Effectiveness of Three Doses of mRNA COVID-19 Vaccines in the Hemodialysis Population during the Omicron Period

Visual Abstract

Abstract

Background

Coronavirus disease 2019 (COVID-19) vaccine effectiveness studies in the hemodialysis population have demonstrated that two doses of mRNA COVID-19 vaccines are effective against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and severe complications when Alpha and Delta were predominant variants of concern. Vaccine effectiveness after a third dose versus two doses for preventing SARS-CoV-2 infection and severe COVID-19 in the hemodialysis population against Omicron is not known.

Methods

We conducted a retrospective cohort study in Ontario, Canada, between December 1, 2021, and February 28, 2022, in the maintenance hemodialysis population who had received two versus three doses of mRNA COVID-19 vaccines. COVID-19 vaccination, SARS-CoV-2 infection, and related hospitalization and death were determined from provincial databases. The primary outcome was the first RT-PCR confirmed SARS-CoV-2 infection, and the secondary outcome was a SARS-CoV-2–related severe outcome, defined as either hospitalization or death.

Results

A total of 8457 individuals receiving in-center hemodialysis were included. At study initiation, 2334 (28%) individuals received three doses, which increased to 7468 (88%) individuals by the end of the study period. The adjusted hazard ratios (aHR) for SARS-CoV-2 infection (aHR, 0.58; 95% confidence interval [CI], 0.50 to 0.67) and severe outcomes (hospitalization or death) (aHR, 0.40; 95% CI, 0.28 to 0.56) were lower after three versus two doses of mRNA vaccine. Prior infection, independent of vaccine status, was associated with a lower risk of reinfection, with an aHR of 0.44 (95% CI, 0.27 to 0.73).

Conclusions

Three-dose mRNA COVID-19 vaccination was associated with lower incidence of SARS-CoV-2 infection and severe SARS-CoV-2–related outcomes during the Omicron period compared with two doses.

Introduction

Individuals receiving maintenance dialysis are at greater risk of mortality after severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection compared with the general population.¹ Previous studies have demonstrated that vaccine effectiveness in the dialysis population with two-dose vaccination is 68%–79% for the prevention of infection and 83% for severe outcomes, including SARS-CoV-2–related hospitalization and death.^2–6 However, these estimates were determined when the B.1.1.7 (Alpha) or B.1.617.2 (Delta) variants were the predominant variants of concern.^2,3,7 In Ontario, by December 13, 2021, over 80% of coronavirus disease 2019 (COVID-19) cases were attributed to Omicron,⁸ and by January 17, 2022, all COVID-19 cases were assumed to be Omicron (B.1 or B.2). Owing to concerns of waning serologic response after two vaccine doses,^9,10 individuals receiving dialysis in Ontario, Canada, were prioritized for a third dose of mRNA vaccine by December 2021. Serologic studies have found that a third dose of an mRNA vaccine significantly increases SARS-CoV-2 antibody levels.^11–14

Studies in the general population have reported lower vaccine effectiveness against Omicron infection, with a decrease from 93% to 70%.¹⁵ Three doses of mRNA COVID-19 vaccines provided additional protection against Omicron, but the vaccine effectiveness estimates were lower than those for Delta and waned quickly over time.¹⁶ The objective of this study was to determine the vaccine effectiveness of three versus two doses of mRNA COVID-19 vaccines in the in-center hemodialysis population. A secondary objective was to determine the association of prior SARS-CoV-2 infection with the risk of a future SARS-CoV-2 infection.

Methods

Study Population

This retrospective cohort study included all adults receiving outpatient maintenance in-center hemodialysis in the province of Ontario, Canada, as of December 1, 2021 (index date). Individuals receiving hemodialysis were identified using provincial health administrative databases, including the Ontario Renal Reporting System (ORRS), a mandatory provincial reporting system for regional kidney programs administered by the Ontario Renal Network.¹⁷ Individuals required a valid Ontario Health Insurance Plan (OHIP) number to permit linkage across provincial databases. OHIP is a single-payer insurance plan for residents of Ontario.

Exclusion criteria included individuals younger than 18 years; receiving acute dialysis; receiving home dialysis; starting dialysis after December 1, 2021; ChAdOx1 (Oxford-AstraZeneca) vaccine recipients (1.4% during the study period); unvaccinated; or had received only one mRNA vaccine dose.

Primary Exposure

The primary exposure was the receipt of a third dose of BNT162b2 or mRNA-1273. Vaccination information was obtained from the COVaxON database. This database includes information on all COVID-19 vaccinations administered in the province.⁶ In Ontario, third doses were first administered in August 2021 to a subset of individuals receiving dialysis who were also immunocompromised or in congregate living settings, such as long-term care homes. Individuals aged 70 years or older were prioritized beginning November 3, 2021, and the remaining dialysis population as of December 2, 2021.

The study period was between December 1, 2021, and February 28, 2022, and their vaccination status could change (from two doses to three doses) during the follow-up period as a time-varying exposure. Follow-up time extended up to 6 months after third dose of vaccination.

Outcomes

The primary outcome was laboratory-confirmed SARS-CoV-2 infection, defined as a positive RT-PCR result, obtained through data from the Ontario Laboratories Information System.

The indications for testing in-center hemodialysis patients in Ontario included (1) symptoms consistent with COVID-19, (2) known contact with an individual or patient with COVID-19 within the prior 10 days, or (3) admission to the hospital. Asymptomatic testing did not occur. The secondary outcome was a SARS-CoV-2–related severe outcome, defined as either hospitalization or death. Hospitalizations were identified using the Canadian Institute for Health Information's Discharge Abstract Database (CIHI-DAD) and Same Day Surgery (CIHI-SDS) database. For a hospitalization to be attributed to SARS-CoV-2, we required a positive RT-PCR test to have occurred within 14 days before or at any time during the admission. International Classification of Diseases—10th revision (ICD-10) hospital discharge codes were reviewed to determine whether the admission was primarily related to COVID-19. COVID-19–related deaths were identified using the Registered Persons Database (RPDB) and defined as death within 30 days after a positive test.

Individuals were followed from study entry to the first positive SARS-CoV-2 RT-PCR test (for the primary outcome analysis) or first severe outcome (for the secondary outcome analysis), discontinuation of dialysis (due to non–COVID-19–related death, recovered kidney function, transfer out of province, kidney transplant, or withdrawal from dialysis), transfer to a home dialysis modality, receipt of a fourth dose of an mRNA vaccine, or until the end of the study period. Individuals were followed for 30 days postinfection, with the time of the severe outcome being the earliest date of hospitalization or death.

Covariates and Baseline Characteristics

Demographic information included age, sex, and ethnicity. Ethnicity information is reported in the ORRS by each regional kidney program at the time of registration on the basis of charting by clinical staff who could ask individuals to self-identify ethnicity if necessary but are not mandated to do so. Baseline characteristics also included years on dialysis, diabetes, cardiac disease (history of myocardial infarction or congestive heart failure), cancer (including skin cancer), and Charlson Comorbidity Index using ICD-10 codes in the OHIP or CIHI-DAD databases (Supplemental Table 1). All individuals were given a base score of two on the Charlson Comorbidity Index for advanced CKD. The lookback period for baseline characteristics was 5 years from the date of study entry. Residency in long-term care was determined from physician billing codes in the OHIP database in the 6-month period before study entry. We linked residential postal codes to 2016 Statistic Canada Census data or LHIN Crosswalk to determine neighborhood income quintiles and Public Health Unit region.¹⁸ COVID-19–related variables included history of SARS-CoV-2 infection confirmed by RT-PCR test (dating back to March 2020), time from prior infection, number of prior RT-PCR tests, and time between each mRNA vaccine dose.

Statistical Analysis

Individuals were grouped as having received two versus three doses as of February 28, 2022, to compare the characteristics of the study population. They were summarized using standard statistical methods for categorical and continuous variables. Differences between groups were reported as standardized differences, with values >0.10 considered clinically relevant.¹⁹

For calculation of vaccine effectiveness, individuals could enter this study having received two or three doses. Receipt of the third dose was modeled as a time-varying exposure; therefore, an individual could contribute follow-up time to both the two-dose and three-dose groups if their status changed during follow-up. We considered the second-dose and the third-dose vaccination effective 7 days after administration.²⁰ Rates of SARS-CoV-2 infections and severe outcomes were reported per 100,000 patient-days.

Cox proportional hazards models using the counting process extension were used to determine the unadjusted and adjusted hazard ratio (HR) and 95% confidence interval (CI). The model was adjusted for age, sex, ethnicity, Public Health Unit region, Charlson Comorbidity Index, long-term care residence, cumulative time on dialysis, income quintile, and prior SARS-CoV-2 infection. Models were evaluated to test the proportional hazards assumption and model fit. If variables violated the proportional hazards assumption, stratified models were used. The cumulative incidence of infection and the hazard over time were plotted on the basis of the receipt of two or three doses.

In the primary analysis, individuals were followed until a SARS-CoV-2 infection or until follow-up ended. If an individual had a SARS-CoV-2 infection but did not have a severe outcome, they were censored 30 days after the infection in the severe outcome analysis.

A subgroup analysis was completed by prior SARS-CoV-2 infection, vaccine type, and time from vaccination to assess waning. Vaccine effectiveness was calculated as (1 – HR) ×100%.⁶ Interaction terms were included as a test for significance when examining whether there was heterogeneity in vaccine effectiveness in these subgroups.

All statistical analyses were conducted using R, version 3.6.1, and an a priori level of statistical significance was set at 0.05. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies (Supplemental Table 2).

Ethical Considerations

Ontario Renal Network, a part of Ontario Health, is a provincial agency that funds and manages services for individuals with CKD. Data collection and analysis were in accordance with Ontario Health's legislative authority under the Ontario Personal Health Information Protection Act, 2004.

Results

The study cohort included 8457 individuals receiving in-center hemodialysis in Ontario as of December 1, 2021. On that date, 2334 (28%) individuals had received three doses, and by study completion, 7468 (88%) individuals had received a third dose. Of those who received two doses, 98% received the same type of mRNA vaccine (BNT162b2 or mRNA-1273). Of the individuals who received three doses, 86% received the same type of vaccine (64% BNT162b2, 22% mRNA-1273, Supplemental Table 3). SARS-CoV-2 infections and vaccination status are presented in Supplemental Figure 1.

Study population characteristics as of February 28, 2022, are presented in Table 1. Data were missing in <1% for all variables except ethnicity, which was unknown in 4% of individuals. A total of 627 (7%) individuals had a history of prior SARS-CoV-2 infection, and they were less likely to receive a third dose. Individuals who had received only two doses as of February 28, 2022, were younger, less likely to be White, and more likely to be receiving dialysis for less than 1 year. There were regional differences in the frequency of three-dose vaccination. The frequency of SARS-CoV-2 RT-PCR testing was high, with over 50% of individuals tested ≥3 times before study entry.

Table 1.

Characteristics of the in-center hemodialysis population in Ontario, Canada, as of December 1, 2021, with vaccine status as of February 28, 2022

Characteristics	Two Doses	Three Doses	Standardized Difference
Individuals, n (%)	989 (12)	7468 (88)
Female, n (%)	424 (43)	2959 (40)	0.07
Age, yr, n (%)
18–49	205 (21)	728 (10)	0.31a
50–64	324 (33)	1891 (25)	0.16a
65+	460 (47)	4849 (65)	0.38a
Median (Q1–Q3)	63 (53–74)	70 (60–78)
Ethnicity, n (%)
Asian	56 (6)	578 (8)	0.08
Black	120 (12)	485 (6)	0.19a
Indian sub-continent	122 (12)	693 (9)	0.10a
White	462 (47)	4399 (59)	0.25a
Other non-Whiteb	183 (19)	1025 (14)	0.13a
Unknown	46 (5)	288 (4)	0.04
Vintage, yr, n (%)
<1	284 (29)	1762 (24)	0.12a
1 to <2	153 (15)	1355 (18)	0.07
2–5	282 (29)	2454 (33)	0.09
>5	270 (27)	1897 (25)	0.04
Median (Q1–Q3)	2.5 (1–5)	2.6 (1–5)
Long-term care residency, n (%)	53 (5)	488 (7)	0.05
Public Health Unit regions, n (%)
Central East	55 (6)	537 (7)	0.07
Central West	153 (16)	1274 (17)	0.04
Durham	49 (6)	332 (4)	0.02
Eastern	37 (4)	543 (7)	0.16a
North	78 (8)	581 (8)	0.00
Ottawa	53 (5)	472 (6)	0.04
Peel	171 (17)	658 (9)	0.25a
Southwest	87 (8)	675 (9)	0.01
Toronto	253 (26)	1807 (24)	0.03
York	41 (4)	513 (7)	0.12a
Missing	12 (1)	76 (1)
Diabetes, n (%)	536 (54)	4181 (56)	0.04
Cancer, n (%)	73 (7)	665 (9)	0.06
Cardiac disease, n (%)	363 (37)	2638 (35)	0.03
Charlson Comorbidity Index,c mean (SD)	4 (2)	4 (2)	0.00
Income quintiles, n (%)
1 (lowest)	347 (35)	2282 (31)	0.10a
2	232 (23)	1643 (22)	0.03
3	186 (19)	1399 (19)	0.00
4	126 (13)	1170 (16)	0.08
5 (highest)	95 (10)	937 (13)	0.09
Missing	3 (0.3)	37 (0.5)
RT-PCR tests before study start, n (%)
0	122 (12)	1289 (17)	0.14a
1	210 (21)	1323 (18)	0.09
2	134 (14)	953 (13)	0.02
3 or more	523 (53)	3903 (52)	0.01
Prior SARS-CoV-2 (RT-PCR), n (%)	104 (11)	523 (7)	0.12a

SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

^aVaccine status was defined as of February 28, 2022.

^bIncludes Pacific Islander, Aboriginal, Middle Eastern/Arab, Latin American, and other/multiracial. These categories are based on the classification scheme of the Canadian Organ Replacement Register and the Ontario Renal Reporting System.

^cCharlson Comorbidity Index includes chronic respiratory diseases, chronic heart diseases, hypertension, diabetes, immunocompromising conditions due to underlying diseases or therapy, autoimmune diseases, CKD, advanced liver disease, dementia/frailty, and history of stroke or transient ischemic attack.²⁵

The mean follow-up time was 82 days (±17) for individuals receiving two doses and 68 days (±28) receiving three doses. Of individuals who received two doses (N=989), 85% had a cumulative time of greater than 6 months from their second dose (Supplemental Table 4).

The unadjusted rates of SARS-CoV-2 infection and severe COVID-19 outcomes by vaccine status are presented in Table 2. The rates of SARS-CoV-2 infection were 150 and 141 per 100,000 patient-days in the two-dose and three-dose vaccination exposure groups, respectively. The proportional hazards assumption for the extended Cox model was satisfied for two-dose and three-dose exposure status. Stratified models were used for analyses where public health unit and long-term care variables violated the proportional hazards assumption.

Table 2.

SARS-CoV-2 infections, related hospitalization, and death by vaccine exposure

Outcomes	Two Doses		Three Doses		Hazard Ratio (95% CI)a
	N (% of cases)	Rate (100,000 days)	N (% of cases)	Rate (100,000 days)	Unadjusted	Adjusted
SARS-CoV-2 infectionsb	257	150	688	141	0.53 (0.45 to 0.61)	0.58 (0.50 to 0.67)
Severe outcomec	56 (22)	32	92 (13)	18	0.35 (0.25 to 0.49)	0.40 (0.28 to 0.56)
Hospitalizationsd	51 (20)	29	79 (12)	16	0.34 (0.24 to 0.49)	0.40 (0.28 to 0.58)
Mortalitye	15 (6)	8	29 (4)	6	0.32 (0.17 to 0.60)	0.33 (0.17 to 0.61)

Reference is two doses and no prior infection. SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; CI, confidence interval; HRs, hazard ratios.

^aData are shown as HRs (95% CI). HRs were adjusted for age, sex, ethnicity, Public Health Unit region, Charlson Comorbidity Index, long-term care residences, cumulative time on dialysis, income quintile, and prior SARS-CoV-2 infection.

^bIncludes SARS-CoV-2 infections that occurred ≥7 days after the third dose of vaccine.

^cSevere outcome is the composite of hospitalization or death related to SARS-CoV-2.

^dHospitalization were attributed to SARS-CoV-2 if a positive RT-PCR test occurred 14 days prior or during the admission.

^eDeath within 30 days of a positive RT-PCR test.

The risk for SARS-CoV-2 infection was lower with three doses of vaccine compared with two doses: adjusted hazard ratio (aHR), 0.58 (95% CI, 0.50 to 0.67) (Table 2). This difference between the adjusted HR and infection rates per 100,000 days may be attributed to the change in HR (Supplemental Figures 1–3).²¹ Compared with those who received two doses, individuals with three doses had a lower risk of severe outcomes (aHR, 0.40; 95% CI, 0.28 to 0.56), hospitalizations (aHR, 0.40; 95% CI, 0.28 to 0.58), and mortality (aHR, 0.33; 95% CI, 0.17 to 0.61) (Table 2). Other risk factors for infection were of non-White ethnicities, age 18–49 years, long-term care facility residence, and higher comorbidity index (Supplemental Table 5).

In total, 130 individuals were hospitalized with a positive SARS-CoV-2 RT-PCR, of whom 67 (52%) had COVID-19 as the most responsible diagnosis (ICD-10 U071). Six individuals had U071 as a later diagnostic code, but the primary condition was consistent with COVID-19 (cough, respiratory failure, and pneumonia) (Supplemental Table 6).

Effect of Prior COVID-19 Infection, Vaccine Type, and Timing of Vaccination

Prior SARS-CoV-2 infection was associated with a lower risk of reinfection during the Omicron period (aHR, 0.44; 95% CI, 0.27 to 0.73) compared with those without prior infection. In a subgroup analysis (Table 3), aHR for three doses with prior infection was 0.77 (95% CI, 0.39 to 1.54) versus aHR without prior infection was 0.57 (95% CI, 0.57 to 0.66). The interaction between prior infection and vaccine effectiveness was not significant (interaction P value = 0.469), and confidence intervals were wide. The lowest risk of SARS-CoV-2 was among those with prior SARS-CoV-2 infection and a third dose (aHR, 0.17; 95% CI, 0.12 to 0.25) compared with those with two doses and no prior infection (Supplemental Table 7).

Table 3.

Subgroup analysis for SARS-CoV-2 infection on the basis of prior SARS-CoV-2 infection and vaccine type

Outcomes	Two Doses		Three Doses		HRs (Unadjusted)a (95% CI)	HRs (Adjusted)a (95% CI)
Subgroup	Infections, N	Rate (100,000 days)	Infections, N	Rate (100,000 days)
No prior infection	240	154	653	144	0.51 (0.44 to 0.59)	0.57 (0.48 to 0.66)
Prior infection	17	109	35	102	0.72 (0.40 to 1.31)	0.77 (0.39 to 1.54)
mRNA vaccine typeb
All BNT162b2	257	156	479	151	0.42 (0.36 to 0.48)	0.50 (0.43 to 0.59)
All mRNA-1273			139	124	0.32 (0.26 to 0.40)	0.34 (0.27 to 0.42)
Two BNT162b2 and one mRNA-1273			45	122	0.35 (0.25 to 0.48)	0.38 (0.27 to 0.54)

Reference is two doses. SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; HRs, hazard ratios; CI, confidence interval.

^aData are shown as HRs (95% CI). HRs were adjusted for age, sex, ethnicity, Public Health Unit region, Charlson Comorbidity Index, long-term care residences, cumulative time on dialysis, and income quintile.

^bmRNA vaccine subtypes reported for three doses. Combinations of mRNA vaccine described for the third dose: all BNT162b2, all mRNA-1272, and two BNT162b2/one mRNA-1273. Additional vaccine combinations made up a minority and are not reported.

The aHR for SARS-CoV-2 infection for three doses of BNT162b2 was 0.50 (95% CI, 0.43 to 0.59), while aHR for three doses of mRNA-1273 was 0.34 (95% CI, 0.27 to 0.42); however, this was not significantly different (interaction P value = 0.176). There was also no significant difference for two BNT162b2 and one mRNA-1273: aHR, 0.38 (95% CI, 0.27 to 0.54, interaction P value = 0.176) (Table 3).

Waning was further investigated, with results in Table 4. Individuals ≤3 months after a third dose had an aHR of 0.54 (95% CI, 0.46 to 0.64), with reference to the two-dose group at ≥6 months. This HR of 0.54 at ≤3 months after third-dose vaccination was lower in comparison with >3 to 6 months (aHR, 0.92; 95% CI, 0.71 to 1.18). This difference was statistically significant (P < 0.01).

Table 4.

SARS-CoV-2 infection rate by time since vaccination in months

	Follow-Up Time (d)	Number of Infected Cases	Rate per 100,000 Person-days	Unadjusted HR (95% CI)	Adjusted HRa (95% CI)
Three doses
≤3 mo	419,011	567	135.32	0.49 (0.42 to 0.58)	0.54 (0.46 to 0.64)
>3 to <6 mo	68,403	121	176.89	0.8 (0.64 to 1.01)	0.92 (0.71 to 1.18)
≥6 mo	25	0	0	NB: Excluded from analysis	NB: Excluded from analysis
Two doses
≤3 mo	4891	13	265.79	1.23 (0.7 to 2.15)	1.15 (0.65 to 2.05)
>3 to <6 mo	29,172	39	133.69	0.96 (0.68 to 1.36)	0.93 (0.66 to 1.32)
≥6 mo	137,390	205	149.21	Ref.	Ref.

SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; HR, hazard ratio; CI, confidence interval; Ref., reference.

^aHRs were adjusted for age, sex, ethnicity, Public Health Unit region, Charlson Comorbidity Index, long-term care residences, cumulative time on dialysis, income quintile, and prior SARS-CoV-2 infection.

Discussion

This study reveals that a third dose of mRNA vaccine, compared with two doses, was associated with a lower risk of SARS-CoV-2 infection and severe outcomes during the Omicron period. Severe COVID-19 outcomes were lower during the Omicron period relative to earlier waves of the pandemic, with less than one third of patients requiring hospitalization and <10% case fatality.²² Additional significant findings include that prior SARS-CoV-2 infection was protective during the Omicron period. There is also evidence that waning of protection occurs 3–6 months after receipt of a third dose of mRNA vaccination.

Despite rising community rates of Omicron during our study period, the third dose was associated with a 42% reduction in SARS-CoV-2 infection and a 60% reduction in severe outcomes compared with two doses. Our results are congruent with other studies demonstrating the significance of a third vaccine dose in the hemodialysis population, with some important differences.^12,23,24 Spensley et al. found an adjusted vaccine effectiveness of 50% against Omicron infection in individuals who received a third dose versus partial vaccination (one or two doses).²⁴ Montez-Rath et al. had similar findings, with a 50% reduction in Omicron infection in individuals with three doses compared with those who were unvaccinated.¹² Both Montez-Rath and Spensley analyzed receipt of vaccine as a baseline variable. We compared the third-dose group with individuals receiving two doses, which is a clinically relevant comparison, and analyzed the vaccine as a time-varying covariate and therefore were able to account in the large change in vaccine status over time.

The primary analysis in Ashby et al. found a 51% reduction in hospitalizations after receipt of a third dose of mRNA vaccine compared with two doses. There was no clear reduction in SARS-CoV-2 infection compared with two doses.²³ We did find a reduction in both infection and severe outcomes with receipt of a third dose, with a larger sample size (n=8457 versus 1126) and province-wide data, which likely increases generalizability.

Spensley et al. evaluated the effect of prior infection. They found that prior infection alone, and with one or two doses, did not change the likelihood of reinfection.²⁴ Our study differs in that prior infection, independent of vaccine status, was associated with a lower risk of reinfection. We found no significant difference in vaccine effectiveness on the basis of prior infection, although the aHR was higher in the prior infection group with confidence intervals crossing one. To emphasize the value of vaccination with prior infection, we demonstrated the lowest risk of infection was in individuals with a third dose and prior infection.

Our study was also novel in that it examined the waning of vaccine effectiveness during the Omicron period. In the general population, large studies with sufficient power have demonstrated that vaccine effectiveness wanes quickly with the Omicron variant. A study in the general population in the United Kingdom found vaccine effectiveness waned rapidly from 20 weeks after the second dose of any vaccine and decreased significantly 10 weeks after receipt of the third dose during the Omicron period.¹⁶ We further analyzed waning as a time-dependent covariate. aHRs were lower in individuals at ≤3 months after a third dose compared with >3 to 6 months. This difference was significant, indicative of waning of vaccine effectiveness after 3 months during the Omicron period. Furthermore, 85% of individuals who received only two doses throughout the study were more than 6 months from their second dose, at which point immunity was likely waning.

A strength includes that it is a large study using administrative data that captures the entire hemodialysis population of Ontario, Canada (population approximately 15 million), their vaccine status, and COVID-19 cases. During the Omicron period, given the exponential rise in infections, there was a decrease in RT-PCR testing capacity, but in-center hemodialysis patients continued to be tested because of frequent visits.

A limitation of this study is the lack of a specific variant of concern testing. However, Omicron was the overwhelmingly predominant variant of concern during the study period on the basis of epidemiologic trends, as outlined in the Methods section. Other studies, including Ashby et al., have demonstrated with variant testing that Omicron was predominant by early January.²³ Regarding testing, while it is possible that some patients performed rapid antigen testing at home that would not be captured reliably, given the high frequency of RT-PCR testing in the dialysis population and infection control measures, we do not think that a large number of infections would have been missed. Further limitations included the observational design and the potential for residual confounding. Our vaccination effectiveness estimates may differ from other designs, including test-negative designs, which rely on patient-reported symptoms.

In conclusion, in this large study involving all in-center hemodialysis patients in Ontario, a third dose of mRNA vaccine was associated with a 42% lower risk of SARS-CoV-2 infection (relative to two doses) and 60% lower risk of severe outcomes during the Omicron period. A third dose conferred benefit even in those with previous SARS-CoV-2 infection, justifying the proactive strategy of prioritizing third doses in hemodialysis patients. This study highlights that waning of immunity occurred at 3–6 months after receipt of a third dose during the Omicron period, and the hemodialysis population continues to be at high risk and should be prioritized for all available COVID-19–related interventions. Ongoing studies are needed to evaluate vaccine effectiveness in response to emerging variants of concern and the timing of additional doses in response to the waning of vaccine effectiveness in this vulnerable population.

Supplementary Material

cjasn-18-491-s001.pdf ^{(565KB, pdf)}

Disclosures

M. Atiquzzaman reports employment with Providence Health Care. P.G. Blake reports honoraria from Baxter Global and Otsuka Australia, an advisory or leadership role as medical director of Ontario Renal Network (a paid role), and serving on the Editorial Board of American Journal of Nephrology. S. Balamchi, J. Ip, D. Thomas, and A. Yeung are salaried employees of Ontario Health. M.A. Hladunewich reports research funding from Calliditas Therapeutics, Chemocentryx, Chinook, Ionis, Pfizer, and Roche; honoraria from UpToDate; and other interests or relationships as medical lead for Glomerular Disease Ontario Renal Network. J.A. Leis reports employment with Sunnybrook Health Sciences Centre. J.A. Leis has received payment for expert testimony as requested by hospitals of the Ontario Hospital Association, Ministry of Attorney General of Ontario, and Seneca College. A. Levin reports employment with BC Provincial Renal Agency and Providence Health Care; consultancy agreements with AstraZeneca, Bayer, Boehringer Ingelheim, Chinook Therapeutics, GSK, Janssen, Kidney Foundation of Canada, NIH, Otsuka, and REATA; research funding from AstraZeneca, Boehringer Ingelheim, Canadian Institute of Health Research (CIHR), CITF (Canadian Immunology Task Force), GSK, Health Research BC, Kidney Foundation of Canada, MOH BC, and Shared Care BC; honoraria from AstraZeneca, Bayer, GSK, Janssen, and NIH; advisory or leadership roles for AstraZeneca, Boehringer Ingelheim, CADTH, Chinook Therapeutics, CITF, GSK, KRESCENT (Kidney Scientist Education Research National Training Program), NIDDK, REATA BC Renal (Exec Director), and Steering Committee Chair CURE Consortium; DSMB for NIDDK, Kidney Precision Medicine, U Washington Kidney Research Institute Scientific Advisory Committee; International Society of Nephrology Research Committee; and other interests or relationships as CREDENCE National Coordinator from Janssen—directed to her academic team, NIDDK CURE Chair Steering Committee, International Society of Nephrology, Canadian Society of Nephrology, Kidney Foundation of Canada Steering Committee ALIGN trial, and DSMB Chair RESOLVE Trial (Australian Clinical Trial Network). E. McArthur reports employment with ICES. K. Naylor reports employment with London Health Sciences Centre. M.J. Oliver is the sole owner of Oliver Medical Management Inc., which is a private corporation that licenses the Dialysis Measurement Analysis and Reporting (DMAR) software system. Oliver Medical Management Inc. is co-owner of a Canadian Patent for DMAR systems. M.J. Oliver is a contracted medical lead at Ontario Renal Network Ontario Health, reports honoraria from Baxter Healthcare, participated in advisory boards for Amgen and Janssen, and reports other interests or relationships with Ontario Health. J. Perl reports grants from the Agency for Healthcare Research and Quality during the conduct of the study; personal fees from AstraZeneca Canada, Baxter Healthcare, DaVita Healthcare Partners, DCI, Fresenius Medical Care, LiberDi, Otsuka, and US Renal Care; research funding and salary support from Arbor Research Collaborative for Health and Agency for Health Research and Quality; speaker bureau for Baxter Healthcare and Fresenius Medical Care; and is on the advisory board for Liberdi, outside of the submitted work. All remaining authors have nothing to disclose.

Funding

This project was funded by the COVID-19 Immunity Task Force (Grant Number: 2122-HQ-000071).

Author Contributions

Conceptualization: Peter G. Blake, Kyla Naylor, Matthew J. Oliver.

Formal analysis: Shabnam Balamchi, Eric McArthur, Doneal Thomas.

Investigation: Matthew J. Oliver.

Methodology: Mohammad Atiquzzaman, Shabnam Balamchi, Peter G. Blake, Stephanie N. Dixon, Eric McArthur, Kyla Naylor, Matthew J. Oliver, Doneal Thomas.

Project administration: Angie Yeung.

Supervision: Matthew J. Oliver.

Writing – original draft: Adeera Levin, Matthew J. Oliver, Sara Wing, Kevin Yau.

Writing – review & editing: Mohammad Atiquzzaman, Shabnam Balamchi, Peter G. Blake, Stephanie N. Dixon, Michelle A. Hladunewich, Jane Ip, Jeffrey C. Kwong, Jerome A. Leis, Adeera Levin, Eric McArthur, Kyla Naylor, Matthew J. Oliver, Jeffrey Perl, Sara Wing, Kevin Yau, Angie Yeung.

Data Sharing Statement

All data used in this study are available in this article.

Supplemental Material

This article contains the following supplemental material online at http://links.lww.com/CJN/B641.

Supplemental Table 1. Diagnostic codes.

Supplemental Table 2. Strobe statement.

Supplemental Table 3. Vaccine combination in three-dose group.

Supplemental Table 4. Time between mRNA vaccine doses and time from prior SARS-CoV-2 infection by vaccine exposure.

Supplemental Table 5. Adjusted hazard ratios (aHRs) for risk factors of SARS-CoV-2 infection.

Supplemental Table 6. Hospitalization of maintenance hemodialysis patients, summary of hospital codes.

Supplemental Table 7. Risk of SARS-CoV-2 by prior infection and number of mRNA COVID-19 vaccine doses.

Supplemental Table 8. Reason for study exit.

Supplemental Figure 1. Evolution of SARS-CoV-2 cases over time and receipt of third dose of COVID-19 vaccine during the study period from December 1, 2021, to February 28, 2022, during which Omicron was the dominant variant of concern.

Supplemental Figure 2. Hazard of SARS-CoV-2 infection over the study period (December 1, 2021, to February 28, 2022) demonstrating an increase in the hazard of SARS-CoV-2 over time during the Omicron wave.

Supplemental Figure 3. Cumulative incidence of SARS-CoV-2 by receipt of two or three vaccine doses.