Impact of viral load at admission on the development of respiratory failure in hospitalized patients with SARS-CoV-2 infection.

The aim of our study was to elucidate if SARS-CoV-2 viral load on admission, measured by real-time reverse transcriptase-polymerase chain reaction (rRT-PCR) cycle threshold (Ct) value on nasopharyngeal samples, was a marker of disease severity. All hospitalized adult patients with a diagnosis of SARS-CoV-2 infection by rRT-PCR performed on a nasopharingeal sample from March 1 to March 18 in our institution were included. The study population was divided according to the Ct value obtained upon admission in patients with high viral load (Ct < 25), intermediate viral load (Ct: 25-30) and low viral load (Ct > 30). Demographic, clinical and laboratory variables of the different groups were analyzed to assess the influence of viral load on the development of respiratory failure during admission. Overall, 455 sequential patients were included. The median Ct value was 28 (IQR: 24-32). One hundred and thirty patients (28.6%) had a high viral load, 175 (38.5%) an intermediate viral load and 150 (33%) a low viral load. Advanced age, male sex, presence of cardiovascular disease and laboratory markers such as lactate dehydrogenase, lymphocyte count and C-reactive protein, as well as a high viral load on admission, were predictive of respiratory failure. A Ct value < 25 was associated with a higher risk of respiratory failure during admission (OR: 2.99, 95%IC: 1.57-5.69). SARS-CoV-2 viral load, measured through the Ct value on admission, is a valuable tool to predict the development of respiratory failure in COVID-19 inpatients.

Introduction

The 2019 novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) epidemic has quickly spread worldwide. As of December 2020, the worldwide outbreak of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), responsible for the viral pneumonia known as COVID-19, has caused more than 71 million infections and 1,597,000 deaths [1]. Most patients experience only minor symptoms that can be monitored at home. However, some individuals will present a severe respiratory distress syndrome, needing intensive care unit (ICU) admission and early respiratory support with significant morbidity and mortality [2]. Taking into account the current limited efficacy of pharmacological therapy, an adequate timely respiratory support is a cornerstone of COVID-19 management. Therefore, identification of early markers, ideally obtained on admission, able to predict respiratory failure, is essential to face the SARS-CoV-2 in a pandemic situation, where respiratory support facilities may be on the verge of being overwhelmed. Previous studies have shown that lymphopenia, elevated serum levels of lactate dehydrogenase (LDH) or proinflammatory cytokines such as interleukin (IL)-6 or IL-1 [3, 4] are markers of disease severity. On the other hand, several works have reported that patients with severe pneumonia needing mechanical ventilation have very high viral loads for extended periods of time [5, 6]. Despite this, to date, few studies have related the initial viral load to the outcome of COVID-19 patients [7].

The aim of our study was to assess the association between the viral load measured by real-time reverse transcriptase–polymerase chain reaction (rRT-PCR) cycle threshold (Ct) value on upper respiratory tract samples on admission and development of respiratory failure in patients who were admitted to our institution with a confirmed SARS-CoV-2 infection.

Methods

Study population

We performed a retrospective analysis from prospective automated medical records of all hospitalized adult patients (aged over 18) with a confirmed diagnosis of SARS-CoV-2 infection by rRT-PCR assay performed on nasopharyngeal exudate on admission, from March 1 to March 18. The study was developed at University Hospital “12 de Octubre”, a large 1300-bed hospital that serves a population of 450,000 inhabitants in southern Madrid (Spain). The study protocol was approved by the University Hospital 12 de Octubre Clinical Research Ethics Committee (reference 20/117), and a waiver of informed consent was granted, due to the retrospective nature of the observational design.

Microbiological methods

Nasopharyngeal samples were collected with flocked swabs in UTM™ viral transport medium (Copan Diagnostics, Brescia, Italy) and processed by automatized extraction and specific polymerase chain reaction methods [8]. Nucleic acid extraction was performed using the MicrolabStarlet IVD platform and the STARMag 96 × 4 Universal Cartridge Kit (Seegene, Seoul, South Korea) or NucliSENS EasyMAG instrument (bioMerieux, Marcy l’Etoile, France). For rRT-PCR the LightCycler 480 System instrument II (Roche Life Science, Indianapolis, IN, USA) performing the test TaqMan 2019nCoV assayKit v1, provided by Thermosfisher Scientific that amplifies three different viral regions in singleplex reactions was used. As a measure of relative quantification, the cycle threshold value (Ct) obtained in the amplification of the N gene was recorded. The RT-PCR Ct value represents the first PCR cycle in which the fluorescent signal for the target is greater than the minimal detection level [9]; thus, the lower Ct value represents the higher viral load in the sample.

Study definitions

For the purpose of the study, we arbitrarily divided the study population into tertiles based on the quantitative value of Ct. A high viral load was defined by a Ct < 25, intermediate viral load between 25 and 30 and low viral load if Ct was > 30.

The primary endpoint was development of respiratory failure, which was defined as the need for mechanical ventilation, non-invasive positive pressure ventilation or invasive mechanical ventilation. This was also the consideration with patients who developed a PaO2/FiO2 level < 200 during admission even if mechanical ventilation was not initiated. Cardiovascular disease was defined as the presence of coronary heart disease, heart failure, stroke and hypertension. Chronic lung disease was defined as the presence of chronic obstructive pulmonary disease, asthma or severe obstructive sleep apnea. Immunosuppression was defined as the presence of any of the following: active malignant neoplasia, autoimmune disease, solid organ transplantation, HIV infection, use of steroids or chemotherapy. Sepsis, septic shock and organ dysfunction were defined according to the terms proposed recently by the Third International Consensus Definitions for Sepsis and Septic Shock [10], including the Sequential Organ Failure Assessment (SOFA) score and the National Early Warning Score (NEWS). Acute respiratory distress syndrome (ARDS) was defined according to the American-European Consensus Conference on ARDS [11].

Statistical methods

Continuous variables were described with mean ± standard deviation (SD) or median and interquartile range (IQR) and compared with analysis of variance (ANOVA), Kruskal-Wallis test, Student’s t test or Mann-Whitney U test, as appropriate. Categorical parameters were notes in absolute numbers and percentage and compared with the Chi2 test or the Fisher exact test. The association of the viral load and the primary endpoint was assessed by multivariate analysis with logistic regression.

Principal component analysis was performed to obtain an overall idea of the data and the interrelationships among the different categorical variables (supplemental material).

Statistical analysis was performed with IBM SPSS Statistics for Windows, Version 25.0 (Armonk, NY: IBM Corp and SAS/STAT 10.1; SAS Institute Inc., Cary, NC).

Results

We included 455 patients with a confirmed SARS-CoV-2 infection. Mean age was 64.9 (SD: 18.1) and 255 (56%) were male. The median Ct value was 28 (IQR: 24–32). One hundred and thirty patients (28.6%) had a high viral load, 175 (38.5%) an intermediate viral load and 150 (33%) a low viral load.

Baseline comorbidities and clinical findings present on admission are shown in Table 1, as well as the differences between the three groups according to their viral load. Patients with high viral load at admission were of advanced age and mostly Caucasian and had more comorbidities. The time from symptom onset to positive rRT-PCR ranged from 0 to 19 days with a median of 5 days (IQR: 3–7); and was lower in patients with high viral load compared to patients with intermediate and low viral load (4 [IQR: 2–6]) vs. 6 [4-8], respectively; P < 0.001). There were no differences between the groups regarding the presence of fever, cough or dyspnoea. Treatments administered during admission were mainly lopinavir/ritonavir in 60.4% of patients, hidroxicloroquine in 74.1%, antibiotics in 91.4% and corticosteroids in 25.1%; only 36 patients (7.9%) received tocilizumab and one patient (0.2%) remdesivir. Corticosteroids were more frequently prescribed in patients with high viral load than in patients with intermediate and low viral load (31.8% vs. 23.4% and 16.7%, respectively; P = 0.026). Treatment with lopinavir/ritonavir was more frequently prescribed in patients with low and intermediate viral load than in patients with high viral load (69.9% and 64 .5% vs. 51%, respectively; P = 0.010).

Table 1

Baseline demographic, clinical characteristics and laboratory parameters at admission

Characteristics	All patients(n = 455)	Ct value (viral load)			P valuea	P valueb
		< 25(n = 130)	25–30(n = 175)	> 30 (n = 150)
Age (mean ± SD)	64.9 ± 18.1	69.1 ± 18.5	65.8 ± 17.4	60.1 ± 17.7	< 0.001	0.002
Sex male (n, %)	255 (56)	79 (60.8)	102 (58.3)	74 (49.3)	0.051	0.199
Caucasian race (n, %) Hispanic race (n, %)	344 (75.6) 92 (20.2)	112 (86.2) 15 (11.5)	128 (73.1) 38 (21.7)	104 (69.3) 39 (26)	0.001 0.003	0.001 0.004
Charlson index score (median, IQR)	3 (1–5)	4 (2–6)	3 (1–5)	2 (0.75–4)	< 0.001	< 0.001
Previous comorbid conditions
Cardiovascular disease1   Chronic lung disease2   Diabetes mellitus   Immunosuppression3   Chronic renal disease   Chronic liver disease   Obesity 4   Current or former smoker	205 (45.1) 98 (21.5) 76 (16.7) 89 (19.6) 29 (6.4) 22 (4.8) 136 (33.4) 115 (25.3)	70 (53.8) 37 (28.5) 28 (21.5) 37 (28.5) 10 (7.7) 4 (3.1) 42 (35.6) 36 (27.7)	79 (45.1) 34 (19.4) 29 (16.6) 39 (22.3) 15 (8.6) 11 (6.3) 50 (32.5) 49 (28)	56 (37.3) 27 (18) 19 (12.7) 13 (8.7) 4 (2.7) 7 (4.7) 44 (32.6) 30 (20)	0.006 0.037 0.048 <0.001 0.075 0.571 0.624 0.128	0.017 0.023 0.080 0.002 0.466 0.339 0.552 0.453
Days of symptoms prior to positive rRT-PCR (median, IQR)	5 (3–7)	4 (2–6)	5 (3–7)	7 (4–8.25)	< 0.001	< 0.001
Clinical findings
Fever (T > 38 °C)   Cough   Dyspnoea   Tachypnea 5   SpO2 < 90% air room   Diarrhoea   Myalgia   Vomiting   SOFA score   NEWS score   Chest X-ray abnormal findings	218 (47.9) 358 (78.7) 240 (52.7) 120 (27.6) 92 (20.3) 82 (18) 159 (34.9) 39 (7.9) 1 (0–2) 3 (1–5) 366 (80.4)	59 (45.4) 102 (78.5) 66 (50.8) 36 (29) 31 (23.8) 22 (16.9) 39 (30) 11 (8.5) 2 (0–3) 3 (1–5) 94 (72.3)	87 (49.7) 136 (77.7) 94 (53.7) 51 (30.7) 33 (19) 32 (18.3) 55 (31.4) 16 (9.1) 1 (0–2) 3 (1–5) 138 (78.9)	72 (48) 120 (80) 80 (53.3) 33 (22.9) 28 (18.7) 28 (18.7) 65 (43.3) 9 (6) 1 (0–2) 2 (1–4.25) 134 (89.3)	0.683 0.741 0.679 0.244 0.294 0.709 0.017 0.427 0.041 0.198 < 0.001	0.495 0.942 0.593 0.684 0.229 0.700 0.162 0.784 0.086 0.311 0.006
Laboratory parameters
Lymphocytes (× 103 cells/μl)   LDH (U/l) 6   GOT (U/l)   GPT (U/l)   CPK (U/l)   TnT (U/l) 7   C-reactive protein (mg/dl)   Ferritin (mg/dl) 8   D-dimers (ng/ml) 9	0.9 (0.6–1.2) 326.5 (265–408) 32 (24–50) 25 (17–40) 86 (52–176.7) 10.5 (5.8–21.4) 7.7 (3.1–14.9) 699 (335–1357) 664 (418–1220)	0.8 (0.6–1.2) 317 (245.2–383) 31 (23–45) 23 (15–31) 101 (61.7–180) 12.7 (7–30) 6.9 (2.7–15) 665 (314–1335) 611 (350–1175)	0.8 (0.6–1.2) 322 (265–426) 32 (25–50) 25 (17–39) 87 (48–194) 10.6 (6.6–18) 7.5 (3–14) 675 (328–1263) 684 (457–1346)	0.9 (0.7–1.2) 348 (282–432) 37 (25–58) 29 (19–50) 78.5 (47–173) 7.4 (5–20) 8.3 (3.8–15) 794 (365–1466) 667 (441–1308)	0.310 0.004 0.049 0.001 0.239 0.012 0.503 0.376 0.206	0.671 0.002 0.026 0.001 0.119 0.011 0.453 0.488 0.085

Results are expressed as mean ± standard deviation (SD) or median with interquartile range (IQR) or as absolute value (percentage); Ct cycle threshold, rRT-PCR real-time reverse transcription polymerase chain reaction, SOFA Sequential Organ Failure Assessment, NEWS National Early Warning Score

1Cardiovascular disease was defined as the presence of coronary heart disease, heart failure, stroke and hypertension

2Chronic lung disease was defined as the presence of chronic obstructive pulmonary disease, asthma or severe obstructive sleep apnea

3Immunosuppression was defined as the presence of any the following: active malignant neoplasia, autoimmune disease, solid organ transplantation, HIV infection, use of steroids, or chemotherapy. Use of steroids was defined as (1) more than 20 mg/day of oral prednisone during 7 days or longer or (2) less than 20 mg/day during a minimum of 3 months

4Obesity was defined as body mass index ≥ 30, data available in 407 patients

5Tachypnea was defined as breathing rate > 20 beats per minute; data available for 434 patients

6LDH lactate dehydrogenase; data available for 364 patients (upper limit of LDH in the local laboratory is 225 mg/dl)

7Troponin T: data available for 257 patients

8Ferritin: data available for 402 patients

9D-dimers: data available for 270 patients

aAcross all Ct value groups. P value for trend is used when appropriate

bBetween high viral load group (Ct < 25) and other groups combined

Laboratory values in the first 24 h from admission are shown in Table 1. Most patients presented lymphopenia as well as elevated acute phase reactants such as ferritin, C-reactive protein (CRP) and D-dimers, without significant differences between groups.

Respiratory failure occurred in 161 patients (35.4%) after a median of 9 days (IQR 6–11), and 120 (26.4%) patients died during hospitalization (Table 2). According to the viral load on admission (Table 2), the patients with high viral load developed respiratory failure more often (42.3% vs. 32.6%, P = 0.051) and experienced higher mortality at 30 days (33.8% vs. 23.4%, P = 0.022). Intensive care unit (ICU) admission rate, however, was higher in the group of patients with intermediate and low viral load (12.3% vs. 6.2%, P = 0.054). There were no differences in the presence of septic shock, acute kidney injury, venous thrombosis, hepatitis or major adverse cardiovascular events (MACE) between groups.

Table 2

Outcome of patients with SARS-CoV-2 infection according viral load at diagnosis

	All patients(n = 455)	Ct value (viral load)			P valuea	P valueb
		< 25(n = 130)	25–30(n = 175)	> 30 (n = 150)
Need for supplemental oxygen ARDS ICU admission Length of ICU stay (days) Non-invasive mechanical ventilation Invasive mechanical ventilation Days of invasive ventilation Prone position Septic shock Acute kidney injury (AKI) Venous thrombosis Hepatitis 1 MACE event 2 Length of hospital stay (days)	311 (68.4) 132 (29) 48 (10.5) 13 (7–18) 29 (6.4) 46 (10.1) 12.5 (7–17) 39 (8.6) 12 (2.6) 69 (15.2) 8 (1.8) 32 (7) 5 (1.1) 9 (6–14)	94 (72.3) 43 (33.1) 8 (6.2) 15 (11–18) 8 (6.2) 8 (6.2) 15 (10–17.2) 7 (5.4) 3 (2.3) 23 (17.7) 2 (1.5) 6 (4.6) 3 (2.3) 10 (6–17)	121 (69.1) 53 (30.3) 21 (12.3) 13 (5.5–17) 12 (6.9) 19 (10.9) 13 (7–16) 19 (10.9) 8 (4.6) 32 (18.3) 4 (2.3) 16 (9.1) 1 (0.6) 9 (5–13)	96 (64) 36 (24) 19 (12.7) 12 (7–23) 9 (6) 19 (12.7) 11 (7–20) 13 (8.7) 1 (0.7) 14 (9.3) 2 (1.3) 10 (6.7) 1 (0.7) 8.5 (6–13.2)	0.134 0.092 0.084 0.533 0.946 0.075 0.488 0.360 0.346 0.045 0.873 0.544 0.203 0.214	0.251 0.227 0.054 0.420 0.903 0.077 0.258 0.125 0.781 0.342 0.822 0.202 0.118 0.085
Respiratory failure Days of symptoms to RF 3 (median, IQR) Overall in-hospital mortality	161 (35.4) 9 (6–11) 120 (26.4)	55 (42.3) 8 (6–10) 44 (33.8)	64 (36.6) 9(6–11) 44 (25.1)	42 (28) 9(6.7–12.2) 32 (21.3)	0.012 0.426 0.019	0.051 0.280 0.022

Results are expressed as median with interquartile range (IQR) or as absolute value (percentage). ICU intensive care unit, ARDS acute respiratory distress syndrome, RF respiratory failure, rRT-PCR real-time reverse transcription polymerase chain reaction

1Hepatitis was defined by AST and/or ALT levels > 5 times the ULN

2Major adverse cardiovascular event (MACE) was defined as nonfatal stroke, nonfatal myocardial infarction and cardiovascular death

3Days from the onset of symptoms to development of respiratory failure

aAcross all Ct value groups. P value for trend is used when appropriate

bBetween high viral load group (Ct < 25) and other groups combined

We performed an analysis of the variables associated with the development of respiratory failure during admission (Table 3). Parameters associated with a higher likelihood of respiratory failure were age, the presence of comorbidities, current or former smoking habit, the presence of dyspnoea or tachypnea on admission, pathological chest X-ray, lymphopenia, a high CRP value and LDH. Patients who developed respiratory failure had a high viral load on admission in greater proportion than those who did not (34.2% vs. 25.5%, P = 0.05). In a multivariate analysis (Table 4) adjusted for age, sex, cardiovascular disease, chronic lung disease, immunosuppression, current or former smoker, the presence of dyspnoea, abnormal chest x-ray findings on admission, severe lymphopenia ≤ 0.7 × 103 cells/μl, LDH ≥ 350 U/l and C-reactive protein ≥ 6 mg/dl, a high viral load was independently associated with increased risk of respiratory failure (adjusted odds ratio 2.99; 95% CI: 1.57–5.69; P = 0.001). Likewise, a high viral load was significantly associated with in-hospital mortality (OR: 2.04, 95%IC: 1.44–4.00, P = 0.037) (data shown in supplemental material).

Table 3

Risk factors for respiratory failure during admission

Characteristics	Respiratory failuren = 161	Non-respiratory failuren = 294	P value	OR	IC 95%
Age ≥ 60 years	132 (76.4)	158 (53.7)	< 0.001	2.78	1.81–4.28
Sex male (n, %)	109 (67.7)	146 (49.7)	< 0.001	2.12	1.42–3.17
Caucasian race (n, %)	127 (78.9)	217 (73.8)	0.228
Hispanic race (n, %)	31 (19.3)	61 (20.7)	0.704
Charlson index score ≥ 2	135 (83.9)	172 (58.5)	< 0.001	3.68	2.28–5.95
Previous comorbid conditions
Cardiovascular disease 1	96 (59.6)	109 (37.1)	< 0.001	2.51	1.69–3.72
Chronic lung disease 2	43 (26.7)	55 (18.7)	0.047	1.58	1.00–2.50
Diabetes mellitus	34 (21.1)	42 (14.3)	0.062
Immunosuppression 3	43 (26.7)	46 (15.6)	0.004	1.96	1.23–3.14
Chronic renal disease	14 (8.7)	15 (5.1)	0.133
Chronic liver disease	10 (6.2)	12 (4.1)	0.311
Obesity 4	60 (39.7)	76 (29.7)	0.038	1.56	1.02–2.38
Current or former smoker	57 (35.4)	58 (19.7)	<0.001	2.23	1.45–3.44
Days of symptoms prior to positive PCR (median, IQR)	5 (3–7)	5 (3–7)	0.896
Clinical findings at admission
Fever (T > 38 °C)	81 (50.3)	137 (46.6)	0.449
Cough	131 (81.4)	227 (77.2)	0.301
Dyspnoea	112 (69.6)	128 (43.5)	<0.001	2.96	1.97–4.45
Tachypnea 5	70 (45.8)	50 (17.8)	<0.001	3.90	2.51–6.06
Diarrhoea	24 (14.9)	58 (19.7)	0.201
Myalgia	45 (28)	114 (38.8)	0.021	0.61	0.40–0.93
Vomiting	17 (3.7)	19 (4.2)	0.122
Chest X-ray abnormal findings	141 (87.6)	225 (76.5)	0.005	2.16	1.30–3.71
Baseline laboratory findings
Lymphocytes ≤ 0.7 × 103 cells/μl	87 (54)	80 (27.3)	<0.001	3.13	2.09–4.68
LDH ≥ 350 U/l 6	103 (64.8)	80 (27.5)	<0.001	4.85	3.20–7.35
C-reactive protein ≥ 6 mg/dl	133 (82.6)	126 (43.3)	<0.001	6.21	3.89–9.90
High viral load (Ct < 25)	55 (34.2)	75 (25.5)	0.051	1.51	0.99–2.30
Intermediate viral load (Ct 25–30)	64 (39.8)	111 (37.8)	0.676
Low viral load (Ct > 30)	42 (26.1)	108 (36.7)	0.021	0.61	0.40–0.93

Results are expressed as mean with standard deviation (SD), median with interquartile range (IQR) or as absolute value (percentage). LDH lactate dehydrogenase, Ct cycle threshold

1Cardiovascular disease was defined as the presence of coronary heart disease, heart failure, stroke and hypertension

2Chronic lung disease was defined as the presence of chronic obstructive pulmonary disease, asthma or severe obstructive sleep apnea

3Immunosuppression was defined as the presence of any the following: active malignant neoplasia, autoimmune disease, solid organ transplantation, HIV infection, use of steroids or chemotherapy. Use of steroids was defined as (1) more than 20 mg/day of oral prednisone during 7 days or longer or (2) less than 20 mg/day during a minimum of 3 months

4Obesity was defined as body mass index ≥ 30, data available in 407 patients

5Tachypnea was defined as breathing rate > 20 beats per minute; data available for 434 patients

6LDH: data available for 450 patients

Table 4

Multivariate analysis of patients for the occurrence of respiratory failure

Variablen = 450	Multivariate analysis
	OR	95% IC	P value
Age ≥ 60 years	2.02	1.16–3.53	0.013
Sex male	1.66	1.02–2.72	0.042
Cardiovascular disease 1	1.93	1.16–3.21	0.011
Dyspnoea	2.46	1.50–4.03	<0.001
Lymphocytes ≤ 0.7 × 103cells/μl	2.26	1.41–3.63	0.001
LDH ≥ 350 U/l 3	3.00	1.79–5.05	<0.001
C-reactive protein ≥ 6 mg/dl	3.39	1.93–5.95	<0.001
Low viral load (Ct > 30)	Ref.
Intermediate viral load (Ct 25–30)	1.81	1.02–3.22	0.044
High viral load (Ct < 25)	2.99	1.57–5.69	0.001

OR odds ratio, IC confidence interval, LDH lactate dehydrogenase, Ct cycle threshold

1Cardiovascular disease was defined as the presence of coronary heart disease, heart failure, stroke and hypertension

3LDH: Data available for 450 patients

Discussion

According to our results, patients with a high viral load of SARS-CoV-2 on admission, measured by a Ct value of the rRT-PCR below 25, were more than twice as likely to develop in-hospital respiratory failure compared to patients with lower viral loads, regardless of other previously described severity parameters.

Viral nucleic acid detection by RT-PCR assays is the gold standard for the diagnosis of COVID-19. Using this technique, we can obtain an indirect viral load value (Ct) easily and immediately after diagnosis. This parameter has been correlated in previous studies with the viral ability of spread and with a longer persistence of the virus in respiratory samples [12, 13]. Also, it has been shown that patients with severe disease have significantly higher viral loads than patients with mild disease [14, 15], and Zou and cols. reported that patients admitted to ICU had detectable viral RNA in nasopharyngeal exudate after 10 days from symptom onset [16]. However, few studies have attempted to elucidate whether there is a relationship between viral load at diagnosis and a worse evolution of the disease.

In our study, patients with a high viral load at admission were of advanced age and had more comorbidities than those with lower viral loads, something that has already been reported in previous studies with SARS-CoV-2 [12, 14]. Indeed, elderly, fragile patients may present a higher expression of ACE2 receptors which serve as a cell entry for SARS-CoV-2 [17] and may also have an impaired immunity, all this contributing to higher viral loads on admission and a more severe disease. Likewise, patients with a higher viral load presented lymphopenia, elevated LDH and CPR more frequently, events that have been related with greater severity of the disease [6, 18]. Finally, patients with a Ct < 25 had a significantly shorter period of symptoms prior to diagnosis suggesting greater severity of the illness.

Besides a high viral load at admission, other factors such as advanced age, comorbidity, lymphopenia, LDH or C-reactive protein were associated with a higher risk of respiratory failure. Remarkably, despite that our patients with a high viral load developed respiratory failure more frequently than those with a lower viral load, the rates of admission to the ICU and mechanical ventilation are lower due to the limited access of these patients to the ICU at the time of maximum healthcare pressure in our centre. Although this fact probably contributed to increase mortality in group, a high viral load remained as an independent factor after adjusting for age and comorbidity in the multivariate analysis.

Our findings align with other previously published studies. In the study by Magleby et al. [7], the presence of a high viral load measured by Ct was independently associated with mortality and with the risk of intubation during hospitalization. They also found that patients with higher viral loads developed myocardial infarction, congestive heart failure and acute kidney injury more frequently than those with lower viral loads. However, this is a fact that may be caused by multiple factors related to hospitalization, and this finding is not replicated in our work. A recently published work also found a relationship between high viral load on admission and a composite outcome of death, intubation or ECMO during hospitalization [19], and Westblade et al. reported a higher mortality in patients with high viral load in the oncological population [20]. Yu X. et al. [21] reported in a limited number of patients that those with high viral loads in sputum had more severe disease than those with lower viral loads. On the other hand, a recently published study [22] did not find an association between initial viral load and clinical outcome in a cohort of patients, mostly non-hospitalized with mild disease, a population not comparable with ours (Fig. 1).

Fig. 1

Probability of respiratory failure during hospitalization among patients with high, intermediate and low viral load

An important question to be answered is whether the viral load measured in nasopharyngeal exudate is a good indicator of that present in lung tissue, since the lower respiratory tract specimens have significantly higher SARS-CoV-2 viral RNA levels than nasal and throat swab specimens [23, 24]. However, there seems to be a good correlation as previously reported [25], and the nasopharyngeal sample is easier, quicker and widely available.

Our study presents several limitations. Firstly, this is a single-centre observational cohort study and includes only patients from the first days of the pandemic in Spain, with high hospital occupancy rates; therefore, mortality and respiratory failure rates could decrease in the current situation. However, it does not necessarily signify that the relationship between higher viral load and worse prognosis is modified. Secondly, it only includes nasopharyngeal samples, whose quality is determined by the ability to collect the sample and may not adequately reflect viral load in lung tissue, as previously noted. However, this sample is much easier to obtain than sputum or lower respiratory tract samples and is accessible in all centres and provides quick results for decision-making.

Thirdly, SARS-CoV-2 RT-PCR assays are multiple and vary between centres; some authors have questioned whether the relationship between viral load and worse evolution is consistent with different diagnostic assays. However, studies using different diagnostic platforms have corroborated this relationship, although there are variations in the cut-off points of Ct [20].

In addition, the rRT-PCR does not distinguish between viable and non-viable virus. Some studies [26] have shown that positive viral culture rate decreases with increasing Ct values. Despite these data, RT-PCR is the gold-standard for diagnosis worldwide due to its sensitivity and specificity and the impossibility of performing viral culture techniques in most centres [27].

In conclusion, in our experience, SARS-CoV-2 viral load, measured by the Ct value of the rRT-PCR in nasopharyngeal swabs on admission, is a viable prognostic marker for the development of respiratory failure. It is easy to obtain and a widely available tool that could help to select those patients who would benefit from a closer follow-up and an in-time and appropriate respiratory support.

(DOCX 25 kb)