During the coronavirus disease 2019 (COVID-19) pandemic, many clinicians and infection control programs began incorporating severe acute respiratory coronavirus virus 2 (SARS-CoV-2) cycle threshold (Ct) measures into their treatment and isolation decision-making, using them as semiquantitive proxies for viral burden. Reference Rao, Manissero, Steele and Pareja1–Reference Rhee, Baker and Klompas4 The Ct value indicates the number of polymerase chain reaction (PCR) amplification cycles required to detect a target nucleic acid sequence; thus, Ct value is inversely proportional to the quantity of virus. The Ct value is not a perfect measure of viral burden; it can vary widely depending upon sample site, sample quality, assay type, assay sensitivity and may not have a linear relationship with analyte quantity. Reference Jamal, Farooq and Bidari5–Reference Wolfel, Corman and Guggemos7 Nonetheless, it gives an approximate, semiquantitative measure of viral burden that can help clinicians differentiate between early versus late infections, particularly when serial samples with sequential Ct values are available, and it may provide an approximate measure of a patient’s likelihood of responding to antiviral treatment as well as their potential contagiousness. Reference Singanayagam, Patel and Charlett8–Reference Rhee, Kanjilal, Baker and Klompas10 SARS-CoV-2 Ct values of <25–32 are typically interpreted as indicating active viral replication and potential contagiousness, based on studies of viral culture and secondary infection rates. Reference Singanayagam, Patel and Charlett8,Reference Cohen, Kleynhans and Moyes11,Reference Eyre, Futschik and Tunkel12
Measures of viral burden have the potential to also inform the management of patients with non–SARS-CoV-2 respiratory viral infections, which are also commonly diagnosed using molecular assays. However, few data are available regarding viral replication dynamics for influenza, RSV, and other non–SARS-CoV-2 pathogens, much less how they correlate with stage of infection or contagiousness. Thus, we conducted a retrospective analysis to describe temporal patterns in Ct values for non–SARS-CoV-2 respiratory viruses relative to the patient’s day of symptom onset. We anticipate that these viral dynamics will provide useful background information to clinicians and researchers to build familiarity with Ct dynamics for non–SAR-CoV-2 viruses and that these insights may inform future studies on how Ct values might be incorporated into clinical decision making regarding patient treatments and duration of isolation.
Methods
We conducted a retrospective observational study at Brigham and Women’s Hospital, an academic hospital with 803 beds in Boston, Massachusetts. We included hospitalized patients and nonhospitalized emergency room patients aged ≥18 years with at least 1 positive PCR assay for influenza (A and B combined), RSV, human metapneumovirus (HMPV), parainfluenza (types 1–4 combined), rhinovirus, or adenovirus between January 2022 and March 2023 performed on nasopharyngeal swabs or aspirates using the following assays: Cepheid Xpert Xpress SARS-CoV-2/Flu/RSV (Cepheid, Sunnyvale, CA), Hologic Panther Fusion AdV/HMPV/RV (Hologic, Marlborough, MA), Hologic Panther Fusion Flu A/B/RSV and the Hologic Panther Fusion Paraflu Assay. The cutoff Ct value for test positivity varied by assay but was generally between 40 and 45. We manually reviewed each patient’s clinical records for date of symptom onset and for the following comorbidities: immunosuppression (drug induced, malignancy related, or underlying immune disorder), cardiac disease (including coronary heart disease, but not isolated hypertension), structural lung disease, liver disease, renal disease, or diabetes. We excluded asymptomatic cases and patients with an unclear date of symptom onset (typically due to cognitive impairment or conflicting statements between provider notes). We converted approximate terms into specific time frames: we equated “a couple of days” with 2 days, “few days” with 3 days, and “several days” with 4 days. If a test was run multiple times on the same sample, we used the lowest available Ct value.
We characterized distributions of Ct values relative to days since symptom onset for each respiratory virus. Median intervals from symptom onset until test positivity for patients with Ct values ≥30 were compared between immunocompromised versus immunocompetent patients for each virus using the Mann-Whitney U test. Analyses were performed in R version 4.2.2 software. 13 The study was approved with a waiver of informed consent by the Mass General Brigham Institutional Review Board.
Results
Patient characteristics
There were 1,863 positive results during the study period: 506 influenza, 430 RSV, 502 rhinovirus, 180 parainfluenza, 219 HMPV, and 26 adenovirus cases. The mean age of the patients was 54.5 years (range, 18–99), and 61.3% were female. Intervals from symptom onset until positive test ranged from 0 to 77 days (median, 4 days; interquartile range [IQR], 2–7). We restricted subsequent analyses to patients whose symptoms began within the 21 days preceding their positive testing: 501 of 506 influenza cases, 416 of 430 RSV cases, 475 of 502 rhinovirus cases, 175 of 180 parainfluenza cases, and 209 of 219 HMPV cases. Adenovirus cases were excluded from further analysis due to small numbers (23 of 26 cases).
Ct values in relation to time since symptom onset
Figure 1 illustrates the distribution of Ct values relative to time since symptom onset by virus. For influenza, the median Ct value was 29.8 on the day of symptom onset, dropped to 22.7 on days 1-3, and then increased thereafter, reaching median values >30 by day 8. A similar pattern was evident for RSV: the median Ct value was 32.3 on day of symptom onset, 23.9 on days 1-3, and then typically >30 by day 10. We also detected a similar but less pronounced pattern for HMPV and a slow steady rise in Ct for parainfluenza from a median of 25.1 on day 0 to median values ≥30 by day 6. Ct values tended to be higher and relatively stable throughout the first week after symptom onset for rhinovirus. The median number of days since symptom onset for patients with Ct values of ≥25, ≥30, and ≥35 per virus are shown in Table 1. Individual Ct value trajectories of 3 patients demonstrating some of the variability in individual Ct-value trajectories are described in more detail in Supplementary Fig. 1 (online).
Figure 1. Grouped box plots representing the distribution of Ct values relative to symptom onset for each virus. The boxes represent the interquartile range (difference between first and third quartile, IQR), and the lower and upper limits of the whiskers represent the maximal and minimal value (within a range of 1.5 times the IQR). Dots outside the boxes represent outliers beyond 1.5 times the IQR. The number within each box is the mean Ct value per group. Note. Ct, cycle threshold; IQR, interquartile range.
Table 1. Median Number of Days Since Symptom Onset Among Patients With Ct Values of ≥25, ≥30, and ≥35 With Symptom Onset Within the Preceding 2–21 Days
Note. Ct, cycle threshold; IQR, interquartile range; RSV, respiratory syncytial virus; HMPV, human metapneumovirus.
Association between immune status and Ct values
Immunosuppression rates varied from 10.4% for patients with positive influenza tests, 22.1% for patients with positive RSV tests, 52.6% for rhinovirus, 53.1% for parainfluenza, and 56.9% for HMPV. Median times since symptom onset among patients with Ct values ≥30 were similar for immunocompromised versus immunocompetent patients: median 7.5 versus 7 days for influenza (P = .27), 7 versus 6.5 days for RSV (P = .73), and 5 vs 4 days for HMPV (P = .67), 5 vs 6 days for parainfluenza (P = .82), and 5 versus 7 days for rhinovirus (P = .26). Figure 2 displays the median Ct values grouped by day since symptom onset according to immune status.
Figure 2. Comparison of median Ct values according to immune status grouped by days since symptom onset. Note. Ct, cycle threshold.
The presence of other comorbidities (heart, lung, liver, kidney disease or diabetes) was not associated with longer duration since symptom onset among patients with Ct values ≥30. Details of the comorbidity analysis are provided in Supplementary Table 1 (online).
Most influenza and RSV tests (∼80%) were performed using the Cepheid platform. On comparison with samples that tested positive using the Panther platform; however, Ct value trajectories and median values by day since symptom onset were similar (Supplementary Fig. 2 online).
Discussion
We have documented respiratory viral replication dynamics for influenza, RSV, and other non–SARS-CoV-2 respiratory viruses, using Ct values as proxy measures for viral burden, from the day of symptom onset through the following 3 weeks. In general, we found that viral burden was intermediate on the day of symptom onset, increased to peak levels (low Ct values) over the next 1–3 days, and then steadily declined thereafter (rising Ct values). Ct values were typically >30 by ∼1 week after symptom onset. We noted substantial variability between patients, however, in Ct values relative to days since symptom onset. We observed generally higher Ct values (less virus) and less day-to-day variability for HMPV and parainfluenza, and we observed generally high and steady Ct values for rhinovirus. We did not find a clear association between immunosuppression and viral dynamics, although our study had limited statistical power for subgroup analyses.
The general pattern of low Ct values near symptom onset followed by a steady increase over the ensuing days with substantial patient-to-patient variability broadly mirrors the viral dynamics seen with the SARS-CoV-2 omicron variant. With the original SARS-CoV-2 strain, Ct values tended to peak around the time of symptom onset or even slightly before. Reference He, Lau and Wu14 However, the pattern right-shifted with the omicron variant; viral burden now tends to peak 1–3 days after symptom onset, Reference Zhou, Hu and Zhao15,Reference Frediani, Parsons and McLendon16 which is consistent with our findings.
The potential parallels between Ct dynamics for SARS-CoV-2 and other respiratory viruses is helpful insofar as many clinicians and infection control practitioners have learned how to integrate Ct values into their decision making for patients with positive SARS-CoV-2 tests with regard to treatment (eg, does this patient require an antiviral medication such as remdesivir?) and necessity and duration of isolation (eg, does this patient have an acute infection or residual viral RNA alone from a prior resolved infection? does this patient need to remain on precautions or have they reached a less contagious point and isolation can be lifted?). Ct values do need to be interpreted cautiously; they are proxy measures alone, prone to intersample and interassay variability. They are best interpreted in conjunction with patients’ prior histories, clinical trajectories, and serial Ct values. Nonetheless, notwithstanding these caveats, Ct values have generally provided helpful collateral data to inform the care and management of patients with SARS-CoV-2–positive tests.
Our findings suggest the possibility that Ct values may prove similarly helpful in the interpretation of positive PCRs for influenza, RSV, and other respiratory viruses. Further data and experience are needed, however, to aid in their evaluation. The interpretation of SARS-CoV-2 Ct values are bolstered by viral culture studies correlating Ct values with the probability of recovering culture-viable virus. Reference Wolfel, Corman and Guggemos7,Reference Singanayagam, Patel and Charlett8 They are further aided by studies documenting correlation between patients’ Ct values and the frequency of secondary transmission. Reference Eyre, Futschik and Tunkel12,Reference Bjorkman, Saldi and Lasda17 Few published data are available regarding viral dynamics for non–SARS-CoV-2 respiratory viruses, and few data are available that correlate non–SARS-CoV-2 respiratory virus Ct values with viral culture. Reference Cohen, Kleynhans and Moyes
The limited existing data on influenza, RSV, and rhinovirus do, however, mirror our findings. Influenza A Ct values tend to be lower with less time since symptom onset, Reference Spencer, Chung and Thompson18,Reference Brittain-Long, Westin, Olofsson, Lindh and Andersson19 and titers of infectious virus tend to peak 24–72 hours after symptom onset and remain elevated until about day 5 after symptom onset. Reference Frediani, Parsons and McLendon16,Reference Bell, Nicoll and Fukuda20 A small viral culture-based study showed a 1 log decrease in the viral load of H1N1 influenza A at 3 days after symptom onset. Reference To, Chan and Li21 Other studies have suggested that H1N1 influenza A levels may be highest on the day of symptom onset rather than the day after, as in our study. Reference To, Chan and Li21,Reference Duchamp, Casalegno and Gillet22
To our knowledge, only 1 study has correlated non–SARS-CoV-2 Ct values with probability of transmission, and this study showed that household transmission of influenza was correlated with Ct values <30. Indeed, investigators in this study found that household transmission of influenza was 7-fold higher from people with Ct values <30 versus ≥30. Reference Cohen, Kleynhans and Moyes11 Fuller et al Reference Fuller, Njenga and Bigogo23 did not find a correlation between days since symptom onset and Ct values in a cohort of RSV-infected adults in Kenya, but they did find a correlation among infected children.
Others have observed the relatively weak association between symptom duration and Ct values for rhinovirus. Generally high and steady Ct values suggest that RNA of the virus is present but that the virus is not replicating, possibly indicating colonization rather than acute infection. Others have hypothesized that this finding may be due to the genetic variability of the rhinovirus to increase sample–target mismatch, which diminishes rhinovirus detection. Reference Brittain-Long, Westin, Olofsson, Lindh and Andersson19 The interpretations of all these findings, however, are limited by small sample sizes and substantial patient-to-patient variability.
The high rate of variability between patients, across time, between assays, and between viruses suggest the importance of measuring serial Ct values in patients to clarify likely stage of infection and current viral burden rather than relying on time alone to estimate these parameters.
Our study had several limitations. We used patients’ self-reported intervals since symptom onset garnered through retrospective chart reviews. These may have been subject to recall bias, imprecision, and documentation errors. We excluded asymptomatic patients, which precluded us from describing viral burden before symptom onset, a period that is increasingly recognized as an important contributor to transmission. Reference Cohen, Kleynhans and Moyes11,Reference Huff and Singh24 We created rules to infer concrete time intervals from nonspecific language. We drew Ct values from different molecular assays, which may vary in their amplification characteristics and yield different Ct values for the same sample. We did not detect significant differences between the Cepheid and Panther platforms in a limited set of samples. Patients were not serially sampled in a prospective fashion; rather, we inferred Ct value dynamics based on distributions of Ct values in different patients who tested positive at various time intervals relative to symptom onset. As such, some of the differences in Ct values across different time intervals may have been affected by patient-level variations in viral dynamics. Additionally, we were not able to control for possible confounders, such as personnel collecting the specimen, specimen type, and the adequacy of sampling. Our sample sizes for different intervals varied and in some cases were small, particularly for the day of symptom onset, which increased the variability and uncertainty of our estimates. The immunocompromised patient population was heterogeneous; respiratory virus dynamics among more severely immunocompromised patients may differ more clearly from nonimmunocompromised patients. Finally, we assessed each virus independently, and we presumed that patient symptoms were due to the virus under evaluation without accounting for coinfections or other diagnoses, which may have led to misattribution of some symptoms.
We documented distributions of Ct values as a proxy for respiratory viral burden for influenza, RSV, and other non–SARS-CoV-2 respiratory viruses relative to date of symptom onset. Patterns generally mirrored the dynamics of SARS-CoV-2. Ct values were intermediate on the day of symptom onset, lower on the subsequent few days, and then higher again, generally reaching Ct values >30 after about a week from symptom onset. Results varied widely between patients and between viruses, however, suggesting the importance of measuring serial Ct values to estimate viral burden rather than presuming viral burden based on time since symptom onset alone. These results suggest that Ct values in particular, and quantitative tests in general, may provide valuable information to aid in understanding phase of illness and potential contagiousness for non–SARS-CoV-2 respiratory viruses just as they have for SARS-CoV-2. More data are needed, however, to correlate non–SARS-CoV-2 respiratory virus Ct values with viral culture, contagiousness, and response to therapy.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/ice.2023.286
Acknowledgments
Financial support
No specific funding was obtained for this study. S.E. received nondirected funding from the Swiss Study Foundation for a research stay in the United States.
Competing interests
M.K. and C.R. report grant funding from CDC, AHRQ, and royalties from UpToDate. All other authors report no conflicts of interest relevant to this article.
