Introduction

In February 2020, the World Health Organization (WHO) named the disease caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) coronavirus disease 2019 (COVID-19). The first case occurred in Wuhan, China, in December 2019, and COVID-19 quickly spread worldwide, posing great challenges to the health systems of various countries1,2. SARS-CoV-2 shares a 79.6% nucleotide identity with SARS-CoV and 51.8% identity with Middle East respiratory syndrome coronavirus (MERS-CoV). The latter two pathogens caused regional outbreaks of acute respiratory diseases in 2002–2003 and 2012, respectively, and each disease caused hundreds of deaths3,4,5. Influenza virus is an RNA virus that can cause influenza in humans and animals and belongs to the Orthomyxoviridae family. This family has four genera; however, only genera A and B are clinically relevant to humans6. Influenza has obvious seasonality7. The WHO estimates that the annual influenza season usually leads to approximately 3–5 million critical cases and 290,000 to 650,000 deaths8. Although influenza often leads to significant mortality and morbidity, the public usually thinks that influenza is a trivial disease, similar to the common cold9. In the early stage of the pandemic, COVID-19 was downplayed as only a ‘little flu’ by some people10. There are many similarities between SARS-CoV-2 and influenza virus in terms of transmission and manifestations, but compared with influenza virus, SARS-CoV-2 is more contagious11, and influenza makes diagnosing COVID-19 difficult12.

We aimed to explore and compare the clinical characteristics of COVID-19 and influenza to deepen the understanding of these two diseases and provide some guidance for clinicians to make differential diagnoses. In November 2021, the Omicron variant was first detected and quickly became the main pathogenic strain worldwide, which was quite different from those of previous strains13. The virulence of the Omicron variant has significantly decreased with a lower rate of severe cases of disease, hospitalizations, and deaths, but it is more infectious and transmissible14,15. The non-Omicron strains primarily affects the lungs, whereas the Omicron variant primarily affects the upper respiratory tract16. A study by the University of Hong Kong found that the Omicron variant infects human bronchi more than 70 times faster than the Alpha variant and has a higher replication speed. In contrast, the original strain infects the lung more than 10 times faster than the Omicron variant, and the replication speed is also faster than the latter17. We think that it is not rigorous enough to put the Omicron variant and the non-Omicron strains in the same group when comparing the clinical characteristics of COVID-19 and influenza. Therefore, patients infected with the Omicron variant were not included in our study.

Materials and methods

This meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (PROSPERO registration number: CRD42023397835).

Eligibility criteria

Articles that met the following criteria were included in this meta-analysis: 1) articles with controls; 2) the experimental group and the control group were COVID-19 and influenza patients, respectively; and 3) the data collection time of COVID-19 patients had to be before November 2021, that is, before the outbreak of the Omicron variant of SARS-CoV-213.

The exclusion criteria were as follows: 1) animal studies, 2) case reports, 3) reviews and comments, 4) duplicate data, 5) samples from children only, and 6) sample size less than five.

Information sources and search strategy

We searched PubMed, Embase and Web of Science for articles published before November 25, 2023. There was no restriction on the language of the article to collect useful information on a global scale. The search strategy in PubMed was as follows: ((((((COVID-19[Title]) OR (2019-nCoV[Title])) OR (Coronavirus Disease 2019[Title])) OR (SARS-CoV-2[Title])) OR (2019 Novel Coronavirus Disease[Title])) OR (2019 Novel Coronavirus Infection[Title])) AND ((((Flu[Title]) OR (Influenza[Title])) OR (Influenzas[Title])) OR (Grippe[Title])).

Study selection process

All articles obtained from PubMed, Embase and Web of Science were imported into NoteExpress software. We first deleted duplicate articles through the matching function of the software and then conducted a preliminary screening by reading titles and abstracts to exclude articles that were not related to our study. Finally, through full-text reading, selected articles were further screened to determine which articles could be included in our meta-analysis.

Data selection process and items

Data extraction was performed independently by two authors to ensure the accuracy of the data. Disagreements were resolved by discussion and referred to a third author for a final decision.

The following information on COVID-19 and influenza patients was included in this meta-analysis: 1) major characteristics, including age, sex, body mass index (BMI), length of hospital stay, length of intensive care unit (ICU) stay, number of current smokers, number of patients admitted to the ICU, number of patients who received mechanical ventilation, number of patients who received extracorporeal membrane oxygenation (ECMO) and number of deaths; 2) symptoms, including fever, cough, expectoration, dyspnea, rhinorrhea, sore throat, headache, chest pain, myalgia, nausea, vomiting, diarrhea and abdominal pain; 3) laboratory findings, including lymphocytes, white blood cells, neutrophils, C-reactive protein, creatinine, hemoglobin, platelets, procalcitonin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), lactate dehydrogenase (LDH) and creatine kinase; and 4) comorbidities, including diabetes, hypertension, heart failure, coronary heart disease, chronic renal failure, stroke, cancer, chronic liver disease, asthma, chronic obstructive pulmonary disease (COPD) and immunodepression.

Study risk of bias assessment

The Newcastle–Ottawa quality assessment scale was used to assess the quality and risk of bias of the included articles. A total score of seven or more meant that the article had a low risk of bias and high quality.

Reporting bias assessment

We used funnel plots and Egger’s test to evaluate reporting bias assessment, and a p value < 0.05 indicated the presence of bias.

Statistical analysis

Data were analyzed and evaluated by using odds ratios (ORs), mean differences (MDs) and standardized mean differences (SMDs), where ORs were applied to dichotomous variables, MDs were applied to continuous variables with the same units of measurement, SMDs were applied to continuous variables with different units of measurement, and confidence intervals (CIs) were both set at 95%. For data with only sample size and quartile, we used the transformation formula to calculate the mean and standard deviation18. The I2 statistic was used to quantify heterogeneity: 0–40% may not be important; 30–60% may represent moderate heterogeneity; 50–90% may represent substantial heterogeneity; 75–100% may represent considerable heterogeneity19. Subgroup analysis was used to investigate the source of heterogeneity. A random-effects model was used to estimate the effect value. Stata 14.0 was used for statistical analysis, and a p value of z test < 0.05 indicated that the result was statistically significant.

Results

Study selection

From the three databases, we retrieved a total of 4968 articles and deleted 2398 duplicate articles through NoteExpress software. A total of 2482 articles that were not relevant to our study were excluded by reading the titles and abstracts. Of the remaining 388 articles, 288 were further screened and excluded by reading the full text. If data from multiple articles were from the same research institution and the dates of extraction of patient information overlapped, the data were considered to be at risk of duplication, the article with higher quality was retained, and the rest were deleted. The flow diagram of the article selection process is shown in Fig. 1.

Fig. 1: PRISMA diagram.
figure 1

The flow diagram of the article selection process.

Full size image

Risk of bias in studies

The Newcastle–Ottawa quality assessment scale is described in Supplementary Table 1. We found that most of the articles included in this meta-analysis were of high quality and had a low risk of bias.

Characteristics and results of individual studies

After screening, a total of 100 articles were included in this meta-analysis (Table 1), including eight prospective studies, 78 retrospective studies, and 14 studies that did not specify the design. Data were collected from 27 regions and included 226,913 COVID-19 patients and 201,617 influenza patients. There were 17 articles that specifically studied patients with influenza A, 41 articles in which the type of influenza was A/B, and the remaining 42 articles did not record the influenza subtype. The earliest data for influenza were from 2006, and all the data for COVID-19 were obtained before the outbreak of the Omicron variant of SARS-CoV-2.

Table 1 Characteristics of individual studies.
Full size table

Results of syntheses

Major characteristics (Supplementary Table 2)

The age characteristics20,21,22,23 of four studies, sex characteristics20,21,22,24 of four studies and BMI characteristics23 of one study were screened after matching the experimental group and the control group; thus, their data were not included in this meta-analysis. We found no significant differences in age between patients with COVID-19 and those with influenza (MD = 0.21, 95% CI: −0.99–1.41, I2 = 98.0%, p = 0.730, Supplementary Figure 1). Compared to influenza, COVID-19 was more common among men (OR = 1.46, 95% CI: 1.23–1.74, I2 = 99.0%, p < 0.001, Supplementary Figure 2) and people with a higher BMI (MD = 1.43, 95% CI: 1.09–1.77, I2 = 67.4%, p < 0.001, Supplementary Figure 3). The proportion of current smokers among COVID-19 patients was lower (OR = 0.25, 95% CI: 0.18–0.33, I2 = 55.4%, p < 0.001, Fig. 2) than that among influenza patients. Although the use of ECMO did not differ between the two groups (OR = 1.06, 95% CI: 0.76–1.47, I2 = 62.5%, p = 0.740, Fig. 3), patients with COVID-19 had longer stays in the hospital (MD = 3.20, 95% CI: 2.58–3.82, I2 = 98.2%, p < 0.001, Supplementary Figure 4) and ICU (MD = 3.10, 95% CI: 1.44–4.76, I2 = 96.9%, p < 0.001, Supplementary Figure 5), more patients required mechanical ventilation (OR = 2.30, 95% CI: 1.77–3.00, I2 = 93.7%, p < 0.001, Supplementary Figure 6) and ICU admission (OR = 2.03, 95% CI: 1.72–2.40, I2 = 95.6%, p < 0.001, Supplementary Figure 7), and patients had higher mortality (OR = 2.22, 95% CI: 1.93–2.55, I2 = 94.1%, p < 0.001, Fig. 4).

Fig. 2: Forest plot of differences in current smokers between patients with COVID-19 and those with influenza.
figure 2

ID identity document, OR odds ratio, CI confidence interval.

Full size image
Fig. 3: Forest plot of differences in extracorporeal membrane oxygenation between patients with COVID-19 and those with influenza.
figure 3

ID identity document, OR odds ratio, CI confidence interval.

Full size image
Fig. 4: Forest plot of differences in mortality between patients with COVID-19 and those with influenza.
figure 4

ID identity document, OR odds ratio, CI confidence interval.

Full size image

Symptoms (Supplementary Table 3, Supplementary Figures 8–20)

COVID-19 patients often present with influenza-like symptoms. By comparison, there was no significant difference in the proportion of patients with abdominal pain (OR = 1.06, 95% CI: 0.83–1.36, I2 = 0.0%, p = 0.657), chest pain (OR = 0.92, 95% CI: 0.55–1.53, I2 = 31.8%, p = 0.753), headache (OR = 1.04, 95% CI: 0.72–1.50, I2 = 75.8%, p = 0.835), myalgia (OR = 0.72, 95% CI: 0.49–1.05, I2 = 76.0%, p = 0.09), fever (OR = 0.73, 95% CI: 0.48–1.11, I2 = 85.3%, p = 0.139), nausea (OR = 0.81, 95% CI: 0.40–1.64, I2 = 73.9%, p = 0.551), vomiting (OR = 1.15, 95% CI: 0.51–2.56, I2 = 64.3%, p = 0.739), or dyspnea (OR = 0.76, 95% CI: 0.50–1.17, I2 = 80.2%, p = 0.217) between COVID-19 patients and influenza patients. Patients with COVID-19 were more likely to have diarrhea (OR = 1.65, 95% CI: 1.30–2.08, I2 = 19.6%, p < 0.001), while rhinorrhea (OR = 0.38, 95% CI: 0.19–0.75, I2 = 78.7%, p = 0.005), expectoration (OR = 0.29, 95% CI: 0.22–0.39, I2 = 47.3%, p < 0.001), cough (OR = 0.54, 95% CI: 0.40–0.72, I2 = 78.1%, p < 0.001) and sore throat (OR = 0.64, 95% CI: 0.44–0.93, I2 = 73.0%, p = 0.020) were more common among influenza patients.

Laboratory findings (Supplementary Table 4, Supplementary Figures 21–32)

Our study showed that platelet counts (SMD = 0.42, 95% CI: 0.28–0.56, I2 = 92.2%, p < 0.001) and hemoglobin (SMD = 0.24, 95% CI: 0.07–0.41, I2 = 91.5%, p = 0.006) and ALT (SMD = 0.35, 95% CI: 0.17–0.54, I2 = 88.6%, p < 0.001) levels were higher in COVID-19 patients than in influenza patients, while creatinine levels (SMD = −0.14, 95% CI:−0.03–−0.25, I2 = 75.3%, p = 0.012), procalcitonin levels (SMD = −0.47, 95% CI: −0.73–−0.22, I2 = 81.5%, p < 0.001), neutrophil levels (SMD = −0.48, 95% CI: −0.79–−0.17, I2 = 97.0%, p = 0.003) and white blood cell counts (SMD = −0.26, 95% CI: −0.37–−0.16, I2 = 88.5%, p < 0.001) were significantly lower than in influenza patients. There were no significant differences in lymphocyte counts (SMD = 0.14, 95% CI: 0.00–0.27, I2 = 90.1%, p = 0.05) or C-reactive protein (SMD = −0.05, 95% CI: −0.28–0.18, I2 = 95.6%, p = 0.664), LDH (SMD = 0.05, 95% CI: −0.23–0.33, I2 = 95.4%, p = 0.773), AST (SMD = 0.09, 95% CI: −0.20–0.37, I2 = 94.3%, p = 0.551) or creatine kinase (SMD = −0.09, 95% CI: −0.31–0.14, I2 = 81.9%, p = 0.454) levels between the two groups.

Comorbidities (Supplementary Table 5, Supplementary Figures 33–43)

We compared 11 comorbidities between patients with COVID-19 and those with influenza and found that patients with COVID-19 were more likely to develop diabetes (OR = 1.09, 95% CI: 1.01 –1.17, I2 = 86.2%, p = 0.028). Patients with COVID-19 were less likely to have most of the comorbidities than patients with influenza (heart failure (OR = 0.63, 95% CI: 0.53–0.74, I2 = 96.6%, p < 0.001), coronary heart disease (OR = 0.59, 95% CI: 0.76–0.99, I2 = 0.0%, p = 0.039), stroke (OR = 0.82, 95% CI: 0.69–0.96, I2 = 76.7%, p = 0.015), cancer (OR = 0.62, 95% CI: 0.51–0.75, I2 = 94.0%, p < 0.001), chronic renal failure (OR = 0.70, 95% CI: 0.51–0.98, I2 = 70.3%, p = 0.039), chronic liver disease (OR = 0.62, 95% CI: 0.39–0.98, I2 = 48.0%, p = 0.043), COPD (OR = 0.42, 95% CI: 0.36–0.49, I2 = 82.9%, p < 0.001) and immunodepression (OR = 0.43, 95% CI: 0.33–046, I2 = 93.8%, p < 0.001)); however, there was no significant difference in the prevalence of hypertension (OR = 0.97, 95% CI: 0.89–1.07, I2 = 92.2%, p = 0.571) or asthma (OR = 0.75, 95% CI: 0.55–1.01, I2 = 75.4%, p = 0.058).

Reporting biases

Funnel plots and Egger’s test were used for reporting bias analysis, and we did not find reporting bias in most of the studies (Supplementary Figures 44–131).

Heterogeneity

Some outcomes in our study might have substantial or considerable heterogeneity, and we attempted to perform a subgroup analysis by using the region or influenza subtype as the basis for classification, but unfortunately, we did not find an accurate source of heterogeneity.

Discussion

The COVID-19 pandemic has been the worst global health crisis since the 1918 H1N1 influenza pandemic, and its impact on the global health care system is much greater than that of influenza25. The characteristics of this new respiratory disease are inevitably compared to seasonal influenza because of their similar symptoms and infectious nature. Angiotensin-converting enzyme 2 (ACE2), a functional receptor of SARS-CoV-2, is crucial for the fusion of virus and cell membranes, and its expression and activity directly mediate SARS-CoV-2 infection. ACE2 is highly expressed not only in lung alveolar type 2 cells and gland cells but also in the ileum and colon and less so on the surface of nasopharyngeal cells26,27. The influenza virus mainly binds to alpha2,6-linked cell receptors via sialic acid-linked glycoproteins. The distribution of sialic acid on the cell surface is one of the determinants of host tropism, and these receptors are expressed particularly in the respiratory tract, from the nasopharynx and trachea to the bronchi, with the exception of the alveoli28. Therefore, it is not surprising that patients with influenza in our study were more likely to present with upper respiratory symptoms than patients with COVID-19, and COVID-19 patients experienced a greater occurrence of diarrhea.

In earlier reports, COVID-19 was more common among men, and our study further indicated that patients with COVID-19 had a higher proportion of men than patients with influenza, which may be associated with higher levels of ACE2 expression in men29. To minimize the spread of SARS-CoV-2, maintaining social distance is recommended in many areas, which may increase loneliness, depression, and psychological pressure in some people, thereby adversely affecting dietary habits and reducing exercise, which can lead to an increase in BMI. Obese individuals are more susceptible to SARS-CoV-2, and obesity is an independent risk factor for COVID-1930. It is well known that smokers are mostly men31. Although the proportion of men among patients with COVID-19 was higher, we found that the proportion of current smokers among COVID-19 patients was significantly lower than that among influenza patients. Previous studies have noted a lower rate of SARS-CoV-2 infection in countries with a high rate of smoking32. Farsalinos et al. even proposed a controversial idea of nicotine as a therapeutic option for COVID-1933. Nicotine acts in a similar fashion as the naturally occurring neurotransmitter acetylcholine on nicotinic acetylcholine receptors and may have anti-inflammatory effects34. Patients with COVID-19 spent longer in the hospital and ICU than patients with influenza and were more frequently admitted to the ICU, reflecting the fact that COVID-19 patients require more medical resources and have a more serious medical condition. Moreover, a greater use of mechanical ventilation also suggested that SARS-CoV-2 was more likely to cause lower respiratory symptoms than influenza virus. Comparing the mortality of these two groups of patients worldwide is always challenging. The main reason is that influenza deaths are usually based on estimates, whereas for COVID-19, although there is underreporting, deaths are currently reported as direct counts rather than estimates35. We included clinical data from multiple regions worldwide before the outbreak of Omicron and found that mortality among patients with COVID-19 was much higher than that among patients with influenza.

Diabetes leads to a functional immune deficiency and increased ACE2 expression in humans, leading to an increase in susceptibility to SARS-CoV-2 and exacerbation of the disease36. The above may explain why we observed a higher prevalence of diabetes among patients with COVID-19 than among patients with influenza. However, in our findings, patients with influenza tended to have more comorbidities than patients with COVID-19, which seemed to suggest that SARS-CoV-2 was more likely to infect healthy people. This may be because people with comorbidities were more cautious during the COVID-19 pandemic and were more active in wearing masks and avoiding social contact and thus had a lower incidence of COVID-19 compared to influenza. Oronasal entry of respiratory viruses results in their direct infection in the pulmonary system with the opportunity to enter the blood and then come into contact with extrapulmonary organs through the bloodstream, so patients with COVID-19 or influenza often have some abnormalities in blood parameters37. In the results of our meta-analysis, there were significant differences in some laboratory findings between patients with COVID-19 and patients with influenza, but the relative increase or decrease in the indices could not represent the severity of organ injury or disease, which only provided a certain reference for distinguishing the two viral infections. At present, the gold standard for distinguishing the two diseases is still reverse transcription-polymerase chain reaction (RT‒PCR)38.

In the early stage of the COVID-19 pandemic, the global circulation of influenza virus also decreased. This may be due to the following reasons: 1) To curb the spread of SARS-CoV-2, countries adopted various social isolation measures and strictly enforced social distancing, which also prevented the spread of influenza virus39. 2) During the pandemic, doctors prioritized the detection of SARS-CoV-2 and reduced efforts to detect influenza virus40. 3) SARS-CoV-2 infection may provide some cross-protection against influenza through nonspecific mechanisms, which stimulate antiviral defenses or trigger an adaptive response to secondary infection with influenza41. However, specific policies to control the COVID-19 epidemic, such as banning social gatherings and travel restrictions, are unlikely to be implemented every year. During the COVID-19 epidemic, low levels of influenza activity may have reduced people’s immunity, and cross-reactive antibodies against influenza infection often expire within one year42.

In the future, COVID-19 is likely to become a seasonal epidemic similar to influenza43. It is also very important to prevent the coinfection of these two diseases because this increases the severity of infection. A study by Public Health England showed that the risk of death of patients infected with both diseases was approximately two times higher than that of patients with COVID-19 alone44. Adherence to vaccination against both viruses is the most effective means of reducing morbidity, mortality and economic impact. Although SARS-CoV-2 infection may occur in a fully vaccinated population, vaccinated individuals are still much less likely to become infected, die, or transmit the disease to others than unvaccinated individuals45.

Limitations

This meta-analysis had the following limitations. Most of the articles included in this study were retrospective, and of the 100 articles, only 1822,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62 explicitly noted that the sample did not contain patients with SARS-CoV-2 and influenza virus coinfection, and the coinfection was not described in detail in other articles, so we were unable to determine the absence of coinfection in all samples. Because we could not extract enough data from the included studies, we could not perform subgroup analysis according to SARS-CoV-2 variant or vaccination status of COVID-19 patients. We did not include the data of Omicron patients in this study, and intend to focus on this field in future studies.

Conclusions

There are some differences in the major characteristics, symptoms, laboratory findings and comorbidities between COVID-19 patients and influenza patients. COVID-19 patients often require more medical resources and have worse clinical outcomes.