Anthropometric data and distribution of infections

Of 1065 patients of all departments including those with re-admission, 932 patients remained when selecting those from the internal medicine, neurology and pediatric units, who were not admitted a second time with the same virus within 4 weeks. These patients comprised 446 women (47.9%) and 486 men (52.1%). The median values (quartiles) of age and BMI were 68 (3; 81) years and 25.7 (22.6, 29.8) kg/m2, respectively. Table 1 provides additional data on the distribution of anthropometric indices across infection groups. Most patients (52.9%) showed infection with SARS-CoV-2, while Influenza A and RSV each accounted for about the half of the remaining cases (23.2% and 21.8%, respectively). Patients with Influenza B or with combined infections were rare (see Table 1), therefore the subsequent statistical comparisons comprised only the 912 patients with single infections with SARS-CoV-2, Influenza A, or RSV.

Table 1 Baseline characteristics of all patients

In these 912 patients, sex did not show statistically significant differences between the three infection groups. BMI, which was available only in adults, also was not significantly different. Age, however, differed (p < 0.001), whereby pairwise comparisons showed that all three groups were different from each other (p < 0.001 each); it was lowest in the RSV group. Age distribution is illustrated in Fig. 1, demonstrating the two distinctive peaks at very low and at high age, as observed in 2022/23 [15].

Fig. 1

Distribution of age in hospitalized patients stratified for infection type. Absolute numbers are shown for each age category comprising 5 years. The numbers refer to the whole observation period from August 2023 to February 2024

Similarly, in patients of age ≥ 18 years, age differed between the three infection groups (p < 0.001), but in this case, only between Influenza A and SARS-CoV-2 (p < 0.001), whereas the RSV group was not significantly different from the other two groups. Median values (quartiles) were 78 (70; 84) years for SARS-CoV-2, 73 (60; 82) years for Influenza A, and 78 (57; 84) years for RSV. In patients of age < 18 years (mean age 1.43 years), age also differed between groups (p < 0.001), whereby all three groups were different from each other (p < 0.001 each), with lowest values for SARS-CoV-2 (0.54 years), highest values for Influenza A (3.28 years), and intermediate values for RSV (1.18 years). For the time course of prevalence values in the two age groups, see the Supplemental Figure S1.

Prevalence of symptoms

Self-reported symptoms were recorded only for patients aged ≥ 18 years. The prevalence of cough and dyspnoea showed significant (p < 0.001 each) differences between the three types of infection (Supplemental Table S1). Both showed the lowest values in SARS-CoV-2 and highest values in RSV. Fever also showed a difference between groups (p = 0.004) and high prevalence in patients with SARS-CoV-2. There were no significant differences regarding fatigue, diarrhoea and nausea. When using logistic regression analysis with the symptoms as dependent variable and the additional predictors age and sex, it turned out that significant differences between infections remained (p < 0.05 each) for cough and dyspnoea but not for fever, in which there was, however, a significant dependence on sex (p = 0.017).

Comorbidities

In patients aged ≥ 18 years, there were significant differences between the three groups regarding the frequencies of peripheral artery disease, heart failure, asthma, dementia and state of immunosuppression (p < 0.05 each, Table 2). Heart failure and asthma were most often found in patients with RSV infection, immunosuppression of all kind most often in Influenza A. The sum score of the comorbidities shown in Table 2 did not significantly differ between groups. In order to clarify, to which extent the differences in prevalence were due to the dependence of comorbidities on age, we again performed logistic regression analysis with the comorbidities as outcomes and the type of infection as well as age and sex as predictors. Heart failure depended on age (p < 0.001) and on RSV (p < 0.001), peripheral artery disease on sex and RSV (p < 0.05 each), asthma also on sex and RSV (p < 0.05 each), while dementia was associated with age (p < 0.001) as well as Influenza A (p = 0.016), and this was also true for immunosuppression (p < 0.001 and p = 0.033, respectively). These observations demonstrate that the association between comorbidities and the type of viral infection was not explained by their dependence on sex and age.

Table 2 Prevalence of comorbidities in patients of age ≥ 18 yearsPrimary outcomes

The occurrence of ICU treatment and in-hospital mortality was compared between Influenza A, SARS-CoV-2 or RSV separately in patients of age ≥ 18 years and < 18 years (Table 3). In patients of age ≥ 18 years, neither regarding ICU admission nor regarding in-hospital mortality, the significant differences between the three groups were statistically, despite the fact that values were highest for RSV; however, it has to be considered that this changed in the multivariable analyses (see below). In patients of age < 18 years, case numbers were low, and there was no mortality at all in this age group. When using the exact test according Fisher-Freeman-Halton to account for the low numbers in ICU admissions, a p-value of 0.934 was obtained for comparing the three infection groups.

Table 3 ICU admission and mortality stratified according to age group

To assess whether the start of data collection in August 2023 compared to October 2022 [15] had affected the results, a similar table as Table 3 was established using only data starting in October 2023 (Supplemental Table S2). The results were very similar to those of Table 3, and identical regarding Influenza A and RSV which became prevalent only in November 2023 (see Supplemental Fig. 1).

Treatment characteristics

Table 4 shows results regarding the treatment of patients aged ≥ 18 years. The length of the hospital stay did not differ between groups, and there were also no significant differences regarding the durations of invasive or non-invasive ventilation. However, the frequencies of NIV, low-flow oxygen supply during the hospital stay, and oxygen supply upon admission differed significantly between the three groups (p < 0.05 each). While RSV showed the highest percentages for NIV and low-flow oxygen supply during the hospital stay, it showed the lowest percentage regarding oxygen supply upon admission.

Table 4 Details of clinical treatment in patients of age ≥ 18 years

Table 5 shows analogous data for patients of age < 18 years. The length of the hospital stay differed between the three groups (p < 0.001), with significant differences between RSV versus both SARS-CoV-2 and Influenza A (p < 0.001 each). There were no significant differences regarding ICU admission and invasive or non-invasive ventilation but marked differences regarding oxygen supply. Both high-flow and low-flow supply during the hospital stay, as well as oxygen supply upon admission were highest in the RSV group (p < 0.001 each). As the group of young children was of particular interest, we also analyzed data of patients with age < 3 years separately (Supplemental Table S3). The great majority of these children had RSV infection. Again, in RSV patients, the duration of hospital stay was longest, and this was due to a significant difference of the RSV group versus the other two groups (p < 0.001 each). Moreover, the percentages of all kind of oxygen supply were by far the highest in this group (p < 0.001 each).

Table 5 Details of clinical treatment in patients of age < 18 yearsVital parameters, arterial blood gas and laboratory parameters upon admission

For patients ≥ 18 years, vital parameters upon admission are shown in Supplemental Table S4. Breathing frequency, heart frequency, body temperature, peripheral oxygen saturation and diastolic blood pressure significantly (p < 0.05 each) differed between groups, while systolic blood pressure did not. In addition, there were differences in pH and pCO2 (p < 0.05 each) but not pO2, eGFR, CRP and D-dimer levels. According to the Bonferroni-adjusted post hoc comparisons, SARS-CoV-2 differed from RSV and Influenza A regarding respiratory rate and heart rate, moreover from Influenza A oxygen saturation and diastolic blood pressure, and from RSV regarding pH (p < 0.05 each). Regarding body temperature, RSV was different from the other two groups (p < 0.05 each). For the other parameters, the adjusted pairwise comparisons did not indicate significant differences between specific groups.

Risk factors for ICU admission and in‑hospital death

The data shown in Table 3 suggested that the frequencies of ICU admission and in-hospital mortality were highest in RSV but the unadjusted analyses did not indicate a statistically significant difference. We therefore determined whether the consideration of confounding factors would indicate a higher risk for RSV, especially in patients of age ≥ 18 years. Among the relevant confounding factors could be age, sex, the need for oxygen supply, CRP as marker of inflammation, eGFR, comorbidities (Table 2) and the symptoms listed in Supplemental Table S1.

A logistic regression analysis using as predictors symptoms, age, sex and oxygen supply in addition to indicator variables of infection with Influenza A or RSV (taking SARS-CoV-2 as reference), showed a tendency for increased ICU admission with RSV infection (p = 0.065; OR = 2.4) but not Influenza A (p = 0.173). Among symptoms, only cough was relevant (p = 0.009). In addition, age (p < 0.001) and sex (p = 0.001) were statistically significant, as well as oxygen supply upon admission (p < 0.001).

To assess the reliability of these findings regarding the type of infection, we further added oxygen saturation, CRP levels upon admission and the comorbidities heart failure, peripheral arterial disease and immunosuppression (see Table 2) as predictors. Age, sex, oxygen supply upon admission and cough were confirmed as significant predictors (p < 0.05 each), and in addition fever (p = 0.047) and CRP (p = 0.004). Neither RSV (p = 0.112) nor Influenza A (p = 0.297) were specifically relevant for ICU admission in the presence of these confounders. In these analyses, comorbidities were never significant and thus omitted from the final analysis shown in Fig. 2. The results underline that the type of infection was not a significant predictor for ICU admission if adjusting for multiple patient characteristics, despite a tendency for RSV.

Fig. 2

Results of multiple logistic regression analysis addressing potential risk factors for ICU admission in patients of age ≥ 18 years. Regarding the comparison of the three infections, SARS-CoV-2 was taken as the reference. Odds ratios and 95% confidence intervals are shown. CRP, C-reactive protein, eGFR, glomerular filtration rate estimated from creatinine

The analogous analysis for mortality identified RSV as highly significant risk factor (p = 0.005; OR = 4.4), besides age (p = 0.011), sex (p = 0.007), oxygen upon admission (p < 0.001), cough (p = 0.031) and eGFR (p = 0.045), while Influenza A was not (p = 0.861). Using the extended set of predictors as for ICU admission (see above), age, sex, oxygen upon admission, cough, oxygen saturation and CRP were significant predictors (p < 0.05 each), in addition to RSV (p = 0.005; OR = 4.7), while again Influenza A was not relevant (p = 0.930). Again, comorbidities did not have significant effects in the presence of the other predictors. The results are shown in Fig. 3 and demonstrate that RSV was a significant predictor for elevated mortality after adjustment for multiple patient characteristics, while SARS-CoV-2 and Influenza A were similar in this respect.

Fig. 3

Results of multiple logistic regression analysis addressing potential risk factors for in-hospital death in patients of age ≥ 18 years. Regarding the comparison of the three infections, SARS-CoV-2 was taken as the reference. Odds ratios and 95% confidence intervals are shown. CRP, C-reactive protein, eGFR, glomerular filtration rate estimated from creatinine