be attributed to underinvestment in these neighborhoods and the downstream impact of adverse social determinants of health. Adverse social determinants of health, such as reduced social cohesion,8 could lead to increased psychological stress,8 potentially accelerating cognitive decline.
Our results showed an increased risk of allograft loss in high-segregation neighborhoods. Possible reasons for this association may include limited access to nephrology care and other health-promoting products and services, or increased exposure to unfavorable environmental conditions. Additionally, we found that living in highly segregated neighborhoods may not inherently lead to higher mortality. Individual factors appear to be a more significant predictor of post-transplant mortality than neighborhood racial composition.9
Our study has several strengths, including a large and representative sample of older KT recipients and the use of reliable/sensitive ICD codes for dementia diagnosis.10 However, limitations include that the ACS data may not account for changes in neighborhood or racial composition. Our segregation scores, derived from diversity, offer insights but may not fully capture the intricate dynamics of segregation. We lacked data on family wealth and genetic factors like APOL1 risk variants, which could impact the association between segregation and allograft survival. Moreover, given that dementia is a rare disease, we were limited to a low number of events for statistical power despite a large sample size.
In summary, residential racial/ethnic segregation, an important indicator of structural racism, is an independent risk factor for dementia and allograft loss among older KT recipients. In addition to intervening on transplant recipient modifiable risk factors, interventions may be needed that target the downstream impacts of segregation to reduce post-KT dementia and allograft loss risk.
Supplementary Material
Supplementary File (PDF)
Figures S1–S4, Items S1–S2, Tables S1–S7.
Acknowledgements:
The data reported here have been supplied by the United States Renal Data System (USRDS).
Support:
This work was supported by R01AG077888 (PI: McAdams-DeMarco) from the National Institute on Aging. Authors were also supported by the following grants: K02AG076883 (PI: McAdams-DeMarco), R01AG055781 (PI: McAdams-DeMarco), DP1AG069874 (PI: Szanton, Co-I Thorpe), U54MD000214 (Thorpe), K02AG059140 (Thorpe), P30AG059298 (Thorpe), R01DK114074 (PI: McAdams-DeMarco), R01DK120518 (PI: McAdams-DeMarco), F32AG067642 (PI: Ruck), and K24AI144954 (PI: Segev). The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; or preparation, review, or approval of the manuscript or the decision to submit for publication.
Footnotes
Financial Disclosure: Dr Segev receives speaking honoraria from Sanofi, Novartis, and CSL Behring. Dr McAdams-DeMarco received an honorarium from Chiesi. The remaining authors declare that they have no relevant financial interests.
Disclaimer: The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy or interpretation of the U.S. government.
Contributor Information
Yusi Chen, Department of Surgery, NYU Grossman School of Medicine and NYU Langone Health, New York, New York.
Yiting Li, Department of Surgery, NYU Grossman School of Medicine and NYU Langone Health, New York, New York.
Yi Liu, Department of Surgery, NYU Grossman School of Medicine and NYU Langone Health, New York, New York.
Byoungjun Kim, Department of Surgery, NYU Grossman School of Medicine and NYU Langone Health, New York, New York; Department of Population Health, NYU Grossman School of Medicine and NYU Langone Health, New York, New York.
Jessica M. Ruck, Department of Surgery, Johns Hopkins University School of Medicine, Baltimore, Maryland.
Maya N. Clark-Cutaia, New York University Rory Meyers College of Nursing, New York, New York.
Aarti Mathur, Department of Surgery, Johns Hopkins University School of Medicine, Baltimore, Maryland.
Tanjala S. Purnell, Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland.
Roland J. Thorpe, Jr., Hopkins Center for Health Disparities Solutions, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland.
Deidra C. Crews, Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland.
Sarah L. Szanton, Johns Hopkins University School of Nursing, Baltimore, Maryland.
Dorry L. Segev, Department of Surgery, NYU Grossman School of Medicine and NYU Langone Health, New York, New York; Department of Population Health, NYU Grossman School of Medicine and NYU Langone Health, New York, New York.
Mara McAdams-DeMarco, Department of Surgery, NYU Grossman School of Medicine and NYU Langone Health, New York, New York; Department of Population Health, NYU Grossman School of Medicine and NYU Langone Health, New York, New York.
Data Sharing:
The datasets used and/or analyzed during the current study are available from the United States Renal Data System (USRDS). Per the Data Use Agreement (DUA) between the authors and USRDS, the deposition of data into publicly available repositories is not allowed.
References
- 1.Williams DR, Collins C. Racial residential segregation: a fundamental cause of racial disparities in health. Public Health Rep. 2001;116(5):404–416. doi: 10.1093/phr/116.5.404 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Olufajo OA, Adler JT, Yeh H, et al. Disparities in kidney transplantation across the United States: does residential segregation play a role? Am J Surg. 2017;213(4):656–661. doi: 10.1016/j.amjsurg.2016.10.034 [DOI] [PubMed] [Google Scholar]
- 3.Forrester SN, Gallo JJ, Whitfield KE, Thorpe RJ. A framework of minority stress: from physiological manifestations to cognitive outcomes. Gerontologist. 2019;59(6):1017–1023. doi: 10.1093/geront/gny104 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.McAdams-DeMarco MA, Bae S, Chu N, et al. Dementia and Alzheimer’s disease among older kidney transplant recipients. J Am Soc Nephrol. 2017;28(5):1575–1583. doi: 10.1681/asn.2016080816 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Johansen KL, Chertow GM, Foley RN, et al. US Renal Data System 2020 Annual Data Report: epidemiology of kidney disease in the United States. Am J Kidney Dis. 2021;77(4)(suppl 1):A7–A8. doi: 10.1053/j.ajkd.2021.01.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Iceland J The Multigroup Entropy Index (Also Known as Theil’s H or the Information Theory Index). United States Census Bureau. https://www2.census.gov/programs-surveys/demo/about/housing-patterns/multigroup_entropy.pdf [Google Scholar]
- 7.Pohl DJ, Seblova D, Avila JF, et al. Relationship between residential segregation, later-life cognition, and incident dementia across race/ethnicity. Int J Environ Res Public Health. 2021;18(21):11233. doi: 10.3390/ijerph182111233 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wong R, Wang Y. Role of neighborhood physical disorder and social cohesion on racial and ethnic disparities in dementia risk. J Aging Health. 2022;34(9–10):1178–1187. doi: 10.1177/08982643221101352 [DOI] [PubMed] [Google Scholar]
- 9.Rodriguez RA, Sen S, Mehta K, Moody-Ayers S, Bacchetti P, O’Hare AM. Geography matters: relationships among urban residential segregation, dialysis facilities, and patient outcomes. Ann Intern Med. 2007;146(7):493–501. doi: 10.7326/0003-4819-146-7-200704030-00005 [DOI] [PubMed] [Google Scholar]
- 10.Taylor DH Jr, Ostbye T, Langa KM, Weir D, Plassman BL. The accuracy of Medicare claims as an epidemiological tool: the case of dementia revisited. J Alzheimers Dis. 2009;17(4):807–815. doi: 10.3233/JAD-2009-1099 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary File (PDF)
Figures S1–S4, Items S1–S2, Tables S1–S7.
Data Availability Statement
The datasets used and/or analyzed during the current study are available from the United States Renal Data System (USRDS). Per the Data Use Agreement (DUA) between the authors and USRDS, the deposition of data into publicly available repositories is not allowed.