Streptococcus pneumoniae remains an important cause of morbidity and mortality among children and adults globally despite the availability of effective pneumococcal conjugate vaccines (PCVs) covering ≥10 S. pneumoniae serotypes (1, 2). S. pneumoniae nasopharyngeal carriage is a necessary precursor of invasive pneumococcal disease (IPD), although disease risk is influenced by a combination of host, pathogen, and environmental factors and in the majority of cases pneumococcal carriage does not result in disease (3).

Respiratory virus infection is one host factor influencing S. pneumoniae carriage and disease risk. Influenza virus infection and risk of increased pneumococcal carriage density and disease has been well-established and investigated in animal (4–6), modeling (7), ecological (8–20), and individual-level epidemiologic (21, 22) studies. Ecologic and epidemiologic studies have also associated infection with respiratory syncytial virus (RSV) and, to a lesser extent, rhinovirus, adenovirus, human metapneumovirus, and human parainfluenza virus with increased pneumococcal disease risk (10–12, 15, 17–20, 23–26).

Increased pneumococcal carriage density has been found to be associated with multiple common respiratory viruses, including rhinovirus, adenovirus, influenza virus, RSV, and common human coronavirus, although viruses and populations investigated varied across studies. Most studies were among children with acute respiratory illness (ARI) or pneumonia (21, 27–30), fewer studies included asymptomatic children (30–32), and one study to our knowledge included adults with ARI (21).

To date, studies describing the association between respiratory viruses and S. pneumoniae carriage have been conducted primarily among children and most have evaluated a small number of respiratory viruses co-detected. We aimed to estimate the association between eight common respiratory viruses and semiquantitative S. pneumoniae carriage density and to describe symptom profiles with and without viral co-detection among individuals of all ages in a community-based surveillance study among households in Seattle, Washington, USA before and during the COVID-19 pandemic.

From November 2019–June 2021 we collected 4630 swabs from 1,700 unique individuals in 437 households. Of these, 346 (8%) were positive for S. pneumoniae and 773 (17%) were positive for any respiratory virus (excluding SARS-CoV-2). Viral detection was greatest during the 2019–early 2020 winter season, declined substantially following COVID-19 pandemic restrictions beginning in March 2020, began to increase during winter 2021, and remained high during the spring and summer months of 2021 through the end of the study in June 2021. Conversely, detection of S. pneumoniae was lower and declined to a lesser degree during the COVID-19 pandemic. Among all swabs collected (excluding swabs from household contacts of SARS-CoV-2 cases, which were collected in the second year of the study only), 30% of swabs were positive for any virus (excluding SARS-CoV-2) and 12% were positive for S. pneumoniae prior to March 2020 vs. 10% and 6%, respectively, from March–December 2020 and 21% and 5%, respectively, from January–June 2021, although these proportions are not adjusted for the change from mid-turbinate to anterior nasal swabs or the age distribution over time (Supplementary Figure S1). Trends in viral and S. pneumoniae detection over the study period were generally similar by age group and the age distribution did not change meaningfully over the study period. One exception was that S. pneumoniae detection was greatest among children and prior to the COVID-19 pandemic, the proportion of swabs with any virus only was similar to the proportion with S. pneumoniae and any virus for children aged <5 years, whereas for older children and adults a lower proportion of pre-COVID-19 pandemic swabs had S. pneumoniae with any virus co-detected relative to any virus only (Supplementary Figure S2). Although we detected SARS-CoV-2 in 35 specimens, none of these had S. pneumoniae and SARS-CoV-2 co-detected.

Subsequent analyses were restricted to 346 swabs from 239 unique individuals in 151 households that tested positive for S. pneumoniae (n = 157 swabs (45%) with and 189 (55%) without one or more viruses detected) (Table 1). The median time between swab collection for individuals with multiple swabs included in the analysis was 42 (range: 1–540 days, interquartile range: 14–103). Approximately half (49%) of participants were male, majority (79%) were White, and most were children (40% children aged <5 years, 31% children aged 5–17 years, and 29% adults aged 18–64 years). Over half (68%) of included households had a child aged <5 years and median household size was 4 individuals (range: 3–7). During the entire study period a median of one swab (range: 1–11) per participant and one swab (1–17) per household were positive for S. pneumoniae. Over half (57%) of S. pneumoniae-positive swabs without viral co-detection were baseline swabs collected at enrollment, relative to 31% of swabs with viral codetection and 19% of swabs with and 12% of swabs without viral co-detection were collected from household contacts of SARS-CoV-2 positive cases.

Among all S. pneumoniae-positive swabs, rhinovirus (n = 87; 25%) was the most frequently detected virus, followed by common human coronavirus [n = 30 (9%)], adenovirus [n = 23 (7%)], influenza A/B virus [n = 8 influenza A and n = 9 influenza B virus (5%)], RSV A/B [n = 9 RSV A and n = 5 RSV B (4%)], human metapneumovirus (n = 6 (2%)), human parainfluenza [n = 5 (1%)], and enterovirus [n = 4 (1%)]. Among swabs with S. pneumoniae and at least one virus, only 17% had ≥2 viruses co-detected (25 with 2 viruses co-detected and 1 with 3 and 4 viruses co-detected each) and these were mostly (69%) among children aged <5 years. The distribution of viruses co-detected with S. pneumoniae varied by age group (Supplementary Figure S3). The proportion of all S. pneumoniae-positive swabs with common human coronavirus detected was greater among adults relative to children and no swabs among adults had S. pneumoniae co-detected with human parainfluenza.

Among symptomatic individuals, days since symptom onset were similar for specimens with and without viral co-detection (median: 3 days, Table 2). Among individuals reporting any symptoms, difficulty breathing, cough, and runny rose were more commonly reported among specimens with viral co-detection (15%, 80% and 93% vs. 8%, 57%, and 51%, respectively) whereas ear pain and headache were less commonly reported among specimens with compared to without viral co-detection (3% and 26% vs. 16% and 41%, respectively). Fatigue, fever, chills or sweats, muscle and/or body aches, and sore throat were similar between specimens with and without viral co-detection (Table 2 and Supplementary Figure S4). Reported symptoms profiles differed by age group (Supplementary Table S1). Cough, fever and runny nose were more commonly reported among symptomatic children aged <5 (75%, 59%, and 90%, respectively) and 5–17 years (77%, 63%, and 71%) compared to adults aged 18–64 years (60%, 36%, and 68%) while difficulty breathing, headache, and muscle and/or body aches were more commonly reported among adults (24%, 68%, and 40%) compared to children aged <5 (12%, 6%, and 10%, respectively) and 5–17 years (6%, 40%, and 23%). Ear pain and fatigue were reported for a similar proportion of symptomatic adults (12% and 76%) and children aged 5–17 years (11% and 74%), relative to younger children (2% and 47%).

Similar proportions of testing events resulted in absence from work or school (43% vs. 49% for swabs with and without viral co-detection, respectively) and medical care seeking (16% vs. 24%) among those reporting any symptoms. Among those eligible and with known influenza vaccination status, 83% of specimens with and 76% of specimens without viral co-detection were from individuals that had received that current season's influenza vaccine at the time of sample collection. Due to the change in swab type during the study, 75% of specimens with S. pneumoniae and any virus detected were mid-turbinate swabs compared to 66% of specimens with only S. pneumoniae detected.

On average, S. pneumoniae Crt values were 2.67 units (95% CI: 1.55, 3.80, P < 0.001) lower (i.e., semiquantitative nasal carriage density was higher) for testing events with compared to without viral co-detection adjusted for age and swab type (Figure 1 and Supplementary Table S2). Of specific viruses evaluated, S. pneumoniae semiquantitative carriage density was higher with viral co-detection for enterovirus, RSV A/B, adenovirus, rhinovirus, and common human coronavirus compared to swabs without viral co-detection (P < 0.01 for each). No significant differences in S. pneumoniae semiquantitative carriage density were detected with or without influenza A/B virus, human parainfluenza and human metapneumovirus or between mid-turbinate and anterior nasal swabs. S. pneumoniae semiquantitative carriage density decreased significantly with age (P < 0.001). When analyses were stratified by age category, S. pneumoniae Crt values were significantly lower (corresponding to higher S. pneumoniae semiquantitative carriage density) with compared to without viral co-detection among children aged <5 years (2.52 units, 95% CI: 0.93, 4.11, P = 0.002) and 5–17 years (2.96 units, 0.90, 5.02, P = 0.005), but not among adults aged 18–64 years (1.08 units, −0.90, 3.05, P = 0.29) (Figure 2 and Supplementary Table S2).

Results were similar when restricting to symptomatic episodes. The only exception was human parainfluenza virus, where S. pneumoniae semiquantitative carriage density was significantly higher with human parainfluenza virus detection for specimens collected from symptomatic individuals (P < 0.001) but not for all specimens (P = 0.25), although among symptomatic individuals, human parainfluenza virus was only detected in one specimen. No difference in S. pneumoniae semiquantitative carriage density was detected by days since symptom onset (Supplementary Figure S5 and Supplementary Table S2). Results were also similar between the period prior to and post-implementation of COVID-19 restrictions and when excluding repeat swabs conducted by participants within 30 days (Supplementary Figures S6–S7 and Supplementary Table S2). No difference in S. pneumoniae semiquantitative carriage density was detected comparing individuals who had vs. had not received the current season influenza vaccine at the time of testing when comparing specimens with and without influenza A/B virus (Supplementary Table S2).

In a study of remote self-collection of nasal swab specimens in families with young children in the household setting, our analysis found that multiple common respiratory viruses were associated with greater concurrent S. pneumoniae semiquantitative nasal carriage density among children but not adults. Results were similar when restricting to specimens from symptomatic individuals. We also found that symptom profiles differed between individuals with specimens with and without viral co-detection, although we are unable to ascribe the etiology of the symptoms for individuals with swabs without any viruses detected. Overall, difficulty breathing, cough, and runny rose were reported more commonly among specimens with viral co-detection whereas ear pain and headache were reported less commonly among specimens with viral co-detection compared to specimens without. Although we tested for SARS-CoV-2, none of the SARS-CoV-2-positive specimens had S. pneumoniae co-detected.

Our findings of greater S. pneumoniae semiquantitative nasal carriage density with any viral co-detection among children are consistent with another US study in healthy children (30) and our findings for specific viruses are broadly consistent with other settings, although most studies evaluated only a few respiratory viruses. In this study, lower S. pneumoniae Crt values (corresponding to higher semiquantitative nasal carriage density) were associated with enterovirus, RSV A/B, adenovirus, rhinovirus, and common human coronavirus; we did not detect a difference for influenza A/B virus, human parainfluenza, or human metapneumovirus. Our findings for rhinovirus, adenovirus, and common human coronavirus are consistent with other studies which have also associated these viruses with increased pneumococcal carriage density (21, 27–29, 31, 32). Findings have been variable for RSV, which was associated with increased pneumococcal carriage density in this study, among hospitalized children in Vietnam (28), and among children with ARI in Kenya (29), but not among healthy children attending daycare in the UK (32). Influenza virus infection is a long-established risk factor for IPD (8, 9, 14, 22) but we did not find influenza A/B virus detection to be significantly associated with lower S. pneumoniae Crt values in this study. This finding for influenza virus is not consistent with the Vietnam study among hospitalized children (28) and the South Africa study among individuals of all ages hospitalized with ARI (21), but is consistent with the UK study of healthy children attending daycare, which found no association with influenza virus (32). Variability across studies, however, may be attributable to differences in study populations (e.g., healthy individuals vs. those with ARI) and sample collection methods (e.g., nasal vs. nasopharyngeal swabs). Similar to the UK study among children attending day care, we did not detect a difference in pneumococcal carriage density for human parainfluenza virus. None of these prior studies evaluated the association with enterovirus, which we found to be associated with higher semiquantative pneumococcal carriage density, or human metapneumovirus, which we did not find to be associated with semiquantative pneumococcal carriage density.

Our study found that viral co-detection was associated with increased S. pneumoniae semiquantitative nasal carriage density among children aged <5 years and 5–17 years, but not among adults aged 18–64 years and that increasing age was associated with lower S. pneumoniae semiquantitative nasal carriage density overall. One other study investigating S. pneumoniae carriage and viral co-detections among individuals of all ages hospitalized with ARI in South Africa found that adults had less pneumococcal nasopharyngeal colonization compared to children overall, but this study did not estimate the association between viral detection and carriage by age group. In addition, carriage and IPD serotype distributions, IPD disease burden, and pneumococcal vaccine uptake differ between the US and South Africa and thus these populations may not be directly comparable (21). The age effect observed in this study may be related to lower detection of S. pneumoniae among adults compared to children as S. pneumoniae carriage density is generally lower among adults and we used nasal rather than more sensitive nasopharyngeal swabs and did not collect oropharyngeal swabs among adults, as recommended by the WHO (41, 42). Alternatively, this may reflect differences in the processes leading to increased pneumococcal carriage between children and adults, a potential impact of a different serotype distribution between children and adults, differences in prior S. pneumoniae infection and vaccination histories, or differences in the distribution of respiratory viruses co-detected. Importantly, although eligible, our analysis did not include any adults aged ≥65 years and this pattern may not be consistent among older adults at higher risk for IPD (43).

This study was conducted over two years from November 2019–June 2021. Similar to other reports, we observed a COVID-19 pandemic nonpharmaceutical interventions-associated decline in respiratory viruses in 2020 and an out-of-season resurgence of respiratory viruses during the spring and summer months of 2021 while pneumococcal carriage declined to a lesser degree over the study period (44–49). This pattern was observed in both Israel and France, where surveillance data showed declines in the incidence of both respiratory viruses, such as RSV and influenza, and IPD following COVID-19 pandemic nonpharmaceutical interventions but no significant change in pneumococcal carriage rates, suggesting that the observed decline in IPD may be attributable to the decline in respiratory viruses rather than a change in pneumococcal carriage rates (44, 45). Our finding that RSV and common human coronavirus were associated with greater concurrent S. pneumoniae semiquantitative nasal carriage density is consistent with this hypothesis that infection with common respiratory viruses may increase risk for pneumococcal disease. However, enterovirus, adenovirus, and rhinovirus, which we found to be associated with greater concurrent S. pneumoniae semiquantitative nasal carriage density in this study, all persisted during the COVID-19 pandemic. Nonpharmaceutical interventions may have had a lesser impact on enterovirus, adenovirus, and rhinovirus due to these being non-enveloped viruses that are environmentally stable, have prolonged shedding, and are often asymptomatic or only mildly symptomatic (50). Therefore, the association between respiratory virus infection and pneumococcal disease risk may be pathogen-specific.

There were several limitations to our study. First, we cannot rule out that differences in pneumococcal carriage density found in our study are due to other underlying differences between individuals, such as genetic, immune, and environmental factors, including pneumococcal vaccination. In a study of children with ARI in Kenya where samples were taken before, during, and after viral infection, RSV and rhinovirus were associated with increased pneumococcal nasopharyngeal carriage density, but this variation was small relative to the variation in carriage density observed across individuals, leading the authors to conclude that viral co-infection may play only a small role in pneumococcal carriage density (29). Our study did not routinely collect more than one specimen per illness episode and thus we are unable to similarly estimate the variation in carriage density between individuals, evaluate changes in pneumococcal carriage density over time, or assess risk for subsequent pneumococcal disease. Second, symptom data were not collected from individuals at baseline or from household contacts of SARS-CoV-2 positive cases and thus we did not collect specimens from any individuals who confirmed not having any ARI symptoms at the time of sample collection. We therefore could not compare differences in the impact of viral co-detection on pneumococcal carriage density between symptomatic, pre-symptomatic, and asymptomatic individuals or between those with viral infection vs. those with viral detection. Our self-reported symptom data may also have underestimated symptoms that are not directly observable among young children. For example, ear pain and headache were infrequently reported among symptomatic children aged <5 years relative to older children. Third, we tested self-collected mid-turbinate and anterior nasal swabs rather than nasopharyngeal swabs, which are the current gold standard for detecting pneumococcal carriage in children, and we did not collect nasopharyngeal in addition to oropharyngeal swabs among adults, the current gold standard (41). Therefore, we likely underestimated detection of S. pneumoniae, such as in low S. pneumoniae density specimens (51). Indeed, our overall positivity rate for S. pneumoniae was very low (8%). In addition, rhinorrhea induced by viruses may have contributed to greater S. pneumoniae density in these relatively superficial samples. Specimen quality may have also been variable due to self-collection, however we have previously shown this method of self-collection to result in a high rate of adequate specimen quality (37). We also relied on S. pneumoniae Crt values as a proxy of semiquantitative carriage density rather than using the more precise gene copies/mL and our assay for detecting S. pneumoniae used proprietary primer sets and therefore data regarding the accuracy of the assay, including in comparison to real world nasopharyngeal samples with culture-based detection, are limited. Fourth, we switched from mid-turbinate to anterior nasal swabs for the second year of the study, although we were able to adjust for swab type in our analyses and did not find a difference in S. pneumoniae Crt values by swab type. Fifth, we tested for ten (and analyzed eight) common respiratory viruses and therefore cannot rule out that some swabs with S. pneumoniae only detected did not also contain a respiratory virus for which we did not test and cannot determine the etiology of symptoms for individuals with swabs without respiratory viruses detected. Some rhinovirus detections may also reflect enteroviruses due to cross-reactivity in our assay. Sixth, due to relatively small sample sizes we conducted stratified analyses for all viruses combined by age group rather than testing for interactions between age and viral co-detection in the main analysis. Seventh, our study population had relatively high rates of influenza vaccination and a very low rate of smoking and may not be generalizable to other populations. Eighth, our study included the period following implementation of nonpharmaceutical interventions to limit transmission of COVID-19, when expected seasonal peaks in respiratory viruses did not occur (50) and overall detection of viruses in the study was low, particularly during this period. Despite this, we did not find a meaningful difference in results when analyses were stratified by pre- and post-implementation of COVID-19 restrictions. Finally, S. pneumoniae positive specimens were not serotyped and we did not collect data on pneumococcal vaccination. Therefore, we could not assess the potential role of pneumococcal vaccination or specific serotypes on the association between viral co-detection and pneumococcal carriage density, an important area for continued study as some S. pneumoniae serotypes have higher invasiveness potential (52–56) and virus-S. pneumoniae associations may be serotype-specific (4, 18, 28, 57).

Despite these limitations, this longitudinal, community surveillance study included both children and adults aged <65 years of age, allowing us to evaluate the association between viral co-detection and pneumococcal carriage by age, which to our knowledge has not been previously studied for most respiratory viruses included in this study. In addition, our study was conducted in a household setting whereas the majority of prior studies evaluating the relationship between S. pneumoniae carriage and viral co-detection have been among hospitalized individuals. We also included data from the pre-COVID-19 period, during the COVID-19 pandemic, and during the out-of-season resurgence of respiratory viruses following the lifting of pandemic restrictions. Our data included eight common respiratory pathogens allowing for virus-specific estimates of the association between virus co-detection and pneumococcal carriage. Finally, we collected detailed data on symptom profiles, allowing comparisons of swabs with and without viral co-detection that is not possible for many studies relying on routinely collected surveillance and laboratory data.

In conclusion, several common respiratory viruses were associated with greater concurrent S. pneumoniae semiquantitative nasal carriage density among children but not adults before and during the COVID-19 pandemic in a community-based household setting. These findings suggest that some common respiratory viruses may increase risk of subsequent pneumococcal disease in children. The potential impact of vaccines against respiratory viruses, including influenza vaccines and future RSV vaccines, for preventing pneumococcal disease should be explored.