COVID-19 vaccine uptake trends in SARS-CoV-2 previously infected cancer patients

Highlights

• •Assessment of COVID-19 vaccination rate in a vulnerable patient population.
• •COVID-19 vaccine uptake among SARS-CoV-2 previously infected cancer patients is low.
• •Demographic and socioeconomic factors are associated with COVID-19 vaccine uptake.
• •COVID-19 vaccine education should be tailored for patients based on these factors.

Abstract

Purpose

Cancer patients are at high risk of developing severe illness from SARS-CoV-2 infection, but risk is lowered with receipt of COVID-19 vaccine. COVID-19 vaccination uptake among previously infected cancer patients may be influenced by an assumption of natural immunity, predicted weak immune response, or concerns about vaccine safety. The objective of this study was to evaluate COVID-19 vaccine uptake trends in cancer patients previously infected with SARS-CoV-2.

Materials and Methods

Medical records of 579 sequential cancer patients undergoing active treatment at Levine Cancer Institute who tested positive for COVID-19 between January 2020 and January 2021 were evaluated. Patients who died prior to vaccine eligibility were excluded from the analysis. Demographic, clinical, and COVID-19 related characteristics were analyzed to identify prognostic factors for COVID-19 vaccine uptake as this information could be important for health policy design for future pandemics.

Results

Eighty-one patients died prior to the availability of COVID-19 vaccines. The acceptance rate of COVID-19 vaccination among 498 previously infected cancer patients was 54.6%. Of the patients with known vaccination dates, 76.8% received their first vaccine by April 17th, 2021. As of November 30, 2021, 23.7.% of eligible patients were boosted. In univariate models, older age, female sex, higher income, solid tumor cancer type, and hormone therapy were significantly associated with higher vaccine uptake, while Hispanic/Latino ethnicity was significantly associated with lower vaccine uptake. In a multivariable model, age (OR 1.18, 95% CI 1.10–1.28; p < 0.001), female sex (OR 1.80, 95% CI 1.22–2.66; p = 0.003), and higher income (OR 1.11, 95% CI 1.01–1.22; p = 0.032), were predictive of COVID-19 vaccine uptake.

Conclusions

Overall, vaccine uptake was low among our cohort of previously infected cancer patients. Older age, female sex, and higher income were the only variables associated with COVID-19 vaccine uptake within this vulnerable patient population.

Introduction

Patients with active cancer are at increased risk for severe disease with SARS-CoV-2 infection. Multiple studies have confirmed that COVID-19 patients undergoing active cancer treatment have poor clinical outcomes, including higher risks of hospitalization and death, as compared to patients without cancer or patients with no recent cancer treatment [1], [2], [3], [4]. Early in the COVID-19 pandemic, the Levine Cancer Institute of Atrium Health formulated a proactive SARS-CoV-2 surveillance testing protocol and innovative patient care strategies to ensure continued and safe delivery of cancer care during the pandemic [5], [6]. We also established a data registry to collect clinical data of on-treatment patients who had a laboratory-confirmed SARS-CoV-2 infection, as this information could be important for health policy design for future pandemics.

Due to their vulnerability, cancer patients were initially prioritized for COVID-19 vaccination in the United States when vaccines became available under Emergency Use Authorization (EUA) in December 2020 [7], [8]. However, vaccine eligibility criteria were constantly evolving in 2021 and distribution patterns differed by state. For example, North Carolina switched to an age-based vaccination strategy (≥65 years were eligible) in January 2021. Cancer patients and other adults with increased risk of severe illness became eligible for vaccination in North Carolina on March 17, 2021.

Cancer patients infected with SARS-CoV-2 during 2020 and who survived their COVID-19 disease, should have been able to receive at least an initial dose of COVID-19 vaccine by mid-April 2021 in North Carolina. However, factors influencingvaccine uptake included the perceived speed of vaccine development, concern about short- and long-term vaccine side effects, lack of trust in COVID-19 related information, misinformation, and an assumption of natural immunity, if previously infected [9], [10]. Additional vaccine uptake factors specific to cancer patients in active treatment included the concern about the potential for interference with cancer treatment efficacy and for reduced immunological response due to immunocompromised status, especially for patients with hematological malignancies [11], [12].

Herein, we present the evaluation of COVID-19 vaccine uptake trends in cancer patients who were infected with SARS-CoV-2 between January 2020 and January 2021. Our aim was to identify factors associated with vaccine acceptance in previously infected patients who were undergoing active cancer treatment, potentially to help in designing health policy decisions for future pandemics. We hypothesized that COVID-19 disease severity, cancer type (solid tumor versus hematological), race/ethnicity, and household income would be associated with vaccination uptake. We also sought to determine if vaccine uptake was aligned with local and national waves of SARS-CoV-2 infection or with the FDA-approval date of the Pfizer-BioNTech vaccine.

Materials and Methods

3. Study Population.

This study received an expedited review and waiver of informed consent by the Atrium Health Institutional Review Board. Retrospective and prospective review was performed on adult (>18 years) cancer patients who tested positive for SARS-CoV-2 infection by COVID-19 PCR between January 1, 2020 and January 31, 2021 and received treatment at the Levine Cancer Institute within 12 months prior to COVID-19 diagnosis. Inclusion criteria for active treatment included patients receiving at least one of the following: surgical resection, hormone therapy, immunotherapy, chemotherapy, radiation therapy, transplant cellular therapy, targeted drug therapy, cryoablation, or radiofrequency ablation. Patients who died on or before April 17th, 2021 were excluded from the analysis, since we could not confirm if they had access to the vaccine prior to their death.

Study variables

Patient demographics, cancer-related clinical characteristics, and co-morbidities were captured from the electronic medical record. Clinical characteristics pertaining to the patient’s COVID-19 disease were also collected, including symptoms at presentation, symptoms at the peak of disease, hospitalization, maximum oxygen requirement, COVID-19-related treatment, and COVID-19 vaccination status. Vaccination status, including receipt of a booster dose, and the patient’s status (alive or deceased) at last clinical follow-up were recorded as of November 30, 2021. Publicly available data from the U.S. Census was used according to patient zip code [13]. Median 2018 household income by census tract was cross-walked to the zip code level and linked to patient zip code as a measure of the area surrounding the patient [14].

Eligibility dates for first vaccine doses were unknown, as the COVID-19 vaccination prioritization criteria in North Carolina were in flux from late 2020 through early 2021. All cancer patients in North Carolina became eligible for vaccination on March 17, 2021, irrespective of age or occupation. Therefore, we used April 17, 2021, which included a one-month grace period, as an arbitrary cutoff date to define “early” versus “late” vaccine uptake in our cohort. Since vaccine roll-out started in December 2020 at our institution and within the state, most vaccine access and roll-out related operational issues had been addressed by April 2021. Vaccine uptake was defined as receival of at least one vaccine dose by the study end date of November 30, 2021. Booster eligibility was determined by the completion date of the initial vaccination series. Patients who received the Janssen COVID-19 vaccine were eligible for a booster after two months while patients who received either the Pfizer-BioNTech BNT162b2 or Moderna mRNA-1273 vaccine series were eligible for a booster after six months [15].

Statistical analysis

Demographic and clinical characteristics were summarized with frequencies and percentages or medians and interquartile ranges (IQR), as appropriate. To account for subjects who had more than one cancer diagnosis, the variables of cancer type, cancer site, and cancer treatment were treated as individual binary variables. The primary endpoint, vaccination status, was analyzed as a binary variable using the November 30, 2021 cutoff date. To evaluate the relationship of demographic and clinical characteristics with vaccination status, univariate logistic regression models were estimated. Significant variables were included in a multivariable model, where backwards selection was used with a significance level of 0.15 to stay, to determine the final multivariable model. All statistical analyses were conducted in SAS Version 9.4 (SAS Institute, Inc., Cary, NC, USA) with a significance level of 0.05.

Results

Patient cohort

This study included 579 consecutive patients who were in active treatment for a hematological or solid malignancy at the Levine Cancer Institute and who tested positive for SARS-CoV-2 by PCR prior to the availability of COVID-19 vaccines (Fig. 1). Eighty-one patients died prior to vaccine availability in North Carolina and were excluded from the analysis. Of these 81 deaths, 35 were due to cancer progression and 25 were due to COVID-19 or COVID-19 related complications. Thus, a total of 498 patients were included in this analysis.

Fig. 1.

Consort diagram. 1Positive SARS-CoV-2 PCR test between March 20, 2020 and January 03, 2021. ²Death prior to April 18, 2021.

The cohort of 498 patients was predominantly female (61.4%) with a median age of 62 years (IQR 51.9–71.8) at first COVID-19 positive specimen collection (Table 1). Patients were 71.9% white, 22.3% black, 1.4% Asian and 5.8% were of Hispanic or Latino ethnicity. Cancer types were 79.8% solid tumor malignancy and 20.2% hematological malignancy. The most common cancer sites were breast (30.7%), genitourinary (10.8%), gastrointestinal (8.7%), and lung (8.0%). Common treatments included chemotherapy (49.2%), hormone therapy (27.7%), radiation therapy (24.9%), surgery (21.5%), and immunotherapy (16.3%). Approximately 50% of patients had multiple comorbidities with hypertension (42.4%), obesity (34.9%), anxiety and depression (20.3%), and diabetes (18.1%) being the most common.

Table 1.

Demographics and clinical characteristics of SARS-CoV-2 previously infected cancer patients.

Characteristic	Total Patients N = 498
Age, years (at first COVID-19 positive specimen)
Median	62.1
Interquartile Range	(51.9–71.8)
Sex
Female	306 (61.4%)
Male	192 (38.6%)
Race
Asian	7 (1.4%)
Black	111 (22.3%)
White	358 (71.9%)
Other	7 (1.4%)
Unknown	15 (3.0%)
Ethnicity
Non-Hispanic/Latino	461 (92.6%)
Hispanic/Latino	29 (5.8%)
Not specified	8 (1.6%)
Cancer type (N = 526)*
Solid tumor	420 (79.8%)
Hematology	106 (20.2%)
Cancer site (N = 573)*
Brain	5 (0.9%)
Breast	176 (30.7%)
Gastrointestinal	50 (8.7%)
Genitourinary	62 (10.8%)
Gynecologic	36 (6.3%)
Head and Neck	22 (3.8%)
Lung	46 (8.0%)
Lymphoma	32 (5.6%)
Melanoma	25 (4.4%)
Myeloid Cancers	29 (5.1%)
Other/Not Listed	24 (4.2%)
Plasma Cell	42 (7.3%)
Sarcoma	5 (0.9%)
Skin	19 (3.3%)
Cancer treatment†
Surgical resection	107 (21.5%)
Hormone therapy	138 (27.7%)
Immunotherapy	81 (16.3%)
Chemotherapy	245 (49.2%)
Radiation therapy	124 (24.9%)
Transplant cellular therapy	5 (1.0%)
Targeted drug therapy	60 (12.0%)
Cryoablation	3 (0.6%)
Radiofrequency ablation	1 (0.2%)
Number of comorbidities
0	98 (19.7%)
1	153 (30.7%)
2	96 (19.3%)
3	70 (14.1%)
≥4	81 (16.3%)
Smoking
Never smoker	290 (58.2%)
Former smoker	166 (33.3%)
Current smoker	39 (7.8%)
Unknown	3 (0.6%)
Symptomatic at COVID-19 diagnosis
No	218 (43.8%)
Yes	269 (54.1%)
Unknown	11 (2.2%)
Severity of COVID-19 symptoms (n = 329)
Mild	264 (80.2%)
Moderate	25 (7.6%)
Severe	12 (3.7%)
Critical	15 (4.6%)
Unknown	13 (4.0%)
Inpatient hospitalization
No	380 (76.3%)
Yes	88 (17.7%)
Unknown or NA	30 (6.0%)

Data are n (%). Due to rounding, not all variables may add up to 100%.

*Some patients had multiple cancer types (N = 526) and multiple cancer sites (N = 573).

†Percentages do not sum to 100%, as some patients received multiple treatments. A binary analysis was performed for each treatment.

At the time of COVID-19 diagnosis, 43.8% of all patients were not exhibiting symptoms and these infections were identified during routine screening prior to an office visit or procedure. Of those who were asymptomatic at the time of diagnosis, 21.6% subsequently developed symptoms. During the course of infection, 66.1% of all patients exhibited COVID-19 symptoms. Of these patients, symptoms were mild in 80.2%, moderate in 7.6%, severe in 3.7%, critical in 4.6%, and unknown in 4.0% according to the World Health Organization (WHO) criteria. The rate of hospitalization for the entire cohort of SARS-CoV-2 infected patients, was 17.7%. Of the hospitalized patients, 9.0% were admitted to an intensive care unit.

Vaccine uptake

A total of 272 previously infected cancer patients received at least one dose of COVID-19 vaccine prior to November 30, 2021, resulting in an uptake rate of 54.6% in this population. Of the patients with known vaccination dates, 76.8% received their first COVID-19 vaccine dose “early” before April 17, 2021, which was a date that served as a descriptive marker for temporal vaccine uptake behavior (Table 2). Vaccination dates were unknown for 17.6% of patients. All three vaccines that were approved by the FDA for Emergency Use Authorization (Pfizer-BioNTech BNT162b2, Moderna mRNA-1273, and Janssen COVID-19 vaccine) were represented in our cohort.

Table 2.

Time to first COVID-19 vaccine dose and booster status.

Vaccine Variable	Vaccinated Patients N = 272
First dose received by April 17, 2021*
Yes	172 (76.8%)
No	52 (23.2%)
Unknown	48
First dose manufacturer
Pfizer-BioNTech	154 (56.6%)
Moderna	53 (19.5%)
Janssen	6 (2.2%)
Unknown/Missing	59 (21.7%)
Booster received† (n = 228)
Yes	54 (23.7%)
No	130 (57.0%)
Unknown	44 (19.3%)

*Study-specific cutoff date for timely vaccination. All cancer patients in North Carolina.

were eligible for vaccination on March 17, 2021.

†Patient was eligible for booster as of November 30, 2021.

The patient status at last clinical follow-up (as of November 2021) is shown in Fig. 2. The death rate among unvaccinated and vaccinated patients was 14.6% and 2.6%, respectively. Of the 272 vaccinated patients, four (1.5%) were deceased prior to booster availability and 40 (14.7%) were not eligible for a booster as of November 30, 2021. Of those eligible for a booster (n = 228), the booster acceptance rate was 23.7% (n = 54) (Table 2).

Fig. 2.

COVID-19 vaccine uptake and patient status at last clinical follow-up.

The timeline and trend of COVID-19 vaccine uptake in our cohort of previously infected cancer patients is shown in Fig. 3A. For comparison, COVID-19 vaccine uptake within the general population of North Carolina and the United States, sourced from publicly available CDC data [16], are shown in Fig. 3B and 3C, respectively. The vaccine uptake peak in April 2021 coincided with the alpha variant wave of infection while the peak in August 2021 coincided with the delta variant wave. The Pfizer-BioNTech vaccine was FDA-approved on August 23rd, 2021 [17].

Fig. 3.

Vaccine uptake timeline for the cohort of previously infected cancer patients (A), the North Carolina general population [16] (B), and the United States general population [16] (C).

Variables associated with COVID-19 acceptance

Univariate logistic regression was performed to assess the demographic and clinical characteristics associated with COVID-19 vaccine acceptance in our cohort of previously infected cancer patients. Demographic variables associated with COVID-19 vaccine acceptance were older age (p < 0.001), female sex (p = 0.029), and higher median household income by zip code (p = 0.048) (Table 3). Solid tumor cancer type (p = 0.034) and hormone therapy (p = 0.020) were the only clinical variables associated with COVID-19 vaccine acceptance. Hormone therapy was administered to subjects with breast cancer (82.6%), prostate cancer (13.0%), ovarian cancer (1.4%), uterine cancer (1.4%), gastric cancer (0.7%), and hepatocellular carcinoma (0.7%). Hispanic/Latino patients were less likely to accept a COVID-19 vaccine (p = 0.016). By multivariable logistic regression analysis, the significant variables for vaccine acceptance among previously infected cancer patients were age (OR 1.18, 95% CI 1.02–1.28; p < 0.001), female sex (OR 1.80, 95% CI 1.22–2.66; p = 0.003), and median household income by zip code (OR 1.11, CI 1.01–1.22; p = 0.032) (Table 4).

Table 3.

Univariate logistic regression analysis to identify patient characteristics associated with receiving the COVID-19 vaccine.

Characteristic	Odds Ratio (95% CI)	P value
Age, 5 years (at first COVID-19 positive specimen)	1.17 (1.10, 1.26)	<0.001
Sex		0.029
Male	ref
Female	1.50 (1.04–2.16)
Race		0.392
White	ref
Asian	2.18 (0.42–11.41)
Black	1.33 (0.86–2.05)
Other	0.66 (0.14–2.97)
Unknown	0.58 (0.20–1.67)
Ethnicity		0.016
Non-Hispanic/Latino	ref
Hispanic/Latino	0.29 (0.13–0.68)
Not specified	0.77 (0.19–3.13)
Median Zip Code Household Income (2018), $10,000	1.10 (1.00, 1.20)	0.048
Number of comorbidities		0.328
0	ref
1	1.34 (0.81–2.23)
2	1.39 (0.79–2.45)
3	1.42 (0.77–2.63)
≥4	1.92 (1.05–3.50)
Severity of COVID-19 symptoms		0.259
Asymptomatic	ref
Mild/moderate	0.95 (0.65–1.38)
Severe/critical	1.94 (0.80–4.68)
Cancer type
Solid tumor	1.69 (1.04–2.76)	0.034
Heme	0.87 (0.57–1.34)	0.524
Cancer therapy
Surgical resection	0.93 (0.61–1.43)	0.752
Hormone therapy	1.61 (1.08–2.41)	0.020
Immunotherapy	1.11 (0.69–1.79)	0.669
Chemotherapy	0.73 (0.51–1.04)	0.078
Radiation therapy	0.89 (0.59–1.34)	0.570
Transplant cellular therapy	1.25 (0.21–7.54)	0.808
Targeted drug therapy	0.87 (0.51–1.50)	0.625

Table 4.

Multivariable logistic regression analysis to identify patient characteristics associated with receiving the COVID-19 vaccine.

Characteristic	Odds Ratio (95% CI)	P value
Age, 5 years (at first COVID-19 positive specimen)	1.18 (1.10, 1.28)	<0.001
Sex		0.003
Male	ref
Female	1.80 (1.22, 2.66)
Ethnicity		0.067
Hispanic/Latino	ref
Non-Hispanic/Latino	2.33 (0.97, 5.57)
Not specified	0.47 (0.04, 5.51)
Median Zip Code Household Income (2018), $10,000	1.11 (1.01, 1.22)	0.032

Discussion

To our knowledge, our study is the first to report the rate of COVID-19 vaccine uptake among cancer patients previously infected with SARS-CoV-2. We thought this to be an important issue as we did not know whether prior infection was associated with reluctance to have vaccination in this population of patients. Vaccine uptake among cancer patients is known to be lower than that within the healthy population [18]. Vaccine refusal is often due to an assumption by patients of reduced efficacy related to their disease and/or concern about negative interactions with cancer therapy [19]. For COVID-19 vaccines, refusal reasons also include a lack of confidence in vaccine development, distrust of regulatory agencies, and concern about adverse events [20], [21], [22].

The 54.6% COVID-19 vaccine uptake rate in our cohort of previously infected cancer patients is lower than the 71.6% uptake rate recently reported in a cohort of 208 cancer patients [21]. However, only 6% of patients in that study had a previous history of SARS-CoV-2 infection. In surveys conducted in France, Poland, Korea, and Italy between November 2020 and April 2021, reports of cancer patients’ willingness to receive a COVID-19 vaccine ranged from 67% to 83% [20], [23], [24], [25]. A small percentage of these patients (0.3% to 15.7%) reported previous SARS-CoV-2 infection on the surveys. Given reports of weaning immunity after natural infection [26] and the fact that cancer patients on chemotherapy are less likely to mount a robust immune response [27], the reason for low vaccine uptake in our cohort remains unclear. We did not conduct a survey to determine the reasons for vaccine acceptance and refusal in our cohort, as our study design relied solely on electronic medical record review of cancer patients that had tested positive for SARS-CoV-2. We did note that the higher death rate in the unvaccinated cohort (14.6% vs 2.6%) was mostly due to disease progression. It is plausible that patients with disease progression did not prioritize the COVID-19 vaccine. However, the overall number of deaths due to disease progression in the unvaccinated cohort was low and cannot completely explain the low vaccine uptake in our study.

By univariate logistic regression analysis, vaccine acceptance was significantly associated with older age, female sex, higher income, solid tumor cancer type, and hormone therapy. On multivariable logistic regression analysis, older age, female sex, and higher median income by zip code remained significantly associated with COVID-19 vaccine acceptance. Older age has been associated with poor COVID-19 clinical outcomes [28], [29] and for that reason, COVID-19 vaccine rollout prioritized older age and other high-risk groups [30]. Nationally, the CDC COVID-19 vaccine data tracker shows that the highest COVID-19 vaccine uptake occurred in the 50 to 64 age category but was lower in the 65 to 74 age category [16]. The correlation between older age and COVID-19 vaccine acceptance has been documented within the general population [31] and has also been reported in one cohort of cancer patients [21].

Data from the U.S. Census Bureau Household Pulse Survey (bi-weekly cross-sectional survey) identified female sex as a predictor of hesitancy for COVID-19 vaccination [32]. A similar finding was seen in a survey of older U.S. adults (>65 years) regarding COVID-19 vaccination, where women were less willing to receive a COVID-19 vaccine [33]. However, when we look at real-world data, the CDC COVID-19 vaccine tracker shows that as of March 2022, women outpace men in vaccine uptake (52.5% vs 47.8%, respectively) [16]. Our data are in line with the national trend, where female sex is associated with higher COVID-19 vaccine uptake. This finding highlights the impact of patient demographics on vaccine uptake and the importance of tailoring patient education around this gender gap.

Higher median income, as estimated at the population level from zip code data, was associated with higher vaccine uptake. This highlights the role communities and neighborhoods play in predicting health behaviors in cancer care [34], [35]. It is important to note that our analysis was performed at zip code level and is not reflective of individual economic data. COVID-19 vaccines were provided at no cost and our study included patients with access to healthcare (irrespective of insurance status), yet economically disadvantaged areas had lower vaccine uptake. Economic status is an important social determinant of health [36] and low-income communities should be prioritized for receival of health literacy regarding COVID-19 vaccination, as COVID-19 remains a public health emergency of international concern [37].

Overall, the trend of vaccine uptake within our cohort of previously infected cancer patients occurred earlier than in the general population, which peaked in mid-April 2021 during the alpha variant wave. The uptake peak in late January 2021 within our cohort is attributable to older cancer patients, as all individuals ≥ 65 years of age, irrespective of health status, became eligible in North Carolina for vaccination on January 14th, 2021. By April 17th, 2021, which was our study cutoff date for timely vaccination, systems and resources were in place within North Carolina for vaccination; therefore, vaccine access should not have been a contributing factor. Our data support the notion that patients who wanted to be vaccinated, did so early and did not wait. Incidentally, a small vaccine uptake peak is seen in our cohort and the general population near the end of August 2021. This could be attributable to FDA-approval of the Pfizer-BioNTech vaccine and/or the delta variant wave of infection, which both occurred at that time. We anticipated that the vaccination rate would increase during the study period due to the availability of more safety data; however, it did not.

Cancer Care teams should prioritize education of their patients on the importance of vaccination, irrespective of prior SARS-CoV-2 infection. Notably, recommendation of vaccination by the patients’ primary physician or oncologist was shown to improve the willingness of patients to get vaccinated from 61% to 91% [25]. However, it appears that only a small minority of physicians (<5%) are initiating conversations with their cancer patients about the importance of vaccination [24].

Our study does have limitations in framing broader health policy. It is limited to a single cancer institute notwithstanding our location in 25 sites in two states. Our data also lack specific social determinants such as education and employment status. The COVID-19 data registry that we established in March 2020 included patients in active treatment who tested positive for SARS-CoV-2, since the goal was to understand the impact of SARS-CoV-2 infection on cancer patients. Therefore, we are unable to determine the vaccine uptake rate of a control group of cancer patients with no history of SARS-CoV-2 infection and not on active treatment.

In conclusion, COVID-19 vaccine uptake is relatively low among our cohort of previously infected cancer patients. Our study highlights the role that demographic and economic factors play in important healthcare decisions. Cancer Care teams and those responsible for health policy should consider these variabilities and tailor patient education about vaccination accordingly.

Funding.

This research was supported in part by grants from the Leon Levine Foundation and the Kerry and Simone Vicar Foundation but did not receive any specific grant from other funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.