tetano
Editor, Senior Moderator
Eur J Pain
. 2025 May;29(5):e70021.
doi: 10.1002/ejp.70021. Predictive Ability of Previous Pain and Disease Conditions on the Presentation of Post-COVID Pain in a Danish Cohort of Adult COVID-19 Survivors
Brian Duborg Ebbesen[SUP] 1 2 [/SUP], Jakob Nebeling Hedegaard[SUP] 3 [/SUP], Simon Grøntved[SUP] 3 4 [/SUP], Rocco Giordano[SUP] 1 5 [/SUP], César Fernández-de-Las-Peñas[SUP] 6 [/SUP], Lars Arendt-Nielsen[SUP] 1 2 7 [/SUP]
Affiliations
Background: Even though many post-COVID pain risk factors have been identified, little is known about the predictive profiles of these risk factors for the development of post-COVID pain.
Methods: Data was collected from two separate questionnaires assessing demographics, pre-existing medical comorbidities, pain history, and post-COVID pain experience. Socioeconomic data and COVID-19 RT-PCR test results were collected from Danish registries. The study cohort (n = 68,028) was stratified into two groups reporting pre-COVID pain (n = 9090) and no pre-COVID pain (n = 55,938). Forward-selection prediction models were employed to identify predictor profiles for post-COVID pain in the full study cohort (Model 1) and the stratified groups with (Model 2) and without (Model 3) pre-COVID pain from 58 potential risk factors.
Results: Model 1 achieved a 5-fold cross-validated AUC (cvAUC) of 0.68. Use of pain medication, stress, high income, age, female gender, and weight were the top predictors contributing to 97% of the model performance. Model 2 (cvAUC = 0.69) identified use of pain medication, breathing pain, stress, height, physical activity, and weight as the top predictors contributing to 98.6% of model predictive performance. Model 3 (cvAUC = 0.65) identified stress, female gender, weight, higher education, age, high income, and physical activity as the top predictors contributing to 98.5% of model predictive performance. Height was unique to Model 2, while being female and higher income were unique to Model 3.
Conclusions: The study highlights potential important predictors, and further research is needed to describe these in detail. The results may apply to the understanding of post-viral pain sequelae after other viral infections.
Significance statement: The explorative study investigates the predictive ability of a battery of pre-COVID risk factors potentially associated with the development of post-COVID pain. This article presents the profiles of predictors of interest in COVID-19 survivors with and without pre-COVID pain. The results will contribute to the understanding of patient profiles that might develop post-COVID pain conditions and provide a first step towards focused clinical predictive research.
. 2025 May;29(5):e70021.
doi: 10.1002/ejp.70021. Predictive Ability of Previous Pain and Disease Conditions on the Presentation of Post-COVID Pain in a Danish Cohort of Adult COVID-19 Survivors
Brian Duborg Ebbesen[SUP] 1 2 [/SUP], Jakob Nebeling Hedegaard[SUP] 3 [/SUP], Simon Grøntved[SUP] 3 4 [/SUP], Rocco Giordano[SUP] 1 5 [/SUP], César Fernández-de-Las-Peñas[SUP] 6 [/SUP], Lars Arendt-Nielsen[SUP] 1 2 7 [/SUP]
Affiliations
- PMID: 40186415
- PMCID: PMC11971649
- DOI: 10.1002/ejp.70021
Background: Even though many post-COVID pain risk factors have been identified, little is known about the predictive profiles of these risk factors for the development of post-COVID pain.
Methods: Data was collected from two separate questionnaires assessing demographics, pre-existing medical comorbidities, pain history, and post-COVID pain experience. Socioeconomic data and COVID-19 RT-PCR test results were collected from Danish registries. The study cohort (n = 68,028) was stratified into two groups reporting pre-COVID pain (n = 9090) and no pre-COVID pain (n = 55,938). Forward-selection prediction models were employed to identify predictor profiles for post-COVID pain in the full study cohort (Model 1) and the stratified groups with (Model 2) and without (Model 3) pre-COVID pain from 58 potential risk factors.
Results: Model 1 achieved a 5-fold cross-validated AUC (cvAUC) of 0.68. Use of pain medication, stress, high income, age, female gender, and weight were the top predictors contributing to 97% of the model performance. Model 2 (cvAUC = 0.69) identified use of pain medication, breathing pain, stress, height, physical activity, and weight as the top predictors contributing to 98.6% of model predictive performance. Model 3 (cvAUC = 0.65) identified stress, female gender, weight, higher education, age, high income, and physical activity as the top predictors contributing to 98.5% of model predictive performance. Height was unique to Model 2, while being female and higher income were unique to Model 3.
Conclusions: The study highlights potential important predictors, and further research is needed to describe these in detail. The results may apply to the understanding of post-viral pain sequelae after other viral infections.
Significance statement: The explorative study investigates the predictive ability of a battery of pre-COVID risk factors potentially associated with the development of post-COVID pain. This article presents the profiles of predictors of interest in COVID-19 survivors with and without pre-COVID pain. The results will contribute to the understanding of patient profiles that might develop post-COVID pain conditions and provide a first step towards focused clinical predictive research.