Correlation of severity & clinical outcomes of COVID-19 with virus variants: A prospective, multicentre hospital network study

The SARS-CoV-2 coronavirus pandemic is considered one of the most severe and highly transmissible events in recorded history, with the Omicron variant being the dominant strain1. India reported its first case of COVID-19 in January 20202 and the infections rapidly expanded due to community transmission34. After the outbreak, the virus started mutating due to random replication errors and different variants of concern (VoCs) started appearing in various parts of the world posing a continued huge burden on public health56. The virus variants have different characteristics which might incite variable host immune and inflammatory responses. The variability in the human host responses to such variants may result in differences in the clinical trajectory of the disease7. The clinical course and outcome of the disease caused by SARS-CoV-2 depend on both viral characteristics and host factors. The pathogenicity and virulence of viruses are determined by their genetic make-up and genomic changes can markedly influence their infectivity and confer a selective survival advantage8. The various host factors associated with the clinical course of COVID-19 caused by SARS-CoV-2 include the demographic profile of the infected individuals, pre-existing comorbid diseases and vaccination status910.

The Delta variant led to a large surge in India from March to May, 2021 with apparently a much higher severity of COVID-19 and mortality11. However, whether the apparent increase in mortality was a result of a much higher number of cases or the virus was intrinsically more virulent could not be determined definitively. The subsequent emergence of the Omicron variant in 2022 led to a widespread surge of infections in the population. The Omicron variant is supposedly less virulent, but the severity of the disease caused by it could have been affected by other host factors such as prior infection or vaccination12.

There are no prospective studies to ascertain the relative severity and outcomes due to infections by various variants. The present, prospective multicentre study was therefore planned under the aegis of the Indian SARS-CoV-2 Genomics Consortium (INSACOG), which was established jointly by the Ministry of Health and Family Welfare, Department of Biotechnology, Council of Scientific and Industrial Research and Indian Council of Medical Research. The objectives of the study were to study the molecular epidemiology of SARS-CoV2, assess the severity of COVID-19 and the clinical outcomes among individuals infected with different VoCs and identify the outcome predictors.

Material & Methods

This was a longitudinal, prospective cohort study conducted at 14 hospitals across the country from September 2021 to December 2022 and coordinated by the Translational Health Science and Technology Institute (THSTI), Faridabad, India (Supplementary Table I and

All adults (18 yr and above) who came to the hospital with a suspicion of COVID-19 infection and tested positive for SARS-CoV-2 by the reverse transcription–polymerase chain reaction (RT-PCR) method were eligible to participate in this study (Fig. 1). All willing participants/legally authorized representative signed an informed consent to participate in this study. The participant population consisted of two cohorts. Cohort 1 enrolled all consecutively recruited individuals who reported and tested positive for SARS-CoV-2 infection in the testing centres of the study sites and were not hospitalized initially, rather treated as outpatients. Prospective follow up was done on these individuals to study their clinical course and need for hospitalization. Cohort 2 enrolled individuals who tested positive for SARS-CoV-2 and were hospitalized for COVID-19. This group was studied for the risk factors for moderate-to-severe disease and clinical outcomes.

Close

Fig. 1

Study flow.

Export to PPT

The enrolment period was from September 2021 to December 2022. For, Cohort 1 (testing centre enrolment), the follow up was also done at the end of three months from the date of testing positive for COVID-19. This follow up was done telephonically to document the current status of the participant (alive or dead) to estimate all-cause mortality, the short-term effects of COVID-19. If a participant complained of any symptoms in between, then they were called in for an in-person hospital follow up as part of routine clinical care. Data from such visits were also captured in the study-specific case report forms. For Cohort 2 (hospitalized individuals), the information about the course of the disease during hospitalization and the outcome of hospitalization was captured at the end of hospitalization. A telephonic follow up was also conducted at three months from the date of discharge.

Data captured included detailed clinical history, demographic details, duration of illness, occupation, history of exposure to a SARS-CoV2-infected individuals, comorbidities, vaccination history, past infection with SARS-CoV-2, RT-PCR CT (cycle threshold) value and prior treatment. Severe COVID-19 was defined as either SpO₂ <92 per cent or requiring treatment with O₂ supplementation or intensive care unit (ICU) care or death as a hospitalization outcome. The clinical data were collected in the electronic case record forms at the clinical sites by a dedicated research staff. The entered data were transmitted in real time to the central server at THSTI, Faridabad. REDCap (a browser-based meta-data-driven Electronic Data Capture platform) was used for the documentation of data by the sites and a Python script was developed to extract the required data for the analysis.

The nasopharyngeal swab samples that were collected in the viral transport media at the testing site were sent under optimal conditions to the linked IGSL for viral genome sequencing to detect SARS-CoV2 variants. Only samples which were positive for SARS-CoV-2 by RT-PCR, preferably with a CT value of 28 or less were used for whole genome sequencing on Illumina (San Diego, California) or Oxford Nanopore Technology (Oxford, UK) platforms as recommended by manufacturer13. The sequences were quality filtered and analyzed as described earlier14. Next clade and Pangolin pipelines were used for lineage assignment15.

Statistical analysis: The sociodemographic characteristics and clinical outcomes of the cohorts were described as median with interquartile range (IQR) for the quantitative variables and as frequency and percentage for the categorical variables. The following risk factors were analyzed for clinical outcomes: age, sex, comorbidity (defined as the presence of any one of the following medical conditions: diabetes, hypertension, heart disease, thyroid problems, cancer, chronic kidney, or lung or liver disease and other neurological disorders), vaccination status including booster (precautionary) dose, prior SARS-CoV-2 infection and virus variant. Multivariable logistic regression analysis was performed as a complete case analysis to estimate the association between the virus variant and the two outcomes of interest i.e., severity and mortality. For severity as the outcome, all the participants of the cohort were considered. For mortality, the analysis in the subset of hospitalized participants were conducted. Data were adjusted for age, sex, previous history of COVID-19, vaccination status, viral load at the time of diagnosis and comorbidity. We also introduced an interaction term between vaccination status and comorbidity in the multivariable model to adjust for potential interaction. To avoid any violation of temporality of association (reverse causality), only those predictors which were sure to have preceded the outcome (i.e. severity), such as age, sex, vaccination status, variant and comorbidity status, were evaluated.

All the data were collected on RedCAP software (Vanderbilt University, Nashville, Tennessee, USA), cleaned in Microsoft Excel and analyzed in STATA version 15 (StataCorp., Texas, USA).

Results

A total of 4972 participants gave consent to be included in the study from all 14 study sites. Of these, 958 (24%) participants belonged to the ‘hospitalized cohort’ and 4014 belonged to the non-hospitalized outpatient cohort. Of them, 3397 provided eligible samples for viral sequencing. The study flow is shown in Figure 1. The site-wise break-up of participants enrolled and samples sequenced is provided in Supplementary Table II.

Clinical characteristics of the participants: SARS-CoV-2 infection was seen across all the age groups. However, the predominant age group was 18-30 yr (42.5%) in the non-hospitalized cohort and >60 yr in the hospitalized cohort. The median age was 36 years (IQR: 27, 54). In the hospitalized cohort, nearly 85 per cent of the participants were symptomatic, with fever being the most common symptom in two-third, 50 per cent had cough and one-third had body aches. Other symptoms were headache, breathlessness, loss of smell/taste, nausea/vomiting, diarrhoea and abdominal pain. Around 76.5 per cent of the hospitalized population had at least one comorbidity, which included hypertension (42.69%), diabetes (36.53%), heart disease, cancer and thyroid problems in 6-12 per cent of the participants. Other comorbidities observed include kidney, liver and lung disease and other neurological disorders. 74.6 per cent of the non-hospitalized group did not have any comorbidity (Table I).

Evolution of virus variants over time: Of the 2723 samples sequenced, the majority were of Omicron variant (69.7%) followed by Delta variant (6.9%). In 12.9 per cent of samples, no specific lineage could be assigned and 10.5 per cent of samples had low RNA concentration and the sequencing could not be done in these. Only two variants dominated throughout the entire study period: Delta variant was confirmed in 188 and Omicron in 1897 patients. The Omicron subvariants BA.1, BA.2, BA.4 and BA.5 were observed across all time intervals during the study, and the XBB subvariant was observed only from September 2022 onwards (Fig. 2).

Close

Fig. 2

Evolution and Omicron sub-variants observed over time.

Export to PPT

Clinical characteristics of COVID-19 among individuals infected with Delta & Omicron variants: There was no significant difference between individuals infected with the Delta variant and those infected with the Omicron variant (Table II). The only significant difference was in the comorbid conditions between the two groups.

Severity of COVID-19 & virus variant: Overall, 146/2723 (5.4%) participants had severe COVID-19. All the parameters of severe disease, i.e. the rate of hospitalization, need for oxygen therapy, ICU admission and need for mechanical ventilation, were significantly higher in the Delta-infected individuals compared to Omicron-infected individuals. The mortality was also significantly higher due to the Delta infection (Table III).

Virus variant & severity of disease during different periods: The severity was higher from September to December 2021 and then declined subsequently over the whole of the calendar year 2022, when the rate of severity was much lesser at 4-5 per cent. When evaluated by lineage, the severity seen from September to December 2021 was also associated with a higher incidence of infections with the Delta variant (Table IV). The most prevalent variant during the subsequent periods was Omicron.

Risk factors & outcomes of COVID-19: The variations in the rate of severity over the course of the study period could be attributed to multiple risk factors. We evaluated the independent risk factors of severity during the entire study period (Table V). The presence of comorbidity was the most dominant risk factor for the severity, with an adjusted odds ratio (AOR) of 10.92 [95% confidence interval (CI): 5.32, 22.38]. This increase in odds was independent of the age, sex, status of the vaccination and prior COVID-19 infection. The other independent risk factors were age, number of doses of vaccine and the virus variant. It is interesting to note that participants aged 60 yr and above had three times more risk of severe COVID-19 as compared to those <30 yr (95% CI: 1.62, 6.19). The protection offered by vaccination was high among those who received two doses as compared to those who received a single dose (92 vs. 72%) when compared to the unvaccinated group, but this did not increase further with the booster dose. Infection with the Delta variant increased the risk of severe COVID-19 by 2.5-fold (95% CI: 1.52, 4.2) as compared to Omicron. The rates of severity were similar among the sub-variants of Omicron (Supplementary Table III). Among the Omicron subvariants, no subvariant was found to be associated with a more severe disease course. It was also observed that there was no association of CT value with the risk of severe COVID-19 (AOR: 1, 95% CI: 0.96, 1.04).

Risk factors for mortality among hospitalized participants: The variations in the mortality rate observed during the study could be attributed to several risk factors. We conducted an analysis to identify the independent risk factors associated with mortality among the hospitalized participants. As with the severity of COVID-19, participants older than 60 yr were >2.5 times higher at risk of death as compared to those aged <30 yr. Furthermore, receiving two vaccine doses provided higher protection (92%) compared to a single dose when compared to the unvaccinated group, but this increased protection did not extend to the booster dose. The other important risk factors for mortality were the Delta variant and the presence of comorbidity, with an aOR of 1.64 (95% CI: 0.73, 3.66) and 3.14 (95% CI: 0.68, 14.54), respectively (Table VI).

Discussion

In a first-of-its-kind hospital network connected to advanced biotechnology laboratories across India, we conducted a comprehensive molecular epidemiological investigation of the SARS-CoV-2 pandemic over two SARS-CoV-2 surges. The study correlated the clinical outcomes of the incident cases of SARS-CoV-2 infections with the virus variants identified in these participants in a longitudinal cohort study. This study documented the transition in the pandemic from being driven by the Delta variant to the Omicron variant over the course of time. In participants who reported complaints suggestive of COVID-19, the percentage of severe disease declined over the study period from 18 per cent in the first quarter (September-December 2022) to ~6 per cent throughout the year 2023. When evaluated for the risk factors of severity, the study identified comorbidity as the strongest risk factor, independent of vaccination status and the variant involved. This was followed by age, type of variant (Delta vs. Omicron) and vaccination status.

The prospective design of the study enabled us to pick up the transition in the course of the pandemic due to the change in the dominant variant. This transition in SARS variants was also associated with a decline in the severity of the illness which we infer, is a result of a complex interplay between the host and viral factors. Over the course of two years, the vaccine coverage increased to 70 per cent of the adult population as of December 202216. This important public health achievement seems to be the primary driving force in the reduction in the severity of infections. We demonstrate that infection with the Delta variant of the virus increased the risk of severity by 2.5-fold while Omicron infection was milder. This reduction in severity was independent of the vaccination status, which suggested an inherent reduction in the virulence of the Omicron virus variant. This inference is supported by experiments that have shown a reduction in severe illness in animal models infected with the Omicron variant compared to the Delta variant17. In line with evidence from other parts of the world, the severity in our cohort was not different between various sub-lineages of Omicron, e.g. BA.5, BA.2.38, BA.2.75, XBB, etc1819. Age, presence of comorbidity and lack of vaccination remained the major independent risk factors of severity, emphasizing the need for non-pharmacological preventive measures in the population20.

It is also important to consider other potential confounders that may have changed over the course of the pandemic and could bias the observed associations. Some of these confounders could be the availability of hospital beds which was an issue during the peak wave of Delta infection. However, the bed availability increased considerably during the second half of 2021 and 2022 and thus unlikely to affect the rate of hospitalization in the present study. Another potential confounder could be COVID test-seeking behaviour in the population, especially during the Omicron wave because it was perceived to be a mild illness after the initial few weeks. Thus, many mildly symptomatic cases might have been missed. Patient preference for the type of medical care could be another confounder, but we included both public and private hospitals and did not find much difference between the two groups (data not presented).

The prospective design, representativeness of various geographical regions of the country and viral genome sequencing are the major strengths of the study and gave us a unique opportunity to evaluate the severity of the illness across the SARS-CoV-2 variants. The limitations of the study include: (i) disproportionately low number of Delta infections compared to Omicron. This was because the Delta variant was on the decline by the time active recruitment in the study started; (ii) we could not assign virus lineage in a proportion of cases due to either technical failure or a high CT value; (iii) a lower number of some of the Omicron subvariants which could have resulted in non-significant results across these subvariants and (iv) a precise number of the approached participants but who did not give consent is not available.

Setting up a prospective cohort study across multiple hospital sites in India to evaluate the severity and outcomes of SARS-CoV-2 infection and correlate them with virus variants presented several operational challenges, especially because the hospital sites were in the midst of the COVID pandemic and the emphasis was on meeting the clinical demands of patient care. However, after our initial proposal to the clinical sites, most of them responded favourably. Issues such as funding, ethical clearance, storage and shipping of samples to the sequencing sites were some of the important issues that required additional efforts. A significant contributor to the success was online periodic meetings. Accurate data collection from various sites through standardized procedures was also challenging. Online training was provided to the study personnel and implemented a robust data management system to maintain data quality. The coordinating site presented the findings periodically to the INSACOG about the circulating virus variants and their severity.

Thus, the observations of this prospective cohort study periodically informed the national pandemic response, emphasizing the utility of such timely contemporary studies which provided real-time data from the hospital network closely linked to the biotechnology laboratories capable of viral genome sequencing in the country. As was the case with the Omicron variant, the emergence of any future variant can be detected and well characterized clinically and biologically through continued surveillance using such a nationwide platform. Sustaining such hospital-laboratory networks is important for the nation’s preparedness for future pandemics.