COVID-19 Infection and Outcomes in Newborn Screening Cohorts of Sickle Cell Trait and Sickle Cell Disease in Michigan and Georgia

Abstract

The sickle cell mutation increases morbidity in those with sickle cell disease (SCD) and potentially sickle cell trait, impacting pulmonary, coagulation, renal, and other systems that are implicated in COVID-19 severity. There are no population-based registries for hemoglobinopathies, and they are not tracked in COVID-19 testing. We used COVID-19 test data from 2 states linked to newborn screening data to estimate COVID outcomes in people with SCD or trait compared with normal hemoglobin. We linked historical newborn screening data to COVID-19 tests, hospitalization, and mortality data and modeled the odds of hospitalization and mortality. Georgia’s cohort aged 0 to 12 years; Michigan’s, 0 to 33 years. Over 8% of those in Michigan were linked to positive COVID-19 results, and 4% in Georgia. Those with SCD showed significantly higher rates of COVID-19 hospitalization than the normal hemoglobin Black cohort, and Michigan had higher rates of mortality as well. Outcomes among those with the trait did not differ significantly from the normal hemoglobin Black group. People with SCD are at increased risk of COVID-19–related hospitalization and mortality and are encouraged to be vaccinated and avoid infection. Persons with the trait were not at higher risk of COVID-related severe outcomes.

Individuals living with sickle cell disease (SCD) have been included in high-risk designations for poor COVID-19 outcomes.1–3 Early results of mortality data suggest that COVID-19 may not have increased mortality among the SCD population in the United States, apart from the oldest populations.4 Limited analyses of clinical data and an SCD/COVID registry have confirmed that COVID-19 poses an important health threat to those living with SCD.5,6 However, most studies are limited to individuals with SCD who are hospitalized for COVID-19 and case studies of fewer than 10 patients.7,8 Sickle cell trait (referenced as “trait” herein), the carrier status for SCD, is not a tracked condition, and people with the trait often are unaware of their status yet may be at higher risk for poor outcomes than those with normal hemoglobin due to differences in coagulation.9–11

People with SCD or trait may be at risk for increased morbidity from COVID-19. Viruses such as influenza H1N1 are more severe in SCD and are associated with an increased rate of SCD complications, including pain and acute chest syndrome.12,13 COVID-19 is a respiratory disease complicated by a hyperinflammatory response that can lead to multiorgan failure and severe coagulopathy.14 The sickle cell mutation induces pathophysiologic changes and clinical risk factors that increase morbidity in SCD and likely in trait.15 A chronic inflammatory thrombotic state with pre-existing, often undetected target organ injury characterizes SCD and may accelerate the morbidity of COVID-19.16–19 Renal impairment occurs in both SCD and trait, and renal function is susceptible to rapid decline following an acute illness like COVID-19.20–25 Those with SCD or trait are at risk of dehydration secondary to renal hyposthenuria, which may be a risk factor for COVID-19.22 When dehydration is combined with acidosis and/or hypoxemia, acute rhabdomyolysis and multiorgan failure can occur in both sickle states.26

COVID-19 is characterized by a silent hypoxia disease.27 Unphysiologic hypoxemia in trait, especially with hyposthenuria and physiological stress, causes an increased risk of rhabdomyolysis, sudden death, and vaso-occlusive complications.28,29 Because COVID-19 causes silent hypoxia, people with SCD or trait can have sudden drops in their oxygen saturation without the characteristic respiratory response of dyspnea or tachypnea. The undetected hypoxemia will induce polymerization with increased phosphatidylserine, leading to an inflammatory thrombotic state.30,31 This induces tissue infarction and amplification of the existing organ injury. Because of this, those with trait who are not being monitored with pulse oximetry may be at particular risk for severe complications from COVID-19.

An understanding of COVID-19 outcomes for people with these comorbid conditions is important. SCD impacts organs that COVID-19 also targets, such as the lungs and heart. Sickle cell trait’s health impacts are not well-researched, but the condition is known to affect individual and population health.32 Newborn screening provides the confirmation of SCD or trait status, and in some states, these records are linkable to COVID-19 test results and outcomes, providing an opportunity to evaluate COVID-19 and SCD or trait among a population-based, confirmed cohort. This analysis investigates Georgians and Michiganders with newborn screening-confirmed SCD or trait, COVID-19 positive test results, COVID-related hospitalization, and COVID-related mortality compared with the newborn screened population with normal hemoglobin.

MATERIALS AND METHODS

Our state-level population-based cohorts included all persons with newborn screening results showing normal hemoglobin, SCD (confirmed using hemoglobin electrophoresis), or trait in Michigan born 1987 through 2019, and in Georgia from 2008 through 2020. Data regarding COVID-19 positive test results, hospitalizations, and mortality were obtained from the Michigan Disease Surveillance System (MDSS) and the Georgia State Electronic Notifiable Disease Surveillance System (SendSS). This reflected all reported COVID-19 cases from March 2, 2020 through November 30, 2021, the end date coinciding with the onset of the Omicron variant of COVID-19 in the United States. Demographic data (age, sex, race, and ethnicity) and hemoglobin type were obtained from the Michigan and Georgia newborn screening programs. In Michigan, ethnicity was not available before 2003. The data sources used in this study were obtained from the Michigan Department of Health and Human Services and the Georgia Department of Public Health.

MDSS and SendSS data were used to identify all individuals in each state born between January 1, 1987 through December 31, 2019 (Michigan) and between January 1, 2008 through December 31, 2020 (Georgia), with at least 1 reported COVID-19 positive test result. These individuals were linked to newborn screening records. The final study population consisted of individuals with a positive COVID-19 test result that had a newborn screening result of normal hemoglobin, SCD, or trait.

The main exposure in the analyses was hemoglobin status; outcomes included infection, hospitalization, and mortality related to COVID-19. Confirmed COVID-19 infection was defined by a positive PCR test reported to the MDSS or SendSS. Georgia also accepted case reports through electronic lab reporting, fax, and calls from providers; it is mandatory for all positive COVID-19 test results from a laboratory to be reported to the health departments in each state. Michigan included individuals with multiple positive test results if each positive result was at least 90 days after the previously reported positive result, indicating reinfection. Reinfection data were not available in Georgia due to delayed capturing, missing information, and incomplete reporting. Therefore, each person could contribute only 1 positive test result.

The MDSS data set provides multiple indicators of hospitalization and mortality for those with positive test results. Michigan performs linkages with vital records to populate their mortality variables; therefore, deaths that were not indicated to be due to COVID-19 on the death certificate were attributed to COVID-19 if they occurred within 28 days of a positive COVID test. In Georgia’s SendSS data, the case was considered hospitalized if the individual was hospitalized at the time the case was reported or when the case was interviewed. The Georgia numbers do not capture hospitalizations that occur after a case was reported and, as such, may underestimate actual hospitalizations. Georgia deaths were recorded for individuals with confirmed COVID-19 who (1) were reported as deceased by health care providers or medical examiners/coroners, (2) were identified by death certificates with COVID-19 indicated as the cause of death, or (3) had evidence that COVID-19 contributed to the death.

Data were analyzed by state. For all individuals in the study population, proportions and frequencies were calculated for demographic characteristics of age (categorical), sex, race, and ethnicity. In the case of missing sex data from newborn screening in Michigan, the sex information from the COVID-19 database was used in linked cases. These characteristics were further stratified by hemoglobin status. The proportion of hospitalization and mortality for those with a reported COVID-19 positive test among each hemoglobin group was also calculated. The proportion of reinfections (positive test results at least 90 days apart, as defined above) among each group was assessed in Michigan.

Logistic regression models with generalized estimating equations were used with Michigan data to assess the relationship between COVID-19 hospitalization (yes/no) and hemoglobin status (SCD versus normal hemoglobin; trait versus normal hemoglobin) as well as COVID-19 mortality and hemoglobin status (SCD versus normal hemoglobin; trait versus normal hemoglobin). Generalized estimating equations were used to account for the correlation within individuals contributing multiple infections. As Georgia data are missing reinfection information, a standard logistic regression model was used to determine the relationship between COVID-19 hospitalization (crude and adjusted) and hemoglobin status, as above. Models for both states’ data were subset to include Black race-only individuals due to missing race data in all hemoglobin groups and adjusted for sex and age as these variables confound the relationship between hemoglobin status and COVID-19 outcomes.33 Odds ratios and 95% confidence intervals were calculated for each comparison.

Use of these data was determined to be exempt from review as a public health surveillance activity by the Georgia State University, the University of Michigan, and the Michigan Department of Health and Human Services institutional review boards. All analyses were performed using SAS 9.4.

RESULTS

Among COVID-19 positive test results in the age group parameters, 63% were linked to an individual with newborn screening results in Georgia (n = 70,394) and 59.4% in Michigan (n = 385,181) (Figs. 1A and 1B). Among 0 to 12-year-old children, Michigan’s linkage rate was similar to Georgia’s. Among the persons in the newborn screening data, 4.1% in Georgia were found among the positive COVID case data, and 8.1% in Michigan, reflecting the larger age range included in that state. In Georgia, 6.5% of those with SCD and 4.1% of those with the trait had positive COVID-19 test results. In Michigan, the proportions of persons identified as having SCD or trait who had positive COVID-19 test results were 11.2% and 6.6%, respectively (Tables 1 and 2).

FIGURE 1.

Study population linkages, COVID-19 infections, and newborn screening results, (A) Georgia and (B) Michigan*. *Probable cases are defined as at least 1 of the following: Meets clinical criteria AND epidemiologic linkage with no confirmatory laboratory testing performed for SARS-CoV-2. Meets presumptive laboratory evidence. Meets vital records criteria with no confirmatory laboratory evidence for SARS-CoV-2.

TABLE 1.

Demographics, Hospitalization, and Mortality of COVID-19 Test Positive Cases (March 2, 2020-November 30, 2021) and a Newborn Screening Result of normal Hemoglobin, Sickle Cell Disease, or Sickle Cell Trait (2008-2020; Georgia)

	Normal hemoglobin			Sickle cell disease			Sickle cell trait
	Births	COVID case		Births	COVID case		Births	COVID case
Georgia	n	n	%	n	n	n	n	n	%
Age
0–4 years	490,319	18,635	3.8	659	50	7.6	17,581	711	4.0
5–12 years	1,154,554	49,315	4.3	1615	97	6.0	39,166	1586	4.1
Sex
Male	838,298	34,281	4.1	1109	74	6.7	28,720	1181	4.1
Female	802,850	33,669	4.2	1158	73	6.3	27,862	1116	4.0
Unknown	3725	—	0.0	7	—	0.0	165	—	0.0
Race
Black	539,240	24,318	4.5	2073	139	6.7	49,077	2008	4.1
White	868,591	34,581	4.0	26	<11	—	3928	163	4.2
Asian	62,534	2011	3.2	47	<11	—	169	<11	—
American Indian/Native	2373	77	3.2	<11	—	—	42	—	0.0
Native HI or OPI	5796	192	3.3	—	—	—	79	—	0.0
Multiracial	30,605	1050	3.4	43	<11	—	1218	44	3.6
Unknown	135,734	5721	4.2	82	<11	—	2234	76	3.4
Ethnicity
Hispanic	193,193	9774	5.1	47	<11	—	2600	129	5.0
Not Hispanic	1,216,818	55,763	4.6	1781	139	7.8	45,252	2104	4.7
Unknown	234,862	2413	1.0	446	<11	—	8895	64	0.7
TOTAL	1,644,873	67,950	4.1	2274	147	6.5	56747	2297	4.1
Hospitalization*	—	735	1.1	—	35	23.8	—	27	1.2
Death	—	<11	—	—	—	—	—	—	—

^*In Georgia, each individual only contributes 1 infection in these analyses.

TABLE 2.

Demographics, Hospitalization, and Mortality of COVID-19 Test Positive Cases (March 10, 2020-November 30, 2021) and a Newborn Screening Result of Normal Hemoglobin, Sickle Cell Trait, or Sickle Cell Disease (1987-2019: Michigan)

	Normal hemoglobin			Sickle cell disease			Sickle cell trait
	Births	COVID Case		Births	COVID Case		Births	COVID Case
Michigan	n	n	%	n	n	%	n	n	%
Age
0–4 years	424,231	19,768	4.7	244	21	8.6	6,842	292	4.3
5–12 years	884,679	66,526	7.5	472	35	7.4	14,704	904	6.2
13–17 years	618,379	64,229	10.4	291	42	14.4	9301	809	8.7
18–33 years	2,038,258	171,392	8.4	1139	142	12.5	31,709	2124	6.7
Sex
Male	2,007,501	159,199	7.9	1064	116	10.9	31,557	1939	6.1
Female	1,916,389	161,659	8.4	1040	124	11.9	30,283	2174	7.2
Unknown	41,657	1057	2.5	42	<11	—	716	16	2.2
Race
Black	659,875	42,101	6.4	1585	226	14.3	50,829	3353	6.6
White	2,678,231	233,470	8.7	41	<11	—	4066	297	7.3
Asian/PI	65,607	2776	4.2	<11	<11	—	118	<11	—
American Indian/Native	15,235	1158	7.6	<11	<11	—	56	<11	—
Arab Descent/Middle Eastern†	118,588	11,265	9.5	14	<11	—	832	71	8.5
Multiracial	135,769	10,487	7.7	30	<11	—	2822	194	6.9
Unknown	292,242	20,658	7.1	474	<11	—	3833	205	5.4
Ethnicity‡
Hispanic	124,592	9294	7.5	14	<11	—	1026	77	7.5
Not Hispanic	1,297,745	92,973	7.2	629	58	9.2	19,258	1136	5.9
Unknown	2,543,210	219,648	8.6	1503	181	12.0	42,272	2916	6.9
Total Persons	3,965,547	321,915	8.1	2146	240	11.2	62,556	4129	6.6
	—	n	% of Infections	—	n	% of Infections	—	n	% of Infections
Hospitalization*	—	2123	0.7	—	45	18.4	—	72	1.7
Death	—	136	<0.1	—	<11	—	—	<11	—
Reinfections*	—	3227	1.0	—	<11	—	—	36	0.9
Total Infections	—	325,142	—	—	245	—	—	4165	—

^*Individuals in Michigan data may have more than one hospitalization or reinfection due to COVID-19.

^†Births 1987 to September 2003 classified as Middle Eastern; births after September 2003 classified as Arab Descent.

^‡Births 2003-2019 only.

The proportion of persons with COVID-19 requiring hospitalization was low in these young cohorts, except for those with SCD; 24% of cases were hospitalized among those with SCD in Georgia and 18% in Michigan. There were no deaths among the SCD or trait groups in Georgia, and the counts of deaths in these 2 groups from Michigan were too small to report (n <11); however, mortality odds ratios are modeled for Michigan below.

Those with SCD in Georgia had a 20.2 (CI 13.6, 30.0) and in Michigan a 14.8 (CI 10.4, 21.1) increase in odds of hospitalization compared with normal hemoglobin (Black race-only, Table 3). There were no COVID-related deaths among those with SCD in Georgia; adjusted odds of COVID-related death in Michigan were 11.2 (CI 3.4, 37.3) times the comparison group. Given the wider age cohort of newborn screening, Michigan was able to examine the effect of age and sex on hospitalization and mortality as well. With every year's increase in age, the odds of hospitalization are 0.95 (CI 0.94, 0.96), and the odds of death are 0.92 (CI 0.89, 0.96). Females have 1.5 (CI 1.3, 1.7) times greater odds of hospitalization compared with males. In contrast, among Black race individuals with COVID-19 positive test results who were linked to newborn screening results in both Georgia and Michigan, the crude and adjusted odds of hospitalization were not significantly different for those with the trait compared with those with normal hemoglobin. Similarly, crude and adjusted odds of mortality among those with the trait in Michigan were not higher than the normal hemoglobin group (among Black race individuals); there were no deaths among the trait group in Georgia.

TABLE 3.

Crude and Adjusted^‡ Odds Ratios for COVID-19 Hospitalization and Mortality Among Black Race Individuals With (1) Sickle Cell Disease and (2) Sickle Cell Trait as Compared With Normal Hemoglobin (Georgia and Michigan)

	Crude				Adjusted‡
	Hospitalization		Death		Hospitalization		Death
	OR	CI	OR	CI	OR	CI	OR	CI
Georgia*
Normal hemoglobin	Ref	—	—	—	Ref	—	—	—
Sickle cell disease	20.8	14.0, 30.8	—	—	20.2	13.6, 30.0	—	—
Sickle cell trait	1.0	0.7, 1.5	—	—	1.0	0.7, 1.5	—	—
Michigan†
Normal hemoglobin	Ref	—	Ref	—	Ref	—	Ref	—
Sickle cell disease	14.7	10.5, 20.6	12.1	3.7, 39.3	14.8	10.4, 21.1	11.2	3.4, 37.3
Sickle cell trait	1.2	0.9, 1.5	1.6	0.7, 3.9	1.2	0.9, 1.5	1.6	0.7, 3.8

^*Logistic models; there were no deaths among those with SCD or trait in Georgia.

^†Logistic models with generalized estimating equations to account for reinfections.

^‡Adjusted for sex and age.

We then also compared hemoglobin groups of all races combined (those with unknown or missing race were excluded) to examine the effect of race on this analysis of COVID-19 outcomes. Models were adjusted for age and sex: there were no deaths in Georgia, but adjusted odds of hospitalization were 21.7 (CI 14.8, 31.9) for those with SCD and not significant for those with the trait compared with those with normal hemoglobin. In Michigan, the adjusted odds of COVID-19–related hospitalization for those with SCD were 33.6 (CI 23.9, 47.2) and for the trait were 2.7 (CI 2.1, 3.4) compared with those with normal hemoglobin, and the adjusted odds of COVID-19–related mortality were 25.6 (CI 7.9, 82.9) for SCD and 3.3 (CI 1.5, 7.6) for the trait.

DISCUSSION

SCD is a severe disease that includes organ damage and a high risk of early mortality. This is the first large-scale, population-based look at SCD, traits, and COVID-19. We found that individuals with SCD are at a significantly increased risk of hospitalization and mortality due to COVID-19 as compared with those with normal hemoglobin, even when accounting for race. However, there is no increased risk observed among those with the trait. These results are even more striking considering the younger age cohort of individuals included in this study; as such, our results are likely underestimating the COVID-19 mortality rate among all groups in the study population. These results emphasize that people with SCD should be encouraged to be vaccinated for the virus and take precautions against exposure, and care providers should be aware that this is a group at high risk of mortality if infected with COVID-19.

The health-seeking behaviors and medical monitoring for those with SCD during the COVID-19 pandemic likely differ from those of the normal hemoglobin group and those with the trait, as has been shown in other severe chronic conditions.34 Therefore, the higher proportions of positive COVID-19 test results may point to more frequent health care encounters and testing compared with the general population versus a true higher rate of infection. In addition, the risk of serious infection is high among children with SCD, which necessitates hospitalization for even mild illness if there are signs of infection such as fever. High odds of hospitalization may be explained as a precaution in many cases; however, the reasons for hospitalization in these data are unknown. Michigan’s data also show that the adjusted odds of mortality among those with SCD are 15.9 times higher than the normal hemoglobin group, suggesting that COVID-19 should be considered a high-risk infection for those with SCD.

The implications of COVID-19 infection for those with the trait are challenging to determine—many with the trait are unaware of their status, as are their health care providers. Hoogenboom’s review (2021) of 11 studies of persons with known trait status found widely varying effects of COVID-19–related hospitalization and mortality among small cohorts and case studies.7 Singh (2021) also found no evidence of increased risk among 449 persons testing positive for COVID-19 with known trait status in a registry.6 Merz (2021) used blood samples to determine trait status among 166 Black persons with positive COVID-19 tests and found no evidence of more severe outcomes than those with normal hemoglobin.35 In Georgia and Michigan’s young cohorts of people with confirmed trait status, we found no evidence of increased odds of hospitalization or mortality. However, in other analyses, sickle cell trait appears to be a significant risk factor for morbidity in older individuals (mean age 64 years). In a recent comprehensive report on sickle cell trait by the Veterans Administration using biobank and medical records, increased morbidity, especially in renal disease, was observed.36 Given the higher rate of severe illness among the US Black population compared with the general population, vaccination and avoidance of the virus are recommended for those with the trait.33

Apart from the immediate implications to those with SCD, traits, and their health care providers, this work has other important indications. Population-wide public health surveillance is a critical need for diseases such as SCD. While there are limited registries and some states are tracking SCD/COVID-19 cases, there are no known population-wide efforts at this time.6 Clinical registries draw primarily from patient populations that are under the care of specialists such as hematologists. Adults with SCD have been found to have low rates of encounters with hematologists; therefore, they are likely to be missed in these data sources.37 To track the health care and outcomes of people with this severe, life-threatening disease should be a top state and federal priority. The 11-state Sickle Cell Data Collection Program, funded and led by the Centers for Disease Control and Prevention, is investigating COVID-19 infection rates and outcomes in multiple states across the lifespan, demonstrating the importance of a public health surveillance system for SCD.

The impact of race on those with COVID-19 infection is important; across the US, rates of hospitalization and mortality related to COVID-19 have been significantly higher for Black populations compared with White or to all populations.33 In our data, we see a slightly higher adjusted odds ratio of hospitalization for those with SCD in the all-races cohort compared with those with normal hemoglobin in Georgia, but a doubling of the odds in the Michigan cohort and a nearly 10-fold increase in odds of mortality. This suggests the effect of race is an important one, but when we remove this by comparing only the Black population cohorts, we see that SCD is still a significant risk factor for both hospitalization and death related to COVID-19.

This work points to important gaps for those with the trait that should be addressed. Research into the health implications of trait is deficient and should be promoted. If the condition is benign, there is a benefit for those with the trait knowing so. And if it does have health implications, such as vulnerability to emerging viruses, this is important to investigate.

While this work begins to fill gaps in information around SCD, traits, and COVID-19, it has limitations that could be addressed with better surveillance or in future research. First, the studied populations in both Georgia and Michigan are younger than the populations that have been most impacted by COVID-19. Tracking confirmed cases of older adults with SCD and trait is important to better understand the true impacts of the virus on these groups; a recent publication by Verna et al suggests that COVID-19 has a more severe impact on those with traits who have certain comorbid conditions, such as kidney disease, which warrants further investigation in larger cohorts.36 Second, while newborn screening data are of high value in confirming hemoglobin status, these cases are not tracked after birth; they may have moved out of state or died. Therefore, a case incidence rate cannot be determined. Third, the Georgia numbers do not capture hospitalizations that occur after a case was reported and, as such, may underestimate actual hospitalizations. Fourth, the data sources in this study do not contain information on morbidity, which may impact the magnitude of the associations between hemoglobin status and COVID-19 outcomes, particularly when considering that organ damage is more prevalent at a younger age among people with SCD. Finally, only COVID-19 cases reported to the states’ departments of public health were available to be linked. The recent availability of at-home testing means that the cases are likely under-reported (although this is likely to be true across all groups, not biasing the results).

These data and analyses are however the first large-scale, population-based look at SCD, traits, and COVID-19, and as such add significantly to what is known about these conditions when comorbid. The hemoglobin status and COVID-19 test results are confirmed, and the findings, even among this young group from only 2 states, are significant and point to an important high health risk for those with SCD. It should reinforce the health care providers' need to increase vaccination rates, which may be limited by families with vaccine hesitancy.38

FIGURE 2.

Crude and adjusted³ odd ratios for COVID-19 hospitalization and mortality among Black race individuals with (1) sickle cell disease or (2) sickle cell trait as compared with normal hemoglobin.¹ Logistic models; there were no deaths among those with SCD or trait in Georgia.² Logistic models with generalized estimating equations to account for reinfections.³ Adjusted for sex and age.