Obesity and age are transmission risk factors for SARS-CoV-2 infection among exposed individuals

Abstract

The coronavirus disease (COVID-19) pandemic has occurred in Massachusetts in multiple waves led by a series of emerging variants. While the evidence has linked obesity with severe symptoms of COVID-19, the effect of obesity on susceptibility to SARS-CoV-2 infection remains unclear. Identification of intrinsic factors, which increase the likelihood of exposed individuals succumbing to productive SARS-CoV-2 infection could help plan mitigation efforts to curb the illness. We aim to investigate whether obese individuals have a higher susceptibility to developing productive SARS-CoV-2 infection given comparable exposure to nonobese individuals. This case–control study leveraged data from the Mass General Brigham's (MGB) electronic medical records (EMR), containing 687,813 patients, to determine whether obesity at any age increases the proportion of infections. We used PCR results of 72,613 subjects who tested positive to SARS-CoV-2 or declared exposure to the virus independently of the result of the test. For this study, we defined susceptibility as the likelihood of testing positive upon suspected exposure. We demonstrate evidence that SARS-CoV-2 exposed obese individuals were more prone to become COVID positive than nonobese individuals [adjusted odds ratio = 1.34 (95% CI: 1.29–1.39)]. Temporal analysis showed significantly increased susceptibility in obese individuals across the duration of the pandemic in Massachusetts. Obese exposed individuals are at a higher risk of getting infected with SARS-CoV-2. This indicates that obesity is not only a risk factor for worsened outcomes but also increases the risk for infection upon exposure. Identifying such populations early will be crucial for curbing the spread of this infectious disease.

Significance Statement.

Large-scale studies have focused on the negative impact of coronavirus disease (COVID-19) on obese individuals, increasing the risk of worse outcomes and mortality. Here, we provide insight into the relationship between obesity and an increased risk of developing a productive SARS-CoV-2 infection. We leveraged the information of 72,613 subjects who tested positive for SARS-CoV-2 or declared exposure to the virus independently of the result of the test as a proxy for direct exposure to SARS-CoV-2. This study highlights how obesity plays an important role as an intrinsic risk factor for the spread of COVID-19. This new evidence could aid in redistributing resources like vaccines toward individuals at higher risk for positivity and, hence, would also help to curb the spread of illness.

Introduction

The SARS-CoV-2 coronavirus outbreak occurred in December of 2019 (1). With an initial estimated reproduction number (R₀) of 2–3 (2, 3), it rapidly spread worldwide, escalated to a pandemic by March 2020, and has been continuously evolving. SARS-CoV-2 is transmitted through close contact between humans (4). Consequently, hand hygiene (5), use of masks (6), and social distancing have proven to be efficacious as personal prophylactic measures. The development of multiple vaccines has helped mitigate the severity of subsequent waves, but their effects depend on the efficacy of vaccine distribution within each community (7). Furthermore, newly emerging mutants have shown different capacities to spread and varying resistance to vaccinations (7). Identification of at-risk populations which have increased susceptibility to developing productive infection, where viral replication is active, upon suspected exposure to SARS-CoV-2 carriers is, therefore, critical to counter the spread of the disease.

Previous studies described the clinical spectrum of the disease, from asymptomatic to mildly symptomatic (with fever, cough, and fatigue) and finally in critical cases leading to severe acute respiratory syndrome (SARS) and death (1, 8). Preexisting conditions, such as type 2 diabetes mellitus (T2DM), hypertension, obesity, and other cardiovascular disease increase the risk of severe symptoms (9–12). An early report of 5,700 coronavirus disease (COVID-19) patients in New York City showed that 41.7% of hospitalized patients listed obesity as a co-morbidity (13). A meta-analysis showed that obesity and COVID-19 mortality are correlated (14). One probable reason is that obesity and COVID-19 share similar inflammatory processes, which may aggravate the infection (15). Another factor that could contribute to higher mortality is that morbidly obese patients have difficulty with airway management (16).

Angiotensin I converting enzyme type 2 (ACE2) is a surface receptor that physiologically cleaves angiotensin I and pathologically acts as a receptor for SARS-CoV-2 to infect cells (17). While strategies to counter ACE2-Spike interaction have been tested (18) in evaluating SARS-CoV-2 infectivity, little is known about how metabolic diseases impact SARS-CoV-2 susceptibility. Prior evidence (19) exists that intrinsic factors such as obesity increase severity of outcome. The impaired rapidity and poor-quality immune response (20) in patients with metabolic syndrome could contribute to weakened SARS-CoV-2 viral clearance (21). In this regard, there remains the need to identify whether obesity is associated with increased risk of SARS-CoV-2 infection. We hypothesize that obese individuals have higher susceptibility to developing productive SARS-CoV-2 infection given comparable exposure to nonobese individuals.

In this study, we used an electronic medical record (EMR)-based approach to evaluate the rate at which an individual with suspected exposure to a SARS-CoV-2 infected individual tests positive for the virus using real time PCR assay. We assessed susceptibility to infection by using a self-reported measure: suspected risk of exposure as a proxy of contact with the SARS-CoV-2 virus. Such proxies have been used in orthologous fields, including physics and social sciences to test hypotheses where direct measurements are not feasible (22, 23). In our analysis, we included 72,613 individuals with suspected exposure counting 18,447 individuals who tested positive. We identified obesity as an intrinsic factor that elevated risk of developing SARS-CoV-2 infection upon suspected exposure across all ages in our patient population.

Materials and methods

All patient data were obtained from the Mass General Brigham (MGB) COVID-19 Data Mart. This study was conducted under IRB protocol number 2023P001104 approved by MGB. The protocol was not deemed exempt from review, and explicit informed consent from participants was not required. This data repository includes all EMRs for 687,813 subjects tested from March 2020 to March 2022 for COVID-19 in one of the MGB facilities regardless of the test result. To avoid the bias that vaccinated subjects may introduce in our analysis, we limited the time frame of our study to those patients tested before 2021 January 25 (340,694 subjects), when the number of fully vaccinated individuals was below 1% in the state of Massachusetts (24). Since the goal of this study was to assess the intrinsic susceptibility factor to SARS-CoV-2 infection, we limited our tested population to those who declared exposure to a (suspected) SARS-CoV-2 positive individual, regardless of test outcome. To assess exposure, we relied on the presence of the ICD10 codes Z20.822 (Contact with and [suspected] exposure to COVID-19), Z20.828 (Contact with and [suspected] exposure to other viral communicable disease), Z20.89 (Contact with and [suspected] exposure to other communicable diseases) on the patient's clinical records as a proxy of direct exposure. To appraise the correspondence between the population under study and general population, we correlate the incidence of COVID-19 in our sample and the state-wide data reported by the government of Massachusetts (24). To account for potential geographical heterogeneity in our analysis, we segmented the population into one of the 14 counties in Massachusetts using the reported residence ZIP code. The assessment of the diagnosis of T2DM and hypertension was done using ICD-10 codes E11 and I10, respectively. We imposed the restriction that an updated record of body mass index (BMI) must be present within 2 months before or after the test date. Therefore, the BMI measured closest to the date of COVID-19 testing was used.

A logistic regression model was used for computing the odds ratio for susceptibility between obese and nonobese exposed subjects. We corrected for multiple potential confounding factors: age, sex, county of residence, testing week, T2DM, and hypertension. Independently, to study the effect of each of those confounding factors, we fitted logistic linear models on different subsets of the sample stratifying by age, sex, T2DM, hypertension, and testing week. We used Pearson's product-moment correlation to check the correlation between the number of positive cases reported by the Massachusetts government and by MGB at the town level on a logarithmic scale. To establish a relation between these two quantities, we fitted their logarithmic values on a linear model. We used cross-correlation analysis to evaluate the correlation between the weekly evolution of positive cases reported by the Massachusetts government and by MGB. We assume a P-value <0.05 as a metric of significance for all the tests, and we provide a 95% confidence interval around every estimated measure.

Results

Description of study population

Overall, 72,613 individuals were included in this study based on the ICD10 codes, Z20.822, Z20.828 and Z20.89, the availability of BMI data, and the availability of confounding factor records (Fig. 1). We limited the period of study from 2020 March 1 to 2021 January 25, since the fully vaccinated population during this period was below 1% of the Massachusetts population (24). The age distribution of the exposed patients in our study were predominantly adults (13–19 years: 2,057 [3.5%]; 20–39 years: 17,940 [24.7%]; 40–64 years: 28,825 [39.7%]; and >64 years: 40,153 [30%]) with some pediatric patients (<13 years: 1,568 [2.2%]) also included. Our study population contained diverse individuals in terms of sex (42,678 [58.8%] women and 29,935 [41.2%] men) as well as race (Asian—2,494 [3.4%]; Black—4,864 [6.7%]; Hispanic—8,955 [12.3%], White—52,595 [72.4%]).

Fig. 1.

Subject inclusion criteria. Red boxes represent discarded subjects. The blue boxes represent patients used in the study. COVID-19 tests are PCR based. BMI information is limited to within 2 months before or after the time of first testing. Confounding variables are age, sex, place of residence, presence of diabetes mellitus type 2 (T2DM), or essential hypertension.

Correlation between the prevalence of COVID-19 and obesity in the state of Massachusetts

To assess the degree to which our data represents that of the entire state of Massachusetts, we used Massachusetts government's reported information regarding case rates for COVID-19. We compared the geographic distribution of those cases at town level between the Massachusetts government data (Fig. 2A) and MGB data (Fig. 2B). As expected, the MGB cases are distributed around the large urban areas in the eastern part of the state, where most MGB facilities are located, having less representation in the western area of the state. We observed a good correlation (r = 0.76 [95% CI: 0.71–0.80]) between the geographical distribution of cases reported by the state and MGB at the town level on a logarithmic scale (Fig. 2C). Another essential aspect to consider is the agreement between the temporal evolution of cases reported by the state and by MGB. After performing a cross-correlation analysis on the weekly evolution of positive cases reported by both sources (Fig. 2D), we documented a high correlation (r = 0.97 [95% CI: 0.96–0.98]) without any temporal displacement (lag = 0).

Fig. 2.

Comparison of Massachusetts vs. MGB. A), Number of positive tests of SARS-CoV-2 positivity in Massachusetts (MA), and B), in Mass General Brigham (MGB) dataset per town. Densities are plotted in a logarithmic scale C), Correlation between number of positive cases per postal code reported by MA and recorded on our MGB dataset. Area of the dots is proportional the town's population. Color represent the percentage of MA reported positives has been recorded in MGB database. The dashed line depicts the log–log linear model that describes the relation between these two quantities. D), Evolution of the number of positive cases per week reported by the MA (top) and by MGB (bottom).

Susceptibility to SARS-CoV-2 infection upon exposure to infected individuals

In our total study population, 18,447 individuals out of 72,613 suspected exposures tested positive for COVID-19 (25.4%) (Table 1). We observed similar rates of SARS-CoV-2 positivity upon exposure across all five age groups tested (<13 years: 25.3%; 13–19 years: 28.0%; 20–39 years: 26.8%; 40–64 years; 26.2% and >64 years: 22.8%) suggesting that age was not an intrinsic factor for susceptibility to infection upon suspected exposure to an infected individual. Both men (26.4%) and women (24.6%) had a comparable positivity rate upon exposure.

Table 1.

Characteristics of the study population.

	Exposed (n = 72,613)	SARS-CoV-2 (+) (n = 18,477)	OBESE (n = 24,438)	SARS-CoV-2 (+) + OBESE (n = 7,239)
Sex
Female	42,678 (58.8%) [34.2%]	10,540 (57.1%) [39.7%]	14,581 (59.7%)	4,186 (57.8%)
Male	29,935 (41.2%) [32.9%]	7,907 (42.9%) [38.6%]	9,857 (40.3%)	3,053 (42.2%)
Age
3–12	1,568 (2.2%) [21.7%]	397 (2.2%) [29.0%]	341 (1.4%)	115 (1.6%)
13–19	2,507 (3.5%) [15.6%]	703 (3.8%) [21.3%]	391 (1.6%)	150 (2.1%)
20–39	17,940 (24.7%) [32.2%]	4,810 (26.1%) [37.6%]	5,775 (23.6%)	1,810 (25.0%)
40–64	28,825 (39.7%) [39.0%]	7,560 (41.0%) [45.7%]	11,231 (46.0%)	3,453 (47.7%)
>64	21,773 (30.0%) [30.8%]	4,977 (27.0%) [34.4%]	6,700 (27.4%)	1,711 (23.6%)
Race
White	52,595 (72.4%) [31.9%]	11,295 (61.2%) [36.3%]	16,762 (68.6%)	4,103 (56.7%)
Hispanic	8,955 (12.3%) [43.8%]	4,055 (22.0%) [47.0%]	3,922 (16.0%)	1,907 (26.3%)
Black	4,864 (6.7%) [46.3%]	1,489 (8.1%) [50.2%]	2,250 (9.2%)	747 (10.3%)
Asian	2,494 (3.4%) [15.4%]	493 (2.7%) [21.1%]	385 (1.6%)	104 (1.4%)
Other	2,869 (4.0%) [30.5%]	888 (4.8%) [34.9%]	875 (3.6%)	310 (4.3%)
Declined	836 (1.2%) [29.2%]	227 (1.2%) [30.0%]	244 (1.0%)	68 (0.9%)
Hypertension
N	47,377 (65.2%) [28.4%]	12,080 (65.5%) [34.5%]	13,441 (55.0%)	4,172 (57.6%)
Y	25,236 (34.8%) [43.6%]	6,367 (34.5%) [48.2%]	10,997 (45.0%)	3,067 (42.4%)
T2DM
N	62,903 (86.6%) [30.7%]	15,406 (83.5%) [36.1%]	19,339 (79.1%)	5,568 (76.9%)
Y	9,710 (13.4%) [52.5%]	3,041 (16.5%) [54.9%]	5,099 (20.9%)	1,671 (23.1%)

(%) Account for the percentage representation of each sub-population within each category. [%] Display the proportion of the obese population in each subgroup as a percentage.

Temporal evolution of positive cases per week and negative cases with declared exposure is reported in Fig. 3A. We observed an evident change in the proportions between these two metrics in June 2020. This abrupt change is not natural; a change in recollection policy induces it: starting on June 22nd, instead of including all negative cases, for each positive added to the database, a random sample of three patients who tested negative in the same week was added. The change in this policy does not impact the methodology followed in this study.

Fig. 3.

Temporal evolution. A), Evolution of the number of patients tested positive (red) and negative with a reported exposure (blue) per week over the period of study. B), Distribution of obese (red) and nonobese (gray) per week in negative tested individuals with a reported exposure (top) and with a positive test (bottom). C), Odds ratio (OR) of susceptibility of obese vs. nonobese per week correcting for age, sex, county, and DM2 and hypertension status. Lines describe the 95% CI of the OR while the dot encodes the expected value. Black lines represent OR with a P-value smaller than 0.05 while gray represent OR with a value larger than 0.05.

Obesity as an intrinsic susceptibility risk factor for COVID-19 upon exposure to infected individuals

While previous reports have shown that obesity serves as a risk factor for worsened outcomes in COVID-19 patients, we analyzed whether obesity is an intrinsic risk factor for susceptibility to the SARS-CoV-2 infection. We used the World Health Organization (WHO) guidelines of BMI >30 to identify adult individuals as obese or nonobese, while we used the Center for Disease Control (CDC) and the American Academy of Pediatrics criteria (growth curve >95 percentile) for pediatric obesity determination (25).

Our study population included 24,438 obese individuals which made up 33.7% of individuals (Table 1). This is in line with the prevalence of obesity in the United States at 42.4% and for Massachusetts at 23.0%. We had a comparable distribution of obesity across ages tested (<13 years 21.7%; 13–19 years: 15.6%; 20–39 years: 32.2%; 40–64 years; 39.0% and >64 years: 30.77%) with the highest rates in middle-aged adults. We observed similar rates of obesity in both sexes (women—34.2% and men—32.9%). Our population had varying rates of obesity across race with low prevalence in Asians (15.4%) and increased prevalence in Black (46.3%) and Hispanic (43.8%) individuals.

Regarding the evolution over time, we observed a consistent ratio between obese and nonobese in positive and negative exposed populations (Fig. 3B) when compared weekly.

To determine whether obesity is an intrinsic risk factor for SARS-CoV-2 infection in our study population of individuals with suspected exposure, we compared the positivity rate of SARS-CoV-2 in obese individuals to the positivity rate of SARS-CoV-2 in nonobese individuals. Using a logistic regression model and correction for several confounding factors (see Materials and methods and Table S1), we observed that obesity had an odds ratio of 1.34 (Fig. 4), indicating 34% higher odds of SARS-CoV-2 positivity in the obese population.

Fig. 4.

Odds ratio. Obesity odds ratio for SARS-CoV-2 (+) individuals vs. patients who tested negative after a self-reported contact with a suspected positive subject. The analysis is performed over the complete sample and over different stratifications based on age, sex, and status of T2DM and hypertension. Black lines represent OR with a P-value smaller than 0.05 while gray represent OR with a value larger than 0.05.

Stratifying for demographic, temporal, and clinical factors associated with obesity for SARS-CoV-2 positivity among individuals with suspected exposure

We compared demographic factors such as age, sex, and county of residence to determine whether they can explain the increased prevalence of SARS-CoV-2 positivity in obese individuals. Age is also known to affect the immune response to viral infection and its outcomes (9). Ageing impacts both innate and adaptive arms of the immune system to impair control of viral infections (21). Obesity was a statistically significant predictor of increased SARS-CoV-2 infection as determined by COVID-19 positivity by PCR in all age groups ranging from OR = 1.79 in 13 to 19 years old to OR = 1.13 in ages >64 (Fig. 4). We also observed that obesity is a significant predictor of SARS-CoV-2 infection after exposure both in both men (OR = 1.39) and women (OR = 1.31).

Regarding the temporal consistency of these results, we show (Fig. 3C) significantly increased susceptibility in obese individuals across the duration of the pandemic in Massachusetts evaluated in this study.

These results suggest that obesity remains an intrinsic susceptibility factor for SARS-CoV-2 positivity in exposed individuals when correcting for demographic variables. We also compared two clinical diagnoses related to obesity to determine whether clinical factors or underlying co-morbidities could explain the increased odds for SARS-CoV-2 positivity in exposed individuals. We chose hypertension and T2DM as the two factors since they are co-morbidities commonly associated with obesity. T2DM increases the risk of more frequent and severe viral infections such as COVID-19. A 2020 report on 5,700 COVID-19 patients in New York includes hypertension (3,026; 56.6%) as the most common comorbidity (13). These clinical factors were also chosen due to the standardized diagnosis within our EMR system. First, we observed that obesity remained an intrinsic susceptibility factor in nondiabetic individuals (OR = 1.39). However, obesity was not a statistically significant risk factor in diabetic individuals (OR = 1.08). While this may indicate that obesity does not contribute significantly to increasing the odds of SARS-CoV-2 positivity in individuals with diabetes, we observed an elevated risk for SARS-CoV-2 positivity in both patients with essential hypertension and those without essential hypertension (Fig. 4). Using the same approach, we also computed the odds ratio for hypertension (OR = 0.93) and T2DM (OR = 1.44) as risk factors.

Discussion

Identifying external risk factors such as close contact, limited air supply, and airborne spread contributed to developing the measures employed to mitigate the spread and exhaustion of hospitalization-related resources of the COVID-19 pandemic. These include social distancing and masking among others as per WHO Infection prevention and control in the context of COVID-19 guideline. However, internal factors that are not readily modifiable may require more attention and reprioritization of resources. This study identified obesity as an internal risk factor for the SARS-CoV-2 virus infection which could aid in the redistribution of resources toward individuals at higher risk for positivity.

In this study, we evaluated how different demographics and clinical outcomes influence the susceptibility of obese individuals to develop productive SARS-CoV2 infection. Remarkably, the odds ratio for SARS-CoV-2 susceptibility in obese individuals is lowered in three sub-groups, (i) older age groups, (ii) diabetic patients, and (iii) hypertensive patients. This variation can be associated with a compromised immune response, thereby increasing vulnerability to infections. In such contexts, the impact of obesity on susceptibility is less pronounced due to the influence of these factors. For instance, the reduction in odds ratio in diabetic patients can be attributed to a similar risk of diabetic patients to test positive for SARS-CoV-2 (OR = 1.44) compared with the risk of obesity. Conversely, hypertensive patients do not have an increased susceptibility to SARS-CoV-2 infection (OR = 0.93) by itself; the high prevalence of diabetes in hypertensive patients (Table S2) could be responsible for the observed reduction in odds ratio for this sub-population. These results suggest a complex interplay among risk factors. As a counterpart, in populations characterized by younger age, absence of diabetes, or hypertension, the primary role of obesity in influencing susceptibility becomes more pronounced. This could indicate that obesity is a major risk factor by itself to increase susceptibility to SARS-CoV-2 infection.

The advent of vaccination, widespread masking, and use of sanitary practices have decreased incidence and severity of COVID-19 cases. However, the emergence of the delta and omicron variants necessitates retrospective analysis to identify at-risk populations to divert resources toward to control the spread of the disease. Obesity in younger demographics has been alarmingly increasing in recent years and is projected to reach nearly 50% prevalence by 2030 (26). Our identification of obesity across all age groups, races, and sexes within Massachusetts warrants subsequent analysis in other cohorts to detect susceptible populations that could contribute to the spread of the disease.

Our sample population within the state of Massachusetts leverages the EMR of individuals with suspected exposure to COVID-19 and uses this information as a proxy of true viral exposure to determine the risk of productive infection. Using this approach to identify intrinsic risk factors, we complement studies that have helped characterize co-morbidities that explain the elevated risk for worsening outcomes. Using the COVID-19 Data Mart compiled from the MGB EMR system allows for geographically centered analysis by providing local information that can shape public policy toward combating the spread of COVID-19.

There are several limitations of this study. Firstly, this study uses a self-reported metric of suspected exposure to use as a proxy of true exposure to viral particles to identify at-risk individuals. It is therefore a subjective parameter to designate exposure. Secondly, all patients are from the MGB healthcare network, which may not be representative of the entire Massachusetts population. Thirdly, our data came from EMR, which may present inaccuracies regarding diagnoses and patient information due to medical transcription errors. Lastly, although we adjusted our analysis for the most common confounding factors (see Materials and methods), there may be underlying factors related to susceptibility that we are not considering which may alter the study's results. Therefore, the conclusions made may not be applicable to a wider and more diverse population. However, this study provides key insights that help identify additional intrinsic factors in other regions.

Our data were intentionally limited to a timeframe before vaccination became widespread in Massachusetts. This aids to identify intrinsic factors without accounting for the heterogeneity in vaccination response. Future studies can compare intrinsic factors within vaccinated individuals of mixed immunization statuses to utilize a personalized medicine approach in predicting where future breakout cases will emerge.

Our study identified obesity as an intrinsic risk factor for the spread of COVID-19 in individuals with comparable exposure risk. Studies evaluating infection upon direct exposure and altered expression of genes frequently associated with obesity, such as acute inflammation, coagulation, and apoptosis have also been associated with infection of coronaviruses (27–29). Other studies also suggested that adipocytes could act as a reservoir for SARS-CoV-2 virus (30–32). The presence of large amounts of adipose tissue (33) in obese individuals could explain the increase in severity (34) and susceptibility to infection. Based on our observation that among diabetic individuals, the presence of obesity does not increase the susceptibility toward productive infection given comparable exposure. This suggests that impaired metabolic state of immune cells within cardiometabolic vascular disease setting could contribute toward COVID-19 susceptibility (20). Future mechanistic studies evaluating shared signaling pathways in obese individuals could lead to the identification of drug targets that can be used to assail the infectivity of SARS-CoV-2.

Supplementary Material

Supplementary material is available at PNAS Nexus online.

Funding

The study was supported in part by research grants from Kowa Company, Ltd, Nagoya, Japan (A11014 to M.A.) and the National Institutes of Health (R01HL126901 and R01HL149302 to M.A.; R01HL136431, R01HL141917, and R01HL147095 to E.A.; and K25HL150336 to A.H.). Kowa was not involved in the study other than funding.

Author Contributions

J.T.M., S.C., J.L.D., and M.A. conceived the project. J.T.M., S.C., and J.L.D designed the methodology. J.T.M. and R.F.F. performed the formal statistical analysis and managed the data used in this work. J.T.M., S.C, J.L.D., and D.S.V.P. interpreted the results of the analysis. J.T.M., S.C., and J.L.D. designed the visualizations. J.T.M., S.C., J.L.D., R.F.F., and D.S.V.P drafted the original manuscript. J.T.M., S.C., J.L.D., A.H., D.S.V.P., E.A., R.M., and M.A. reviewed and edited the manuscript. M.A. supervised the project and provided funds and resources.

Data Availability

In compliance with ethical standards and to ensure the confidentiality of EMR, the raw data supporting the conclusions of this article cannot be made publicly available. The data contain sensitive and confidential patient information that is protected under privacy laws and regulations. However, to facilitate transparency and reproducibility, we are providing the code used for data analysis and a detailed summary of the data structure. These resources are available at . This repository includes comprehensive documentation on the code, along with a conceptual outline of the data structure, minus any confidential or identifiable patient information. We believe that these materials will assist other researchers in understanding our methodologies and in replicating similar studies where feasible. This repository includes comprehensive documentation on the code, along with a conceptual outline of the data structure, minus any confidential or identifiable patient information. We believe that these materials will assist other researchers in understanding our methodologies and in replicating similar studies where feasible.