A New Look At CytoDyn's Severe-To-Critical COVID-19 Trial

Mar. 17, 2021 4:39 PM ETCytoDyn Inc. (CYDY)ABBV, BMY, BMYMP, JNJ, MRK, PFE526 Comments
Vision and Value profile picture
Vision and Value
1.51K Followers

Summary

  • Trial data is supportive of an EUA: Leronlimab outperformed every approved or recommended drug in critically ill patients.
  • Shorts have a flawed analysis that doesn’t correct for patient age distribution.
  • The clinical outcome of length of hospitalization was statistically significant (p=.005).
  • Dexamethasone is recommended as a standard of care based on age-adjusted meta-analysis data.

Intensive care in the hospital, COVID-19
Photo by JazzIRT/E+ via Getty Images

CytoDyn Inc. (OTCQB:CYDY) recently announced the clinical trial results of its monoclonal antibody, leronlimab, in treating severe to critically ill patients with COVID-19. A deeper look at the data shows how important age was as a factor in mortality. Once age is factored in, the data suggests that leronlimab works well in COVID-19 severe and critical patients, but there were a few issues with the trial data that will be addressed. This closer look at the data shows that leronlimab most likely will be approved or issued a conditional Emergency Use Authorization (EUA) sometime soon. After all, doctors are demanding that the FDA approve the drug as they have seen firsthand that it works.

Digging Into the Data

The results across various measurements such as mortality and hospital stay, even when stratified for disease severity, patient baseline characteristics, and other factors were consistent. The only drawbacks were: 1) the loss of 10 patients in the intent to treat (ITT) population exacerbated what was already a slightly underpowered study and 2) baseline characteristics (primarily age), varied between the placebo and the leronlimab groups. This skewed the study by greatly lowering the mortality rate in favor of placebo. Age is such a strong variable in mortality for COVID-19, with one large study finding that:

the mortality rate of COVID-19 was 8.1 times higher (95%CI = 7.7, 8.5) among those 55 to 64 years, and more than 62 times higher (IRR = 62.1; 95%CI = 59.7, 64.7) among those ages 65 or older.”

Given the very confounding factor of age and the significant difference in age subgroups in CytoDyn’s CD12 trial, it is important to stratify the groups by age to see how leronlimab performed when accounting for various factors that may have been dissimilar at baseline. On the investor call, the company indicated that older people were enrolled in the leronlimab arm by a ratio of 3:1. Why? On Aug. 25, 2020, when CytoDyn announced the 50% enrollment checkpoint (195 Patients) of its CD12 trial, the pandemic was just cooling down after its second peak. Essentially, recommendation of Drug Safety Monitoring Board (to continue the trial with the original protocol-defined sample size and power to achieve the primary endpoint) was based on the patient data collected mostly during the second wave of the pandemic. However, as the trial continued with the second half of patient enrollment, the third massive wave of the pandemic hit all across the United States. This tsunami of hospitalized patients may have easily disturbed the distribution of the patient population in a blinded placebo-controlled trial.

Source: The COVID Tracking Project

The average age was almost the same in both arms (58.82 vs 58.51) which means much younger patients could have offset the much older patients. The critical sub-group of the study had a population of 62, which is small for a phase 3, yet it still was statistically significant — the effect on a larger group is generally a lower p-value. However, peer-reviewed analysis on other drug trials for COVID-19 such as the powerful immunosuppressive steroid, dexamethasone, utilized age adjustments to more accurately quantify the therapeutic effects. Dexamethasone was approved and is recommended for use in severe-to-critical COVID-19. If the rationale for its approval used age stratification, shouldn’t leronlimab’s analysis?

These age adjustments were not post-hoc analyses; they were prespecified by CytoDyn as part of the Statistical Analysis Plan (SAP) at the beginning of the trial. It's inaccurate to call this “data mining”, and there is no reason to be concerned about bias being introduced in typical post-hoc analysis of various subgroups.

As short sellers flock to slam CytoDyn’s stock, claiming that leronlimab didn’t work and that CytoDyn is disingenuously spinning the clinical trial results in its favor, it's important to keep a cool and rational perspective on CytoDyn’s recent CD12 trial results. It is fairly clear, given the congruency of the leronlimab CD12 trial data and the magnitude of therapeutic benefit seen therein, that leronlimab works. P values are not everything, and it is extremely important to qualitatively analyze data — across various sets of data, subgroups, magnitudes, and other factors, to make conclusions. In fact, a fairly recent publication in Nature protested against the incorrect use of p values to make claims such as “no difference was observed,” which can result in the “dismissal of crucial effects.”

Source: Amrhein, Valentin, et al. Scientists rise up against statistical significance. Nature 567, 305-307 (2019).

It is first important to gain some perspective by taking a closer look at what other COVID-19 treatments have been approved or recommended for treatment for those hospitalized patients, and what kind of therapeutic benefit these therapies offer patients. One category of EUA-approved therapies for COVID-19 are neutralizing antibodies, which help prevent the virus from replicating by directing immune responses against the virus. Since neutralizing antibody treatment is sometimes recommended in the early course of disease before significant and morbid symptoms develop, but are not recommended for treatment in later stages of disease, no comparison of leronlimab to these will be made. On the other hand, remdesivir (antiviral) and dexamethasone (immunosuppressive steroid) are approved for varying later stages of the disease, for preventing viral replication and suppressing the immune system (from overactivation), respectively.

Dexamethasone

First, dexamethasone is recommended for severe and critical COVID-19 patients based on a meta-analysis of various clinical trials which included thousands of critically ill patients:

“Dexamethasone reduces mortality in critically ill patients with COVID-19 according to a meta-analysis that aggregated seven randomized trials and included data on 1,703 critically ill patients. The largest trial in the meta-analysis was the RECOVERY trial, whose subgroup of mechanically ventilated patients was included. For details about the meta-analysis and the RECOVERY trial, see the Corticosteroids section. Because the benefits outweigh the potential harms, the Panel recommends the use of dexamethasone in hospitalized patients with COVID-19 who require invasive mechanical ventilation or ECMO (AI).”

Shown below is a chart from the RECOVERY trial, which shows mortality as a function of time for each patient population according to disease severity at baseline. While the overall population barely did any better with dexamethasone, it was observed that with increasing disease severity, dexamethasone became more and more helpful. In fact, if administered too early (when viral replication might still be a problem and an overactive immune system might not be a problem), dexamethasone is detrimental. However, in later stages of the disease, inflammation is the primary problem, and dexamethasone appears to have a more robust effect. Based on the CD10 trial, it is clear that there is no early immunosuppressive effect as is with the case of dexamethasone. The timing of administering leronlimab makes it easier to administer and poses no risk to the prescribing clinician according to available data.

Source: The RECOVERY Collaborative Group. Dexamethasone in hospitalized patients with Covid-19. N Engl J Med 2021;384:693-704.

This study had over 6,400 patients enrolled and ultimately used an age-adjusted analysis for its final reporting of total reduction in mortality numbers, since the dexamethasone group, despite having 6,400 patients enrolled in the entire trial, still had an average age increase of 1.1 years compared to placebo. For an idea of how much of a difference a mere average difference of 1.1 years can make in mortality reduction, we can look at the reduction in mortality in the severely ill population before and after it was adjusted for age. Before the adjustment, dexamethasone reduced severely ill patients’ mortality by 2.9%, whereas after the adjustment this reduction in mortality ballooned to 4.2%, a relative improvement of ~45% just on age adjustment.

The critically ill subgroup benefitted from dexamethasone with 29.3% versus 41.4%. The age adjustment made less of a difference in this population. Regardless, one can see how just a 1.1-year difference in the patient populations made a ~45% relative difference in severe patients’ mortality; and given how strongly the age variable affects mortality, it is absolutely critical to use subgroup analyses to confirm treatment effects across patient subgroups. Presumably, these adjustment data were important in securing recommendation of dexamethasone use for severe patients (on noninvasive oxygen). And therefore, the age stratification CytoDyn used doesn’t even closely resemble “data mining”, but represents a standard and important part of the analysis, especially when there were 33% of >65-year-old patients enrolled in the leronlimab arm versus only 23% in the placebo arm. That is enough to make a profound difference in the trial results.

Adjusting CD12 Mortality Rates For Dexamethasone Use and Severity, Using RECOVERY Data

A recent analysis by another author does a good job of outlining how the age adjustment might have also put the CD12 trial’s placebo group’s overall mortality rate in-line with other studies such as the RECOVERY trial: “In taking this corrective action the overall mortality in the leronlimab arm drops to 18.91% and the placebo arm rises to 25.6%.” This is roughly in-line with other critical-to-severe populations such as in the RECOVERY trial, where the critical population had 29.3% and 41.4% mortality rates with and without dexamethasone (important to look at both since many CD12 patients received dexamethasone), and the severe population had 23.3% and 26.2% mortality rates with and without dexamethasone. This gives weighted averages of 24.5% and 29.3% for the groups with and without dexamethasone, critical and severe together.

To estimate critical and severe adjusted mortality rates in the CD12 trial placebo arm based on the rates of dexamethasone use, one can take the weighted averages of the patient populations in the RECOVERY trial that used dexamethasone in the critical and severe arms based on the proportion of patients that used dexamethasone in the entire CD12 trial (61%). This uses the RECOVERY trial’s mortality rates and the CD12 trial’s proportion of dexamethasone use. Then, using those dexamethasone-use-adjusted mortality rates for both critical and severe patient populations, one can adjust for the proportion of critical and severe patients in the CD12 trial, and obtain an overall mortality rate for CD12 that should have been observed had the patient stratification been correct. This is all based on the RECOVERY trial mortality rates, just adjusting for proportion of dexamethasone use and proportion of severe and critical patients.

Thus, if age-adjusted properly, the CD12 trial might have had approximately 25.9% mortality rate across critical and severe, which is almost exactly in line with the other author’s age adjustment. A further adjustment could be made for age, but what is clear is that CD12’s trial arm age differences skewed the results. It appears leronlimab did indeed reduce mortality across patient subgroups, and that in the overall age-adjusted population, the reduction in mortality was 25.6% to 18.9%, a total reduction of 6.7% or a relative reduction of about 26%, above and beyond benefits observed from dexamethasone use.

Remdesivir

It is also important to see what kind of effect remdesivir had on COVID-19. Remdesivir was approved for COVID-19, and its primary analysis showed a 3.8% reduction in mortality in the entire population, 11.4% mortality with remdesivir and 15.2% mortality with placebo. This hazard ratio was statistically relevant but did not reach statistical significance, even with 1,062 patients enrolled, and showed absolutely no difference with the critically ill population (see Table S5 of the Supplementary Appendix). Additionally, in the trial, the remdesivir arm had 3% fewer patients age ≥65 (Table S1), 35% remdesivir versus 38% placebo. There is no handicapping of results based on this benefit the remdesivir arm might have gained from this uneven baseline randomization.

Overall, remdesivir appears to benefit patients hospitalized but not in severe condition, as shown by the recovery curves below. It appears to help patients from tipping over to needing more organ support, including high flow oxygen or invasive mechanical ventilation. Note that the remdesivir trial used different notations for baseline ordinal scale, using increasing numbers with disease severity instead of decreasing numbers as used in the CD12 trial.

Source: Beigel, John H. et al. Remdesivir for the Treatment of Covid-19 — Final Report. N Engl J Med 2020; 383:1813-1826.

In addition, the WHO Solidarity Trial (11,330 patients, better baseline patient balance with respect to age) failed to show any difference in mortality using remdesivir on hospitalized patients. Overall, the remdesivir results look somewhat promising, but they do shed an unfortunate light on the fact that, still, dexamethasone is the only drug physicians can turn to help their patients in the worst condition — on the brink of life and death.

Other Therapies

The United Kingdom is now using IL-6 inhibitors Kevzara and Actemra for hospitalized patients, though their road to approval in the U.S. is less clear; the drugs have shown inconsistent results across studies and a general lack of tolerability, seen by slight increases in side effects.

Currently, for patients on the ventilator, only dexamethasone (and in some cases dex+remdesivir) is recommended. In some cases where dexamethasone cannot be used, baricitinib (JAK inhibitor) can replace dexamethasone and be used in conjunction with remdesivir. However, in remdesivir’s best patient population, baricitinib also failed to show any statistically significant improvement in mortality:

“In a subgroup analysis of participants who required supplemental oxygen but who did not receive it through a high-flow device or invasive mechanical ventilation, the rate ratio for recovery was 1.17 (95% CI, 0.98–1.39). There was no statistically significant difference in mortality by Day 28 between the baricitinib and placebo arms in this subgroup (OR 0.4; 95% CI, 0.14–1.14) or in the overall population. Baseline corticosteroid use was an exclusion criterion, and the trial enrolled most participants prior to the public release of RECOVERY data.

Because dexamethasone has been shown to reduce mortality among patients who required supplemental oxygen, clinicians should prioritize the use of dexamethasone in this subgroup. The Panel therefore reserves baricitinib plus remdesivir for the rare circumstances in which corticosteroids are contraindicated (BIIa). It is unknown whether baricitinib would have an additive benefit or adverse effects when given in combination with corticosteroids. Therefore, the Panel recommends against using the combination of baricitinib, dexamethasone, and remdesivir, except in a clinical trial (BIII).”

NIH Recommended Pharmacological Patient Management

Source: Therapeutic Management of Adults With COVID-19, NIH COVID-19 Treatment Guidelines.

By viewing the COVID-19 treatment recommendations by the NIH panel as well as the data generated for key therapies like remdesivir and dexamethasone, it is clear that there is a huge lack of therapies to improve patient outcomes significantly. Remdesivir was approved, but generated poor data in the Solidarity Trial, casting doubt on its efficacy. The other neutralizing antibodies were given EUAs, but the panel’s wording for recommendation, or not, clearly shows that there is insufficient evidence to strongly recommend administration in early disease. In contrast, leronlimab has demonstrated, when adjusted for age, a significant benefit for severely and critically ill patients. It also appears that leronlimab is effective from the moderate to critical stage of disease.

Leronlimab CD12 Trial Data

Not only did leronlimab appear to significantly reduce mortality at day 28 without dexamethasone use, it benefitted patients who had prior dexamethasone use, indicating a potential non-redundant anti-inflammatory or immunomodulatory effect. This analysis is statistically significant and demonstrates that leronlimab would show superiority in a head-to-head trial of dexamethasone. Since dexamethasone is currently the standard of care, it's conceivable that the NIH recommendations in Figure 1 above could very well change to include leronlimab in the near future.

Source: CytoDyn CD12 Executive Summary (8-K)

This statistically significant benefit was also seen in the group of patients (n=309) who had any prior or concomitant COVID-19 treatment, indicating that leronlimab synergizes with other treatments. This data is arguably more important than the overall population (n=384), which included 75 patients that had no treatment protocol. As the standard of care evolves, it is important to know how leronlimab synergizes with standard of care and other treatments. To save someone’s life, a physician wouldn’t administer leronlimab OR another drug; they would administer both.

Source: CytoDyn CD12 Executive Summary (8-K)

One last thing to highlight is how remdesivir is cited as helping people get out of the hospital faster. Leronlimab got critically ill patients out of the hospital about 5.5 days faster than placebo. This is approximately in-line with remdesivir’s improvement in time-to-recovery in its initial trial (but only in the less severe patients), though the WHO’s Solidarity Trial found no difference in times to recovery using remdesivir.

The best conclusion is that leronlimab has a positive effect on severe and critical patients, and has an even more robust effect on severe and critical patient outcomes than other drugs, though the statistical significance is not quite there in the overall population due to the age difference anomaly. Importantly, leronlimab synergizes with other drugs.

Data Analysis Takeaways

This potentially skewed the study by greatly lowering the mortality rate in favor of placebo. Basically, there are three main points:

  • Baseline characteristics in the CD12 trial lowered the mortality rate in favor of placebo, and this is somewhat verifiable based on other studies’ mortality rates. Regardless, the rest of the data looks great.
  • Dexamethasone appears to have used an age stratification for approval, especially for severe patients (on supplemental oxygen). Remdesivir’s differences in age may have actually skewed the drug’s result into its favor.
  • P values aren’t everything. Magnitude of therapeutic effect is also important. One might rather be given a drug that showed a marginally significant p value but a robust magnitude of therapeutic effect (due to a smaller study size) than a drug that achieved an excellent p value but used a much larger study to achieve that p value since the drug’s magnitude of therapeutic effect was small. This would be especially true if the robustly effective drug was known to have a benign safety profile.

Whether or not CytoDyn will need supplemental data, or another trial to support evidence for approval, or nothing at all to gain EUA in various countries, is anyone’s guess. At most, the company expects to enroll 140 patients in a critical COVID-19 trial (CD16) within the next 4 weeks, which would mean only waiting about another 3 months for data and potential approval. The primary endpoint of the CD16 trial is time of hospitalization, which in the CD12 trial was a reduction of 6 days (p=.005). There is, however, a possibility that the trial would build in an interim data point, providing an earlier look. Regardless, what appears almost certain is that leronlimab consistently demonstrated a therapeutic benefit across the range of severe to critical COVID-19 patients and various subgroups.

The question really isn’t whether the drug worked or not. It’s more a question of sample size and number of patients enrolled. For patients’ sake, I hope the regulatory authorities move quickly to approve leronlimab based on this existing data, as continuing to wait months for more data will certainly come with the opportunity cost of saving lives.

Since the CD12 trial missed its primary endpoint, some may dismiss the secondary endpoints as a fluke, but the totality of the trial tells a much different story. There was a rush to finish the trial in order to save lives, so the interim analysis at 75% enrollment was never conducted. Per protocol, they would have been able to change their primary endpoint to time to recovery. Had CytoDyn taken this administrative action, the CD12 trial most likely would have met their primary endpoint and I probably wouldn’t be writing this article. The reality of the situation is that the FDA should be parsing through both the statistically relevant age-adjusted data and the hospital duration so that it can render its verdict on a conditional EUA.

Translating to Business Opportunities

There is another CCR5 blocker being studied for COVID-19 management: cenicriviroc (CVC), which was developed by Tobira Therapeutics. This company was bought by Allergan for $1.7 billion, and Allergan was acquired by AbbVie (ABBV). CVC (CCR5/CCR2b inhibitor) is being studied in the ACTIV IM trial which includes a TNF-a inhibitor, Remicade made by J&J (JNJ), and a costimulatory interaction (CD80/86 to CD28) inhibitor, which is made by Bristol-Myers (BMY). The ACTIV IM trial along with CytoDyn’s CD12 and CD16 trials could bring more interest into the CCR5 immunomodulatory space for COVID-19 and potentially other indications.

There are two main possibilities. One, leronlimab could be licensed to a big pharma that doesn’t have any ownership of a CCR5 inhibitor. Second, leronlimab could be licensed to a big pharma company that has a CCR5 inhibitor and the knowledge about how they work, but wants access to a CCR5 inhibitor with a better pharmacokinetic (PK) and tolerability profile, i.e., leronlimab.

Big pharmas that have CCR5 inhibitors are the aforementioned AbbVie, Pfizer (PFE), which owns maraviroc, and Merck (MRK), which owns vicriviroc. Vicriviroc’s development was discontinued over a decade ago based on HIV results. All of these drugs are small molecules with inferior tolerability and PK profiles. None of them seem to work as well in HIV because of the inferior PK profile, likely inferior receptor occupancy (all receptors need to be blocked to stop HIV infection of new cells), and the fact that HIV might still somehow mutate and bind to CCR5 when it is not being blocked by a larger antibody such as leronlimab.

Either way, these companies might have an interest in getting a head start in COVID-19 treatment with a better CCR5 inhibitor, and leronlimab is their only avenue to do so.

The Short Sellers

One last thing to address is the short sellers, again.

An author on Seeking Alpha named Paulo Santos claimed that since CytoDyn took a look at the Standard of Care group (Leronlimab or placebo plus Prior or Concomitant COVID-19 Treatment) — this yielded statistical significance when age-adjusted — then it would be fair to look at all the patients that fell under the Leronlimab alone or placebo alone group. This claim is totally fair, and the result looks pretty bad:

Wait! Does this mean that leronlimab increases the chances of death if taken alone? No! What the author failed to look at is how many critical and >65-year-old patients were in each group. Since it is a smaller subgroup, there’s a greater chance that these groups are not well matched. So, let's look at the subgroup statistics:

No Standard of Care Groups

Leronlimab

Placebo

(no pharmacological treatment)

Proportion of Critical Patients (%)

7/54 (13%)

1/21 (4.8%)

Proportion of Severe Patients (%)

47/54 (87%)

20/21 (95.2%)

Proportion of >65-Year-Old Patients

10/54 (29.6%)

3/21 (14.3%)

Proportion of Fatalities (%)

19/54 (35.2%)

3/21 (14.3%)

The leronlimab group had over twice as many >65-year-old patients and almost three times as many critically ill patients. The baseline characteristics are nowhere near matched, and the subpopulation sample size is relatively small, so this is really no surprise, nor should it be concerning.

It is also no surprise that this subgroup was not purposefully balanced because the standard of care was introduced after the trial commenced. Therefore, it is likely that many of these patients may have been enrolled early in the trial before remdesivir and dexamethasone, and other therapies became part of the SOC. To refuse SOC for patients just for the sake of evening out the baseline characteristics of this subgroup would be unethical.

Conclusion

The CD12 clinical data is solid when you parse the data, and whether CytoDyn needs to gather more data or CytoDyn will get an EUA in one or more countries is unclear. However, what is clear is that the situation is now de-risked. Before this dataset, investors were betting on strong anecdotal evidence in over 60+ eINDs and the mild to moderate CD10 trial data which showed how the drug performed in the two extremes (i.e., moderate and critical cases). What investors were missing was the meat of the patient population that leronlimab is most likely able to serve. Now, investors have a placebo-controlled study to support their opinions on the drug’s efficacy, as well as its eventual FDA registration. The three most likely outcomes are:

1) one or more EUAs,

2) waiting for CD16 data, where CD12 has already de-risked the situation, or

3) a combination of both 1) and 2) (approval in one country and further data gathering for the FDA)

Approval (full or EUA) and sales could quickly bring in tens to hundreds of millions in revenue for CytoDyn, which would likely help fund its pipeline and would bring a lot of awareness of leronlimab to physicians.

This article was written by

Vision and Value profile picture
1.51K Followers
Investor and small business owner, mechanical engineer

Disclosure: I am/we are long CYDY. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Recommended For You

Comments (526)

To ensure this doesn’t happen in the future, please enable Javascript and cookies in your browser.
Is this happening to you frequently? Please report it on our feedback forum.
If you have an ad-blocker enabled you may be blocked from proceeding. Please disable your ad-blocker and refresh.