SARS-CoV-2 Infection Risk Factors among Maintenance Hemodialysis Patients and Health Care Personnel In Outpatient Hemodialysis Centers

In Brief

Key Points

• Increased risk of SARS-CoV-2 infection was associated with community prevalence.
• Increased risk of SARS-CoV-2 infection was associated with exposure to infected family members and personal infection prevention measures

Introduction

In the United States during the first pandemic wave, the seroprevalence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection designated as coronavirus disease 2019 (COVID-19) among patients on maintenance hemodialysis (HD) was 8.3% (8.0%–8.6%), with regional variation of 3.5% (3.1%–3.9%) in the West to 27.2% (25.9%–28.5%) in the Northeast (¹). Hospitalized patients with ESKD, including those on chronic HD, are at increased risk of death from SARS-CoV-2 infection when compared with those without ESKD (²). Because of the congregate nature of dialysis units, open layout, and physical proximity of patients during HD sessions, patients and health care personnel (HCPs) may be at increased risk of exposure to SARS-CoV-2 (³). To mitigate this risk, the Centers for Disease Control and Prevention (CDC) provided guidance on SARS-CoV-2 infection prevention measures specific to outpatient HD units in March 2020 (⁴). However, the risk of SARS-CoV-2 infection is also influenced by factors outside the HD units, such as local community prevalence of SARS-CoV-2 infection, residence type (e.g., long-term care facility [LTCF]) (⁵,⁶) and household characteristics, exposure to household members and others with SARS-CoV-2 infection (⁷), and personal infection prevention activities (⁷). Studies examining the risk factors for SARS-CoV-2 infection among patients on in-center HD and dialysis HCPs are lacking. We examined the potential source of and risk factors for SARS-CoV-2 infection among patients on in-center HD and HCPs in three outpatient HD units that implemented the CDC SARS-CoV-2 infection prevention measures in March 2020.

Materials and Methods

This study was conducted in three outpatient HD units at Washington University School of Medicine (WUSM) in St. Louis, Missouri and was approved by the WUSM Human Research Protection Office. Adult patients on in-center HD and dialysis center HCPs were enrolled after obtaining informed consent. During September 2020, a survey was administered to patients and HCPs to assess household characteristics, SARS-CoV-2 exposures, personal prevention activities, and history of SARS-CoV-2 infection since March 1, 2020. Previous exposure to SARS-CoV-2 was assessed using the Abbott SARS-CoV-2 IgG serologic assay that targets the viral nucleocapsid protein (Abbott, Abbott Park, IL). The result unit for the SARS-CoV-2 IgG assay is index (signal/calibrator), and a value of ≥1.40 index (signal/calibrator) is considered as positive. The sensitivity and specificity of the Abbott assay are approximately 90%–99% and >99%, respectively (⁷,⁸). Electronic medical records of patients from March through September 2020 were reviewed to obtain relevant clinical information, including dates of SARS-CoV-2 PCR test(s), hospitalizations due to SARS-CoV-2 infection, and comorbid conditions. Cumulative SARS-CoV-2 cases per 1000 population by zip code through the first week of September were obtained from St. Louis University’s Missouri COVID-19 Tracking Project GitHub data repository (https://github.com/slu-openGIS/MO_HEALTH_Covid_Tracking.git) and the Illinois Department of Public Health COVID-19 dashboard (https://www.dph.illinois.gov/covid19/covid19-statistics).

For study purposes, a SARS-CoV-2 case was defined as either history of a positive PCR test documented in the medical record any time from March to September 2020 or a positive IgG serology result at the time of study enrollment (September 2020). Univariate analysis of the patient characteristics among SARS-CoV-2 cases and noncases was performed using Fisher exact tests or univariate logistic regression for categorical variables and Mann–Whitney U tests for continuous variables. Correlation between positive SARS-CoV-2 test results (antibody, PCR, or both) of patients on HD and community SARS-CoV-2 infection prevalence by patient’s home zip code was performed using the Spearman correlation test. To determine potential acquisition of SARS-CoV-2 infection inside HD units, we examined the distribution of cases by dialysis day and shift in each HD unit. Statistical analyses were performed using Python pandas, SciPy, and Matplotlib packages.

Results

Among the three HD units, 226 patients and 39 HCPs were enrolled into the study (Supplemental Table 1). Of the 226 patients, 21 patients (9%) were considered to be SARS-CoV-2 cases on the basis of history of positive PCR or positive serology. During the study period, 130 (58%) patients had greater than or equal to one PCR test performed (114 patients were asymptomatic, and PCR was performed as a screening test prior to a procedure) (Supplemental Table 2). None of the HCPs had evidence of SARS-CoV-2 infection by either PCR or serology.

In univariate analysis, SARS-CoV-2–infected patients were significantly more likely to reside in an LTCF (19% of cases versus 1.5% of noncases; P=0.002), had a household family member (24% versus 3%; P=0.001) or extended family member/friend diagnosed with SARS-CoV-2 infection (19% versus 5%; P=0.03), traveled outside of the local metropolitan area (29% versus 12%; P=0.04), and reported less adherence to mask use outside of the home (14% versus 2%; P=0.02) (Table 1). The SARS-CoV-2 infection prevalence among patients on HD correlated with SARS-CoV-2 infection prevalence in their community as assessed by zip code (r=0.346; P=0.01) (Figure 1). The prevalence per 1000 persons of SARS-CoV-2 infection was significantly lower in the HCPs’ home residence zip codes compared with patients on HD (median: 1.73 [interquartile range [IQR], 1.45–1.97] versus 19.50 [IQR, 18.11–21.39], respectively; P<0.001).

Figure 1.: Correlation between severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection prevalence among patients on maintenance hemodialysis and community SARS-CoV-2 infection cumulative incidence per 1000 population by patients’ home zip code.

Table 1. - Univariate analysis of characteristics of patients on hemodialysis who tested positive or negative for severe acute respiratory syndrome coronavirus 2 infection

Variable	Severe Acute Respiratory Syndrome Coronavirus 2–Infected Case, a N=21 (9%)	Noncase, N=205 (91%)	P Value (Mann–Whitney U Test or Fisher Exact)
Demographics
Women	11 (52.4)	106 (51.7)	>0.99
Age, median [IQR], yr	54 [47–64]	60 [49–69]	0.08
Age (>65 yr)	4 (19.0)	71 (34.6)	0.22
Race (Black)	18 (85.7)	179 (87.3)	0.74
BMI, median [IQR], kg/m2	32.7 [24.6–37.8]	29.13 [25.2–35.6]	0.73
Hemodialysis unit
Facility A	11 (52.5	91 (44.4)	Reference
Facility B	5 (23.8)	90 (43.9)	0.20
Facility C	5 (23.8)	24 (11.7)	0.35
Transportation to and from the hemodialysis unit
By private car (self or family member/friend)	11 (52.4)	111 (54.1)	>0.99
By public transportation	1 (4.8)	38 (18.5)	0.14
By medical transport company	9 (42.8)	55 (26.8)	0.13
Cause of ESKD
Diabetes mellitus	7 (33.3)	89 (43.4)	0.49
Hypertension	14 (66.7)	133 (64.9)	>0.99
Glomerular disease	4 (19.0)	11 (5.3)	0.04
Other etiologies	4 (19.0)	68 (33.2)	0.23
Dialysis vintage, median [IQR], mo	37 [20–66]	36 [13–79]	0.31
Comorbidities
Diabetes mellitus	10 (47.6)	110 (53.7)	0.65
Hypertension	21 (100.0)	193 (94.1)	0.61
Cerebral vascular accident	2 (9.5)	19 (9.3)	>0.99
Cardiovascular disease	9 (42.8)	72 (35.1)	0.49
Neurologic disease (excluding cerebral vascular accident)	2 (9.5)	9 (4.4)	0.27
COPD/asthma	6 (28.6)	39 (19.0)	0.39
Smoking	1 (4.8)	22 (10.8)	0.70
Active hepatitis C	1 (4.8)	8 (3.9)	0.59
Cirrhosis	1 (4.8)	4 (2.0)	0.39
HIV with CD4<200	1 (4.8)	1 (0.5)	0.18
Active malignancies	1 (4.8)	10 (4.9)	>0.99
Medications 30 d prior to serology test (no. of patients on dialysis)
Immunosuppressive medications	2 (9.5)	26 (12.7)	>0.99
Hypertension medications	20 (95.2)	40 (80.5)	0.14
Cardiovascular disease medications	18 (85.7)	168 (82.0)	>0.99
Diabetes medications	5 (23.8)	62 (30.2)	0.62
Asthma/COPD medication	3 (14.3)	50 (24.4)	0.42
No. of hospitalizations since March 1st, 2020, median [IQR]	1 [0–1]	0 [0–2]	0.13
SARS-CoV-2 exposures
Household member tested for SARS-CoV-2	11 (52.4)	57 (27.9)	0.03
Household member diagnosed with SARS-CoV-2	5 (23.8)	6 (2.9)	0.001
Extended family/friend positive for SARS-CoV-2	4 (19.0)	10 (4.9)	0.03
Travel outside of the St. Louis area	6 (28.6)	24 (11.8)	0.04
Resided in long-term care facility	4 (19.0)	3 (1.5)	0.002
Household size
Lives alone	1 (4.8)	55 (26.8)	Reference
Two or more people	8 (38.1)	144 (70.2)	0.45
No. of generations in the household
One generation	4 (19.0)	102 (49.8)	Reference
Two generations	10 (47.6)	70 (34.1)	0.05
Three generations or more	3 (14.3)	26 (12.7)	0.17
Household density (household size per no. of bedrooms)	1 [1.0–2.3]	1 [0.8–1.2]	0.27
First quartile (0–0.75)	4 (19.0)	55 (26.8)	Reference
Second and third quartiles (0.75–1.2)	10 (47.6)	93 (45.4)	0.77
Fourth quartile (>1.2)	3 (14.3)	52 (25.4)	>0.99
SARS-CoV-2 infection prevention behaviors
Reported mask use at dialysis centers
All of the time	21 (100.0)	196 (95.6)	Reference
Less than all of the time	0 (0.0)	8 (3.9)	>0.99
Reported social distancing at dialysis center (% adherence)
All of the time	16 (76.2)	140 (68.3)	Reference
Less than all of the time	5 (23.8)	64 (31.2)	0.62
Reported hand hygiene at dialysis center prior to sitting on dialysis chair
All of the time	19 (90.5)	174 (84.9)	Reference
Less than all of the time	2 (9.5	30 (14.5)	0.75
Reported mask use during daily life outside of home
All of the time	14 (66.7)	155 (75.6)	Reference
Less than all of the time	3 (14.3)	45 (22.0)	0.77
Sometimes (about 50% of the time) or less	3 (14.3)	4 (2.0)	0.02
Visit to other health facilities (doctor offices, dentist offices) between March and September 2020
First tertile 0–1 health visits	6 (28.6)	75 (36.6)	Reference
Second tertile 2–5 health visits	9 (42.9)	63 (30.7)	0.42
Third tertile >5 health visits	6 (28.6)	66 (32.2)	>0.99
Visit to public spaces for daily activities between March and September 2020
First tertile 0–6 visits	10 (47.6)	62 (30.2)	Reference
Second tertile 7–28 visits	7 (33.3)	71 (34.6)	0.44
Third tertile >28 visits	4 (19.0)	71 (34.6)	0.10
No. of gatherings with >10 people attending b
First tertile (0 gatherings)	7 (33.3)	105 (51.2)	Reference
Second tertile (1 gathering)	5 (23.8)	35 (17.1)	0.30
Third tertile (>1 gathering)	9 (42.9)	64 (31.2)	0.18

IQR, interquartile range; BMI, body mass index; COPD, chronic obstructive pulmonary disease; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

^aDefined as a positive PCR and/or serology for SARS-CoV-2.

^bTypes of gatherings include house of worship/religious services, funerals, weddings, family/friend gatherings (birthdays, dinners, cookouts, etc.), graduations, concerts or plays/theater productions, eating at a restaurant, visits to a bar, in-person school or classes, protests, and others.

Among the 21 SARS-CoV-2–infected patients, ten (48%) patients had both a positive PCR test in the past and positive IgG antibody in September (eight patients with symptomatic disease; the two asymptomatic patients included an LTCF resident and a worker in an LTCF), with median time from the positive PCR and positive IgG antibody of 128 days (IQR, 85–140) (Supplemental Table 3). Two (10%) asymptomatic patients had prior positive PCR with negative IgG antibody results. For these two patients, the time intervals from positive PCR test to negative IgG antibody test were 124 and 139 days. Nine patients (43%) had positive IgG antibody without a positive PCR test; of these, six had no PCR test performed, and three had a negative PCR test as part of preprocedure screening (median time from latest negative PCR test to positive IgG antibody test was 45 days). Of these nine patients, four reported having contact or living with a known SARS-CoV-2–infected patient, one lived in an LTCF, three reported travel outside of the metropolitan area, and one patient had no obvious SARS-CoV-2 risk factor. No clustering of SARS-CoV-2–infected patients was observed among the three HD units by shift and day of the week (Table 2). Except for one patient, all SARS-CoV-2–infected patients had one or more risk factors for exposure to SARS-CoV-2 outside the HD unit.

Table 2. - Number of severe acute respiratory syndrome coronavirus 2–infected patients by hemodialysis facility, pod, and shift

Center	Pod	Shift	Case Identification	Positive PCR or IgG Positive Only	Know Exposure Outside Hemodialysis Facility	Has Other Risk Factor a
A	A	M/W/F second shift	16	IgG positive	No	No
A	A	M/W/F third shift	4	PCR positive	Yes	No
A	A	T/T/S first shift	2	PCR positive	Yes	Yes
A	A	T/T/S second shift	1, 3	Both PCR positive 1 wk apart	1: Yes	1: Yes
					3: Yes	3: No
A	B	T/T/S first shift	5	PCR positive	No	Yes
A	C	M/W/F first shift	6, 16	One PCR positive	6: No	6: Yes
				One IgG positive 1 mo apart	16: No	16: No
A	C	T/T/S first shift	18	IgG positive	Yes	No
A	Rotate different pods	M/W/F second shift	19	IgG positive	No	Yes
A	Rotate different pods	T/T/S third shift	13	IgG positive	Yes	Yes
B	2	M/W/F first shift	11	PCR positive	Yes	No
B	3	T/T/S first shift	9	PCR positive	Yes	No
B	3	T/T/S second shift	7	PCR positive	No	No
B	1	T/T/S second shift	8, 14	One PCR positive	8: Yes	8: No
				One IgG positive 3 mo apart	14: Yes	14: No
C	3	M/W/F first shift	21	IgG positive	No	Yes
C	4	M/W/F first shift	12	PCR positive	No	Yes
C	1	T/T/S first shift	10, 15	One PCR positive	10: Yes	10: Yes
				One IgG positive 3 mo apart	15: No	15: Yes
C	Not sure	T/T/S second shift	20	IgG positive	No	Yes

M/W/F, Monday/Wednesday/Friday; T/T/S, Tuesday/Thursday/Saturday.

^aOn the basis of univariate analysis in Table 1.

Discussion

A few studies have examined risk factors for SARS-CoV-2 infection among patients on HD (¹,⁶,⁹,¹⁰). However, none of the previous studies included HCPs or reported on household characteristics, SARS-CoV-2 exposures, and personal prevention activities of patients and HCPs while evaluating the risk factors for SARS-CoV-2 infection. Our results indicate that the risk of infection among patients on in-center maintenance HD and dialysis HCPs is associated with the local community prevalence of SARS-CoV-2 infection, diagnosis of SARS-CoV-2 infection among immediate family members or friends, residence in an LTCF, and personal SARS-CoV-2 infection prevention activities, such as adherence to face mask use and travel outside the local metropolitan area. Residence in an LTCF has been identified as a SARS-CoV-2 infection risk factor previously, and our results support this finding (⁵,⁶). Notably, none of the enrolled HCPs had a positive IgG serology result, which suggests that CDC infection prevention guidance can help limit SARS-CoV-2 transmission within dialysis centers.

There are several limitations to this study. Because of the small number of cases, we could not perform multivariate analysis to identify independent risk factors for SARS-CoV-2 infection. We are currently in the process of obtaining follow-up serology and survey data from study participants and may be able to perform multivariable modeling in future analyses. Survey responses regarding reported activities, SARS-CoV-2 exposures, and prevention activities (such as mask use) may be subject to recall bias or “good behavior” bias. Direct observation of behaviors and activities would be optimal. This is not feasible outside of the dialysis centers, but direct observations within the dialysis centers are currently ongoing. Future analyses of these data can be used to validate the survey data regarding in-center COVID-19 prevention practices. Additionally, we did not perform periodic surveillance testing for asymptomatic patients or HCPs using PCR test, and thus, the overall SARS-CoV-2 infection prevalence may be underestimated as some patients and HCPs could have asymptomatic SARS-CoV-2 infection with no detectable antibody response or had not yet developed antibodies. Finally, the majority (87.2%) of patients in our cohort were Black, and thus, this may limit the generalizability of our findings as adult Black patients account for only 35% of all United States patients on dialysis (¹).

Our results suggest that although some SARS-CoV-2 risk factors are not easily modifiable, others are, and adherence to infection prevention practices outside of HD units, such as consistent use of face masks (¹¹), may help decrease the risk of SARS-CoV-2 in patients on HD. We did not find clear evidence for SARS-CoV-2 infection acquisition by patients or HCPs within HD units, suggesting that CDC SARS-CoV-2 infection prevention measures were effective. However, with the ongoing increases in community prevalence and the strong association found here between risk of SARS-CoV-2 and community rates, it is important to continue to examine the effectiveness of CDC SAR-CoV-2 infection prevention measures and risk factors for SARS-CoV-2 among patients on HD and HCPs. The longitudinal follow-up of this study is ongoing and will help address these questions.

Disclosures

C.W. Farnsworth reports consultancy agreements with Roche Diagnostics and research funding from Abbott Diagnostics, Beckman Coulter, and Siemens Healthineers. T. Li reports consultancy agreements with ChemoCentryx, GlaxoSmithKline, and Reata Pharmaceuticals, Inc.; research funding from Aurinia Pharmaceuticals, Bristol-Myers Squibb, Human Genome Science, and Omeros Corporation; and scientific advisor or membership with the American Society of Nephrology (ASN) Education Committee (member), ASN Publications Committee (member), and GlaxoSmithKline (scientific advisor). M. A. Olsen received grant funding from Merck, Pfizer, and Sanofi Pasteur in the past 3 years and worked as a consultant to Pfizer not related to COVID-19. All remaining authors have nothing to disclose.

Funding

This study was funded by Centers for Disease Control and Prevention (CDC) Epicenters Program grant 6 U54CK000482-04. J.H. Kwon is supported by National Institute of Allergy and Infectious Diseases grant 1K23AI137321-01A1.