Healthcare workers transfer the body of a person who died from COVID-19, at a hospital in Kolkata, May 12, 2021. Photo: Reuters/Rupak De Chowdhuri
Mumbai: On May 1, the Drug Controller General of India (DCGI) approved a drug called 2-deoxy-d-glucose (2-DG) for emergency use among people with moderate and severe COVID-19, to help manage the disease.
This drug was jointly developed by researchers at the Institute of Nuclear Medicine and Allied Sciences, which falls under the Defence Research & Development Organisation (DRDO), and the pharmaceutical giant Dr Reddy’s Laboratories.
In line with the DCGI’s approval for favipiravir, itolizumab and Verafin, the approval for 2-DG is based on poor evidence.
2-DG is a modified glucose molecule that has been found to have some therapeutic value as an anticancer and antiviral agent. Research on 2-DG goes as far back as 1956, although it hasn’t been approved to treat any other diseases yet. It is currently mostly used in diagnostic testing and research-related activities.
The Wire Science couldn’t find any preprint or peer-reviewed research paper uploaded by the DRDO and Dr Reddy’s team on 2-DG clinical trials vis-à-vis COVID-19. Instead, we had to rely on publicly available information, like a press release – from the Ministry of Defence! – and clinical trial registrations, to access the quality of the evidence.
No promise of success
In news reports about the 2-DG approval, the most widely used image is from an in vitro study of 2-DG against SARS-CoV-2. In vitro refers to studies performed outside a biological entity – like the human body or humanised mice. Studies conducted inside a biological entity are called in vivo.
(As it happens, according to one preprint paper, members of the Patanjali Research Institute and others suggested the use of 2-DG last year. It was based, of all things, on a computer simulation – in silico.)
This image, which the government shared in a press release, shows that cell cultures in a laboratory without 2-DG had more viral plaques – clear spots indicating cell damage by the virus – compared to the ones with 2-DG. These studies were conducted at the Centre for Cellular and Molecular Biology, Hyderabad.
While these experiments show that 2-DG can inhibit viral growth, they tell us little about its efficacy in humans.
For example, a study published in August 2020 found that around 90 drugs that had been approved by the US Food and Drug Administration had antiviral activity against SARS-CoV-2 – as did ivermectin, hydroxychloroquine, chloroquine, doxycycline, azithromycin and lopinavir. But none of them has been found to have any meaningful effect in human trials with COVID-19 patients.
Setting targets after firing the gun
Based on the results from this in vitro trial, the national drug regulator had approved a phase 2 trial for 2-DG – possibly in May 2020.
According to the press release, “Phase 2a was conducted in six hospitals and Phase 2b (dose ranging) clinical trial was conducted at 11 hospitals all over the country. Phase-2 trial was conducted on 110 patients.”
The Wire Science couldn’t find the corresponding entries for any of these trials in the Clinical Trial Registry of India (CTRI). The reason for this discrepancy is unclear; emails to the principal investigators of both studies hadn’t elicited a response at the time of publishing this article. The only phase 2 trial registered for 2-DG involved 40 patients across 12 sites.
From 2009, the DCGI mandated all human clinical trials in the country to be registered on CTRI in advance. As such, if the 2-DG trials haven’t been registered, they would be in violation of the Indian Council of Medical Research’s ethics guidelines.
Along with making sure that a study has been approved by an ethics committee, registering a trial on the CTRI before it begins also ensures researchers declare what parameters they will be measuring in their study. This is a safeguard against unscrupulous researchers measuring many different parameters and only reporting the ones that support their hypothesis.
The DRDO and Dr Reddy’s conducted the 2-DG phase 3 trial in 220 patients at 26 sites around the country. (The DRDO is listed as the trial’s primary sponsor and Dr Reddy’s the secondary.) However, the CTRI registration of this trial does not mention which parameters the trial researchers plan to measure.
For example, the phase 2 registration says that the trial’s primary endpoint – the main objective of the study – would be to measure the improvement of trial participants on a 10-point scale. The secondary endpoint includes around 15 other measurements, like mortality, improvement in symptoms, time spent with supplemental oxygen, etc.
So the trial will be deemed to be successful only if the researchers measure a significant positive change on the primary endpoints – which in 2-DG’s case the researchers were marking on the 10-point scale.
The CTRI registration of the phase 3 trial, however, doesn’t disclose what the primary endpoints were.
According to the press release, in the phase 3 trial, the participants who got 2-DG had better ‘symptom improvement’ and spent less time receiving supplemental oxygen. But since we don’t know if these two parameters were primary endpoints of the phase 3 trial or just one of the many secondary endpoints, we can’t know if the trial was a success or if the release is only reporting the trial’s favourable findings. Also note that ‘symptom improvement’ and oxygen dependence were secondary endpoints in the phase 2 trial.
Dr S.P. Kalantri, a professor of medicine at the Mahatma Gandhi Institute of Medical Sciences, Sevagram, told The Wire Science that failing to declare which analyses the researchers had decided to perform in advance destroys scientific rigour. “The data analysed in such studies is akin to a fishing expedition — trying to find out if something worthwhile emerges from the data. More often than not, such associations are by chance and are spurious.”
Room for prejudice
Next, consider the results from the phase 2 trial of 2-DG. The primary endpoint of this trial was the number of days required for the patients to score 4 or less on the WHO’s 10-point scale to measure a COVID-19 patient’s clinical status.
There are two problems. First, this is a surrogate measurement that has little clinical value for patients. “The study does not tell us if the drug is capable of preventing ICU admissions, the need for mechanical ventilation or reducing deaths,” Dr Kalantri said – and these are outcomes “that matter to the people”. Instead, the scale only tells us how a patient might do on a subjective scale.
“For a patient with severe disease, improvement in symptoms – like cough, fever – or reduction in oxygen use, is immaterial if the survival does not improve,” Dr Sahaj Rathi, of the Institute of Liver and Biliary Sciences, New Delhi, said. “COVID-19 is a fatal but finite disease: one either survives it or not – the outcome is clear before discharge from hospital in nearly all cases. The outcomes which truly matter are mortality reduction in severe disease and preventing hospitalisation in mild disease.”
Second, this 10-point scale is highly, and famously, prone to bias. There is no one-to-one matching between a patient’s symptoms and their score on the scale. It’s up to the doctors involved in the trial: they decide who gets a 4 and who gets a 6. And when it’s not perfectly possible to justify each decision, there is ample room for prejudicial decision-making.
For example, clinicians who participate in trials that are sponsored by the companies making the drug are paid significant sums of money for their time. This compensation can in turn encourage them to undermine the results in favour of the company.
To shield trials from such bias, they are often blinded: the clinician recording the data doesn’t know which patients got the drug and which didn’t. Blinding is important when researchers are measuring a subjective parameter – like a patient’s score on a 10-point scale.
Both the phase 2 and phase 3 trials of 2-DG did not blind the clinicians.
The results
DRDO and Dr Reddy’s have not published any manuscripts describing the trials either in preprint repositories or scientific journals. The little information that we have about the efficacy of these drugs come from the government press release.
For phase 2, the press release says the researchers reported a “significantly favourable trend”: that the vital signs of those who received 2-DG returned to ‘normal’ 2.5 days sooner on average versus those who didn’t receive 2-DG. (We don’t know what is ‘normal’ here either.) There are two problems here.
First, normalisation of vital signs is one of the 15 secondary endpoints – not a primary endpoint. There is no information in the release about the primary endpoint nor the other 14 secondary endpoints.
Second, the word “significantly favourable” and “significantly higher” have been used to describe the results of the phase 2 and 3 results. The phrase ‘statistically significant’ can’t be a throwaway term. To be statistically significant means a particular measurement is too large to be the result of chance. And to claim a result is significant in this way, researchers typically perform specific statistical calculations to prove their point.
Without seeing these calculations, it’s impossible to say if the use of the term “significantly” in the press release alludes to significance of the statistical variety or the propagandist one. The press release also doesn’t use the word “statistically”.
Further, the DCGI relies on the recommendation of the Subject Expert Committee (SEC), a group of independent experts, to grant COVID-related approval. On October 29, 2020, the SEC asked[footnote]PDF download[/footnote] Dr Reddy’s to add “mortality at 28 days” as one of the efficacy endpoints in the phase 3 trial. But while the press release says the drug was found to be efficacious in the phase 3 trial, it doesn’t say anything about the mortality endpoint.
Safety
Next, a drug’s efficacy is only as good as its safety profile. While there have been several human trials for 2-DG over the years, the drug hasn’t been approved for human use before this month.
According to the CTRI entries, 2-DG’s dosage in its phase 2 trial was 45 mg per kg of body weight in the morning and 18 mg/kg in the evening. But in phase 3, this was increased to 45 mg/kg in the morning and 45 mg/kg in the evening – for a total of 90 mg/kg per day.
A smaller study by researchers in the US, published in September 2010, tested the drug’s effects among 12 cancer patients. They found that 60 mg/kg of 2-DG per day was shown to cause QT prolongation – a severe cardiac condition. As Priyanka Pulla has reported earlier:
The electrical signal that contracts the heart muscles, thus allowing them to pump blood, passes through the heart once every heartbeat and shows up on an ECG as a series of waves labelled P,Q, R, S and T. When the time delay between the Q and T waves becomes longer than it should be, the condition is called long QT syndrome. This syndrome can render the heart beat chaotic in some people. This arrhythmia can in turn lead to a sudden cardiac arrest.
Another study, published in December 2012, found that ingesting 63-88 mg/kg of 2-DG per day could, among other things, increase the person’s blood sugar levels.
The press release also doesn’t say anything about the drug’s safety profile, as ascertained in the phase 2 or phase 3 trials. Phase 3 trials in particular are crucial to understanding any drug’s or vaccine’s long-term safety. (This is one of the reasons the DCGI’s approval for Covaxin without any data from its phase 3 trial proved so controversial.) And of course, it’s impossible to discover any rare side-effects in a trial with only 220 participants; tens of thousands had to have been enrolled instead.
From the in vitro experiments to the phase 3 trials, which were sponsored by DRDO, public money funded 2-DG’s development. And because the trials in particular were so poorly designed, and poorly reported too, 2-DG’s trajectory during the COVID-19 represents a waste of the country’s resources – and a betrayal of the people’s trust in their public institutions.
“There is minimal information about how the trials were conducted, as there is no data or publication in the public domain. This is not a new or proprietary molecule, and this is publicly funded research. The data should be in the public domain,” Dr Rathi said.
Requests for comment sent to the DCGI’s office, the DRDO and Dr Reddy’s hadn’t elicited replies at the time of publishing. This article will be updated as and when any of them respond.
Note: This article was updated at 10:30 am on May 13, 2021, to include Dr Sahaj Rathi’s comments, and at 1:07 pm on May 15, 2021, to include the SEC’s request to Dr Reddy’s.
Note: This article was updated at 6:13 pm on May 21, 2021, to note that 2-DG was approved on May 1. The previous version mistook the date of approval with the date of announcement of the approval.
Ronak Borana is a science communicator based in Mumbai. He tweets at @ronaklmno.