Viral load dynamics and shedding kinetics of mpox infection: a systematic review and meta-analysis

Abstract

Background

Viral load dynamics and shedding kinetics are critical factors for studying infectious diseases. However, evidence on the viral dynamics of mpox remains limited and inconclusive. Thus, we aimed to provide a comprehensive understanding of the viral load and viability of the re-emerged mpox virus since 2022.

Methods

For this systematic review and meta-analysis, we searched PubMed/MEDLINE, Embase and Google Scholar for published articles that are related to mpox viral dynamics up to April 2023.

Results

From 19 studies, 880 samples and 1477 specimens were collected. The pooled median Ct values appeared in the following order: skin lesion [Ct value 21.7 (IQR 17.8–25.5)], anorectal [22.3 (16.9–27.6)], saliva [25.9 (22.5–31.1)], oral [29.0 (24.5–32.8)], semen [29.6 (25.9–33.4)], urine [30.5 (24.6–36.4)], pharyngeal [31.9 (26.5–37.3)], urethra [33.0 (28.0–35.0)] and blood [33.2 (30.4–36.1)]. People living with human immunodeficiency virus (HIV) have a lower Ct value in the skin [skin HIV+, 19.2 (18.3–20.0) vs skin HIV−, 25.4 (21.2–29.0)]. From the Ct values and test day since symptom onset, we identified temporal trends of viral load for each specimen type. Changes in the trend were observed at 4 days in saliva, 5 days in blood, 6 days in skin, 7 days in anorectal, urine, semen and pharyngeal and 8 days in the urethra. We determined optimal Ct cutoff values for anorectal (34.0), saliva (27.7) and urethra (33.0) specimens, where a Ct value above each cutoff suggests minimal viral viability. Using these cutoff values, we derived the duration of viable viral isolation in each specific specimen type (anorectal 19 days, saliva 14 days and urethra 14 days).

Conclusion

Skin lesion, anorectal and saliva samples contained the highest viral load. The peak viral load manifests within 4–8 days after symptom onset, and viable virus detection was presumed to cease within 14–19 days from symptom onset in anorectal, saliva and urethral samples.

Introduction

Although initially recognized as an endemic disease limited to certain areas of Africa,^1,2 the global outbreaks of mpox have been unprecedented since 2022.^3,4 This outbreak exhibits unique characteristics in terms of severity, sex distribution and symptoms, including a notable transmission route through sexual contact among men who have sex with men (MSM).⁵ Based on recent genomic analysis, the current outbreak is more likely caused by superspreading events with a single infection source, wherein infected individuals had extensive contacts within a short period, combined with global travel patterns.^6–8 The low genomic variability of circulating strains supports this hypothesis.^6–8

Several studies have investigated the viral dynamics of mpox. Quantitative polymerase chain reaction (PCR) was commonly used to measure viral loads in various specimens, and some studies also employed viral culture to confirm viability.⁹ The patterns and duration of viral clearance were assessed. Although viral load or culture positivity does not directly correlate with mpox infectivity,¹⁰ these factors are considered as indicators when estimating viral infectivity.^11,12

While advancements have been made in developing more sensitive and specific assays for measuring viral load, previous studies have been limited in their focus on specific sites and small sample sizes. Therefore, in this systematic review and meta-analysis, we aimed to provide a comprehensive understanding of the viral load, viability, transmission and shedding dynamics of the mpox virus. By examining specific site characteristics and viral shedding patterns, we proposed a plausible mechanism and suggested reference values for the development of policies aimed at preventing and controlling the outbreak of mpox.

Methods

For this systematic review and meta-analysis, articles on PubMed/MEDLINE, Embase and Google Scholar were searched using the keywords ((“monkeypox” or “MPX” or “MPOX”) and (“DNA” or (“viral” and (“load” or “shedding” or “dynamics” or “kinetics”)))) up to 21 April 2023. This study included cohort studies, cross-sectional surveys, case series and case reports that depicted the clinical features of mpox and the viral load for each specimen type. The selection process of the studies included in our systematic review and meta-analysis is illustrated in the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) 2020 diagram.¹³ Our systematic review and meta-analysis protocol were registered with PROSPERO (registration number CRD42023421311).

Literature search strategy

Two independent researchers (HK and WC) conducted separate searches and reviews of the titles, abstracts and full texts of published studies in accordance with the PRISMA 2020 guidelines.¹³ Another researcher (DKY) participated and confirmed any disagreements in screening.

We evaluated all the existing related observational studies, including retrospective or prospective cohort studies, cross-sectional studies, case series and case reports. Duplicate records, review articles, meta-analyses and studies that omitted the Ct value or reported it in a different unit were excluded. The study selection process according to the PRISMA 2020 guidelines is summarized in Figure S1.

Eligibility checkpoints

This systematic review and meta-analysis collected and analysed cohort studies, cross-sectional studies and case-series/reports that included the following characteristics: age, sex, men who had MSM, history of human immunodeficiency virus (HIV) infection, smallpox vaccination status and viral load of mpox virus in different types of specimens. Specimens were collected from the skin, anus and rectum, saliva, mouth, urine, semen, pharynx (oropharynx or nasopharynx), urethra and blood.

Mpox viral load was inferred from the PCR cycle threshold (Ct) value, which indicates the number of PCR cycles required to obtain a positive result. The Ct value is approximately inversely proportional to the quantity of viral DNA. Generally, a lower Ct value indicates a higher viral load, as there is a higher quantity of viral DNA in the specimen. We referred to the trend in viral load as changes in Ct values over a period of time and analysed the daily Ct value of the subjects from the onset of symptoms to track the temporal trend of viral load in each type of specimen.

Data extraction and analysis

We extracted the following data from each of the collected studies: first author, publication year, country, sample size, study date, study design (cohort, cross-sectional, case series and case report), characteristics (age, sex, MSM, history of HIV infection, smallpox vaccination status, assay type and DNA extraction tools), test day since symptom onset and mpox viral load.

The median and interquartile ranges were recalculated by combining the Ct values and days since symptom onset for each specimen in the included studies.¹⁴ To identify non-linear trends in Ct values over time, we employed local polynomial regression, a non-parametric statistical method.¹⁵ This technique involves estimating the relationship between the variables, Ct values and days since symptom onset by fitting separate polynomial functions to local cluster of data points. The size of each cluster or neighbourhood, was determined by a chosen window width or fraction of the data, which dictated the number of nearby data points used in the estimation process. In our study, we selected a fraction of two-thirds. Subsequently, we conducted weighted linear least squares regression on each subset, systematically moving through the data point by point. This iterative process was repeated until each individual data point had been accounted for.

We applied a mathematical method, called an optimization algorithm technique, to track the optimal Ct value to classify viability from the Ct values of specimens with confirmed viral culture. We iteratively scanned the Ct value as a variable and found the point where the accuracy in the confusion matrix was maximized. This specific Ct value was then defined as the Ct cut-off value. Additionally, we examined the receiver operating characteristic curve (ROC) curve to observe how sensitivity and specificity values changed with varying Ct values. The area under the ROC curve (AUC) was calculated for providing an indication of the model’s ability to differentiate between negative and positive groups.

We defined the day of ending isolation as the point at which virus viability was stochastically low at the intersection of the Ct cut-off value and the Ct value curve over time. We assumed that this marked reduction in viral viability corresponds to a notable decrease in the probability of transmitting the virus to others.

All statistical analyses and generated figures were performed using Excel (version 2021; Microsoft, Redmond, WA, USA) and Python 3 (version 3.10.6; Python Software Foundation, Wilmington, DE, USA), with Pandas (version 1.5.1), Statsmodels (version 0.14.0), Scikit-learn (version 1.0.2), Seaborn (version 0.12.1) and Matplotlib (version 3.6.0).

Assessment of the methodological quality of included studies

For each cross-sectional and cohort study, the adapted Newcastle–Ottawa scale was utilized for quality assessment in Table S1.⁵ The Joanna Briggs quality evaluation instrument was reviewed for each case report and series in Table S2.⁵

Patient and public involvement

No patients were engaged in establishing the research question or outcome measures, nor in developing plans for the study’s design or implementation. No patients were asked to provide feedback on the interpretation or writing of the results. Nonetheless, if requested, we intend to convey the findings of the study to relevant groups.

Results

Study characteristics

Our reference studies had three study designs: 68.4% case-series/report (13/19), 26.3% cross-sectional studies (5/19) and 5.3% cohort study (1/19). The studies were conducted in eight countries (Italy, Spain, Israel, France, Australia, Belgium, USA and India) across four continents (Europe, Asia, Australia and North America) all within or during 2022 (Tables 1 and 2). All studies used real-time PCR (RT-PCR) as a mean of viral load assay.

The 19 reference studies included 880 samples and 1477 specimens, participants consisted of 559 males (63.5%), 11 females (1.3%) and 310 unknown (35.2%). A total of 441 participants (94.0%) out of 469 respondents were MSM, 192 participants (37.8%) out of 508 respondents were living with HIV and 64 participants (18.1%) out of 353 respondents had a history of smallpox vaccination. The median age of the participants was 36.1 years [standard deviation (SD) 2.1 years]. Their clinical presentations are depicted in the table, with rash being the most common (95.8%) symptom (Table 2).

Key findings

We presented the mpox viral load and viability of the samples from case reports, classified by specimen types. As the lower Ct value was related to higher viral DNA in the specimen, the pooled median Ct values and test day since symptom onset appeared in the following order in the individuals: skin lesion [Ct value 21.7 (IQR 17.8–25.5); test day since symptom onset 6.0 (IQR 4.0–10.0)], anorectal [22.3 (16.9–27.6); 7.0 (5.0–12.0)], saliva [25.9 (22.5–31.1); 6.0 (4.0–8.0)], oral [29.0 (24.5–32.8); 3.5 (3.3–3.8)], semen [29.6 (25.9–33.4); 8.0 (6.0–12.8)], urine [30.5 (24.6–36.4); 6.0 (4.0–10.0)], pharyngeal [31.9 (26.5–37.3); 6.0 (4.0–9.0)], urethra [33.0 (28.0–35.0); 8.0 (6.5–13.0)] and blood [33.2 (30.4–36.1); 5.0 (4.0–8.0)] (Table 3).

The skin lesion, anorectal and saliva contained the most viral DNA and the specimens taken from the pharyngeal, urethra and blood had relatively low levels of viral DNA. Among samples with information on viability, samples from anorectal (27 out of 37) and saliva (23 out of 40) had a higher yield of viable virus. By contrast, the virus contained in pharyngeal samples (4 out of 53) was mostly non-viable (Table 3 and Figure 1).

From the Ct values and test day since symptom onset, we found temporal trends of viral load for each specimen. The changes in the trend of viral load were observed over a period of 6 days in skin, 7 days in anorectal, 4 days in saliva, 7 days in urine, 7 days in semen, 7 days in pharyngeal, 8 days in urethra and 5 days in blood (Figure S2 to S4). No significant differences were observed in the temporal trends of any specimen across different age groups (Table S3 and Figure S5). A comparison of viral shedding between HIV-positive and HIV-negative individuals across various specimens revealed notable disparity only in the skin samples (Table S4 and Figure S6). Skin samples from people with HIV showed significantly low Ct value [19.2 (18.3–20.0)] compared to those from people without HIV [25.4 (21.2–29.0)].

We established the optimal Ct cutoff values for anorectal (34.0), saliva (27.7) and urethra samples (33.0) based on the relationship between Ct values and viral culture. Samples with Ct values exceeding these cutoffs were deemed unlikely to yield successful culture results. Subsequently, we calculated the duration of isolation required for each specimen type, resulting in 19.0 days for anorectal samples, 14.0 days for saliva samples and 14.0 days for urethra samples.

Different types of specimens collected showed different trends in viral load. In the curve of skin samples, the Ct value remained below 20 for ~6 days before the curve gradually declined to around 30 by Day 20. Conversely, the anorectal curve exhibited a lower overall viral load compared to skin samples and demonstrated a decrease after reaching its peak on Day 7. The saliva curve maintained a relatively constant Ct value of around 25, with a small peak observed on Day 4. Both urine and semen samples showed peak Ct values of around 28 on Day 7. The pharyngeal and urethra curves remained relatively stable, ranging between Ct values of 30–35, with minimal changes until Day 20. The blood curve followed a similar Ct value range but was plotted up to ~10 days (Figure 2).

Discussion

Findings of our study

Our results showed that mpox viral load was found to be high in skin lesion, anorectal and saliva and low in semen, urine, pharynx, urethra and blood, in decreasing order. While there are limited data on viability in skin lesion samples, a substantial portion of samples from both the anorectal and saliva sources were identified to contain viable viruses.

Consistent with our analysis, previous reports have identified skin, anorectum and saliva as the primary sources of mpox transmission. Given that anogenital lesions were reported to be the most common location of mucocutaneous lesions^16,17 and that many isolated lesions presented at the exact site of sexual contact,¹⁸ these findings strengthen the idea that local contact, encompassing intimate contact in sexual intercourse, anal or oral sex and kissing, was the main mode of transmission in the 2022 outbreak.

No case or evidence supporting the transmission through urine, semen, urethra or blood has been reported thus far.¹⁹ Nevertheless, the potential for sufficient infection through urine or semen remains plausible. There were research findings of viable virus isolation in seminal fluid,²⁰ and data of viable virus isolated from urine samples were also included in our data.

Notably, the diagnostic value of the pharyngeal specimen seemed to be lower than previously understood. The median Ct value for oropharyngeal specimens was 32.4, which is close to the minimum of detection, considering that the US Centers for Disease Control and Prevention recommend immediate re-extraction and re-testing for mpox specimen samples with a Ct value of ~34 or above.²¹ In addition, among 53 oropharyngeal specimens with information on viability, 49 (92.5%) were found to be non-viable. These results suggest that oropharyngeal specimens may contribute little to diagnostic accuracy and raise the need for a revision of the WHO recommendation that oropharyngeal swabs be taken as an additional test specimen along with skin lesion swabs.²² Blood samples showed the highest Ct values, which was consistent with the WHO recommendation that blood samples were not suitable for diagnosing mpox.

We outlined the temporal profile of viral load across different specimen types using local polynomial regression and observed that peak viral load typically occurs within 4–8 days after symptom onset, followed by a gradual decline. This being established, infectivity was presumed to be at its highest within the first week after symptom occurrence. The viral load of the skin lesion exhibited a plateau during the early phase of infection, aligning with previous reports that skin lesions are among the initial symptoms. Following the peak, both skin lesion and anorectal specimen viral loads displayed a steep decrease, suggesting a rapid decline in infectivity from these sources after ~6 and 7 days, respectively. These findings underscored the importance of implementing timely preventive measures and interventions, within a week of symptom occurrence, to curb the transmission of the virus.

Saliva demonstrated an earlier and more pronounced peak in viral load compared to blood. What was particularly intriguing was that, unlike other specimen types, the majority of saliva samples containing viable viruses were collected within a week of symptom onset. As no viable virus had been detected in blood samples thus far¹⁹, it was plausible to propose that the high viral load observed in saliva is predominantly a result of direct oral infection through activities like kissing or oral sex, while shedding of the non-viable viral particles from the bloodstream to saliva may also play a certain role in contributing to the viral load. This notion was supported by the detection of positive saliva PCR results in asymptomatic cases before the appearance of a rash or skin lesions at the time of testing.²³

While the high viral load and presence of viable virus in saliva suggested the potential for transmission through respiratory droplets or aerosols, it was unlikely to be the main mechanism of spread in the 2022 outbreak. One previous study reported a high proportion of mpox DNA in saliva (85%).²⁴ However, the study found viable virus in only two out of 45 mask samples (4.4%) and no viable virus in 44 air filters (0%).²⁵ Furthermore, there had been no documented cases of human-to-human aerosol transmission in mpox.

Our evaluation of the mpox virus revealed that pharyngeal samples, including the oropharynx and nasopharynx, had a low viral load and a minimal likelihood of viability. Studies examining the viral load of mpox have found that a notable percentage of patients (43.6%) only test positive for the virus in skin lesions.²⁵ Additionally, in most observed cases, the first lesion to appear was a rash in the genital and perianal areas. These findings suggested that the respiratory epithelium was not the primary site of viral replication during the 2022 outbreak. Therefore, although saliva contains a high viral load with viability, it is less likely that inhaled respiratory droplets play a significant role as a source of infection and transmission. Instead, we propose that a viable virus presented in saliva primarily initiates transmission and infection when it enters non-intact skin or mucosa through direct contact.

Through the analysis of the relationship between viral load and the viability of the mpox virus, we proposed an optimal Ct value cutoff for anorectal, saliva, and urethra for determining virus viability. Samples exceeding these cutoffs demonstrated limited success in culture. Combined with this cutoff, we were able to calculate the duration of viable virus isolation for each specific specimen type: 19 days for anorectal samples, 14 days for saliva and 14 days for the urethra. After this duration of culture viability, based on our assumption, the likelihood of infection through that particular source is expected to be very low.^9,26 Therefore, conducting additional tests to assess infectivity from that specific source may not be necessary for patients with Ct values above the cutoff or after the specified duration.

The impact of comorbidity with HIV on mpox infection is widely recognized, although there is limited evidence supporting different mpox viral dynamics. Most studies examining this impact included people with well-controlled HIV and high CD4 counts; however, Mitjà et al. reported the greatest disease severity, hospitalization and mortality among individuals with low CD4 counts and high HIV viral load.²⁷ Our results demonstrate a higher viral load only in skin specimens from people with HIV, which aligns with previous findings reporting widespread, large and severe skin lesions in individuals with lower CD4 counts.^27–29

Plausible mechanism

The WHO had designated two major clades of mpox virus: Clade I (Congo Basin) and Clade II (West African), with the latter encompassing subclades IIa and IIb.¹⁷ The 2022 global outbreak primarily involved Clade IIb (sublineage: A.2, B.1).^17,24 In the current outbreak, the anogenital area is the most common site of skin rash presentation.

The genital and anorectal epithelium have a lower level of keratinization and a higher presence of antigen-presenting cells like macrophages and dendritic cells compared to other parts of the skin. Thus, these areas are more susceptible to pathogen acquisition.^17,30 Moreover, possible mucosal abrasion and damage acquired during sexual intercourse may facilitate the inoculation of the mpox virus. Consequently, from a clinical standpoint, the highest transmission rate was observed through sexual contact, despite the potential for infection in other areas.³⁰ This observation supports the notion that, while mpox could theoretically be transmitted via respiratory droplets, engaging in oral sex is likely to pose a higher risk of transmission. Nevertheless, further studies are required to support our speculations and to acquire a fuller understanding of the mode of transmission in mpox virus.

Policy implications

Despite its initial identification in humans in 1970, the endemic nature of mpox across several Central and West African countries has resulted in a limited amount of research on the virus, particularly regarding the Clade IIb that caused the 2022 outbreak.^2,17,31 As mpox continues to spread to new regions via travel, it might be perceived as a novel disease in each area. There is an imperative need for political efforts to establish protocols for diagnosis, treatment and, most importantly, prevention. Policymakers should be able to decide specimen selection and collection timing for diagnosis based on current findings.

Strengths and limitations

To the best of our knowledge, the present systematic review and meta-analysis is the first to examine the viral load dynamics of mpox comprehensively and numerically across different specimen types. Our review extended to the temporal profiles, time-varying viability of the mpox virus and duration of viable virus isolation. These findings provide indirect insights into its mode of transmission and carry implications for preventive measures.

Our study benefits from a large sample size and various specimen types, enabling a thorough understanding of viral dynamics. By establishing the timeframe for peak viral load and the duration of viable virus isolation for different types of specimens, our findings can assist in the development of social restriction protocols for confirmed or suspected individuals, particularly for international travellers. Furthermore, when selecting specimens and determining the optimal collection timing for diagnosis, our research outcomes should be considered as valuable guidance. Moreover, our analysis explored the relationship between viral load and viability, leading to the identification of an optimal Ct value cutoff for predicting the presence of viable viruses.

Nevertheless, our study contains several limitations. First, we have limited specific demographic information for each sample, and there are also instances of partial disclosure of this information. As a result, we were unable to compare the viral dynamics between subgroups based on factors, such as MSM status, HIV infection and smallpox vaccination. These differences could be estimated through references from individual studies. Second, the available results correlating clinical features with viral load are insufficient. We can only indirectly compare these findings with our previous paper. Third, there is limited information on viable cultures. However, based on the available information, some assessment of viability can still be made. Moreover, it is important to consider potential variability between laboratories, as the Ct value can be influenced by multiple factors beyond viral load. These factors include sample quality, PCR assay efficiency, probe design and instrument variability. However, it is worth noting that real-time PCR technology has evolved throughout the COVID-19 pandemic, resulting in little variation.^32,33 In addition, we conducted separate analyses for each manufacturer to ensure comparability between different DNA extraction platforms (Table S5 and Figure S7). Nevertheless, the lack of clear standardization of procedures across laboratories may introduce bias in Ct values favouring studies with larger sample sizes. To improve future studies, it would be advantageous to establish guidelines for the standardization process between laboratories. Fourth, due to the limitations of literature so far, only 5% of cohort study was included in our study. Finally, the direct correlation between mpox infectivity and viral load or culture viability remains uncertain. The clinical significance of our findings and their implications are based on inferences, and further evaluations through prospective studies are necessary.

Conclusion

The present systematic review and meta-analysis offer an insight into the viral dynamics of the mpox virus. This work provides clinicians and researchers with essential reference values, such as Ct value cutoffs and days of ending isolation, to aid in managing precautions, guiding the selection of diagnostic specimens and conducting follow-up observations.

Funding

This research was supported by grants from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health and Welfare, Republic of Korea (grant number: HV22C0233), from the National Research Foundation of Korea (NRF; grant number: RS-2023-00248157) and from the Ministry of Food and Drug Safety (grant number: 21153MFDS601) in 2023. The funders played no role in the study design, data collection, analysis, interpretation, or writing of the manuscript.

Acknowledgements

None.

Author contribution statement

D.K.Y. and W.C. had full access to all the data in the study and took responsibility for the integrity of the data and the accuracy of the data analysis. All authors approved the final version before submission. Study concept and design: W.C. and D.K.Y.; acquisition, analysis or interpretation of data: W.C. and D.K.Y.; drafting of the manuscript: H.K. and W.C.; critical revision of the manuscript for important intellectual content: all authors; statistical analysis: W.C. and D.K.Y. and study supervision: J.I.S., W.C. and D.K.Y. D.K.Y. supervised the study and is guarantor for this study. D.K.Y. is a senior author. J.I.S., W.C., and D.K.Y contributed equally as co-corresponding authors. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.

Authors’ contributions

Hakyoung Kim (Conceptualization, Formal analysis, Investigation, Writing—original draft [equal]), Rosie Kwon (Writing—review & editing [equal]), Hojae Lee (Writing—review & editing [equal]), Seung Won Lee (Writing—review & editing [equal]), Masoud Rahmati (Writing—review & editing [equal]), Ai Koyanagi (Writing—review & editing [equal]), Lee Smith (Writing—review & editing [equal]), Min Seo Kim (Writing—review & editing [equal]), Guillermo F. López Sánchez (Writing—review & editing [equal]), Dragioti Elena (Writing—review & editing [equal]), Seung Geun Yeo (Writing—review & editing [equal]), Jae Il Shin (Supervision [equal]), Wonyoung Cho (Writing—review & editing [equal]), and Dong Keon Yon (Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing—original draft, Writing—review & editing [equal])

Conflict of interest: None declared.

Data availability

All data are provided in the article and in the appendix. The study protocol, statistical code and data set are available from DKY (Email: [email protected]).

Ethics statement

This systematic review article does not require Institutional Review Board approval. Our systematic review and meta-anal-ysis protocol was registered with PROSPERO (registration number CRD42023421311).

Informed consent

Not applicable.

References