Effective post-exposure prophylaxis of Covid-19 is associated with use of hydroxychloroquine: Prospective re-analysis of a public dataset incorporating novel data

Introduction

There have been (as of November 29, 2020) over 61.8 million cases of Covid-19 and over 1.4 million deaths worldwide,¹ about one fifth of them within the USA.² Early interest developed in deploying hydroxychloroquine (HCQ) and in March 2020 the Food and Drug Administration (FDA) issued an Emergency Use Authorization (EUA).³ Lacking randomized clinical trial (RCT) data, many observational reports emerged which were unfavorable (with exceptions⁴) to the use of HCQ.⁵ With toxicological concerns, on April 24, FDA cautioned⁶ against using HCQ outside hospital or trial settings.

HCQ became highly controversial, it being suggested that “to some extent the media and social forces — rather than medical evidence — are driving clinical decisions and the global Covid-19 research agenda.”⁷ Against this background, on June 15, FDA revoked³ HCQ’s EUA, citing only two just-published substantive RCTs. The findings of the RECOVERY Trial⁸ announced June 5 was cited as offering “persuasive evidence of a lack of benefit of HCQ in the treatment of hospitalized patients.”

The second and only study,⁹ addressing prevention examined post-exposure prophylaxis (the “PEP” study) with HCQ in 821 asymptomatic adults with a high or moderate risk household or occupational exposure to Covid-19. Subjects received HCQ (1.4g first day, then 600 mg daily for 4 more days) or folate (USA) or lactose (Canada) placebo. The study concluded that “…HCQ did not prevent illness […] when initiated within 4 days after […] exposure” (HCQ 11.8% vs. placebo 14.3%; RR 0.83, 95%CI 0.58-1.18, p=0.35).

We¹⁰ and others have criticized the study’s interpretation. Since this was a pragmatic trial, with typically greater heterogeneity and smaller effect sizes than in an explanatory trial,¹¹ powering the study to detect a 50% reduction in Covid-19 may have been over-ambitious, especially given its early termination.¹² An arguably¹³ clinically meaningful reduction of 17% was similar to that associated with dexamethasone.¹⁴ The authors suggested¹⁵ that the study was primarily powered to collect data quickly under pandemic conditions rather than to meet specific clinical goals.

Non-statistically significant signals of HCQ-associated efficacy included an overall effect apparently driven by a 31% reduction among household cohabitees. There were age-dependent reductions (<= 35 years, 36%) found in other analyses¹⁶ to be statistically significant. The use of folate as placebo and the ex-protocol use of zinc and Vitamin C may have been confounding (Supplement). With a reduction of 49% associated with early (“Day 1”) HCQ prophylaxis, we¹⁰ and others¹⁷ found a negative association between treatment lag and reduction of Covid-19.

We conjectured that post hoc exploratory re-analysis of the PEP study dataset would inform a time- and age-nuanced approach to Covid-19 using HCQ, testable in prospective studies. Our objectives were to define: (a) any time-dependent, or (b) any age-dependent effects associated with HCQ and, (c) any influence of gender, exposure type, use of zinc, ascorbic acid or folate on the study outcomes.

Methods

Results

Considering shipping schedules, we estimated¹⁰ that within each of the reported strata for “Time from exposure to enrollment” (1 to 4 days), there were overlapping variations in time from exposure to drug receipt. For example, some “Day 1” (range 1.4-4.4 days) and “Day 4” (range 4.4-7.4 days) subjects may have received drug after the same interval.

New data (9/9 revision) provided at our request broadly confirmed these estimates revealing a reduction in Covid-19 associated with HCQ given within 2 (elapsed time) days of exposure (RR 0.35, 95%CI 0.13 – 0.93; p=0.0438) but not later (RR 0.98, 95%CI 0.67 – 1.45).¹⁰ Recognizing limitations (Supplement) to these estimates, further detail was requested and provided (10/6 revision). The PEP study protocol had intended to enroll only those receiving drug within 4 days from exposure, assuming overnight shipping. We found that 332 and 95 subjects (52% of all subjects) received medication one or two days later than this respectively, with 152/821 (19%) subjects receiving drug outside of the intended 4-day window (Table S 3).

We stratified subjects according to the interval between exposure and drug receipt. We found a reduction in Covid-19 associated with HCQ when received “Early” between 1 and 3 (elapsed time) days after exposure from 16.5% to 9.6% (RR 0.58, 95%CI 0.35 – 0.97; p=0.044; NNT 14.5) but not after three days (“Late”) (RR 1.22, 95%CI 0.72 – 2.04) (Table 1). We did not detect differences in the severity of symptoms for both Early and Late prophylaxis cohorts (Supplement). A comparison of the demographic and clinical characteristics of the two groups shows a largely conserved balance between the Early and Late cohorts (Table 2, Table S 1).

Adopting the same age strata as the PEP study, we found in the Early cohort non statistically significant Risk Ratios of 0.53 (18-35 years), 0.52 (36-50 years), and 2.80 (> 50 years). With no a priori reason for selecting these strata, the data are less subjectively supportive of two age strata. We discerned a boundary between 42 and 48 years and on a post hoc exploratory basis we set it at the point where the upper 95% confidence interval for one of the Risk Ratios in the two strata was <1. We found reductions of Covid-19 associated with drug given up to 3 (elapsed time) days after exposure in both younger (18-45 years) (RR 0.53, 95%CI 0.29-0.97; p=0.0448, NNT 11.5) and older (>45 years) (RR 0.75, 95%CI 0-27-2.05) subjects. Within the Early and Late cohorts, further stratification revealed no obvious gender-based differences in this effect (Table S 4).

Considering only subjects reporting no co-morbidities (particularly excluding asthma and co-morbidities classified as “other”), suggested stronger effects associated with HCQ within time- or age-related strata (Supplement).

With a higher baseline rate in household contacts (26.8%) than in HCW (13.8%), the reduction of Covid-19 in household contacts associated with Early HCQ was statistically significant, compared with placebo (RR 0.35, 95%CI 0.13-0.89; p=0.025, NNT 5.7)(Table 4).

Directionally similar results were obtained after excluding subjects who did not contribute to outcome data and subjects who did not take study medication. Stratifying into Early and Late prophylaxis cohorts revealed no discernible effect associated with folate (Supplement). With poorly detailed observational data, there did not appear to be an effect associated with zinc or Vitamin C (Supplement). The use of zinc and ascorbic acid appears balanced between the groups both for the whole cohort and the Early and Late time strata (Table 1).

Discussion

In tackling our primary objective of defining any temporal effect of HCQ, our assumption was that HCQ prophylaxis had been “initiated within 4 days after […] exposure.”⁹ We were joined in this understanding by others,^13,16,17 including the authors of NIH²¹ guidelines and the editorial⁷ accompanying the paper. This assumption requires revision for two main reasons.

Firstly, many participants received medication after the intended overnight delivery or after four days from exposure. A similar issue likely pertains to the companion PET study.¹⁸ Secondly, inconsistent terminology lead to an overestimate by one day of the time from exposure to enrollment or to drug receipt.

Correcting these assumptions results in a statistically significant reduction of 42% of Covid-19 associated with HCQ when received between 1 and 3 (elapsed time) days after exposure, with no effect later than this. The early use of HCQ is supported by estimates for the incubation period of 3-8 days²²

We found an age-dependent, statistically significant reduction of Covid-19 associated with drug given Early. This finding is consistent with a 45% reduction of Covid-19 associated with HCQ pre-exposure prophylaxis (PrEP)²³ in younger patients (HR 0.55, CI 0.32-0.96, p=0.038, combined treatment groups) in a companion study and an observational study involving mainly younger HCW.²⁴ Apparent age-dependent and time stratified effects associated with HCQ may be confounded by increases in incubation period with age.²⁵ Re-analyzing the same PEP dataset without time stratification and confirmed by us (Supplement), Luco¹⁶ described reductions in Covid-19 associated with HCQ in subjects younger than 50 years (RR 0.71, p=0.089) reaching statistical significance (RR 0.63, p=0.029) in the sub-cohort experiencing high-risk exposures.

Small population sizes within the co-morbidity subgroups prompt cautious interpretation. However, the presence of co-morbidities attenuated age- and time-dependent effects associated with HCQ. Although age-related responses associated with HCQ may be dependent on the presence of co-morbidity,¹⁶ excluding patients with co-morbidities did not yield equivalent effects in the younger and older age strata. Asthma and co-morbidities classified as “other” contributed most to attenuating the response associated with HCQ. This finding is supported by the application of Multiple Correspondence Analysis and the Mantel test to the same dataset by Luco¹⁶ who described confounding clinical differences, between study groups particularly regarding asthma and “other” co-morbidities.

Mirroring the original data, we found a significant effect associated with HCQ in household contacts. This result may reflect differences in access to advanced PPE, hygiene training, and the ability to quarantine after exposure. A further difference may originate from the limitation discussed relating to likely multiple high-risk exposures in household contacts. Thus, household contacts in this study may share much with the first responders in the companion PrEP study,²³ where a 64% reduction in Covid-19 associated with HCQ was observed (combined dose groups).

We did not detect an effect of using folate. Due a paucity of data, we could not determine whether there was an effect of zinc or Vitamin C other than noting no differences associated with HCQ in subjects taking neither agent and the entire ITT cohort. Further, based on the apparently balanced use of zinc and Vitamin C in the stratified cohorts, confounding due to these agents seems unlikely.

Our findings are made in the general climate of concern²⁶ for the reliability of publications related to Covid-19 and the continuing controversy and confusion surrounding HCQ.^27,28 This is partly fueled by a widening understanding of Covid-19 pathogenesis and the multiple mechanisms that have been proposed for HCQ.²⁹ It is important therefore in discussing HCQ’s putative actions to do so in relation to a specific stage in the disease.

In hospitalized patients RCT findings³⁰ evincing HCQ’s ineffectiveness are supported by observational studies, notably two related studies^4,31 reporting a significant reduction in mortality associated only with HCQ’s use with zinc. Further confounding within a number of observational studies is related to the possibly synergistic use of steroids.³² At earlier stages, any effect associated with HCQ appears independent of zinc, evinced by the lack of synergy we observed between HCQ and zinc. Further, using zinc may be futile in otherwise healthy, especially younger subjects with no zinc deficiency or dysregulation.

For prophylaxis, understanding differences in timing of intervention, patterns of exposure, the ability to quarantine, testing methods, co-morbidities, and the possibility of multiple rather than single “index” exposures, appear important in reconciling apparently conflicting studies where distinctions between pre- and post-exposure prophylaxis may become blurred.

Our findings are consistent with those of a prospective, non-randomized, study³³ in which asymptomatic subjects (non-HCW, mean age 37.2 years) mostly exposed to laboratory confirmed COVID-19 cases could opt to receive HCQ prophylaxis (800 mg first day, then 400 mg weekly for 3 weeks) or standard care alone (control). There was a reduction of the 4-week incidence of Covid-19 associated with HCQ compared with controls (19.4% vs. 10.6%, p=0.033, NNT =12). A single “index” exposure could not be identified. Nonetheless, subjects were enrolled after the primary case positive report and prophylaxis began at the earliest within 48 hours of knowing about the high-risk contact (D. Dhibar personal communication).

A household randomized trial³⁴ reported in abstract form examined the effect of HCQ given to subjects last exposed to an infected person within 96 hours of enrollment. By day 14, no difference in infection rate was observed between the HCQ and ascorbic acid control groups (aHR 0.99, 95% CI 0.64-1.52). Medication was mailed with an unstated shipping time, and data were neither stratified by age nor time from exposure to drug receipt.

A Spanish cluster-randomized trial³⁵ found a small reduction of Covid-19 associated with HCQ (RR 0.86, 95%CI 0.52-1.42). There was a higher mean age (∼ 48 years) than in the PEP study (∼ 42 years) where we have suggested a stratum boundary at approximately 45 years. There was a similar effect associated with intervention <= 3 days (RR 0.89) or 4-6 days (RR 0.93), but not >= 7days (RR 4.09) after exposure. Notably, there were differences according to a subject’s status by Covid-19 PCR testing at baseline. In PCR-positive subjects there was no effect associated with HCQ (RR 1.02; 95%CI 0.64–1.63), whereas for PCR-negative patients there was a signal for an HCQ-associated effect (RR 0.68, 95%CI 0.34–1.34). A change in PCR status is likely to be a function of the moment beyond which a drug is unlikely to be effective, possibly more accurately than an estimate of time from exposure. Accordingly, this study supports our findings suggesting a beneficial effect associated with HCQ when used early.

Further, the PCR-negative cohort from the Spanish study shares much with not only the PEP study population, but also that of its companion PrEP study.²³ Although the PrEP study was hampered by poor recruitment, once or twice weekly use of HCQ in HCW was associated with reduced development of Covid-19 by 27%, compared with folate placebo (HR 0.73, CI 0.48-1.09, p=0.12, combined groups). Comparison with another PrEP RCT³⁶ in HCW, with a very small study population (n=132) due to early termination and a low incidence of Covid-19, is not meaningful.

Limitations

Our primary time stratification based on newly-acquired data essentially represents the a priori analysis intended by the original PEP study. Nonetheless, this remains a post hoc study, and our examination of various subgroups is subject to well-known statistical challenges.³⁷ These are partially offset by our stringent use of two-sided tests, when directionality in the original data may have justified using one-sided tests. Results should be interpreted cautiously, and the hypotheses generated should be tested in prospective studies sufficiently powered to accommodate multiple comparisons in sub-groups.

Our study retains the limitations acknowledged by its original authors related to the availability and access to testing, the use of a clinical case definition of Covid-19, the reliance on self-reported data and the generally young population studied. There are other limitations. The study poorly represents African-American and Hispanic or Latino populations. The rapidly executed study overcame several logistical challenges to allow the collection of real-world data, having both advantages and disadvantages of a pragmatic design.^11,38 Self-selection bias inherent in this type of study may have been compounded by FDA cautions regarding HCQ.⁶ Unlike similar studies,^34,35 the original PEP study was not cluster-randomized.³⁸

Several limitations relate to the estimation of the interval between exposure and treatment with 24-hour windows of uncertainty on either side. The earlier of these is due to subjects providing only the date of their highest risk exposure. The later window is due to the way the drug receipt times were de-identified, and the unknown interval before the ingestion of the first dose. The original authors (personal communication) attempted to ensure that this was minimized by delivering medication to an address where the participant knew they would be at its expected arrival time. Rather than a single “index” exposure, there were likely other exposures before the “index” contact was diagnosed.

Time-related or other biases may be associated with the exclusion of 100 subjects who were randomized to the PEP study, became symptomatic before medication was received and aggregated into the companion treatment study.¹⁸

The poorly understood relationship between age and susceptibility to Covid-19 provides no a priori justification for defining particular strata especially in a weakly powered and statistically fragile system. Indeed, the same three strata were used in one companion study¹⁸ to the PEP study but not another.²³ Several methods³⁹ have been proposed to define strata which we attempted to accomplish empirically on a post hoc exploratory basis. Lastly, analysis of the effect of risk level is confounded by its changing definitions and the inability to discriminate between nuances within the high-risk category.

Conclusions

Analysis⁴⁰ of the PEP,⁹ companion^23,18 and other³⁵ studies raise no significant safety concerns for using HCQ in the populations studied. Integrating a public dataset with new unpublished data, we found that, especially in younger subjects, hydroxychloroquine was associated with significantly reduced illness compatible with Covid-19 when initiated between 1 and 3 days after a high-risk or moderate-risk exposure. This finding warrants prospective confirmation.

Funding and Conflicts of Interest

There is no external support for this study. The sponsor is entirely responsible for the design and conduct of this study. The sponsor and principal investigator have no financial or other conflicts of interest in the subject matter of this protocol. DMW is the president of Synechion, Inc., that provides services for the medical industry outside of the area of this work. DMW is also the president of KevMed, LLC., that markets medical products outside the scope of this work. DM is the president of ZSX Medical, LLC. that develops surgical devices and a Principal at Third Eye Associates, a technical consulting company. PK and SAS report no conflicts.

Data Sharing

Microsoft Excel files will be made available on request up to one year after publication.