Abstract
Background Hundreds of thousands of deaths have already been recorded for patients with the severe acute respiratory syndrome coronavirus (SARS-CoV-2; aka COVID-19). Understanding whether there is a relationship between comorbidities and COVID-19 positivity will not only impact clinical decisions, it will also allow an understanding of how better to define the long-term complications in the groups at risk. In turn informing national policy on who may benefit from more stringent social distancing and shielding strategies. Furthermore, understanding the associations between medications and certain outcomes may also further our understanding of indicators of vulnerability in people with COVID-19 and co-morbidities.
Methods Electronic healthcare records (EHR) from two London hospitals were analysed between 1st January and 27th May 2020. 5294 patients presented to the hospitals in whom COVID status was formally assessed; 1253 were positive for COVID-19 and 4041 were negative. This dataset was analysed to identify associations between comorbidities and medications, separately and two outcomes: (1) presentation with a COVID-19 positive diagnosis, and (2) inpatient death following COVID-19 positive diagnosis. Medications were analysed in different time windows of prescription to differentiate between short-term and long-term medications. All analyses were done with controls (without co-morbidity) matched for age, sex, and number of admissions, and a robustness approach was conducted to only accept results that consistently appear when the analysis is repeated with different proportions of the data.
Results We observed higher COVID-19 positive presentation for patients with hypertension (1.7 [1.3-2.1]) and diabetes (1.6 [1.2-2.1]). We observed higher inpatient COVID-19 mortality for patients with hypertension (odds ratio 2.7 [95% CI 1.9-3.9]), diabetes (2.2 [1.4-3.5]), congestive heart failure (3.1 [1.5-6.4]), and renal disease (2.6 [1.4-5.1]). We also observed an association with reduced COVID-19 mortality for diabetic patients for whom anticoagulants (0.11 [0.03-0.50]), lipid-regulating drugs (0.15 [0.04-0.58]), penicillins (0.20 [0.06-0.63]), or biguanides (0.19 [0.05-0.70]) were administered within 21 days after their positive COVID-19 test with no evidence that they were on them before, and for hypertensive patients for whom anticoagulants (0.08 [0.02-0.35]), antiplatelet drugs (0.10 [0.02-0.59]), lipid-regulating drugs (0.15 [0.05-0.46]), penicillins (0.14 [0.05-0.45]), or angiotensin-converting enzyme inhibitors (ARBs) (0.06 [0.01-0.53]) were administered within 21 days post-COVID-19-positive testing with no evidence that they were on them before. Moreover, long-term antidiabetic drugs were associated with reduced COVID-19 mortality in diabetic patients (0.26 [0.10-0.67]).
Conclusions We provided real-world evidence for observed associations between COVID-19 outcomes and a number of comorbidities and medications. These results require further investigation and replication in other data sets.
Introduction
As of the 16th of August 2020, the COVID-19 pandemic has resulted in more than 770,000 deaths worldwide over the course of a few months. Although most of the confirmed cases of COVID-19 infection show mild or no symptoms, a global concern exists in ensuring that healthcare systems can cope with those that require hospitalisation, especially with the coming winter pressure and while preparing for a potential second wave. Correct identification of individuals with higher risk of presentation or of death due to COVID-19 will not only help hospitals better identify those in need of hospitalisation but will also assist the community in identifying the vulnerable who require more careful shielding.
Hypertension and diabetes have been shown to be associated with poorer outcome in COVID-19. (Perico, et al., 2020). Furthermore, it has been postulated that certain medications may modulate the immune response to COVID-19 infection, positively or negatively.
We have set out to investigate the association of comorbidities and medications with either presentation with COVID-19 positive diagnosis or death due to COVID-19 in a real-world-evidence (RWE) dataset.
Methods
The dataset
Electronic health record (EHR) data were obtained from the Chelsea & Westminster Hospital NHS Foundation Trust for 5294 patients presented to the Trust’s hospitals between 1 st January and 27th May 2020. The patient population consisted of 1253 COVID-19 positive and 4041 COVID-19 negative patients as diagnosed using viral PCR from swap tests.
For these 5294 patients the dataset included patients’ history of primary and secondary diagnoses recorded in the hospital (from 2004 to 2020) medications administered or prescribed in the hospital (from 2010 to 2020), and death status.
Association analysis of comorbidities with COVID-19 presentation and mortality
We examined the association of comorbidities with COVID-19 positive presentation as well as with inpatient mortality of COVID-19 patients. This is realised by using Fisher’s exact test to compare the outcomes of patients with a given comorbidity (the active arm) and patients without that comorbidity (the control arm). To account for possible confounders, the two arms were matched on age, sex, and the number of admissions of patients in their available history (Figure 1 (b)). To reduce false discoveries, multiple hypothesis testing was conducted using the Benjamini Hochberg (BH) method. Furthermore, a robustness analysis was carried out by running the same test experiment once with 100% of the available data and 20 more times with randomly selected proportions (60%, 70%, 80%, or 90%) of the data. An association is considered significant if its adjusted p-value was smaller than 0.05 in more than 70% of these individual experiments (Figure 1 (a)).
Patients are included in the active arm if a record of the given disease appears at least once in their available history. Disease diagnoses are recorded in our data using the International Classification of Diseases, tenth revision (ICD-10) codes. ICD-10 codes were translated into 19 disease groups following the approach proposed by (Quan, et al., 2005) with some amendments ensuring that hypertension and asthma were individually assessed in the analysis (Table 1).
Analysis of association of medications with COVID-19 presentation and mortality
We investigated the association of pharmacological therapy with COVID-19 positive presentation and inpatient mortality in COVID-19 patients. To assess the association of a given medication with these end points, we utilised the same stringent statistical approach that was described above for the comorbidities analysis, including matching on age, sex, and number of admissions, as well as the robustness analysis with different proportions of the full dataset (Figure 1). However, two further aspects were considered in this analysis; the first was conditioning on comorbidities; that is, the association of each medication with end points was assessed separately for patients with different comorbidities (Figure 2). This is to reduce the possibility of confounding observed association of medications with outcomes by their association with diseases. The definitions of comorbidity ICD-10 codes were based on Table 1.
The second aspect was the incorporation of the time of prescription. This is because the effects of medications can be dependent on the time of admission. Figure 3 illustrates the six lines of analysis that were conducted separately for each medication conditioned on comorbidities. The lines of analysis were designed to handle short-term and long-term medications differently. We define a short-term medication as a medication that was prescribed to the patient within the short period of time of indicated in Figure 3 but never before that (Figure 3: PP-ST1, PP-ST2, M-ST1, and M-ST2). As for long-term medications, they are defined as those which were prescribed at least twice in the patient history, one of which is more recent (Figure 3: PP-LT and M-LT). In all of these lines of analysis, medications with no more than ten patients in their active arm were skipped.
Medications were encoded using British National Formulary (BNF) codes. BNF codes form a hierarchy where medications are classified into BNF chapters, which are sub-classified into BNF sections, which in turn are classified into BNF paragraphs, and so on up to eight levels of depth. In most of the cases, medications in our dataset are recorded with codes up to three or four levels of depth. When counting data records of a given BNF code, all medications of codes in deeper hierarchy levels are counted as well.
Results
Characteristics of the dataset
Figure 4 shows that the COVID-19 positive cohort has higher median age, more males, and patients with more hospital admissions than the COVID-19 negative cohort. A similar trend also appears in the cohort of deceased COVID-19 patients compared to the cohort of COVID-19 patients with no observed death in hospital (Figure 4).
Association of comorbidities with COVID-19 presentation and mortality
We observed that the comorbidities of hypertension and diabetes have statistically significant associations with the diagnosis of COVID-19 (Table 2). Moreover, hypertension, diabetes, congestive heart disease, and renal disease show significant association with higher inpatient mortality in COVID-19 patients (Table 3). These results were obtained after controlling for the effect of age, sex, and number of admissions in patients’ history.
Association of medications with COVID-19 presentation and mortality
Table 4 shows the numbers of medications qualified for testing in each line of analysis given that they were prescribed to more than 10 patients within the relevant time window (Figure 3). This table also shows the number of medications, out of all of those qualified for testing, that showed significant and robust association with reduced presentation with COVID-19 (lines PP-ST1, PP-ST2, and PP-LT) or with reduced inpatient mortality in COVID-19 patients (lines M-ST1, M-ST2, and M-LT).
None of the tested medications was significantly and robustly associated with reduced COVID-19 presentation in our results, whether while considering short-term medications (PP-ST1 and PP-ST2) or long-term medications (PP-LT). Results with different thresholds of robustness are provided in Supplementary File S1.
Table 5 lists medications that passed the test of robust statistical significance. Where a BNF section and one of its BNF paragraphs both appear significant, the one that is less specific is omitted from this table for a more concise display. Nonetheless, all tested medications are listed in Supplementary File S2.
Considering short-term medications (M-ST1 & M-ST2), we observed an association with lower COVID-19 mortality for both diabetic and hypertensive patients for whom anticoagulants, lipid-regulating drugs, and penicillins were administered after their positive COVID-19 test (Table 5).
Biguanides prescribed to diabetic patients after testing positive on COVID-19 were associated with reduced mortality (Table 5). Interestingly, the more general BNF section of drugs used in diabetes, which includes insulins as well as antidiabetic drugs, showed association with reduced COVID-19 mortality for diabetic patients for whom there is an evidence of prescription before and after a positive COVID-19 test (M-LT) (Table 5).
Furthermore, angiotensin-converting enzyme inhibitors (ARBs) and antiplatelet drugs showed association with reduced COVID-19 mortality when administered to hypertensive patients after testing positive on COVID-19 (M-ST1 & M-ST2) (Table 5).
Discussion
We describe the retrospective analysis of 5294 patients presenting to two hospitals in London; 1253 of whom were COVID-19 positive. We observe that having a diagnosis of hypertension or diabetes was associated with a higher presentation with COVID-19, and that hypertension, diabetes, congestive heart failure, and renal disease were associated with a higher chance of in-hospital death for COVID-19 positive patients. Furthermore, we observe associations between a number of medications prescribed after a positive COVID-19 test and reduced mortality; examples include anticoagulants, lipid-regulating drugs, and penicillins for diabetic and hypertensive patients, biguanides for diabetic patients, and ARBs and antiplatelets drugs for hypertensive patients. We also observed an association with reduced COVID-19 mortality for diabetic patients who have been on diabetic drugs even before testing positive on COVID-19.
Co-morbidities associated with COVID-19 have been reported previously in the literature for a number of diseases such as hypertension (Williams & Zhang, 2020; Vizcaychipi, et al., 2020), renal diseases, and diabetes (Wu & McGoogan, 2020; Chen, et al., 2020; Yang, et al., 2020; Wang, et al., 2020; Perico, et al., 2020; Vizcaychipi, et al., 2020). COVID-19 patients with such co-morbidities were more likely to present to hospital which is in line with many NHS triage systems who recognise this group as vulnerable and therefore are more likely to recommend hospital attendance.
Some of our observations of associations of medications are in-line with published literature, such as the protective effect of anticoagulants patients on COVID-19 outcomes (Atallah, et al., 2020; Tang, et al., 2020). On the other hand, some of our observations on medications were at odds with the published literature or published literature gave conflicting results, such as the effect of ARBs (Perico, et al., 2020; Cure & Cure, 2020a; Cure & Cure, 2020b).
For all lines of analyses of comorbidities and medications, matching active and control arms on age, sex, and number of admissions was essential to account for substantial confounding factors. The essence of this is indicated by observing that the distributions of these covariates differed significantly between patients who had positive and negative COVID-19 test results as well as between COVID-19 patients who died and who did not (Figure 4). Matching the number of hospital admissions aims at comparing patients with more comparable medical history and overall health condition.
Limitations
Observed associations should be interpreted with care as they might be attributable to confounding factors. For example, we did not use measures of functional status, e.g. clinical fragility scale, which may have been over-represented in one arm of our analysis. Furthermore, we have not considered interactions between multiple co-morbidities and their association with outcomes. Severity of COVID-19 infection at presentation was not incorporated and no distinction is being made in this analysis between those who have been admitted for an acute episode of some comorbidity in their most recent admission and those who have not despite having that condition chronically. Also, length of symptoms at point of presentation to hospital were not assessed neither was the frailty scale.
Observations in this study are based on a population of patients presenting to hospitals and for whom COVID-19 status has been assessed to be positive or negative. This may cause a bias as it is not comparing individuals belonging to the general community population. Also, this population of patients comes from two London metropolitan hospitals that may not have a similar distribution of comorbidities and features as hospitals in other metropolitan cities or hospitals away from large cities. Generalisability is therefore not assured without further confirmatory studies. Additionally, our data does not include medications prescribed in primary care, resulting in a potentially inaccurate representation of the medication history of patients, especially prior to having a positive COVID-19 test result.
Finally, low data coverage and bias may cause absence of statistically significant associations with COVID-19 outcomes for some comorbidities or medications. Therefore, no conclusions may be drawn for such cases without further investigation.
Conclusions
This study provides an important piece of real-world evidence on associations between co-morbidities and medication prescription, respectively with COVID-19 positivity presenting to hospital and inpatient mortality. Identifying these associations can help in the crucial task of defining the vulnerable groups that may benefit from more stringent social distancing especially as lockdown due to COVID-19, is relaxed. Nonetheless, observations in this study have to be interpreted with caution due to potential bias and confounders, and confirmatory studies will be required to draw reliable conclusions.
Data Availability
_
Compliance with ethical guidelines
All methods were performed in accordance with the relevant guidelines and regulations.
The study was approved by the IG management team of Sensyne Health plc and Chelsea & Westminster NHS Foundation Trust under the Strategic Research Agreement (SRA) and relative Data Sharing Agreements (DSAs) signed by the NHS Trust and Sensyne Health plc on 25th July 2018.
All analyses were conducted on data with no personal identifying information. Therefore, informed consent was waived by the ethics committee of the Chelsea & Westminster NHS Foundation Trust, which provided ethical approval for the study.
Acknowledgements
This work uses data provided by patients and collected by the NHS as part of their care and support. We believe using patient data is vital to improve health and care for everyone and would, thus, like to thank all those involved for their contribution.
Special thanks are due to the Chelsea and Westminster (CW) COVID-19 AICU Consortium, comprising all critical care personnel who were part of the delivery of care during the COVID444 19 pandemic as follows: CW Anaesthetics Consultants, Critical Care Consultants, Trainees & Fellows from ICU, Anaesthesia, and seconded to ICU from other specialities, Surgeons, the supporting Respiratory and ED Physicians, Operating Department Practitioners and CW Critical Care Nurses. This united approach to an unprecedented clinical condition was critical not only to the management of the patients but also to our ability to document and collate the key data in a timely manner to support this analysis.
Also, a special thank you to Trystan Hawkin, Chris Chaney from CWplus, the Planned Care Clinical Division managers, porters, domestic personnel and the CW local community who without hesitation have supported the National Healthcare System throughout the COVID-19 pandemic.