This study was a cross-sectional study conducted in the period between June and October 2023, with data collection from June to August 2023. We searched which of the 60 antineoplastic and supportive drugs on the WHO EML were imported from China National Medical Products Administration (11) (NMPA). Then, according to the manufacturer of the drug imported in China, we went to the Pharmaceuticals and Medical Devices Agency (PMDA), and the DailyMed to find the drug label of the same manufacturer (CSL STYLE ERROR: reference with no printed form). If there is a drug from the same manufacturer that is not available in Japan or the US, we use the Japanese Orange Book or the FDA Orange Book to find the reference listed drug (RLD) labels to replace it (12, 13). The labels must be approved in NMPA and PMDA or the US FDA. Labels must be accessible in the public domain.

We performed a direct comparison of the proportion of all information given to each safety information across the same drug and the proportion of the number of labels with box warnings or warnings to that of all labels in the three countries. For example, the proportion of adverse reaction information was calculated by dividing number of adverse reaction information words by the total number of safety information words. We defined the measures independently of language: for Chinese and English, we counted the number of words, and for Japanese, we counted the number of letters in the safety sections, and then divided this by the total number of words or letters on the label. Each number of safety information words was calculated based on the sections listed in Table 1 from each country.

Pediatric use is often omitted in oncology labels. We narrowed the analysis to drugs with explicit pediatric information (dosage/age) in ≥1 country. We conducted a comparative analysis of pediatric medication information for 19 medicines across three countries, including information such as pediatric medication dosage and age on the label. If there is information on this drug for children, we mark it as 1, and if the trial was not conducted or the safety is unclear, we mark it as 0. Finally, we added up the scores of all 19 medicines in each country, and the higher the score, the more complete pediatric medication information in that country (Table 2).

Warnings convey important safety information in eye-catching fonts on the first page of labels, alerting healthcare professionals and patients to medication risks. We identified a boxed warning (BW) on Chinese, Japanese, and US drug labels through a manual search and measured PBW on all labels.

There is a possibility that the volume of text is different, but the information conveyed is the same (14). To understand the differences in the information provided across the three countries, we read the content of each safety information, gave the complete content a score of 1, the incomplete or missing content a score of 0, and then summed the scores of each item; the higher the score, the more complete the content of each safety information (Table 3).

Descriptive statistics were performed on the data obtained. Data were presented as the mean and standard deviation (SD). One-way analysis of variance (One-way ANOVA) was performed to find out the differences between groups followed by Scheffe’s post hoc test to determine which means were different with a level of significance set at p < 0.05. Data were analyzed using Statistical Analysis Software (SPSS 26.0).

Table 4 shows the proportion of each safety information on drug labels in three countries. The proportion of each safety information was significantly different across the countries except for the proportion of geriatric use information.

The proportion of adverse reaction information on Chinese labels {23.1 (10.2) [mean (standard deviation)]} and US labels (19.0 [8.0]) was higher than on Japanese labels [13.9 (6.9)]. The proportion of contraindication information on Japanese labels [1.4 (1.4)] was higher than on US labels [0.6 (0.8)]. The proportion of precaution information on Chinese labels [17.1 (0.9)] was higher than on US labels [10.8 (7.5)]. The proportion of pregnancy and lactation information on US labels [3.9 (2.1)] was higher than on Japanese labels [2.5 (1.5)]. The proportion of pediatric use information on US labels [2.0 (2.5)] was higher than on Chinese labels [0.8 (0.8)]. The proportion of drug interaction information on Chinese labels [5.0 (4.8)] and Japanese labels [5.2 (4.8)] was higher than on US labels [1.5 (1.8)]. The proportion of overdose information on Chinese labels [1.9 (1.9)] was higher than on Japanese labels [0.2 (0.9)] and US labels [0.9 (1.0)]. The proportion of pharmacology and toxicology information on Chinese labels [6.0 (4.0)] was higher than on Japanese labels [3.8 (2.6)]. The proportion of total safety information on Chinese labels [58.6 (12.6)] was higher than on Japanese labels [40.6 (12.9)] and US labels [43.7 (10.4)] (Figure 1).

Table 5 shows a detailed comparative analysis of the content of each safety information. We sum the scores of each safety information and then get the total score for each item in three countries (Table 3). Chinese labels achieved the highest scores in overdose, pharmacology, and toxicology, suggesting that these sections are more comprehensively labeled. Japan scored highest in contraindications, precautions, and drug interactions, while the US received the highest scores in adverse reactions, pregnancy and lactation, pediatric use, and geriatric use. Overall, Chinese drug label safety information total scores are higher than those in the US and Japan.

Table 6 shows the proportion of the number of labels with BW in China, Japan and the US. PBW on the Japanese labels (85%) was larger than that on the US labels (55%) and Chinese labels (21%). Concordance of the presence or the absence of a BW on the label between China, Japan, and the US was 36% on the total labels.

We compared and analyzed the differences in pediatric medication information for 19 medicines included in the WHO EMLc in three countries (Table 2). Of the 19 drugs, Japan had the most pediatric drug information with 13 drugs, followed by the US with nine drugs and China with eight drugs. And seven out of the 19 drugs have pediatric medication information available in all three countries.

We found significant differences in the proportion of each safety information provided on the same drug label in the three countries, except for geriatric use information. Chinese labels have significantly more proportion in more safety information items than other countries, followed by the US and then Japan, and Chinese labels’ proportion of total safety information is significantly more than the US and Japan. In a previous study conducted on the labels of new molecular entities between the US, the UK, and Canada, it was found that significant differences among the mean proportion of total safety information (PSI) of all three countries (14).

In our study, we also found that there are significant differences among the mean PSI and proportion of many safety information items of all three countries, which is most likely due to the different guidelines set by these authorities for the preparation of drug labels (15–17), as different laws in different countries may influence regulatory decisions (18).

The proportion of each safety information was generated by counting words/characters, which are crude measurements that do not always indicate critical safety issues (5). So we compared the content of each label’s safety information and found that there were indeed differences between labels in three countries. We found that except for precautions, Chinese labels have significantly more words than Japanese labels, but the score is lower than that of Japan, that is, the information integrity is lower than that of Japan, and the other safety information is higher the number of words, the higher the score.

The Japanese guidelines for writing package labels require precautions to be divided into [precautions related to indications], [precautions related to use and dosage], [important basic precautions], [other precautions], [operational precautions], etc., while the Chinese guidelines for writing package labels do not require precautions to be divided, which may lead to a more complete content of precautions in Japanese package leaflets (15, 16).

We went on to study the section related to adverse reactions in China, Japan, and the US. The proportion of adverse reaction information on Chinese labels and US labels was higher than on Japanese labels and the adverse reaction score on Chinese labels and the US labels was also higher than those in Japanese labels. For the drug mesna, the proportion of adverse reaction information on Chinese labels (0.37) and US labels (0.12) was higher than on Japanese labels (0.05), and the Chinese label contained 11 adverse reactions, the US 15, and Japan only 10.

In the pregnancy and lactation, pediatric use, and geriatric use sections, we also found that the US, which had more words, scored higher. We found that the Chinese and Japanese labels contained vague information about the drug details for certain groups, including that the trial was not conducted, that there were no references, that it was not clear, and so on. For example, there is no data on the presence of everolimus or its metabolites in human milk, the effects of everolimus on the breastfed infant, or on milk production. The lack of human data relevant to pregnant and lactating populations, children and the elderly has long been identified as an area of significant public health need (19), mainly because the current patients participating in pre-marketing clinical trials of drugs are highly selected patients, only qualified patients can participate in clinical trials, and there will be a lack of clinical data for special populations (20).

The warning is displayed in a prominent font below the title of the drug label, which is about serious adverse reactions to medicines and their serious safety concerns, covering items such as contraindications and precautions. Our comparison shows that Japan has a higher proportion of BWs, followed by the US and China. In addition, the concordance of the presence or the absence of a BW on the label between China, Japan, and the US was 36% on the total labels, indicating a low level of coordination between the three countries, which could have implications for medical professionals and patients (3).

Both the US and Japan have stricter requirements for warnings, and contraindications or serious warnings, especially those that could lead to death or serious injury, must be included in boxed warnings. China’s Provisions for Drug Insert Sheets and Labels stipulate that drug manufacturers can take the initiative to add warnings to drug specifications or labels, but the regulations are not sufficiently clear on the legal responsibility for regulating warnings and do not stipulate that regulatory authorities, production and management companies have their responsibilities and must do so. The previous study found that the use of febuxostat decreased significantly in the US after the FDA added a boxed warning to the label for febuxostat (21). Physicians may have responded to the FDA’s warning by prescribing AAPs less often and adopting, instead, alternative therapeutic strategies for managing NPSs in elderly individuals with dementia (22). Therefore, many Chinese pharmaceutical companies do not actively add warnings to their drug labels so as not to affect drug sales.

We found that only 7 of the 19 WHO EMLc drugs in three countries had pediatric medication information on their use in children, confirming previous research that pediatric drug use is often missing due to lack of evidence or regulatory delays in label updates, which may put children at higher risk for side effects, adverse drug reactions, and medication errors (23).

We also found differences in pediatric dosing regimens of the same drug in different countries, for example, cytarabine for children with acute leukemia, China uses body surface area to convert drug doses, while the United States and Japan use body weight to convert drug doses. Historically, however, pediatric dosing regimens have often been derived from pharmacokinetic data in adults (24). This approach, although clinically widespread, is not supported by solid empirical evidence and may result in neonates and children being exposed to inappropriate systemic drug levels (25).

While pediatric drug development is international in scope, pediatric drug labeling information is not uniformly shared and is under the control of national regulatory authorities (4), requiring countries to increase cooperation, information sharing, dissemination, discussion, and efforts to reach consensus on regulatory “standards” for quality pediatric medicines.

We acknowledge several limitations of our study. The most important limitation was the generalizability of our findings in antineoplastic and supportive medications to other types of medications. Also, this study included few drug samples, which may have affected the accuracy of the results.