1 Introduction

Over the past 3 years, corona virus disease 2019 (COVID-19) has spread globally and has become a growing public health concern, posing a serious threat to human survival. As of November 2022, over 637 million cases of COVID-19 have been confirmed globally, of which about 1.04% have died [1]. In the context of widespread infection, the possible systemic comorbidities and sequelae of COVID-19 may cause panic.

It has been previously reported that patients with COVID-19 may experience various adverse and persistent symptoms during hospitalization (comorbidity) and after recovery (sequelae or post-acute COVID-19 syndrome), including dyspnea, pain, respiratory diseases, cardiovascular diseases, digestive diseases, neurological disorders, renal failure, diabetes, musculoskeletal diseases, anxiety and depression, etc. [2,3,4,5,6,7,8,9,10,11,12,13,14,15,16] However, the prevalence and severity of COVID-19-related comorbidities and sequelae remain controversial. Large-scale and long-term cohort studies are still needed to demonstrate the incidence of COVID-19-related comorbidities and sequelae, as well as individual differences in susceptibility, which is of great significance for clinical strategy determination.

The aim of this study is to perform a comprehensive outcome-wide association analysis of COVID-19 comorbidities (short-term) and sequelae (long-term), and individual susceptibilities based on the large prospective UK Biobank (UKB) cohort.

2 Material and Methods

2.1 Study Design and Data Source

This prospective population-based outcome-wide association study was conducted in the UKB, which is a national prospective cohort that recruited over 500,000 participants aged 40–69 years from 22 assessment centers across the United Kingdom between 2006 and 2010. UKB obtained ethical approval from the North West Multicenter Research Ethics Committee [17], and has collected extensive detailed baseline and long-term follow-up data, including real-time updated extensive clinical records. SARS-CoV-2 testing results were obtained from Public Health England (PHE), Public Health Scotland (PHS) and Secure Anonymized Information Linkage (SAIL). Hospitalization information of participants from England, Scotland, and Wales were acquired from Hospital Episode Statistics, the Scottish Morbidity Record, and the Patient Episode Database for Wales, respectively. Mortality data were extracted from the National Health Service (NHS) Digital and NHS Central Register.

2.2 Participants and Exposure

A total of 106,760 UKB participants who were tested for SARS-CoV-2 during March 2020 and February 2021 were included. Among the 17,832 participants with at least one positive SARS-CoV-2 test (COVID-19-positve participants), those who were tested with positive result after 2021-2-23 (had limited follow-up time, n = 1041), or withdrew during follow-up (n = 15) were excluded. We further defined the severity of COVID-19 according to the presence of death or hospitalization due to COVID-19. For these 88,928 participants without any positive SARS-CoV-2 test result (COVID-19-negative participants), we excluded those who were reported to die or hospitalize due to COVID-19 (n = 303), died before the pandemic of COVID-19 (2020-1-31, n = 1,095) or withdrew during follow-up (n = 14). Then for each COVID-19-positive participant, we matched up to 4 COVID-19-negative participants according to birth year, sex, and Townsend deprivation index (TDI) by propensity scores. Finally, we excluded COVID-19-negative participants who died before the SARS-CoV-2 test date of their matched COVID-19-positive counterparts. These inclusion and exclusion criteria, and matching process resulted in a final cohort of 16,776 COVID-19-positive participants and 58,281 COVID-19-negative participants (Supplementary Figs. 1–2).

2.3 Outcome and Covariates

Comorbidities and sequelae, defined using 3-digit ICD-10 codes (International Classification of Diseases, 10th revision, excluding codes used for special purposes, injury, poisoning and certain other consequences of external causes, factors influencing health status and contact with health services, as well as external causes of morbidity and mortality) from medical records, referred to newly onset illnesses < 1 month and ≥ 1 month after the diagnosis of COVID-19, respectively [18,19,20]. We then reclassified the eligible (hazard ratio [HR] > 1, p < 0.05, and case number > 10) diseases into more broadly defined comorbidities and sequelae (Supplementary Tables 1–2). In the case of multiple identical records for the same individual, the date of diagnosis was derived from the earliest record. Ethnicity, body mass index (BMI), smoking status, and Charlson comorbidity index (CCI, without age calculation, Supplementary Table 3) were included as covariates [21,22,23]. In subgroup analysis, we also defined another covariate, current age, as the age on January 31st, 2020.

2.4 Follow-Up

The start date of follow-up was the test date for COVID-19-positive participants. For COVID-19-negative participants, it was the same as that for their matched counterparts. Follow-up ended on the day of (1) specific disease diagnosis, (2) death, or (3) end of follow-up (March 31st, 2021), whichever came first. The longest follow-up duration was 14 months.

2.5 Statistics Analyses

Student’s t-tests and Chi-square tests were performed as appropriate to assess the differences among groups. Univariable conditional Cox proportional hazards models were performed to identify eligible diseases for comorbidity and sequela reclassification, and outcome-wide association analyses for the risk of COVID-19-related comorbidities and sequelae were conducted using multivariable conditional Cox proportional hazards models adjusting for ethnicity, BMI, smoking status, and CCI [24]. In the specific analysis of each disease, participants diagnosed with corresponding disease before COVID-19 were excluded. In the analysis of death in COVID-19-positive participants, we also excluded individuals who died on the day of diagnosis. To explore the impact of severity of COVID-19 on COVID-19-related sequelae, we additionally adjusted for age, sex, and TDI. In subgroup analyses, participants were stratified by current age (< 65 and ≥ 65 years), sex (female and male), BMI (< 25 kg/m2 and ≥ 25 kg/m2), smoking status (ever-smoker and never-smoker), and CCI (≤ 1 and ≥ 2).

All analyses were performed using R software (version 3.6.3, https://www.r-project.org/), and a two-tailed p < 0.05 was considered statistically significant.

3 Results

3.1 Participant Characteristics

A total of 16,776 COVID-19-positive participants were included and 58,281 COVID-19-negative participants were matched for comparison (1:4). A total of 2670 participants were hospitalized due to COVID-19, and 1169 participants died of COVID-19. Compared with COVID-19-negative participants, COVID-19-positive participants were younger (64.6 vs. 65.8, p < 0.001), more deprived (TDI: − 0.7 ± 3.3 vs. − 0.9 ± 3.3) and more overweight/obese (normal BMI: 25.5% vs 30.1%, p < 0.001, Supplementary Table 4).

3.2 Descriptive Analysis of COVID-19-Related Comorbidity Burdens

We observed that compared with COVID-19-negative participants, 121 types of comorbidities showed significantly higher incidences in participants with COVID-19 (HR > 1 and p < 0.05, Supplementary Table 5). In the outcome-wide association analysis adjusting for ethnicity, BMI, CCI and smoking status, 47 out of the 51 reclassified comorbidities showed higher risks in COVID-19-positive participants (HR > 1 and p < 0.05, Table 1). Representative comorbidities included lower respiratory infection (incidence: 5.93%, HR = 48.32, p < 0.001), respiratory failure (incidence: 2.02%, HR = 103.02, p < 0.001), electrolyte imbalance (incidence: 2.00%, HR = 9.41, p < 0.001), renal failure (incidence: 1.51%, HR = 8.8, p < 0.001), hypertension (incidence: 1.25%, HR = 2.49, p < 0.001) and other heart disease (incidence: 1.05%, HR = 3.92, p < 0.001, Table 1). However, the incidence rates of the other remaining 40 types of COVID-19-related comorbidities were less than 1%. Besides, among all COVID-19-positive participants who developed COVID-19-related comorbidities, 70.37% of them were reported to have two or more co-occurring comorbidities, and the most common form of co-occurrence was respiratory failure plus lower respiratory infection (Fig. 1 and Supplementary Fig. 3).

Table 1 The incidence and hazard ratio of comorbidity in COVID-19 individuals and matched negative comparisons, adjusted for ethnicity, BMI, smoking status, and CCI
Full size table
Fig. 1
figure 1

Upset plot representing the coexistence of comorbidities in A all COVID-19 patients, B mild (non-hospitalized) COVID-19 patients and C severe (hospitalized) COVID-19 patients, respectively. COVID-19 corona virus disease 2019

Full size image

In addition, the mortality of COVID-19-positive participants was significantly higher than COVID-19-negative participants regardless of COVID-19 severity (COVID-19-positive vs. COVID-19-negative: 7.09% vs. 0.91%, HR = 10.6, p < 0.001; mild COVID-19-positive vs. COVID-19-negative: 1.47% vs. 0.80%, HR = 2.11, p < 0.001; severe COVID-19-positive vs. COVID-19-negative: 37.36% vs. 1.43%, HR = 44.15, p < 0.001, Supplementary Table 6).

3.3 Burden of Comorbidities by COVID-19 Severity

Compared with mild COVID-19 patients (non-hospitalized), severe COVID-19 patients (hospitalized) were more likely to be male, older, deprived, obese, ever-smokers and had higher CCI scores (Supplementary Table 7). As expected, they had higher comorbidity burdens (Supplementary Table 8). Compared with COVID-19-negative participants, severe COVID-19 patients had significantly higher risks of having 48/51 types of COVID-19-related comorbidities, among which the incidence rates of lower respiratory infection (61.68%, HR = 12,351.54, p < 0.001), electrolyte imbalance (13.88%, HR = 54.76, p < 0.001), respiratory failure (13.83%, HR = 1394.47, p < 0.001), hypertension (11.04%, HR = 32.55, p < 0.001) and renal failure (10.86%, HR = 81.82, p < 0.001) were relatively higher. Notably, only severe COVID-19 patients were at an increased risk of hearing loss (incidence: 0.59%, HR = 7.57, p = 0.001). In contrast, mild COVID-19 patients only had 11/51 types of comorbidities with incidence rates less than 0.5%: septicaemia (HR = 2.89, p = 0.020), vitamin deficiency (HR = 2.14, p = 0.043), delirium (HR = 3.66, p < 0.001), lower respiratory infection (HR = 2.41, p < 0.001), respiratory failure (HR = 3.32, p = 0.031), rash and dermatitis (HR = 3.31, p = 0.031), decubitus ulcer (HR = 34.6, p = 0.004), cough (HR = 5.55, p = 0.001), dyspnea and asphyxia (HR = 2.36, p = 0.013), disorientation (HR = 2.18, p = 0.033) and emotional state symptoms and signs (HR = 4.68, p = 0.020, Table 2, and Supplementary Fig. 4). Besides, severe COVID-19 patients had more co-occurrence comorbidities, with lower respiratory infection plus respiratory failure being the most common form. On the other hand, the incidence of co-occurring comorbidities in mild COVID-19 patients was relatively lower, and renal failure plus electrolyte imbalance was the commonest (Fig. 1B ~ C and Supplementary Fig. 3B ~ C).

Table 2 Clinical comorbidity of COVID-19 individuals with different severity (mild/non-hospitalized and severe/hospitalized) and matched negative comparisons, adjusted for ethnicity, BMI, smoking status, and CCI
Full size table

3.4 Subgroup Analyses of COVID-19-Related Comorbidities by Age, Sex, BMI, Smoking Status and CCI

To understand whether the burden of COVID-19-related comorbidities differed among various populations, we further carried out subgroup analyses stratified by current age, sex, BMI, smoking status and CCI, respectively (Fig. 2).

Fig. 2
figure 2

Differences in adjusted hazards ratio of clinical comorbidities among COVID-19-positive participants compared with COVID-19-negative participants stratified by age, sex, BMI, smoking status, and CCI, respectively. Note: Only the top 20 comorbidities in terms of incidence were shown. Analyses were adjusted for ethnicity, BMI, smoking status and CCI as appropriate. COVID-19 corona virus disease 2019, BMI body mass index, CCI Charlson comorbidity index

Full size image

In terms of subgroup analysis for current age, older COVID-19-positive participants (> 65 years) were at a significantly higher risk of 48/51 clinical comorbidities, whereas only 26/51 comorbidities were significant in younger COVID-19 patients (Table 3). Specifically, among younger participants, COVID-19-positive individuals had a higher risk of respiratory failure (HR = 99.05, p < 0.001), vitamin deficiency (HR = 66.51, p = 0.004), delirium (HR = 51.65, p = 0.006), septicaemia (HR = 50.86, p = 0.026), etc. However, the incidences of these comorbidities were less than 1%, except for respiratory failure (incidence: 1.23%). In contrast, 16 out of the 48 comorbidities in elderly COVID-19-positive patients had an incidence greater than 1%, e.g., lower respiratory infection (9.22%), electrolyte imbalance (3.72%), respiratory failure (3.00%), renal failure (2.79%), etc.

Table 3 Clinical comorbidity in COVID-19 individuals and matched negative comparisons by age, adjusted for ethnicity, BMI, smoking status, and CCI
Full size table

Females had higher risks of 36/51 types of COVID-19-related comorbidities, whereas males were at higher risks of having 37/51 types of COVID-19-related comorbidities (Supplementary Table 9). Notably, blood cell disease (incidence: 0.65%, HR = 5.42, p < 0.001), obesity (incidence: 0.61%, HR = 2.34, p = 0.028), hypertension (incidence: 1.95%, HR = 2.77, p < 0.001), cerebrovascular diseases (incidence: 0.13%, HR = 4.74, p = 0.027), vascular disease (incidence: 0.20%, HR = 3.53, p = 0.042), COPD/emphysema (incidence: 0.64%, HR = 1732.86, p = 0.030) and rash and dermatitis (incidence: 0.31%, HR = 33.8, p = 0.006) were observed as significant comorbidities only in males but not in females.

In addition, individuals who were overweight/obese, ever-smoker, or with more comorbidities at baseline (CCI score ≥ 2) had more COVID-19-related comorbidities than their counterparts, respectively (Supplementary Tables 10–12).

3.5 Descriptive Analysis of COVID-19-Related Sequelae Burden

We observed that 14 types of sequelae were positively associated with the infection of COVID-19 (HR > 1 and p < 0.05, Supplementary Table 13). In the outcome-wide association analysis adjusting for ethnicity, BMI, CCI and smoking status, only 6 out of the 11 reclassified more broadly defined sequelae were observed in COVID-19-positive participants, including lower respiratory infection (incidence: 0.10%, HR = 8.33, p < 0.001), immobility (incidence: 0.10%, HR = 4.82, p = 0.001), interstitial pulmonary disease (incidence: 0.12%, HR = 2.4, p = 0.018), fecal abnormalities (incidence: 0.13%, HR = 2.24, p = 0.011), decubitus ulcer (incidence: 0.17%, HR = 1.96, p = 0.020) and urinary incontinence (incidence: 0.21%, HR = 1.81, p = 0.010, Supplementary Table 14), etc. The incidence rates of all these COVID-19-related sequelae were less than 1%.

4 Discussion

In this prospective cohort study, we conducted comprehensive outcome-wide association analyses to identify COVID-19-related comorbidities and sequelae in a large population. Overall, 47 types of COVID-19 related comorbidities that occurred within one month after COVID-19 infection were identified, by incidence from high to low, including lower respiratory infection, respiratory failure, electrolyte imbalance, renal failure, hypertension and other heart disease, etc. We also observed that COVID-19-related comorbidities tended to co-occur, especially in severe COVID-19 patients. Besides, older age, male gender, obese/overweight, smoking history, higher CCI scores and severe COVID-19 were risk factors for experiencing more types of comorbidities. Meanwhile, we identified 6 types of COVID-19 related sequelae that began to appear after one month following COVID-19 infection, such as lower respiratory infection, immobility, interstitial pulmonary disease, fecal abnormalities, decubitus ulcer and urinary incontinence. Nonetheless, the incidence rates of these COVID-19-related sequelae were all relatively low (< 1%). Therefore, the public should be urged not to worry too much about these low-morbidities and low-seriousness sequelae.

Previous studies showed that COVID-19 patients had a higher risk of having sequelae, such as myalgia, sexual dysfunction, hearing loss and disturbances of smell and taste [25,26,27,28], however, these sequelae were not prominent in our results, as we only observed 6 types of COVID-19-related sequelae with low incidences. However, we observed plenty of comorbidities involving multiple organs after COVID-19 infection, such as respiratory, neurological, circulatory and urinary systems. Besides, the majority of COVID-19 patients developing comorbidities showed more than one comorbidity. Thus, comorbidity was a more prominent issue for COVID-19 patients.

It has been established that SARS-CoV-2 can upregulate the expression of the type 2 angiotensin converting enzyme (ACE-2), and can bind ACE-2 receptors on the surface of the host cells for cell entrance in many organs, which may explain the comorbidities we observed in COVID-19 patients, such as hypertension, diabetes and COPD [29,30,31]. COVID-19 is associated with a high inflammatory burden and SARS-CoV-2 can affect the myocardium and cardiac biomarker level and lead to myocarditis and heart failure [32,33,34,35]. Moreover, interleukin-mediated modulation of phosphokinases and phosphatases, NOD-like receptor thermal protein domain associated protein 3 (NLRP3) inflammasome-mediated inflammation and pathological accumulation of amyloid-β are associated with COVID-19 related neurological disorder. Several studies have found abnormalities in brain structures in COVID-19 patients, such as reduced grey matter thickness, tissue-contrast in the cortex and gyrus, and reduced overall brain size [36,37,38,39]. In addition, the activation of the RAS, hemodynamic changes and secondary infection of the urinary tract following COVID-19 infection are associated with the comorbidities and sequelae of urinary system in COVID-19 patients [40,41,42]. Hearing loss was found in the comorbidities of severe COVID-19 patients, which may be associated with brainstem dysfunction resulting from neuroinflammatory mechanisms. Cytokine storm after COVID-19 infection could damage the auditory glial cells and might play a role in hearing loss [43].

We observed that participants with advanced age, male sex, smoking status or excessive obesity were at higher risks of COVID-19-related comorbidities. Old and obese participants were usually characterized by more pre-existing comorbidities, weaker immune defense, and higher levels of proinflammatory cytokines, which may contribute to their more comorbidities [44, 45]. Besides, the discrepancy of COVID-19 related outcomes between male and female participants could be attributed to the differences in sex hormones, expression levels of ACE2 and Transmembrane protease serine 2 (TMPRSS2), and lifestyles [46]. Moreover, smoking is associated with a higher expression level of ACE2 in airway epithelial cells, which may induce the occurrence of COVID-19 related comorbidities and sequelae [47]. These populations should be paid special attention as they were more susceptible to COVID-19-related comorbidities.

Although the advent of our study provided new insight into the comorbidities and sequelae of COVID-19 patients, a few limitations still existed. First, due to the data limitation, we defined the severity of COVID-19 according to the hospitalization status or death cause, which may lead to partial bias but was acceptable [48]. Second, due to the limited data on COVID-19 medications, we could not assess the effect of COVID-19 medications on COVID-19-related outcomes. Third, our observations were mainly of comorbidities and sequelae associated with the SARS-CoV-2 Alpha variant, which was the main strain in the UK between January 31, 2020 and March 31, 2021, but not the Delta and Omicron variants of SARS-CoV-2, which began to emerged and spread in the UK from March and October 2021, respectively. The SARS-CoV-2 Alpha variant was considered relatively more pathogenic but less infectious than Delta and Omicron variant, therefore, the incidence and severity of sequelae of Delta and Omicron variants might be lower than, also be different from, those of Alpha variants [49,50,51,52]. Unfortunately, to date, we are unable to obtain the lagging comorbidities and sequelae data of Delta and Omicron variant for analysis.

In conclusion, 47 types of high-risk comorbidities might occur within one month after COVID-19 infection, especially in patients with older age, overweight/obese, more pre-existing comorbidities and severe COVID-19. And only 6 types of COVID-19-related sequelae appeared after one month following COVID-19 infection, indicating that more attention and health care should be given to these susceptible populations after COVID-19 infection.