Characteristics and outcomes of 627 044 COVID-19 patients living with and without obesity in the United States, Spain, and the United Kingdom

Martina Recalde, Elena Roel, Andrea Pistillo, Anthony G. Sena, Albert Prats‐Uribe, Waheed‐Ul‐Rahman Ahmed, Heba Alghoul, Thamir M. Alshammari, Osaid Alser, Carlos Areia, Edward Burn, Paula Casajust, Dalia Dawoud, Scott L. DuVall, Thomas Falconer, Sergio Fernández‐Bertolín, Asieh Golozar, Mengchun Gong, Lana Yin Hui Lai, Jennifer C. E. Lane, Kristine E. Lynch, Michael E. Matheny, Paras Mehta, Daniel R. Morales, Karthik Natarjan, Fredrik Nyberg, Jose Posada, Christian Reich, Peter R. Rijnbeek, Lisa M. Schilling, Karishma Shah, Nigam H. Shah, Vignesh Subbian, Lin Zhang, Hong Zhu, Patrick Ryan, Daniel Prieto‐Alhambra, Kristin Kostka, Talita Duarte‐SallesInternational Journal of Obesity 2021 · DOI

TL;DR

This study analyzed over 627,000 COVID-19 patients across the US, Spain, and UK, revealing that individuals with obesity (PLWO) are more likely to have pre-existing conditions, experience more severe COVID-19 outcomes, and require higher levels of care, including hospitalization and intensive services. These findings underscore the critical need for targeted preventive strategies and highlight obesity as a significant factor in COVID-19 severity.

What Problem Does This Paper Address?

While it was quickly recognized that obesity might be a risk factor for severe COVID-19, a comprehensive characterization of these patients, including their diverse medical conditions and specific outcomes, was lacking. Previous studies often focused on a limited set of risk factors or were susceptible to biases due to sampling methods. This paper aims to fill that gap by providing a detailed, large-scale comparison of COVID-19 patients with and without obesity, using real-world data from multiple countries to offer a more generalizable understanding of the differences.

Methods

The researchers conducted a multinational cohort study using routinely collected healthcare data from January to June 2020. They employed the Observational Medical Outcomes Partnership (OMOP) Common Data Model to standardize data from six different databases across Spain, the UK, and the US. Two non-mutually exclusive cohorts were defined: one for all patients diagnosed with COVID-19 (clinical diagnosis and/or positive test) and another for hospitalized COVID-19 patients. Patients were followed for 30 days post-diagnosis or hospitalization. Obesity was defined based on an ever-recorded obesity diagnosis, BMI measurement (30-60 kg/m²), or bodyweight measurement (120-200 kg). Standardized Mean Differences (SMDs) were used to compare the prevalence of demographics, medical conditions, and 30-day outcomes between the obesity and non-obesity groups. A federated analysis approach was used to ensure data privacy, where analytical code was run locally within each data source, and only aggregate results were shared.

Datasets

The study utilized data from six healthcare databases:

  • Spain: Information System for Research in Primary Care (SIDIAP)
  • United Kingdom: Clinical Practice Research Datalink (CPRD)
  • United States:
    • Columbia University Irving Medical Center (CUIMC)
    • IQVIA Open Claims
    • Stanford Medicine Research Data Repository (STARR-OMOP)
    • United States Department of Veterans Affairs (VA-OMOP)

These databases provided electronic health records and health insurance claims data.

Key Insights and Ideas

  • Obesity is a pervasive factor in COVID-19 severity: The study strongly suggests that obesity isn't just a risk factor for severe COVID-19 but is associated with a broader spectrum of medical conditions and poorer outcomes.
  • Beyond the usual suspects: PLWO had a higher prevalence of not only obesity-related comorbidities (like hypertension and diabetes) but also over a thousand other health conditions, indicating a more complex health profile.
  • Hospitalized PLWO are younger: Counterintuitively, hospitalized patients with obesity were younger than their non-obese counterparts, suggesting that obesity can lead to more severe disease even in younger individuals.
  • International consistency: The findings were largely consistent across the US, Spain, and the UK, strengthening the generalizability of the results despite differences in healthcare systems and populations.
  • Data harmonization is key: The use of the OMOP CDM and a federated analysis approach allowed for a large-scale, multinational study while respecting data privacy, showcasing the power of collaborative data science in public health emergencies.

Key Results

  • Prevalence: Out of 627,044 diagnosed COVID-19 patients, 33.1% had obesity. Among 160,013 hospitalized patients, 39.9% had obesity, indicating a higher prevalence of obesity in those with more severe disease.
  • Demographics: PLWO were more commonly female in the diagnosed cohort across most sites. Hospitalized PLWO were generally younger than hospitalized patients without obesity.
  • Comorbidities: PLWO consistently had a higher prevalence of comorbidities compared to patients without obesity, with meaningful differences observed for conditions like hypertension, heart disease, and type 2 diabetes across all databases.
  • 30-Day Outcomes:
    • Diagnosed Cohort: Hospitalization rates were higher among PLWO. Fatality rates were also higher in some regions (Spain and US), though not consistently across all databases.
  • Hospitalized Cohort: PLWO more frequently required intensive services (mechanical ventilation, tracheostomy, ECMO) and experienced adverse events like ARDS, heart failure, and sepsis. Acute kidney injury was the only outcome more frequent in patients without obesity, but this difference was not meaningful.

Limitations and Future Work

  • Selection Bias: The authors acknowledge that testing restrictions and asymptomatic cases might lead to underreporting, potentially biasing the COVID-19 case selection.
  • BMI Granularity: The study did not have access to BMI as a continuous variable, preventing the investigation of specific obesity categories (e.g., overweight, class I, II, III obesity) and their differential impact. This might also explain some observed differences between US and European data, as US populations tend to have higher BMIs.
  • Data Quality: As with any study using routinely collected data, there are concerns about data quality and potential oversampling of certain groups.
  • External Factors: Outcomes like hospitalization and intensive service requirements can be influenced by external factors such as bed availability and admission criteria, which may vary across databases.

Future work could involve investigating the impact of different BMI categories, exploring the intersection of sex/gender and obesity in COVID-19 outcomes, and utilizing more advanced statistical methods to account for confounding factors. The ongoing nature of the CHARYBDIS study also promises updated results with longer follow-up periods.

Summary updated

Abstract

We show that PLWO differ from patients without obesity in a wide range of medical conditions and present with more severe forms of COVID-19, with higher hospitalization rates and intensive services requirements. These findings can help guiding preventive strategies of COVID-19 infection and complications and generating hypotheses for causal inference studies.
BibTeX
@article{recalde2021characteristics,
  author = {Martina Recalde and Elena Roel and Andrea Pistillo and Anthony G. Sena and Albert Prats‐Uribe and Waheed‐Ul‐Rahman Ahmed and Heba Alghoul and Thamir M. Alshammari and Osaid Alser and Carlos Areia and Edward Burn and Paula Casajust and Dalia Dawoud and Scott L. DuVall and Thomas Falconer and Sergio Fernández‐Bertolín and Asieh Golozar and Mengchun Gong and Lana Yin Hui Lai and Jennifer C. E. Lane and Kristine E. Lynch and Michael E. Matheny and Paras Mehta and Daniel R. Morales and Karthik Natarjan and Fredrik Nyberg and Jose Posada and Christian Reich and Peter R. Rijnbeek and Lisa M. Schilling and Karishma Shah and Nigam H. Shah and Vignesh Subbian and Lin Zhang and Hong Zhu and Patrick Ryan and Daniel Prieto‐Alhambra and Kristin Kostka and Talita Duarte‐Salles},
  title = {{Characteristics and outcomes of 627 044 COVID-19 patients living with and without obesity in the United States, Spain, and the United Kingdom}},
  journal = {International Journal of Obesity},
  year = {2021},
  doi = {10.1038/s41366-021-00893-4},
}
Figure