Clinical trial registration in India: 12% of drug regulatory trials are not registered, as required by law

Abstract

Background

Amongst other things, the drug regulator, Central Drugs Standard Control Organization (CDSCO), is responsible for regulating clinical trials that underlie drug approvals in India. Since 2009, CDSCO has mandated that all regulatory clinical trials be registered with the Clinical Trials Registry – India (CTRI). In this study, we aimed to determine whether regulatory trials, for which CDSCO had given permission, were registered with CTRI as required. We also aimed to quantify missing CTRI records, if any.

Methods

This study involved regulatory trials, for which CDSCO permission letters were available. The permission letters, available as portable document format (PDF) files, were analysed to extract trial titles and protocol numbers. These data were then used to search the CTRI database, which was initially downloaded on 19 January 2024 and updated in August 2024, for matching records. The matches were confirmed by cross-referencing titles and protocol numbers, ignoring protocol version variations.

Results

Of the 381 trials examined, 335 (88%) had corresponding CTRI records, whereas 46 (12%) did not.

Conclusions

This study highlights that not all regulatory trials approved by CDSCO are registered with CTRI, despite legal requirements. The absence of a significant number of trials from CTRI raises concerns about transparency in the clinical trial enterprise in India. To improve trust and regulatory compliance, it is essential to establish explicit linkages between CDSCO and CTRI records to avoid the challenges of identifying matches. This study contributes to ongoing efforts to increase transparency and accountability in clinical trial registration in India.

Background

Clinical Trials Registry – India (CTRI) is the registry set up by the Government of India that is recognized as a primary registry of the WHO. The number of records in the registry increased from 29 in 2008 [1] to 8969 on 30 June 2017 [2], and to 71 444 records on 29 July 2024 [3]. It is now the third largest registry, in terms of the number of records, following the United States’ ClinicalTrials.gov (503 188 records) and the Chinese Clinical Trial Register, also called ChiCTR (85 723 records), both as of 29 July 2024. We determined this ranking of the registries by inspecting all the primary registries [4] and the data providers [5] to the WHO’s International Clinical Trial Registry Platform [6].

The Indian drug regulator is called the Central Drugs Standard Control Organization (CDSCO). It is headed by the Drug Controller General of India (DCGI). We will use the term regulator to refer to either CDSCO or the DCGI. The regulator has a range of responsibilities, including regulating the importing of drugs, their manufacture, their licensing and the clinical trials that underpin drug approval. These clinical trials are called regulatory trials. In 2009, CDSCO made it mandatory that trials be registered with CTRI [7]. This was reiterated in the New Drugs and Clinical Trials Act, 2019 [8]. Furthermore, since 1 April 2018, all trials need to be registered with CTRI prospectively, and although there have been a few exceptions [9], those trials that have been registered have largely been prospectively registered. Nevertheless, to the best of our knowledge, there are no penalties for non-registration, and also no process for confirming that a given trial has been registered.

The regulator issues a letter of permission to the sponsor to run a trial, and these letters, which we term the permission letters, are in the public domain [10]. The records of all studies registered with CTRI are also public [3]. The regulator permits a particular trial to run, and this trial is then registered with CTRI. As such, both the regulator and CTRI are concerned with a given regulatory trial. However, there is no cross-referencing of the records such that it is clear that the trial permitted by the regulator has a complementary record in the CTRI database. Identifying the correct match – that is, which CTRI record corresponds to a particular permission letter by CDSCO – is non-trivial [11].

In this study, we wanted to identify the matching CTRI record for a set of trials for which we were able to access the permission letters. We wished to determine whether each permission letter had a corresponding CTRI record, and if not, to quantify the missing records. This study contributes to efforts to increase transparency around the clinical trial enterprise in India.

Our study is a mini-audit of data held by CTRI and the Indian drug regulator. Other researchers have similarly compared data on trials registered on two or more platforms such as a registry, a publication or the United States Food and Drug Administration (USFDA). We provide some examples. The first regards trials registered in two registries; Fleminger and Goldacre (2018) audited the completion status of about 10 500 trials registered with both the United States’ ClinicalTrials.gov and the European Union Clinical Trials Register, and found that nearly a third of the records in the latter registry were discrepant from the American records [12]. The second regards trials with records in a registry and with the USFDA; Schwartz et al. examined 100 trials that were tied to particular drug approvals. They found the reported primary outcomes were largely the same in the ClinicalTrials.gov website and in the records of the USFDA [13]. The third regards trials with records in a registry and a publication; a 2020 study of the results of 94 trials, published in high-impact-factor journals, determined that there were one or more discrepancies in 74% of the trials [14]. Finally, the last regards trials with records in a registry, a publication and with the USFDA; Pradhan and Singh (2018) audited 15 pivotal trials of randomly selected new molecular entities (USFDA approved) in the period from 2013 to 2015. The researchers compared the data relating to serious adverse events and mortality in these trials across all three platforms and found significant discrepancies [15].

As mentioned above, since 2009 it has been mandatory that all trials be registered with CTRI. As the CDSCO permission letters used in this study were issued between January 2020 and April 2023, inclusive, the trials mentioned in these letters were required to be registered with CTRI. If it is found that not all regulatory trials have been registered with CTRI, this is a transgression of the law.

Methods

The CDSCO permission letters were issued between January 2020 and April 2023, inclusive, and are available as portable document format (PDF) files [10]. In previous work [11], we had wished to extract specific fields of information from a set of these letters. We downloaded 1000 PDFs on 29 May 2023, but were able to extract text from only 385 of them. Data extraction was carried out according to the protocol detailed in that work [11]. In the study that used 151 letters, for which we readily obtained CTRI matches of the permission letters, we wished to know whether the PIs and sites listed in the permission letters matched those listed in the CTRI records. In this study of the remaining 234 letters, we wish to know whether CTRI matches could be found for the permission letters..

However, in four cases the titles of the trials were not clearly stated in the permission letters. We removed those cases, leaving 230 for this study [16]. The steps taken to identify the trials registered with CTRI that corresponded to these 230 permission letters are outlined in Fig. 1. In brief, the CTRI database was downloaded on 19 January 2024, using the script in Additional file 1. This database was searched for the title of the trial as listed in the permission letter. If no match was found, the protocol number in the permission letter was used to search CTRI. In addition, even when matches were found using the title, we sought to confirm that the permission letter and the CTRI record related to the same study by comparing their protocol numbers. It should be noted that a version number frequently accompanies a protocol number. We ignored the version number when identifying matching protocol numbers.

Fig. 1

The process of identifying the CTRI records that match particular CDSCO permission letters

One author extracted all the text from the 230 PDFs. All subsequent steps were independently performed by at least two authors. The corresponding author subsequently manually confirmed all matches on the CTRI website. Some permission letters did not have CTRI matches. Between 9 and 15 August, the CTRI database was downloaded again, via the protocol detailed in Additional file 2. The non-matching trials protocols were checked against the freshly downloaded CTRI database.

Results

The process of identifying the CTRI records that match particular CDSCO permission letters is outlined in Fig. 1. The particular permission letters at each step and the CTRI number and URL (for matching cases) are listed in Additional file 3.

We successively used the titles and protocol numbers, available in the CDSCO permission letters, to search the CTRI database for counterparts for 230 trials. We found only 184 matches, with 46 not having a match.

When the protocols of the 46 non-matches were checked against the CTRI database after they were downloaded again in August 2024, no further matches were found.

Here, we provide a snapshot of these 46 trials. The therapeutic areas span a wide range. The largest number of trials was for coronavirus disease 2019 (COVID-19) (11 trials), followed by cancer (7 trials). There were three trials each for haemophilia and tuberculosis and two for diabetes. Other therapeutic areas included alopecia, cholera, liver cirrhosis, multiple sclerosis, ophthalmology and osteoporosis, among others. Regarding the sponsors, 26 were domestic companies, 1 an international research institute and 19 multinational companies. In terms of the phase of the trial, 7 were not stated, 6 were phase 1 (including 1/2), 15 were phase 2 (including 2/3), 13 were phase 3 and 5 were phase 4.

This information is presented in Table 1.

Table 1 Some details of the 46 trials with no matches in CTRI

Discussion

Before discussing the results, we wish to point out that the 230 trials were part of a larger set, of which 151 had CTRI matches, and had been studied earlier. When considering the entire dataset of 381 trials, the 46 non-matches accounted for 12% of the trials for which CDSCO had given permission. Therefore, our findings are that all regulatory trials have not been registered with CTRI. Why does this matter?

Transparency is an important tool for building trust. In the area of clinical research, various countries or regions have established registries [17], and the WHO [6, 18], governments [19], philanthropists [20], journals [21] and academics [22, 23] have all played important roles in increasing transparency. Most recently, in October 2024, the Declaration of Helsinki was updated to emphasize that the results of clinical trials have to be made public in a timely manner and to be accurate and complete [24]. Therefore, these efforts supplement efforts at increasing transparency around drug regulation, and this paper relates to these two interrelated areas.

CDSCO has taken a major step towards transparency by making the letters of permission that are issued to sponsors to run particular trials public. This practice started several years ago. Although there were challenges in extracting data from earlier letters, the quality of the PDFs is better now, and it has been possible to do so relatively easily from more recent ones. Similarly, the establishment of CTRI in 2007 was an enormous step towards increasing the transparency of the clinical research enterprise in India.

Other countries also grapple with issues of transparency. The FDA Amendments Act (FDAAA) in the United States [25] requires the registration of most trials, and there has been a big up-tick in such registration since the FDAAA came into force [26, 27]. In the UK, the Health Research Authority (HRA) requires that most trials be registered [28], although in 2018 it was noted that many trials had not been [29]. Therefore, the challenge of ensuring the registration of all trials is not unique to India.

Notably, in the United Kingdom, it is the Health Research Authority, not the sponsor, that now registers clinical trials in the registry ISRCTN. This ensures that all trials for which there is regulatory permission to run will be registered, unless there is an approved deferment [19]. CDSCO could consider such an arrangement.

There are some limitations of this work. First, it involves only 381 trials. It is impossible to say whether a similar percentage of so-called non-matches would be found if we had been able to extract the data from a larger set of permission letters. Although there were challenges in extracting data from earlier letters because of the poor quality of the PDFs, their quality is better now, and it has been possible to do so relatively easily from more recent PDFs. Second, when CDSCO gives permission to run a trial, it lists the title of the study. Usually, sponsors use the same title when they register a study with CTRI. However, this is not mandatory, and there may be some variations in the titles of the same study between the CDSCO and CTRI documents. Furthermore, CTRI records with very similar titles may refer to different trials, as evident from their different protocol numbers, for example. Therefore, it required considerable effort to identify the CTRI counterparts for the trials listed by CDSCO. Although two researchers searched for CTRI matches independently, it is conceivable that some matches were not found. This number is likely to be small. Third, we have no way to determine whether a given trial ever ran. However, since it costs Indian rupees (Rs) 200 000–300 000 to file an application with CDSCO [8], the number of such cases is likely low. Fourth, we did not check any of the other WHO-recognized registries to look for the trials that appear to not have been registered with CTRI, since the research question pertained to CTRI registrations alone.

As recommended earlier [11], it is essential to further increase the transparency of both CTRI and CDSCO by cross-referencing both sets of records, where relevant. Even the United States has not done this, although it recognized the importance of the issue and committed to doing so in 2018 (30). India could take the lead in this matter by making these connections explicit. In case the trial has been registered with CTRI before the permission letter is issued, then the CTRI number should be stated in the permission letter. If the trial has not yet been registered, then CDSCO should issue a second version of the letter stating the CTRI number after it is issued. However, it may be more convenient to simply have a list of all permission letters and the corresponding CTRI numbers on the CDSCO or CTRI website, which can be updated as soon as a trial is registered. This will require cooperation between CDSCO and CTRI. In turn, the permission letter name should be entered in a given CTRI record at the time of registering the trial, or the record can be updated once permission has been granted. There is already a field in the CTRI record entitled ‘Regulatory Clearance Status from DCGI’, and a subfield could be created that lists the name of the permission letter.

It is clear that not all the trials permitted by CDSCO are registered with CTRI. There could be a couple of reasons for this. As mentioned above, even though the sponsor applied to CDSCO for permission to run the trial, their trial may never have run, although we anticipate that the number of such cases was low. Second, even though it is a legal requirement to register a regulatory trial in CTRI, and no deferments are allowed, the sponsor may have neglected to do so. We found some cases where the exact version of the trial mentioned in the permission letter was not registered with CTRI. If the correct versions of some trials were not registered with CTRI, then it is conceivable that for some trials, no registration happened at all.

Conclusions

In this study with a set of 230 trials for which CDSCO had given permission, we wished to identify the matching CTRI record. A total of 46 trials did not have a match in CTRI. The 230 trials were part of a larger set, some of which we had studied earlier. Therefore, overall, 12% of 381 trials did not have a CTRI match. It is important that linkages between CDSCO and CTRI records are made explicit, and all regulatory trials are registered with CTRI, as required by law. Transparency is an important tool for building trust, and this study contributes to efforts to increase transparency around the clinical research enterprise in India.