In December 2019, multiple cases of undiagnosed pneumonia emerged in China. Sequencing on respiratory samples indicated the presence of Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) causing Coronavirus Disease 2019 (COVID-19). ^{1– 3}

The main method of diagnosing COVID-19 is by the isolation of SARS-CoV-2 through Reverse Transcriptase Polymerase Chain Reaction (RT-PCR or PCR) nucleic amplification testing of respiratory samples. ⁴ In a study done by Yu et al. the viral load predicted by the Cycle threshold (Ct) value was higher in sputum samples than nasopharyngeal and throat swabs. Patients who had an increasing viral load and lower CT values had a progressive disease and more severe lung injury. ^{5, 6} Furthermore patients with higher initial viral loads had worse outcomes and more prolonged viral shedding. ⁷ The viral dynamics studied in 18 different patients resembled that of patients with influenza rather than SARS-CoV. ⁸

Given the importance for early identification of patients with severe disease and that a low Ct value is proposed to be inversely proportional to disease severity, ⁷ we aim to describe the different clinical manifestations and disease severities in relation to the initial and subsequent Ct values in patients who were admitted to King Faisal Specialist Hospital and Research Centre-Riyadh (KFSH&RC).

During the study time period, 35 subjects had presented to the KFSH&RC-Riyadh and for whom sufficient follow-up information was available (21 women and 14 men). In total, 28 patients were considered at presentation mild (stage B), 5 were moderate (stage C), and 2 were asymptomatic (stage A). The average age was 49.9±21.9 years. Comorbidities were observed in 68.6% of the cases. All patients had tested positive for SARS-CoV-2 by PCR. Table 2 shows the disease severity progression through the course of 14 days following admission. Table 3 presents symptoms that developed sometime during the course of infection.

There were nine different treatment profiles administered across the 35 subjects when considering the seven different drugs used during the course of illness – HCQ, AZI, Kaletra, Ribavirin, Tocilizumab, IVIG, and other. Of interest is to differentially identify from among the 35 study subjects those who were given the combination of HCQ and AZI as their initial treatment from those who were not given this combination as their initial treatment. The schematic in Figure 1 depicts the course of treatments received by each of the 35 subjects and how it was decided into which of the two treatment groups each was assigned post-hoc. It is seen that 21 (60%) pateints were given HCQ and AZI as their initial treatment. Henceforth, this treatment group will be referred to as the HCQ/AZI group. Those subjects not in the HCQ/AZI treatment group will be denoted the HCQ / AZI ¯ group.

Table 4 displays a comparison of baseline characteristics between the two treatment groups. There was no significant difference in age and gender between the two groups. The frequency of individuals with a past history of smoking in the HCQ / AZI group was significantly less in the HCQ / AZI ¯ group. No differences were observed in the radiological findings or vital signs between the two groups. Among the baseline laboratory values, ALT was lower in the HCQ / AZI group. Table 5 displays a comparison of the baseline clinical severity of the two treatment groups; there is no statistically significant difference in the two.

Table 6 shows a comparison for the number of days to reach negative RT-PCR results between the two treatment groups. From the table it can be seen that 7 of the 21 HCQ / AZI -treated subjects remained positive for 14 days after admission whereas 12 were able to reach a negative status within that time period. There were 2 subjects not fully followed for 14 days post-admission for whom a conclusion could not be reached. Moreover, 9 of the 14 HCQ / AZI ¯ -treated subjects remained positive for 14 days after admission, whereas 5 were able to reach a negative status within that time period. Considered as two ordered multinomial distributions, the two columns did not show a statistically significant difference between the two treatment groups (p = 0.1571) with respect to the time to reach negativity. Figure 2 presents a time-to-negativity actuarial chart for the two treatment groups with no statistically significant difference in the two curves.

The Ct values ranged from a minimum of 11.44 to a maximum of 38.35 across the 5 time points (days 0, 3, 7, 10, and 14), across the two treatment groups, and during the 14 days. There were many days among the 35 subjects for which Ct values were not available (i.e. a PCR test was not carried out). These missing results presented a challenge in trying to evaluate the relative Ct profiles between the two treatment groups. It was noted however that, although the Ct value was missing, a positivity/negativity indication was often present in the medical record when the Ct value was missing. Among those missing both the Ct value and an indication of positivity/negativity, a clinical severity stage value (A, B, C, D, or dead) was present for each of the 5 days. Using this information together with whatever Ct values were present, an attempt at imputing information for missing Ct values was made using the following logic. The logic lead to an ordinal Ct score 1, 2, 3, or 4 for each subject for each of the 5 days (0, 3, 7, 10, and 14), and it will be referred to as the Ct score. The methodology includes the following five steps: 1.

Ct values were firstly regarded as falling within one of four categories:

Note that, even though a Ct value of 40 or more was not seen among the 35 subjects, this category was conceptualized to accommodate a Ct value for a subject who reached a negative RT-PCR test or a status of no disease. 2.

For each day when a subject’s Ct value was available, a score (as described in #1 above) was assigned to the Ct score.

In the end, as a result of the above, a complete Ct score profile was available for 33 of the 35 subjects. A breakdown of these profiles separately for the two treatment groups is as shown in Tables 7 and 8. In reviewing the tables, it can be seen that there are 12 unique Ct score profiles for those treated with HCQ / AZI ¯ , and 14 unique Ct score profiles for those treatment with HCQ / AZI . It was found that 5 of the profiles from Table 7 were the same as 5 of the profiles from Table 8, i.e. 18 unique profiles of Ct scores across all subjects. The 18 profiles are presented in Table 9 and ranked in order of favorability, where a more favorable outcome is one for which a higher score is reached earlier. A rank of 1 is the least favorable profile and larger ranks assigned to a profile indicate a more favorable one. A rank-based comparison was performed on these two sets of the ranks, and the results suggested no statistically significant difference in the two treatment groups ( HCQ/AZI median rank: 11; HCQ / AZI ¯ median rank: 5.5; p = 0.2212 (Kruskal-Wallis)) with respect to the time to a recovering Ct value.

Table 10 describes the change in disease severity from admission to day 14. One patient was classified at admission with stage A and then his disease developed at day 14 to stage D. Around 36% of the patients presented with mild (stage B) severity and then lessened to asymptomatic (stage A). In total, 7 patients presented with mild stage then they were tested negative by day 14. Table 11 provides an overall comparison of the clinical severity on day 14 between the two treatment groups with no statistically significant difference between the two (p > 0.05).

Several lab tests were carried out on the study subjects during admission ( Table 12). These included hemoglobin, creatinine, D-dimer, CD4, IgG, and others. From among the lab tests the following were found to significantly relate to disease severity at admission: D-dimer, troponin, LD, C4, absolute CD8 (p < 0.05).

The current study involves subjects presenting early in the COVID-19 pandemic and much knowledge about the virus and the ensuing disease has been generated since the time of these subjects’ presentations. However, there is still relatively little new knowledge related to the time profile of Ct values and their potential value in understanding the course of disease. This study aims to add some related insights and particularly how treatment efficacy (here a comparison of HCQ and AZI) may be able to be evaluated based on Ct values. Additionally, the study recognizes the challenges associated with timely RT-PCR testing and always documenting the Ct values and proposes an imputation method that may be used with conservative statistical assumptions for analyses of Ct profiles. There are several limitations of this study. It is a retrospective study that has a small sample size and was conducted in a single center.

In this study, a treatment regimen of HCQ and AZI was identified for 21 of 35 subjects testing positive for SARS-Cov2 and admitted to a single tertiary care hospital early in the COVID-19 pandemic.

This study presents the follow-up information gathered on the patients for up to 14 days after admission, specifically days 0 (admission), 3, 7 10, and 14. Of particular interest for the follow-up was the clinical staging of COVID-19 and the Ct value from the RT-PCR. Ct values were scored as 1, 2, 3, or 4 based on the value of the Ct being below 20, 20-30, 30-40, or 40 or more. The follow-up Ct value was not consistently available at each of the four follow-up timepoints following admission, and imputation of missing Ct scores (1-4) was made using an algorithm based on coincident clinical stage and interpolation. Recognizing the assumptions associated with the designed imputation scheme, a conservative, nonparametric (i.e. rank-based) method was used for comparing the two treatment groups with respect to outcomes. A validation study of this method is going to follow.

The patients treated with the HCQ and AZI were similar to those not treated with respect to all demographic, clinical, radiological, and laboratory features with the exception of a higher rate of smoking history among those treated with HCQ and AZI. The follow-up profiles of the two groups were compared with respect to improvement trends (i.e. improving clinical staging and improving Ct scores). There was no evidence of any difference in the sets of profiles for the two treatment groups.