Epidemiology and Outcomes of COVID-19 in Home Dialysis Patients Compared with In-Center Dialysis Patients

• chronic dialysis
• chronic kidney disease
• epidemiology and outcomes
• ESKD
• peritoneal dialysis
• risk factors
• SARS-CoV-2
• COVID-19

Patients on dialysis are particularly vulnerable to coronavirus disease 2019 (COVID-19), with multiple studies describing mortality >20%.^1⇓–3 Although infection rates among patients on dialysis tend to parallel local patterns, this population has a higher rate of COVID-19 compared with the general population; this may be a reflection of increased symptom screening and testing, and a limited ability to achieve physically distancing, particularly given the dependence of most patients on maintenance dialysis in the United States on in-center hemodialysis.⁴ Studies describing COVID-19 in patients receiving home dialysis are lacking but needed, given they share with patients on in-center dialysis similar risk factors for poor outcomes, including possible impaired immunity and high prevalence of comorbid conditions.

Dialysis Clinic, Inc. (DCI) is a national not-for-profit dialysis provider serving approximately 2000 patients on home dialysis (90% receiving peritoneal dialysis, 10% hemodialysis). The company has 116 clinics in 27 states with active home dialysis programs, although some of these clinics have only one or two patients on home dialysis. This retrospective cohort study included all DCI patients on home dialysis with positive testing for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) from February 22, 2020 to December 31, 2020, with outcomes ascertained through February 1, 2021. Patients were diagnosed with COVID-19 via nasopharyngeal or oropharyngeal swab sent for RT-PCR testing. All positive COVID-19 tests were captured, regardless of whether the patient was assessed in the dialysis clinic, at a testing center, or at a hospital, and all patients with a positive SARS-CoV-2 test were considered to have COVID-19.

Demographic, comorbidity, and clinical data were collected from DCI’s electronic health records. To reflect the evolving epidemiology of the COVID-19 pandemic, analyses were conducted for two time periods: February 22 to September 30, 2020 (termed “phase 1”) and October 1 to December 31, 2020 (termed “phase 2”). For each phase, patients on home dialysis with COVID-19 were mapped to their clinics’ counties, which were then used to identify and compare dialysis patients in both home and in-center programs located within the same counties (Figure 1). We conducted multivariable logistic regression analysis to assess risk factors for COVID-19 among patients on home dialysis using a priori–identified covariates. To compare COVID-19 risk by modality, we calculated COVID-19 incidence rate for each modality, indexed to the date of the first COVID-19 case in a clinic’s county, whether home or in-center. We also collected 30-day mortality rates, indexed to the date of COVID-19 diagnosis for patients with COVID-19 and indexed to first COVID-19 case in the clinic’s county for patients without COVID-19. We conducted multivariable analysis with a priori–identified covariates to assess risk factors for mortality among patients on home dialysis with COVID-19.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Derivation of the study population in each phase of the study.

In phase 1, 46 of 1024 (4.5%) patients on home dialysis were diagnosed with COVID-19, deriving from clinics in 31 counties (Figure 1; see Supplemental Figure 1). Among patients on home dialysis, Black race, Hispanic ethnicity, and long-term care facility (LTCF) residence were associated with COVID-19 (Table 1). Patients on in-center dialysis had a significantly higher incidence of COVID-19 that was attenuated after excluding LTCF residents (Table 2). In phase 2, 99 of 1547 (6.4%) patients on home dialysis were diagnosed with COVID-19, deriving from clinics in 52 counties (Supplemental Figure 1). The incidence of COVID-19 no longer differed significantly by modality. COVID-19 prevalence among patients in an LTCF declined from phase 1 to phase 2. Among patients not in an LTCF, COVID-19 prevalence rose among those receiving home dialysis but remained stable among those receiving in-center dialysis (Table 2).

Among patients on home dialysis, six of 46 (13.0%) with COVID-19 died in phase 1, and 12 of 99 (12.1%) with COVID-19 died in phase 2 (Table 2). This compares with 74 of 978 (7.6%) and 127 of 1448 (8.8%) of patients on home dialysis without COVID-19 in phases 1 and 2, respectively (data not shown). Multivariable analysis conducted on the entire cohort (from February 22 through December 31) of patients on home dialysis with COVID-19 found that older age, longer vintage (dialysis duration), and cardiovascular disease were associated with mortality (Table 3).

Among patients on home dialysis, 4.5% had COVID-19 from February 22 through September 2020, with Black race, Hispanic ethnicity, and LTCF residence identified as significant risk factors, whereas from October through December 2020, 6.4% of patients on home dialysis had COVID-19. These analyses reflect the epidemiology of the COVID-19 pandemic in the United States over the past year. Early in the pandemic, testing constraints limited diagnoses to the symptomatic and those with easier access to health care, and COVID-19 was concentrated in LTCFs and urban communities, affecting a greater proportion of Black and Hispanic patients (Supplemental Table 1).⁵

Over time, LTCFs and dialysis facilities implemented and improved infection control practices targeting SARS-CoV-2 transmission, and the pandemic broadened to also affect more rural areas of the country. Thus, the epidemiology of COVID-19 in patients on home dialysis reflects its evolution in the general community, paralleling COVID-19 in the general dialysis population.^{1 , 6} Critically, the COVID-19 prevalence reported here may differ from prior reports, due to both duration of the study period and inclusion of rural areas.^{1 , 2 , 7 , 8} Moreover, COVID-19 prevalence in patients on home dialysis still exceeded that of the general population, which may be for multiple reasons. This population has greater-than-average health care utilization resulting from both ESKD and other medical comorbid conditions, a need for in-person medical care, and—in some patients—the associated need for shared transportation that limits physical distancing ability. Because this patient population has more comorbid conditions, it may also have a lower threshold to seek testing. Of note, epidemiologic patterns influenced by community COVID-19 prevalence, described in the October through December phase, are likely to persist in the coming months.

The early difference in COVID-19 prevalence by modality may be attributable to multiple factors, including limited ability to effectively practice physical distancing and higher numbers and a higher proportion of LTCF residents with COVID-19 among patients on in-center dialysis. Additionally, particularly in phase 1, when diagnostic testing resources were less available, patients receiving in-center dialysis likely encountered more opportunities for screening and testing. In phase 2, more widespread testing in the community provided an opportunity for greater diagnostic parity, although residual disparity may have persisted. In addition, LTCF residence was a significant contributor to in-center COVID-19 prevalence in phase 1; lower COVID-19 rates among LTCF residents in phase 2 appear associated with lower rates of COVID-19 among patients on in-center dialysis. These findings reinforce the importance of infection-control practices and frequent testing, especially in high-risk congregate settings.

Patients on home dialysis likely have greater mortality risk from COVID-19, compared with the 2.2% mortality rate recently estimated among the general population.⁹ Of note, the mortality rate in patients on home dialysis here was less than that reported by other studies of COVID-19 in the general dialysis population, dominated by patients treated in-center,^{1 , 2 , 7 , 8} including an earlier study by this group³; this likely reflects the generally better health of patients on home dialysis compared with the overall dialysis population. Mortality decreased from phase 1 to phase 2, falling by half among the in-center population, which likely reflects expanded testing availability enabling diagnosis of milder cases, improvement in COVID-19 management over time, fewer infections in LTCF residents, and possibly a reduced inoculum burden with widespread masking.¹⁰ Nevertheless, the high rate of poor outcomes in this vulnerable population reinforces the need for vaccine prioritization and continued vigilance in treating these patients. Older age and cardiovascular disease are known predictors of poor outcomes from COVID-19,¹¹ reflected in this study as well.

We acknowledge this study’s limitations. Matching by county may have introduced certain biases, but this restriction critically mitigates the bias of COVID-19’s geographic variation during the study period. Due to the study’s observational design, we cannot exclude confounders when comparing outcomes by modality. It is also possible that the reduced differences in phase 2 compared with phase 1 were, at least in part, due to a larger study population, demonstrating regression to the mean.

In conclusion, COVID-19 epidemiology among patients on home dialysis over time reflects community trends and is similar to the epidemiology among patients on in-center dialysis. This study reinforces previous findings that residence in an LTCF is a significant risk factor for infection. Although the COVID-19 case fatality rate among patients on home dialysis trends lower than that for in-center hemodialysis, it is still high, underscoring that preventing COVID-19 in the dialysis population remains critical.

• Copyright © 2021 by the American Society of Nephrology