The international clinical trials registry platform (ICTRP): data integrity and the trends in clinical trials, diseases, and drugs

Abstract

Introduction: Clinical trials are the gold standard for testing new therapies. Databases like ClinicalTrials.gov provide access to trial information, mainly covering the US and Europe. In 2006, WHO introduced the global ICTRP, aggregating data from ClinicalTrials.gov and 17 other national registers, making it the largest clinical trial platform by June 2019. This study conducts a comprehensive global analysis of the ICTRP database and provides framework for large-scale data analysis, data preparation, curation, and filtering. Materials and methods: The trends in 689,793 records from the ICTRP database (covering trials registered from 1990 to 2020) were analyzed. Records were adjusted for duplicates and mapping of agents to drug classes was performed. Several databases, including DrugBank, MESH, and the NIH Drug Information Portal were used to investigate trends in agent classes. Results: Our novel approach unveiled that 0.5% of the trials we identified were hidden duplicates, primarily originating from the EUCTR database, which accounted for 82.9% of these duplicates. However, the overall number of hidden duplicates within the ICTRP seems to be decreasing. In total, 689 793 trials (478 345 interventional) were registered in the ICTRP between 1990 and 2020, surpassing the count of trials in ClinicalTrials.gov (362 500 trials by the end of 2020). We identified 4 865 unique agents in trials with DrugBank, whereas 2 633 agents were identified with NIH Drug Information Portal data. After the ClinicalTrials.gov, EUCTR had the most trials in the ICTRP, followed by CTRI, IRCT, CHiCTR, and ISRCTN. CHiCTR displayed a significant surge in trial registration around 2015, while CTRI experienced rapid growth starting in 2016. Conclusion: This study highlights both the strengths and weaknesses of using the ICTRP as a data source for analyzing trends in clinical trials, and emphasizes the value of utilizing multiple registries for a comprehensive analysis.

Keywords: DrugBank; ICTRP; clinical trials; hidden duplicates; trends; trends ICTRP.

FIGURE 1

Pre-processing steps of the deduplication process. Hidden duplicates are duplicates not marked by the ICTRP platform as duplicates. To identify them we matched trials with an identical secondary ID, target size, primary sponsor, NIH drug name, or a matched compressed and lowercased intervention of the interventions column (for non-matched records). We also grouped trials that had at least 80% % similarity of concatenated public and scientific titles which was calculated using Levenshtein distance. Trials that did not contain the same EUCTR prefix and did not have similar titles were checked manually. Among this dataset, we then searched for trials with different dosages, phases and conditions and considered them as non-duplicates. All the left trials 833 were considered to be hidden duplicates. The final dataset for further analysis will contain trials with the earliest registration date and/or first alphabetical ID. Below the dashed line, we briefly summarized the databases analysis process. Using DrugBank and NIH databases we identified the exact agents and drug categories used in clinical trials. We then collected statistical data on identified agents, including unique agents for further analysis of general trends in either clinicaltrials.gov or non-clinicaltrials.gov databases. Such analysis also allows us to compare the efficacy of NIH and DrugBank drug databases by comparing the number of identified drugs using each of them.

FIGURE 2

(A) Annual numbers of registered trials between clinicaltrials.gov and non-clinicaltrials.gov. CT—clinicaltrials.gov, non-CT—non-ClinicalTrials.gov. (B) Annual numbers of registered trials between separate registries in the non-clinicaltrials.gov group. Both clinicaltrials.gov and non-clinicaltrials.gov groups almost equaled by the year 2020 with a slight prevalence in the number of trials in the non-clinicaltrials.gov group. Such a result was achieved by a constant increase in trials registered in the non-clinicaltrials.gov databases that started in 2005. EUCTR (34 342 trials), IRCT (27 096 trials), CHiCTR (22 915 trials) and CTRI (21 739 trials) had the highest numbers of trials identified in the ICTRP database. The most rapid increase in the number of trials found in the ICTRP was with the CHiCTR database which almost reached 8 000 trials in 2020. Second and third place in the rate of new trial registration was found to be among CTRI and ISRCTN. Despite EUCTR having the highest number of trials in the ICTRP, we did not identify any notable increase in the number of trials. We also identified a decline in the number of trials registered in jRCT in the 2019—2020 period. CRiS—Clinical Research Information Service of Republic of Korea, ANZCTR—Australian New Zealand Clinical Trials Registry, jRCT—Japan Registry for Clinical Trials, IRCT—Iranian Registry of Clinical Trials, ISRCTN—The International Traditional Medicine Clinical Trial Registry, LBCTR—Lebanese Clinical Trials Registry, PACTR—Pan African Clinical Trials Registry, RPCEC—Cuban Public Registry of Clinical Trials, SLCTR—Sri Lanka Clinical Trials Registry, NTR—Netherlands Trial Register, CHiCTR—Chinese Clinical Trial Registry, TCTR—Thai Clinical Trials Registry, EUCTR—The EU Clinical Trials Register, CTRI—Clinical Trials Registry—India, DRKS—German Clinical Trials Register, REPEC—Peruvian Clinical Trials Registry.

FIGURE 3

Overlapping hidden duplicate groups. An overlap between the hidden duplicates in non- 861 clinicaltrials.gov sources and clinicaltrials.gov is illustrated on the left. A more detailed view of non-clinicaltrials.gov sources is shown on the right side of this figure. As is shown in this figure, most hidden duplicates were found in the non-clinicaltrials.gov databases (87%) while clinicaltrials.gov accounted only for 5%. EUCTR had 82.9% of all hidden duplicates in the non- clinicaltrials.gov group with all other registries (CTRI, DRKS, ISRCTN and PER) ranging from 0.1% to a maximum number of 1.8% of hidden duplicates. EUCTR indicates EU Clinical Trials Register; CT, ClinicalTrials.gov; CTRI, Clinical Trials Registry—India; DRKS, German Clinical Trials Register; ISRCTN, International Standard Randomised Controlled Trial Number Registry; PER, Peruvian Clinical Trial Registry.; Other, Thai Clinical Trials Registry (TCTR), Pan African Clinical Trial Registry (PACTR), Chinese Clinical Trial Registry (CHiCTR), Japan Primary Registries Network (JPRN).

FIGURE 4

Distribution of disease classes between registry groups. The neoplasms category was one of the top conditions mentioned in trials both in clinicaltrials.gov (26.4%) and non-clinicaltrials.gov (19.8%). However, Pathological Conditions, Signs and Symptoms (22.7%) in non-clinicaltrials.gov outscored Neoplasms, while in clinicaltrials.gov Neoplasms were the leading category. Cardiovascular Diseases were in third place in both groups (clinicaltrials.gov—8.7%, non-clinicaltrials.gov—7.5%). Obtained results indicate a little difference in conditions studied in clinical trials in non-clinicaltrials.gov databases and clinicaltrials.gov. Categories with percentages below 2% are not labeled.

FIGURE 5

(A) Top identified drugs and non-drug words in the Intervention field using DrugBank in clinicaltrials.gov and non-clinicaltrials.gov groups. (B) Top identified drugs and non-drug words in the Intervention field using DrugBank in clinicaltrials.gov database. (C) Top identified drugs and non-drug words in the Intervention field using DrugBank in non-clinicaltrials.gov databases. We identified a high number of non-drug words in the non-clinicaltrials.gov group including pea (5 537 trials), fica (5 083 trials), sage (3 242 trials), oat (2 489 trials), water (2 746 trials), rice (1 151 trials), date (1 090 trials), neon (612 trials), honey (512 trials) and corn (508 trials). These words were found in the Intervention field but also marked as drugs in the DrugBank database. The non-drug words were also present among the clinicaltrials.gov database but in smaller number and variety: sage (859 trials), fica (825 trials), oat (480 trials) and water (637 trials). Cisplatin was the drug with most trials in the clinicaltrials.gov group (1 908 trials). Widely represented categories of drugs in both (B,C) were anticancer therapy (cisplatin, cyclophosphamide, paclitaxel and bevacizumab) and anesthetics and analgesics (bupivacaine, ropivacaine, propofol, lidocaine and medetomidine). Among anticancer therapy, in the clinicaltrials.gov group, a notable part was dedicated to monoclonal antibodies (pembrolizumab with 1 153 trials or bevacizumab with 924 trials). Vitamin D and iron were present in both clinicaltrials.gov (783 trials) and non-clinicaltrials.gov (757 trials) groups.

FIGURE 6

NIH categories in (A) all databases, (B) clinicaltrials.gov and (C) non-clinicaltrials.gov databases. Unique NIH categories are represented in (D) all databases, (E) clinicaltrials.gov, (F) non-clinicaltrials.gov databases. Antineoplastics agents were widely represented in clinicaltrials.gov group with 15 309 trials while reaching only 8 873 registries in the non-clinicaltrials.gov group. Both groups had approximately the same number of unique agents with 245 unique antineoplastics clinicaltrials.gov drugs and 203 non-clinicaltrials.gov medications. Non-clinicaltrials.gov group comprised a large number of anesthetics (7 303 trials compared to 3 428 trials in clinicaltrials.gov). However, both groups differed only by 3 unique agents with clinicaltrials.gov having 32 drugs and non-clinicaltrials.gov having 35 unique drugs. In clinicaltrials.gov, anti-infective (4 568 trials) and anti-bacterial agents (3 979 trials) also played a definitive role. Such categories as antidepressive agents (924 trials in non-clinicaltrials.gov and 1 420 trials in clinicaltrials.gov), cardiovascular agents (788 trials in non-clinicaltrials.gov and 1 066 trials in clinicaltrials.gov) or central nervous system agents (864 trials in non-clinicaltrials.gov and 1 294 trials in clinicaltrials.gov) appeared to be slightly underrepresented in comparison to other categories with most trials being identified in clinicaltrials.gov database.

FIGURE 7

Drugs and non-drug words identified in the Intervention field using NIH id in (A) both groups, (B) clinicaltrials.gov database and (C) non-clinicaltrials.gov databases. Vitamin D was consistently present in all databases with 949 trials in non-clinicaltrials.gov and 829 in clinicaltrials.gov. Clinicaltrials.gov’s top drugs were either anticancer therapy [paclitaxel (1 256 915 trials), pembrolizumab (932 trials), docetaxel (888 trials) and bevacizumab (1 042 trials)] or antidiabetics with 915 trials on metformin and 893 on insulin. Anesthetics and analgesics were one of the most abundant groups in non-clinicaltrials.gov databases comprising ether (2 265 trials), lidocaine (954 trials), ropivacaine (1 087 trials), propofol (885 trials), fentanyl (542 trials), ketamine (486 trials) and morphine (383 trials). The two most commonly studied drugs among the non-clinicaltrials.gov group were ether (2 265 trials) and glucose (1 232 trials). In contrast to DrugBank, we did not identify any non-drug words in the clinicaltrials.gov top drugs using NIH. Non-drug words were present in the non-clinicaltrials.gov database but were notably smaller in variety and numbers of trials than in the same group in DrugBank. Among non-drug words we identified levan (923 trials), tempo (441 trials) and curcumin (338 trials).

FIGURE 8

Trials identified using NIH id and DrugBank in (A) clinicaltrials.gov and (B) non-clinicaltrials.gov group. DrugBank identified more trials than NIH id in both groups (31.6% in clinicaltrials.gov and 39.8% in non-clinicaltrials.gov). Both NIH (28.5%) and DrugBank (39.8%) showed better performance among non-clinicaltrials.gov registries (in comparison to 31.6% DrugBank and 25.8% NIH in the clinicaltrials.gov group). Therefore, more trials were found in non-clinicaltrials.gov registries with the None category being remarkably reduced compared to the clinicaltrials.gov group (from 42.6% to 31.7%). In terms of exact number, clinicaltrials.gov group contained more than 140 000 not identified trials, while non-clinicaltrials.gov did not reach 100 000.