THREAD (30Mar2020 BREAKING): 1/X Newly publicized study by Chen evaluating #hydroxychloroquine (HCQ, N=31, 200 mg bid) vs control (N= 31) for 5 days in MILD #covid19 patients, showing improved clinical outcomes and chest CT favoring HCQ.
medrxiv.org/content/10.110…
medrxiv.org/content/10.110…
2/X. Key endpoints were pre-defined time to clinical recovery (TTCR) which included cough and temperature. An evaluation of CT scan was done, although unclear if readers were blinded or not. Importantly viral load was NOT measured (or not reported) unfortunately.
3/X All patients received standard treatment: "oxygen therapy, ANTIVIRAL AGENTS, antibacterial agents, and immunoglobulin, with or without corticosteroids". Didn't describe what antiviral agents were used. Since study unblinded, possible bias in management exists
4/X Demographics- its important to highlight the study population: Of 142 with confirmed COVID19, 80 were excluded (Reasons not given- perhaps they had mod/severe disease and other reasons listed in eligibility)
5/X Demograhics (cont) of 62 randomized, only a subset had fever or cough at baseline, and were clinically evaluated (22 & 17 w/ fever, and 22 & 15 w/ cough, for HQ or control, respectively); Don't know if patients with cough and fever are the same.
6/X Results demonstrated an improvement by about 1 day in the cough or fever subset. So there is no clinical data (TTCR) for the 35-40% of patients in study. 
7/X Importantly CT scans were done on all patients. These did suggest a fairly impressive improvement for HCQ group. This is important objective endpoint. Main criticism is that not clear if CT reader was blinded. If not there could be bias in how CT were read.
8/X Authors also reported that for larger proportion of HCQ patients, "pulmonary inflammatory had be partially absorbed". Not a radiologist, so not sure how that was determined. Authors are implying this is less risk for ARDS (inflammatory condition leading to severe disease)
9/X Most important data from this study for me is CT scan improvement, and the fact that 0/31 in HCQ progressed to severe disease vs 4/31 in control arm. This is consistent with frequency of progression to severe disease reported thus far without treatment
10/X One can debate how important~ 1 day improvement in cough and fever (not unlike benefit of tamiflu w influenza), but if progression to severe disease and transmission can be lessened it is is significant. I wish viral load could have been reported.
11/X My assessment: I keep wanting to find data that will blow me away, and not necessarily finding that yet. However, I do think all data so far continue to support the benefit of HCQ, particularly in mild patients. To me this is most convincing pilot study thus far.
