Risk of COVID-19 Disease, Dialysis Unit Attributes,and Infection Control Strategy among LondonIn-Center Hemodialysis Patients

Ben Caplin ,1,2 Damien Ashby,3 Kieran McCafferty,4 Richard Hull,5 Elham Asgari,6 Martin L. Ford,7,8 Nicholas Cole,9

Marilina Antonelou ,1,2 Sarah A. Blakey,3 Vinay Srinivasa ,4 Dandisonba C.B. Braide-Azikwe,7 Tayeba Roper ,6

Grace Clark,2 Helen Cronin,7 Nathan J. Hayes,10 Bethia Manson,2 Alexander Sarnowski,5 Richard Corbett ,3

Kate Bramham ,7,8 Eirini Lioudaki ,7,8 Nicola Kumar,6 Andrew Frankel,3 David Makanjuola,9 Claire C. Sharpe ,7,8

Debasish Banerjee ,5 and Alan D. Salama,1,2 on behalf of the Pan-London COVID-19 Renal Audit Group*

AbstractBackground and objectives Patients receiving in-center hemodialysis treatment face unique challenges during thecoronavirus disease 2019 (COVID-19) pandemic, specifically the need to attend for treatment that prevents self-isolation. Dialysis unit attributes and isolation strategies that might reduce dialysis center COVID-19 infectionrates have not been previously examined.

Design, setting, participants, & measurementsWe explored the role of variables, including community diseaseburden, dialysis unit attributes (size and layout), and infection control strategies, on rates of COVID-19 amongpatients receiving in-center hemodialysis in London, United Kingdom, betweenMarch 2, 2020 and May 31, 2020.The two outcomes were defined as (1) a positive test for infection or admission with suspected COVID-19 and (2)admission to the hospital with suspected infection. Associations were examined using a discrete time multileveltime-to-event analysis.

Results Data on 5755 patients dialyzing in 51 units were analyzed; 990 (17%) tested positive and 465 (8%) wereadmitted with suspected COVID-19 between March 2 and May 31, 2020. Outcomes were associated with age,diabetes, local community COVID-19 rates, and dialysis unit size. A greater number of available side rooms andthe introduction of mask policies for asymptomatic patients were inversely associated with outcomes. Noassociation was seen with sex, ethnicity, or deprivation indices, nor with any of the different isolation strategies.

Conclusions Rates of COVID-19 in the in-center hemodialysis population relate to individual factors, underlyingcommunity transmission, unit size, and layout.

CJASN 16: 1237-1246, 2021. doi: https://doi.org/10.2215/CJN.03180321

IntroductionPatients with kidney failure and those receiving KRTare among the highest-risk groups for severe acuterespiratory syndrome coronavirus 2 (SARS-CoV-2)–related disease and death (1). Alongside kidneydysfunction, this group of patients exhibits a largeburden of risk factors for coronavirus disease 2019(COVID-19), including over-representation of Blackand minority ethnic groups and socioeconomic dis-advantage, as well as high rates of diabetes andcardiovascular disease. Furthermore, those receivingin-center hemodialysis cannot shield or self-isolate.National data confirm that patients with kidneyfailure have been disproportionately affected byCOVID-19, with 822 of the 36,437 COVID-19–

associated deaths reported in England as of July 15,2020 having occurred in patients with established kid-ney disease (2, 3). The majority of these have been inpatients receiving dialysis treatments, and at least 623occurred in patients on in-center hemodialysis (2),representing 2% of all COVID-19 deaths in the UnitedKingdom, despite making up ,0.1% of the popula-tion. However, rates of infection in London hemodial-ysis units varied widely, from just a few cases to 30%of patients.In London, dialysis is organized geographically,

with seven tertiary National Health Service (NHS) kid-ney centers providing local care to patients on the basisof their home address. Each center provides long-termdialysis in units colocated within the main hospital, at

1Department of RenalMedicine, UniversityCollege London, London,United Kingdom2Renal Services, Royal FreeLondon NHS FoundationTrust, London, UnitedKingdom3Kidney and TransplantServices, Imperial CollegeHealthcare NHS Trust,London, United Kingdom4Renal Service, BartsHealth NHS Trust, London,United Kingdom5Renal Department, St.George’s UniversityHospitals NHS FoundationTrust, London, UnitedKingdom6Kidney Services, Guy’sand St. Thomas’NHSFoundation Trust, London,United Kingdom7Department of RenalMedicine, King’s CollegeHospital NHS FoundationTrust, London, UnitedKingdom8Faculty of Life Sciences &Medicine, Kings CollegeLondon, London, UnitedKingdom9SouthWest Thames Renaland Transplantation Unit,Epsom and St. HelierUniversity Hospitals NHSTrust, London, UnitedKingdom10Renal Services, NorthMiddlesex UniversityHospital NHS Trust,London, United Kingdom

Correspondence: Dr. BenCaplin, Department ofRenal Medicine,University CollegeLondon, 2nd Floor, RoyalFree Campus, RowlandHill Street, London NW32PF, United Kingdom.Email: b.caplin@ucl.ac.uk

www.cjasn.org Vol 16 August, 2021 Copyright © 2021 by the American Society of Nephrology 1237

Article . Published on August 4, 2021 as doi: 10.2215/CJN.03180321CJASN ePress

another hospital, or in standalone community locations (sat-ellite units), and each unit varies in size (from six to 66stations).In line with published guidelines, during the first phase

of the pandemic, all dialysis units attempted to identify,isolate, and, where possible, test patients arriving for treat-ment with symptoms consistent with COVID-19 (4). Nursescaring for the patients wore personal protective equipmentas per NHS guidelines (5) only when delivering direct careto patients with known or suspected disease. In addition,kidney centers across London introduced masking policiesat different times and used varied patient segregation strat-egies (for isolation and cohorting of patients with knownor suspected disease) to reduce spread of disease. We setout to identify risk factors associated with the burden ofdisease during the first phase of the COVID-19 pandemic.

Materials and MethodsStudy PopulationAll prevalent London-based patients on in-center hemo-

dialysis on March 2, 2020 were included in the analysis.Patients receiving short-term hemodialysis therapy for AKIwere excluded. Patients were assigned to the unit wherethey dialyzed on March 2nd, except where that unitbecame a test-positive isolation unit, in which case theywere assigned to the first unit they dialyzed in aftertransfer to an unaffected unit. Cumulative incidence ofCOVID-19 test positivity and hospital admission rates werecollected from March 2 to May 31, 2020. Approximately23% of our whole population has already formed the basisof a previous report (6). This work was conducted underapproval of the St. George’s Research Ethics Committee(reference no. 283130).

Data Sources and DefinitionsPseudoanonymized demographic and clinical data along

with COVID-19 test results and hospital admission dateswere collected from electronic records. Dialysis unit charac-teristics and other unit and center data were collected bystructured questionnaire (Supplemental Material). Commu-nity COVID-19 cases and deprivation indices on the basisof a 37-variable score reflecting aspects of deprivationassessed by middle layer super output area, a geographichierarchy used to describe small area statistics, were cap-tured from publicly available datasets (3,7). For these varia-bles, we linked the postcode for each patient to the mediandeprivation index for all output areas within that postcodeand the weekly total of all reported COVID-19 cases aver-aged across all of the output areas within that postcode.

OutcomesAll estimates of infection were subject to limitations dur-

ing the study period. Testing for SARS-CoV-2 was under-taken using nasopharyngeal swabs followed by RT-PCRperformed independently in each of the seven differentcenters. Availability of swabs was limited during the earlyphase of the pandemic, so no testing took place in asymp-tomatic patients and was done variably in those with mildsymptoms that would not require hospitalization. Rates ofadmission with clinically suspected COVID-19 are less

likely to be affected by test availability, although rates willvary according to local practice. We, therefore, examinedthe rates of two outcomes over the 13-week follow-upperiod: (1) the first of either a positive SARS-CoV-2 test oradmission with suspected COVID-19 and (2) admissionwith suspected COVID-19.

AnalysesAs infections occurred at different time points across

London for the primary analysis, we defined the time atrisk as the time (in weeks) from first positive test at eachdialysis unit. We then used a complementary log-log multi-level discrete time survival model (8) to estimate hazardratios (HRs) associated with individual- and community-level factors, along with dialysis unit physical attributesand policies. Patients were nested within dialysis units,which were in turn nested within kidney centers to accountfor clustering of outcomes.

Models were developed sequentially on the basis ofprespecified candidate factors known or suspected to beassociated with outcomes. At each stage, a complete caseanalysis was performed. We initially examined associationsbetween individual-level demographic and clinical factors(age, sex, ethnicity, diabetes, and deprivation index) andoutcomes. All clinical and demographic variables wereretained in subsequent models. We then added estimatesof time-dependent community transmission (lagged by 0,1, 2, and 3 weeks), followed by physical attributes of thedialysis units. Variables either that improved fit (as judgedby the 22 log likelihood) or where the 95% confidenceinterval (95% CI) for the HR did not include unity wereretained (although for time-lagged variables, only the timelag with the largest effect size was retained). Where two ormore correlated explanatory variables were associated withoutcome when examined individually (e.g., prevalentpatients and number of stations), the variables with theleast missing data or that led to the greatest improvementin model fit were retained.

The above then formed the basis to examine individualassociations between the various isolation and deisolationstrategies, as well as patient and staff masking policies astime-dependent variables (the latter three variables againwere lagged by 0, 1, 2, and 3 weeks). For the analysis of iso-lation strategies, responses were grouped with a view ofproviding the most useful comparisons as outlined in Sup-plemental Table 1.

Finally, we presented the coefficients for all of theretained potentially explanatory variables (i.e., not includ-ing contemporaneous staff masking) that were associatedwith the outcome from the unit attributes and maskinganalyses in a multivariable model.

To understand the relative contribution of each group ofvariables that was associated with outcome (individual riskfactors, community disease, unit attributes, and masking),we examined the area under the receiver operating curve(AUROC) with each model. Finally, we explored the differ-ence in absolute risks associated with the introduction of apolicy of asymptomatic patient masking 2 weeks before-hand by estimating the predicted marginal risk with andwithout the mask policy.

1238 CJASN

Table1.

Patien

tde

mog

raph

ics,ou

tcom

es,a

ndco

mmun

itycases

Cha

racteristic

Cen

ter

A,n

574

2B,n

564

5C,n

532

5D,n

511

79E,n

513

39F,

n584

0G,n

568

5Total,

n557

55

Patients

Sex,

men

460(62%

)40

5(63%

)20

8(64%

)71

7(64%

)79

4(59%

)53

5(64%

)40

7(59%

)35

26(61%

)Ethnicity

Asian

101(14%

)50

(8%)

83(25%

)35

3(30%

)58

2(43%

)65

(8%)

27(4%)

1261

(22%

)Black

200(27%

)27

4(42%

)10

9(33%

)33

5(28%

)35

6(27%

)11

9(14%

)29

1(42%

)16

84(29%

)White

252(34%

)24

5(38%

)83

(25%

)28

7(24%

)38

6(29%

)49

7(59%

)25

3(37%

)20

03(35%

)Other

189(25%

)76

(12%

)50

(15%

)18

1(15%

)15

(1%)

159(19%

)11

4(17%

)78

4(14%

)Age

,yr,mean(SD)

64(15)

62(15)

64(16)

61(15)

65(14)

66(15)

62(15)

63(15)

Diabe

tes

370(50%

)26

1(40%

)12

9(40%

)39

4(33%

)61

9(46%

)35

1(42%

)27

7(40%

)24

01(42%

)Po

stcodemed

ian

dep

riva

tion

index

rank

,mean(SD)

13,280

(662

1)12

,442

(625

4)19

,172

(624

5)10

,519

(516

0)13

,497

(563

9)19

,978

(754

1)13

,055

(656

4)14

,041

(687

3)

Outcom

esTestpo

sitive

orad

mission

168(23%

)10

1(16%

)27

(8%)

208(18%

)30

5(23%

)84

(10%

)97

(14%

)99

0(17%

)

Admission

105(14%

)40

(6%)

18(5%)

68(6%)

148(11%

)50

(6%)

36(5%)

465(8%)

Com

munitycases,

cases/middle

laye

rsu

per

outputarea

,mea

n(SD)

Week1

0.0(0.0)

0.0(0.0)

0.0(0.0)

0.0(0.1)

0.0(0.1)

0.0(0.0)

0.0(0.0)

0.0(0.0)

Week2

0.3(0.5)

0.3(0.4)

0.6(0.5)

0.1(0.2)

0.4(0.5)

0.2(0.2)

0.4(0.5)

0.3(0.4)

Week3

1.7(1.1)

2.1(1.1)

2.4(1.3)

1.2(0.7)

2.4(1.2)

1.2(1.1)

2.0(1.4)

1.8(1.3)

Week4

4.6(1.9)

4.8(2.0)

5.2(1.4)

4.0(1.4)

5.5(2.2)

3.3(2.2)

4.7(2.3)

4.6(2.1)

Week5

5.6(2.2)

7.5(1.5)

6.5(2.0)

6.5(1.2)

6.6(2.7)

5.3(2.9)

6.5(2.5)

6.3(2.4)

Week6

4.3(1.6)

6.0(1.5)

5.0(1.4)

5.6(1.5)

5.2(1.8)

5.0(2.0)

5.1(1.8)

5.2(1.8)

Week7

2.9(1.1)

3.8(1.0)

3.6(1.7)

2.9(1.2)

4.5(1.6)

5.0(1.7)

3.1(1.1)

3.8(1.6)

Week8

2.5(1.7)

2.5(0.8)

2.0(1.3)

2.0(1.0)

3.3(1.4)

3.8(1.6)

2.1(1.0)

2.7(1.5)

Week9

1.4(0.9)

2.2(0.9)

2.0(1.1)

1.7(1.1)

2.6(1.3)

3.1(1.4)

1.8(1.2)

2.2(1.3)

Week10

0.6(0.7)

1.0(0.5)

0.6(0.6)

1.1(0.9)

1.2(0.9)

1.3(0.9)

0.8(0.7)

1.0(0.9)

Week11

0.3(0.5)

0.8(0.4)

0.6(0.6)

0.2(0.4)

0.8(1.0)

1.0(0.9)

0.6(0.6)

0.6(0.8)

Week12

0.2(0.3)

0.7(0.8)

0.3(0.3)

0.2(0.2)

0.4(0.4)

0.7(0.8)

0.4(0.9)

0.4(0.6)

Week13

0.1(0.2)

0.2(0.7)

0.1(0.2)

0.1(0.2)

0.2(0.2)

0.3(0.4)

0.4(1.0)

0.2(0.5)

Value

sareN

(%)u

nlessothe

rwisestated

.

CJASN 16: 1237–1246, August, 2021 COVID-19 in London In-Center Hemodialysis Patients, Caplin et al. 1239

Sensitivity AnalysesWe conducted identical models using calendar week

follow-up time rather than weeks from the first test-positive case to explore whether the alignment of outbreaksfor the analysis had introduced substantial bias. The a wasset at 0.05. As this was an exploratory study, no adjust-ments for multiple testing were made. All analyses wereperformed in Stata v.16.

ResultsPatient Population, Center Policies, and OutcomesAfter exclusions (Supplemental Figure 1), we analyzed

data on 5755 patients in 51 individual dialysis units fromseven kidney centers; 1339 (23%) of these were previouslyreported (6). Mean age was 63 years (SD of 15), 2229 (39%)were women, and 2401 (42%) had a diagnosis of diabetes(Table 1). There was a broad range of ethnicities repre-sented among the patient population, with 1261 (22%)Asian (predominantly South Asian), 1684 (29%) Black, and2003 (35%) White. Units dialyzed between eight and 338patients (Table 2, Supplemental Figure 2). Isolation strate-gies by center are illustrated in Figure 1 (with further detail

in Supplemental Table 2). Introduction of mask-wearingpolicies occurred over the first 7 weeks. Over the course ofthe 13-week follow-up, the outcome of test positivity orhospital admission occurred in 990 patients (17%) (Figure2), and almost half of these, 465 (8%), met the admissionoutcome. However, these figures encompassed substantialbetween-unit variation, with first cases occurring betweenweek 1 and week 6 and between 0% and 44% of patientsreaching the outcome of test positivity or admission (Figure3). Across all centers, following the first swab-positive case,the HR peaked in the third week for the outcome of testpositivity or admission (HR, 2.1; 95% CI, 1.9 to 2.4 versusweek 1) (Supplemental Figure 3) and the fourth week foradmission alone (HR, 2.8; 95% CI, 2.3 to 3.3 versus week 1)(Supplemental Figure 3). When examining both outcomesby calendar date, estimates were less precise, but the peakfor both outcomes occurred in the fifth week of follow-up.

Clinical and Demographic Variables and Estimates ofDisease in Patients’ Local Communities

As expected, in the clinical and demographic model forboth time to test positivity or admission and admissionalone, age (HR, 1.01; 95% CI, 1.00 to 1.01 and HR, 1.02; 95%

Treatment of patients known to be SARS-CoV2 NP swab positive

Patients arriving for usual dialysis treatment with symptomsconsistent with Covid-19

Treatment of patients withsymptoms consistent with Covid-

19 but with a single negativeSARS-CoV2 NP swab

Return of patients withconfirmed Covid-19 whose symptoms have resolved to

usual dialysis

Side room Physical separation(if no side room)

Different shift

Off-site

Off-site

Physical separationalongside cases(if no side room)

Side room Physical separationfrom all others

(if no side room)

Side room Usual dialysis

1 week vs 2 weeks

Negative NPswab required

Figure 1. | Isolation and deisolation strategies for management of patients with clinically suspected or confirmed coronavirus disease2019 (COVID-19). NP, nasopharyngeal; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

1240 CJASN

0

02-M

ar

09-M

ar

16-M

ar

23-M

ar

30-M

ar

06-A

pr

13-A

pr

20-A

pr

27-A

pr

04-M

ay

11-M

ay

18-M

ay

25-M

ay

10

20

30

patie

nt c

ases

by

dial

ysis

cen

tre

(N/d

ay)

general population cases (N/10000/w

eek)

40

50

60

0

3

2

1

4

5

6

7

8

centreGFEDCBA

Figure 2. | Epidemic timeline in patients on in-center hemodialysis across London. Patient cases (first of positive test or admission) byrenal center expressed as actual numbers per day. Cases in the general population are available weekly by location and calculated bypostcode per 10,000 population. The whole population average (dashed line) is weighted by the number of patients per postcode.

Figure 3. | Geographic distribution of cases by dialysis unit by week for the first 8 weeks of the pandemic and at the final follow-upweek. Each in-center hemodialysis unit is represented by a circle. Circle size indicates the number of patients dialyzed. Circle color indi-cates the cumulative proportion of patients who tested positive for COVID-19 or were admitted with suspected COVID-19. Underlyinggreen intensity reflects the number of cases in each middle layer super output area (not necessarily reflective of patients’ home locality asmost, but not all, patients dialyze in their closest unit). Three units are not represented as they fall outside of the area of the map. Datesrepresent the first day of the week represented. MSOA, middle layer super output area.

CJASN 16: 1237–1246, August, 2021 COVID-19 in London In-Center Hemodialysis Patients, Caplin et al. 1241

CI, 1.01 to 1.03 per year, respectively) and presence of dia-betes (HR, 1.20; 95% CI, 1.05 to 1.38 and HR, 1.22; 95% CI,1.00 to 1.49, respectively) were associated with outcomes(Table 3). There was also a strong association with esti-mates of disease in patients’ local community, moststrongly in the concurrent week (HR, 1.17; 95% CI, 1.13 to1.21 and HR, 1.14; 95% CI, 1.08 to 1.19 per case per middlelayer super output area in patients’ home postcode, respec-tively) (Table 3).

Dialysis Unit CharacteristicsAfter adjustment for individual-level variables and com-

munity cases, a number of dialysis unit characteristics wereassociated with the risk of the outcomes, specifically unitsize and layout (Table 3). These included number of dialy-sis stations (HR, 1.04; 95% CI, 1.00 to 1.28 and HR, 1.02;95% CI, 1.01 to 1.03 per station for test positivity or admis-sion and admission alone, respectively) and dialysis unitarea per station (HR, 0.92; 95% CI, 0.87 to 0.98 and HR,0.89; 95% CI, 0.83 to 0.96 per m2 for test positivity or admis-sion and admission alone, respectively) as well as distancebetween dialysis stations and number of side (isolation)rooms per station for the test positivity outcome but notthe admission alone outcome. This was also the case forboth outcomes when examining the association with num-ber of prevalent patients. Many of these characteristicswere strongly correlated (Supplemental Figure 4) (e.g.,smaller units with fewer stations typically had relativelymore side rooms; therefore, only the number of prevalentpatients and number of side rooms per station wereretained for the modeling of infection control strategies,masking, and staff illness).

Isolation Strategies and Mask PolicyWhen examining the various units’ isolation strategies

(after adjustment for clinical and demographic factors,community cases, and physical unit characteristics), wefound no evidence for a difference in either outcome withany of (1) treating all known positive patients on a separatesite; (2) the different isolation approaches applied to newlysuspected cases attending the dialysis unit; or (3) deisola-tion policies, either for those who had tested positive (bytime or requirement for a negative swab) or symptomaticpatients who had tested negative (Table 3). We observed areduction in admissions (but not test-positive or admissionoutcome) following institution of patient masking 2 weekspreviously (HR, 0.64; 95% CI, 0.44 to 0.93). No similar effectwas seen for staff masking in our primary analyses; rather,there was a positive association with both outcomes withinstitution of staff masking in the concurrent week (HR,1.47; 95% CI, 1.12 to 1.92 and HR, 1.71; 95% CI, 1.10 to 2.65for test positivity or admission and admission alone,respectively).

The final multivariable model (Table 4) demonstratedage, diabetes, number of patients in the unit, and estimatesof community transmission for the current week as posi-tively associated with both outcomes. Number of siderooms (for the test-positive outcome only) and introductionof patient masking 2 weeks previously (for admission only)were inversely associated with outcomes. Sensitivity analy-ses using calendar week rather than time from first case

Table2.

Unit-levelc

haracteristics

Cha

racteristic

Cen

ter

A,n

55

B,n

57

C,n

55

D,n

511

E,n

58

F,n5

8G

n57

Total,n

551

Unit

characteristics

Prev

alen

tpa

tien

ts,m

edian(IQR)

189(76–17

2)80

(62.5–92

.5)

82(34–82

)12

3(47–

132.5)

144(116–20

6)13

0(75.5–

131)

103(67–

105)

100(70–13

2)Stations

,med

ian(IQR)

29(16–30

)17

(12–20

)18

(6–18

)20

(18–27

)24

.5(23–45

)24

(19.5–25

.5)

20(20–

28)

20(18–26

)Sideroom

s,an/

20stations

,med

ian(IQR)

1.7(1.7–2.5)

3.3(0–4.0)

2.7(2.2–6.7)

0.0(0.0–1.0)

0.0(0.0–0.3)

3.3(3.1–4.6)

3.0(2.1–5.0)

2.2(0.0–3.3)

Waiting

room

,bm

2pe

rstation,

med

ian(IQR)

1.0(0.7–1.3)

1.25

(0.7–1.7)

1.6(0.5–4.0)

0.9(0.8–1.0)

1.4(1.1–1.7)

1.1(0.4–1.6)

—1.30

(0.4–4.0)

Dialysisarea,b

m2pe

rstation,

med

ian(IQR)

7.7—

10.3

(9.0–10

.8)

12.9

(9.2–14

.6)

13.6

(11.9–14

.9)

11.2

(10.2–14

.4)

12.1

(10.3–

12.3)

—11

.2(9.6–12

.9)

Stationdistanc

e,cmean(SD)

2.49

(0.69)

2.14

(0.37)

2.39

(0.58)

2.68

(0.77)

2.21

(0.41)

2.07

(0.45)

—2.39

(0.59)

Masking

Week(orrang

e)of

introd

uction

ofstaff

masks

forall

46

(1–6)

34

(1–6)

(2–6)

(1–6)

Week(orrang

e)of

introd

uction

ofpa

tien

tmasks

forall

56

(2–6)

37

(1–6)

(2–4)

(1–7)

Value

sareN

(%)u

nlessothe

rwisestated

.IQR,interqu

artile

rang

e;—

,not

available.

a Dataareav

ailablefor46

units.

bDataareav

ailablefor31

units.

c Dataareav

ailablefor33

units.

1242 CJASN

demonstrated broadly consistent findings, although a num-ber of coefficients differed and 95% CIs were generallywider for time-dependent effects (Supplemental Table 2).The AUROC analysis (Supplemental Table 3) suggestedthat community transmission explained a substantial pro-portion of the variation in the model for the test-positive oradmission outcome, whereas individual factors were moreimportant for the admission outcome. The modeled abso-lute risk reduction associated with a policy introduction ofasymptomatic patient masking 2 weeks previously was inthe range 0.3%–0.4% during the weeks when most

admissions occurred (Supplemental Table 4), although theproportion of the follow-up period when most centerspracticed different patient masking policies was limited tojust a few weeks.

DiscussionWe describe the individual and community factors, dial-

ysis unit, masking policies, and isolation strategies associ-ated with risk of COVID-19 in London patients on

Table 3. Individual associations with study outcomes

CharacteristicPositive Test or Admission Admission

Hazard Ratio 95% Confidence Interval Hazard Ratio 95% Confidence Interval

Clinical and demographic factors, mutually adjustedAge, per yr 1.01 1.00 to 1.01 1.02 1.01 to 1.03Men 0.94 0.82 to 1.07 1.04 0.85 to 1.26Ethnicity

Asian 1.12 0.93 to 1.36 1.05 0.79 to 1.38Black 1.11 0.93 to 1.33 1.16 0.90 to 1.51Other 0.90 0.71 to 1.13 0.75 0.52 to 1.08

Diabetes 1.20 1.05 to 1.38 1.22 1.00 to 1.49Median IMD rank, per 100 1.00 1.00 to 1.00 1.00 1.00 to 1.00

Community case burden adjusted for all clinical and demographic factors, cases per output area in home postcodeCommunity cases, week 0 1.17 1.13 to 1.21 1.14 1.08 to 1.19Community cases, week 21 1.11 1.06 to 1.14 1.11 1.05 to 1.16Community cases, week 22 1.01 0.97 to 1.05 1.05 0.99 to 1.11Community cases, week 23 0.90 0.85 to 0.94 0.96 0.89 to 1.02

Unit attributes individually adjusted for all clinical and demographic factors and community disease burdenPrevalent patients, n 1.00 1.00 to 1.01 1.00 1.00 to 1.01Stations,a n 1.01 1.00 to 1.03 1.02 1.01 to 1.04Dialysis area,b m2 per station 0.92 0.87 to 0.98 0.89 0.83 to 0.96Waiting room size,c m2

per station1.02 0.99 to 1.05 1.01 0.99 to 1.04

Station distance,d m2 0.68 0.49 to 0.94 1.03 0.67 to 1.59Side rooms,e n/20 stations 0.88 0.82 to 0.94 0.90 0.80 to 1.01

Isolation policies individually adjusted for clinical and demographic factors, community disease burden, unit prevalentpatients, and side roomsa

Isolation of test-positivepatients off site

0.88 0.68 to 1.13 0.78 0.53 to 1.16

Isolation of symptomaticarrivals off site or in SR

1.10 0.76 to 1.57 0.79 0.42 to 1.49

Return of test-negative patientsto usual dialysis

0.89 0.65 to 1.20 0.89 0.48 to 1.65

Return of test-positive patientsafter 2 wk (versus 1 wk)

0.87 0.65 to 1.15 0.64 0.39 to 1.06

Return of test-positive patientsafter testing negative

0.96 0.72 to 1.29 0.77 0.47 to 1.27

Masking policies individually adjusted for clinical and demographic factors, community disease burden, unit prevalentpatients, and side roomsa

Staff masks, week 0 1.47 1.12 to 1.92 1.71 1.10 to 2.65Staff masks, week 21 1.06 0.84 to 1.34 1.41 0.97 to 2.04Staff masks, week 22 1.04 0.82 to 1.32 0.99 0.70 to 1.39Staff masks, week 23 1.12 0.86 to 1.46 0.95 0.65 to 1.38Patient masks, week 0 1.07 0.86 to 1.32 1.03 0.75 to 1.41Patient masks, week 21 0.92 0.73 to 1.15 0.84 0.60 to 1.17Patient masks, week 22 0.87 0.66 to 1.15 0.64 0.44 to 0.93Patient masks, week 23 0.89 0.68 to 1.16 0.69 0.45 to 1.04

P, 0.05 for coefficients in bold. Data are available on n55316 unless stated. IMD, Index of Multiple Deprivation; SR, side room.an55217.bn53730.cn53536.dn54436.en54928.

CJASN 16: 1237–1246, August, 2021 COVID-19 in London In-Center Hemodialysis Patients, Caplin et al. 1243

in-center hemodialysis during the first phase of the pan-demic between March and May 2020. To our knowledge,aside from registry data, which lacks this degree of granu-larity (9), this report encompasses the largest number ofcases of COVID-19 in hospital-based patients on dialysisand is the first to examine the evolution of the disease inindividual dialysis units over time. This analysis providesimportant insights into the role of endemic disease, unitcharacteristics, mask policies, and isolation strategies onCOVID-19 transmission in patients on hemodialysis.Our data confirm that patients on dialysis are at high riskof symptomatic illness and hospital admission fromCOVID-19, with almost 20% of the patients on prevalentdialysis reaching the primary outcome of test positivity oradmission. This proportion is similar to those reportedfrom other European and UK centers (6, 10). However, it issubstantially higher than those reported in France (9, 11)and China (12), albeit over shorter time frames. Consistentwith the published data on patients without kidney disease(13) and patients with kidney failure (9), both age and dia-betes were strong risk factors for outcomes. Unlike otherstudies, we did not observe associations with ethnicity (14)or measures of deprivation. However, unlike others, weadjusted for community burden of disease, and these nega-tive findings were apparent even in those models adjustedsolely for individual-level characteristics. Therefore, it ispossible that the COVID-19 disease risk associated withdialysis treatment overwhelms the effect of these social/demographic factors.The importance of local community COVID-19 disease

as reflected in the time-dependent rates of disease nearpatients’ homes is highlighted in this study. The AUROCanalysis suggests that community disease burden explainsa substantial proportion of the variability in test positivityin this cohort, although only those in the general popula-tion with disease severity requiring hospitalization werebeing tested for COVID-19 during this period, making itdifficult to draw specific conclusions about the level ofcommunity transmission that associates with a particularrisk to patients on dialysis. These findings suggest that

disease in dialysis units is not occurring as isolated“outbreaks” within treatment centers, but rather is depen-dent on introductions from contacts, in patients’ homes,and from local communities. Consistent with this is theobservation that larger units suffered a higher burden ofdisease, with more potential direct and indirect contactsbetween patients and others in the community.

Nonetheless, a number of (inter-related) dialysis unitcharacteristics were also associated with the outcomes, sug-gesting that within-unit spread is also a factor. The numberof side rooms per dialysis station was strongly associatedwith both outcomes in all models and suggests that thecapacity to isolate patients with suspected or confirmeddisease is protective against onward transmission. Simi-larly, measures of distancing between patients, estimatedby total dialysis area or between-station distance, wereboth inversely associated with outcomes (although notindependent of other unit factors). Consistent with theimportance of asymptomatic transmission, the introductionof a mask policy for all patients was associated withreduced risk of admission. Although specific to this studypopulation and not generalizable, the modeled estimatessuggested that introducing a policy of asymptomaticpatient mask wearing was associated with a small absolutereduction in probability of admission. Furthermore,although overall a small effect (at least partly because theintroduction of mask policy varied only a few weeksbetween centers), this association was observed after a2-week lag, consistent with the reported time from infec-tion to hospitalization. The lack of a similar associationwith the test-positivity outcome may reflect lack of poweror an alternative biologic explanation. Taken together, theabove findings suggest that transmission from those with-out clinically suspected disease is an important factor indetermining COVID-19 burden in dialysis units (15).

Importantly, none of the varying isolation strategies usedby the different kidney centers were associated with out-comes. These policies were instituted in addition to univer-sal precautions mandated for the care of symptomaticpatients across all centers, so the power to detect what may

Table 4. Multivariable coefficients

CharateristicPositive Test or Admission Admission

Hazard Ratio 95% Confidence Interval Hazard Ratio 95% Confidence Interval

Age, per yr 1.01 1.00 to 1.01 1.02 1.01 to 1.03Men 0.95 0.83 to 1.09 1.03 0.84 to 1.41Ethnicity

Asian 1.06 0.88 to 1.29 1.00 0.75 to 1.33Black 1.08 0.90 to 1.30 1.12 0.86 to 1.45Other 0.88 0.70 to 1.13 0.71 0.49 to 1.04

Diabetes 1.16 1.04 to 1.41 1.21 0.99 to 1.48Median deprivation index rank, per 100 1.00 1.00 to 1.00 1.00 0.98 to 1.02Community cases, week 0, case per

output area in home postcode1.15 1.12 to 1.19 1.15 1.09 to 1.22

Side rooms, n/20 stations 0.88 0.83 to 0.95 0.88 0.77 to 1.01Prevalent patients, n 1.00 1.00 to 1.01 1.00 1.00 to 1.01Patient masks, week 22 0.82 0.64 to 1.04 0.64 0.44 to 0.93

P50.05 for coefficients in bold. Mutually adjusted. n54928

1244 CJASN

be small additional protective effects is likely to be low.These data suggest that community levels of COVID-19,along with a small contribution of unit factors and mask-ing, determine how the infection burden evolves. The rele-vance of these data for planning for future waves of diseasemay be modified by (1) unrestricted/asymptomatic testingthat is now available, (2) the use of surgical face masks byasymptomatic patients and those treating them, and (3) thepotential effect of vaccination or acquired immunity in thepopulation.This work has a number of limitations, including the

observational design, which restricts the conclusions thatcan be drawn about causality. Data were collected retro-spectively, and many of the dialysis unit–level and kidneycenter–level factors were on the basis of questionnaires,which could have been interpreted differently between cen-ters, despite the validation checks performed. We used adiscrete time survival approach and, therefore, cannotaccount for the effect of recovery or reinfection on diseasetransmission, but given the short time frame and, presum-ably, an initially fully susceptible population, this isunlikely to affect our conclusions. Although we used twodifferent outcome measures with broadly similar results,we will have missed large numbers of asymptomatic caseswith the test-positive outcome, as serologic studies confirmthat large numbers of patients in London dialysis unitsseroconverted who were either never tested or tested nega-tive during the early phase of the pandemic (16).In this comprehensive study of COVID-19 disease in the

dialysis units across London between March and May2020, we found that individual factors and local commu-nity case rates along with dialysis unit size and layout asso-ciate with the risk of disease. Masking of asymptomaticpatients was associated with a reduction in admissions,supporting the concept that transmission from asymptom-atic patients is important, but no differences were seenbetween the different infection control strategies aimed atmanaging those with confirmed or clinically suspected dis-ease. This work confirms the large burden of COVID-19experienced by the dialysis population and highlights theneed to prioritize this vulnerable group in strategies aimedat reducing spread, protecting individuals, and/or treat-ment of established disease.

DisclosuresM. Antonelou reports receiving research funding from Medical

Research Council. D. Ashby reports receiving honoraria from Fibr-ogen. D. Banerjee reports receiving research funding from the Brit-ish Heart Foundation; receiving grants from AstraZeneca and Kid-ney Research UK; and receiving honoraria from AstraZeneca,Pfizer, and Viforpharma. S.A. Blakey reports employment withWest London Renal and Transplant Centre. K. Bramham reportsconsultancy agreements with Alexion; receiving honoraria fromAlexion and Otsuka; and serving as a scientific advisor or memberof Alexion. B. Caplin reports consultancy agreements with LifeArcand receiving research funding from AstraZeneca and grants fromColt Foundation, Medical Research Council, and Royal Free Char-ity outside the submitted work. R. Corbett has been issued PatentWO2017148836A1: “A device for maintaining vascular con-nections.” M.L. Ford reports other interests/relationships with theAstraZeneca advisory board, which led to the publication of the

paper. A. Frankel reports receiving research funding from Boeh-ringer Ingelheim/Lilly Alliance and receiving honoraria fromAstraZeneca, Boehringer Ingelheim/Lilly Alliance, Merck Sharp &Dohme, Napp Pharmaceuticals Limited, and Novo Nordisk. R.Hull reports consultancy agreements with AstraZeneca, Pharmo-cosmos UK Ltd., and Travere Pharmaceuticals; speakers bureaufor Napp Phamaceuticals; serving as an elected council member ofRenal Association, UK; and other interests/relationships with JointSpecialist Committee Renal Medicine, Royal College Physicians,London. K. McCafferty reports receiving research funding fromAstraZeneca and receiving honoraria from Bayer, Napp, Pharma-cosmos, and Vifor Fresenius. A.D. Salama reports receivingresearch funding from Chiesi and Natera; receiving honorariafrom AnaptysBio, AstraZeneca, Hansa Medical, and Vifor Pharma-ceuticals; and serving as Nephrology Dialysis Transplantation Editorand a UK Renal Association Executive Member. C.C. Sharpereports consultancy agreements with Novartis Pharmaceuticals;receiving honoraria from Napp Pharmaceuticals; serving as anEditor for BMC Nephrology and as a trustee and treasurer for theRenal Association; and speakers bureau for Napp Pharmaceuticals.All remaining authors have nothing to disclose.

FundingNone.

AcknowledgmentsThe authors acknowledge the key role of the clinical nursing andmedical staff who also contributed to the Pan London Covid RenalAudit.D. Ashby, D. Banerjee, K. Bramham, B. Caplin, A. Frankel, E.Lioudaki, D. Makanjuola, K. McCafferty, A.D. Salama, and C.C.Sharpe designed the study; D. Ashby and B. Caplin analyzed thedata; M. Antonelou, E. Asgari, D. Banerjee, S.A. Blakey, D.C.B.Braide-Azikwe, K. Bramham, G. Clark, N. Cole, R. Corbett, H. Cro-nin, M.L. Ford, N.J. Hayes, N. Kumar, D. Makanjuola, B. Manson, K.McCafferty, T. Roper, A.D. Salama, A. Sarnowski, and V. Srinivasacollected the raw data; and all authors approved the finalmanuscript.

Supplemental MaterialThis article contains the following supplemental material online

at http://cjasn.asnjournals.org/lookup/suppl/doi:10.2215/CJN.03180321/-/DCSupplemental.

Supplemental Table 1. Isolation and deisolation strategies formanagement of patients with clinically suspected or confirmedCOVID-19 by center.

Supplemental Table 2. Sensitivity analysis—models using calen-dar week follow-up.

Supplemental Table 3. ROC curve analysis of models.Supplemental Table 4. Modeled absolute risks of admission

associated with a policy of asymptomatic patient masking 2 weeksfollowing introduction by week.

Supplemental Figure 1. Study flow chart.Supplemental Figure 2. In-center hemodialysis units dialyzing

test-negative patients in London during the COVID-19 pandemiccolored by responsible renal center.

Supplemental Figure 3. Hazard ratios by week using the twodifferent follow-up scales.

Supplemental Figure 4. Ethnicity distribution and associationsbetween unit-level characteristics.

Supplemental Material. Data collection tool and STROBE checklist.

CJASN 16: 1237–1246, August, 2021 COVID-19 in London In-Center Hemodialysis Patients, Caplin et al. 1245

References1. Valeri AM, Robbins-Juarez SY, Stevens JS, Ahn W, Rao MK,

Radhakrishnan J, Gharavi AG, Mohan S, Husain SA: Presenta-tion and outcomes of patients with ESKD and COVID-19. J AmSoc Nephrol 31: 1409–1415, 2020

2. UK Renal Registry: COVID-19 surveillance report for renalcentres in the UK: All regions and centres. Bristol, United King-dom, UK Renal Registry, 2020

3. Department of Health: Coronavirus (COVID-19) in the UK,2020. Available at: https://coronavirus.data.gov.uk. AccessedJune 16, 2020

4. Basile C, Combe C, Pizzarelli F, Covic A, Davenport A, KanbayM, Kirmizis D, Schneditz D, van der Sande F, Mitra S: Recom-mendations for the prevention, mitigation and containment ofthe emerging SARS-CoV-2 (COVID-19) pandemic in haemo-dialysis centres. Nephrol Dial Transplant 35: 737–741, 2020

5. Department of Health: Guidance on infection prevention andcontrol for COVID-19, 2020. Available at: https://www.gov.uk/government/publications/wuhan-novel-coronavirus-infection-prevention-and-control#history. Accessed July 1, 2020

6. Corbett RW, Blakey S, Nitsch D, Loucaidou M, McLean A,Duncan N, Ashby DR; West London Renal and TransplantCentre: Epidemiology of COVID-19 in an urban dialysis center.J Am Soc Nephrol 31: 1815–1823, 2020

7. Office for National Statistics: Index of Multiple Deprivation(December 2019) Lookup in England, 2019. Available at:https://geoportal.statistics.gov.uk/datasets/index-of-multiple-deprivation-december-2019-lookup-in-england. Accessed June16, 2020

8. Austin PC: A tutorial on multilevel survival analysis: Methods,models and applications. Int Stat Rev 85: 185–203, 2017

9. Couchoud C, Bayer F, Ayav C, B�echade C, Brunet P, ChantrelF, Frimat L, Galland R, Hourmant M, Laurain E, Lobbedez T,Mercadal L, Moranne O; French REIN registry: Low incidenceof SARS-CoV-2, risk factors of mortality and the course of ill-ness in the French national cohort of dialysis patients. KidneyInt 98: 1519–1529, 2020

10. Alberici F, Delbarba E, Manenti C, Econimo L, Valerio F, PolaA, Maffei C, Possenti S, Lucca B, Cortinovis R, Terlizzi V,Zappa M, Sacc�a C, Pezzini E, Calcaterra E, Piarulli P, GueriniA, Boni F, Gallico A, Mucchetti A, Affatato S, Bove S, BracchiM, Costantino EM, Zubani R, Camerini C, Gaggia P, Movilli E,Bossini N, Gaggiotti M, Scolari F: A report from the BresciaRenal COVID Task Force on the clinical characteristicsand short-term outcome of hemodialysis patients with SARS-CoV-2 infection. Kidney Int 98: 20–26, 2020

11. Keller N, Chantrel F, Krummel T, Bazin-Kara D, Faller AL,Muller C, Nussbaumer T, Ismer M, Benmoussa A, Brahim-Bouna M, Beier S, Perrin P, Hannedouche T: Impact of first-wave COronaVIrus disease 2019 infection in patients on hae-moDIALysis in Alsace: The observational COVIDIAL study.Nephrol Dial Transplant 35: 1338–1411, 2020

12. Xiong F, Tang H, Liu L, Tu C, Tian JB, Lei CT, Liu J, Dong JW,Chen WL, Wang XH, Luo D, Shi M, Miao XP, Zhang C: Clini-cal characteristics of and medical interventions for COVID-19

in hemodialysis patients in Wuhan, China. J Am Soc Nephrol31: 1387–1397, 2020

13. Williamson EJ, Walker AJ, Bhaskaran K, Bacon S, Bates C, Mor-ton CE, Curtis HJ, Mehrkar A, Evans D, Inglesby P, Cockburn J,McDonald HI, MacKenna B, Tomlinson L, Douglas IJ, RentschCT, Mathur R, Wong AYS, Grieve R, Harrison D, Forbes H,Schultze A, Croker R, Parry J, Hester F, Harper S, Perera R,Evans SJW, Smeeth L, Goldacre B: Factors associated withCOVID-19-related death using OpenSAFELY. Nature 584:430–436, 2020

14. Ioannou GN, Locke E, Green P, Berry K, O’Hare AM, Shah JA,Crothers K, Eastment MC, Dominitz JA, Fan VS: Risk factors forhospitalization, mechanical ventilation, or death among10,131 US veterans with SARS-CoV-2 infection. JAMA NetwOpen 3: e2022310, 2020.

15. Jones NR, Qureshi ZU, Temple RJ, Larwood JPJ, Greenhalgh T,Bourouiba L: Two metres or one: What is the evidence forphysical distancing in COVID-19? BMJ 370: m3223, 2020

16. Clarke C, Prendecki M, Dhutia A, Ali MA, Sajjad H, Shivaku-mar O, Lightstone L, Kelleher P, Pickering MC, Thomas D,Charif R, Griffith M, McAdoo SP, Willicombe M: High preva-lence of asymptomatic COVID-19 infection in hemodialysispatients detected using serologic screening. J Am Soc Nephrol31: 1969–1975, 2020

Received:March 2, 2021 Accepted: April 27, 2021

B.C. and D.A. contributed equally to this work.

* The Pan-London COVID-19 Renal Audit Group authors are Mar-ilina Antonelou, Elham Asgari, Damien Ashby, Debasish Banerjee,Sarah Blakey, Kate Bramham, Rosalind Brewster, DandisonbaBriade-Azikwe, Hugh Cairns, Ben Caplin, Paramit Chowdhury,Grace Clark, Nicholas Cole, Richard Corbett, Andrea Cove-Smith,Helen Cronin, Maria Davari, Abigail Dias, Kevin Evans, SuzanneForbes, Martin Ford, Andrew Frankel, Sahana Gnansampanthan,Catriona Goodlad, Nathan Hayes, Heidy Hendra, Richard Hull,Nicola Kumar, Eirini Lioudaki, Marina Loucaidou, David Makan-juola, Bethia Manson, Kieran McCafferty, Daniel McGuinness,Aegida Neradova, Mysore Phanish, Katherine Price, Ravi Rajakar-iar, Alex Rankin, Tayeba Roper, Noam Roth, Alan Salama, Alexan-der Sarnowski, Cassim Schott, Claire Sharpe, Vinay Srinivasa,Damir Tandaric, Sujanita Thyagarajan, Gisele Vajgel, and GregorYoung.

Published online ahead of print. Publication date available atwww.cjasn.org.

See related Patient Voice, “Keeping Dialysis Patients Safe,” andeditorial, “Evidence-Based Practices to Reduce COVID-19 Trans-mission in Dialysis Facilities,” on pages 1141–1142 and 1146–1148,respectively.

1246 CJASN