Adverse event profile differences among long-acting gonadotropin-releasing hormone analogs: A real-world, pharmacovigilance study

Dysregulation at any level of the hypothalamic-pituitary-gonadal (HPG) axis leads to, or exacerbates, a range of hormone-dependent diseases. Given its crucial role as a regulator of the HPG axis, agonist and antagonist analogues of gonadotropin-releasing hormone (GnRH) have proven efficacy in treating these conditions. Additionally, GnRHa plays a pivotal role in assisted reproductive therapies [ 1 ]. GnRHa are available as rapid-acting or long-term depot preparations. The long-acting preparations available include: leuprolide, goserelin, histrelin, buserelin and triptorelin. They were first approved in the following order: buserelin (1984), leuprolide (1985), goserelin (1989), triptorelin (early 1990s), and histrelin (2005). These compounds have a high affinity for the pituitary Luteinizing Hormone-Releasing Hormone (LHRH) receptor and are resistant to enzymatic degradation. Long-acting GnRHa have been commonly used for the medical treatment of prostate cancer, breast cancer, precocious puberty, endometriosis, adenomyosis, also for the prevention of chemotherapy-induced premature ovarian failure in gynecological cancer [ 2 – 5 ]. Although long-acting GnRHa have a wide range of applications, the reported adverse reactions during the treatment of these diseases are also not uncommon ( Fig 1 ). In children with CPP receiving GnRHa therapy, multiple adverse events have been documented in clinical reports, including vaginal spotting/bleeding [ 6 ], subcutaneous nodules with sterile abscess formation [ 7 ], slipped capital femoral epiphyses (SCFE) [ 8 ], pseudotumor cerebri (PTC) [ 9 ] and hypertension [ 10 ]. When treating prostate cancer, the adverse reactions of long-acting GnRHa reported including life-threatening anaphylaxis [ 11 ], follicular mucinosis and mycosis-fungoides-like drug eruption [ 12 ], cardiovascular toxicities [ 13 ]. Moreover, multiple clinical evaluations have found many adverse reactions containing leukocytoclastic vasculitis [ 14 ], vasomotor symptoms and bone mineral density loss [ 15 ], low sexual desire, abnormal vaginal bleeding, and liver function damage [ 16 ] in the treatment of endometriosis with GnRHa. Long-acting GnRHa have been reported corresponding adverse reactions in various organ systems of the human body for different indications. The main affected organs include the pituitary gland, bones, cardiovascular system and reproductive organs, etc. In summary, GnRHa exhibit various adverse reactions in the treatment of various indications. However, due to the scarcity of cases, rigorous qualification standards, and constrained duration for monitoring and follow-up, clinical trials or case reports cannot fully reflect the safety profile of long-acting GnRHa in the real-world. FAERS, a public pharmacovigilance database that is utilized for collecting post-marketing safety data submitted to the FDA all over the world, serves as a valuable resource compensating for the limitations of clinical trials. Several studies have utilized this database with comparable methodologies, such as analysis of sildenafil [ 17 ], Everolimus [ 18 ], topotecan [ 19 ], acetylsalicylic acid [ 20 ] and a comparison of vinorelbine and vincristine [ 21 ]. These precedents support the feasibility and relevance of conducting our research on long-acting GnRHa using the FAERS database. Therefore, we utilized FAERS to evaluate the safety characteristics of long-acting GnRHa and provide a practical and secure reference for their clinical applications.

The population characteristics is presented in Fig 3 and S2 File . The reports of Long-acting GnRHa were submitted with an increasing trend year by year, mainly from 2019 to 2023 ( Fig 3A ). As for the age in reports ( Fig 3B ), the proportion of null was almost always the highest (except for buserelin). A higher proportion of the reports in leuprolide (23%) and goserelin (19%) were patients of 61–80 years old, while in triptorelin (15%) and histrelin (37%) was ≤ 18, and patients of 19–45 years old occupied the highest proportion (37%) in buserelin reports. In terms of role code ( Fig 3C ), these five drugs are predominantly identified as the primary suspect drugs in the reports, with the exception of buserelin, which was consistently classified as a secondary suspect drug in 100% of the cases. As showed in Fig 1D , AEs of these five drugs were reported simultaneously in prostate cancer with leuprolide owning the highest proportion, and goserelin was as the main drug reported for breast cancer. The reports of these five drugs were mostly submitted from the United States, while reports of buserelin primarily from Germany ( Fig 3E ). In terms of gender ( Fig 3F ), the proportion of female patients in buserelin, histrelin and triptorelin was higher than that of male patients, while the opposite results occurred on goserelin, leuprolide. The observed gender disparity in GnRHa utilization patterns may primarily stems from differences in approved indications and clinical practice norms. Goserelin and leuprolide are predominantly prescribed for prostate cancer (male-prevalent), whereas buserelin, histrelin, and triptorelin are more frequently used in CPP (female-predominant) and endometriosis. This may also explain why AEs related with histrelin and triptorelin are reported in a larger proportion of patients younger than 18 years of age. Furthermore, the 3-month depot formulation of leuprolide aligns with prostate cancer management protocols requiring sustained testosterone suppression. When it comes to clinical outcomes ( Fig 3G ), patients receiving goserelin demonstrated significantly higher mortality (29.3%) compared to other GnRHa agents, while patients receiving buserelin demonstrated significantly higher rates of hospitalization initial or prolonged (44.2%). (A) Date received, p  < 0.001 a ; (B) Age in report, p   0.05 a ; (D) Top 10 indication, p  < 0.001 a ; (E) Top ten reporter country, p  < 0.001 a ; (F) Gender, p   0.05 a . a: Fisher’s exact test; b: One-way ANOVA; c: Pearson Chi-Square. We have highlighted statistically significant values in boldface for emphasis. The numbers in Fig 3A and 3D represent signal number. The numbers in Fig 3B , 3C , 3E , 3F and 3G represent signal number percent. The signals unrelated to drugs or disease progression were excluded, such as general disorders and administration site conditions, product issues, social circumstances, investigations, injury, poisoning and procedural complications, congenital, familial and genetic disorders, and surgical and medical procedures. After excluding specific signals (repetitive, non-SOC included, IC025 < 0), the number of significant signals for each agent was as follows: leuprolide (402), goserelin (285), triptorelin (152), histrelin (35), and buserelin (19). The top twenty AEs with the highest frequency and strongest signal intensity for these products were analyzed. As shown in Fig 4 and S3 File , the most frequently reported AE with leuprolide was hot flush, whereas the most common AE with goserelin was malignant neoplasm progression. Ovarian hyperstimulation syndrome (OHSS) was the most frequently reported AE for both triptorelin and buserelin, whereas cutaneous rash emerged as the predominant AE associated with histrelin. The AEs with the strongest signal intensities were as follows: urinary tract toxicity (ROR = 490.89) for leuprolide, malignant spinal cord compression (ROR = 3958.65) for goserelin, ovarian hyperstimulation syndrome (ROR = 719.85) for triptorelin, delayed menarche (ROR = 16901.8) for histrelin, and noninfective oophoritis (ROR = 26614.6) for buserelin. Likewise, as shown in Fig 4 and S3 File , ovarian hyperstimulation syndrome (n = 181, ROR = 719.85) demonstrated both a prevalence and a robust signal strength for triptorelin and buserelin. The x-axis displays the normalized AE frequency (original values divided by 50) for leuprolide, goserelin, and triptorelin, while the y-axis represents the different AEs. The bubble size reflects the ROR value with log2 (ROR). We have presented the top twenty AEs with the highest frequency using bubbles, where each bubble corresponds to an AE, appearing on the far right of the x-axis, and the top twenty AEs with the highest signal strength are shown in the biggest bubbles. And AEs that display both high frequency and strong intensity are shown in big blue bubbles appearing to the right side of the x-axis. To further analyze the differences among the five long-acting GnRHa agents, the comprehensive signal mining results at the SOC level are shown in Fig 5 and S4 File . As shown in Fig 5 , the proportion of AEs in each SOC system was different. Reproductive system and breast disorders possessed the greatest proportion in AEs of the five drugs, followed by neoplasms benign, malignant and unspecified (including cysts and polyps), renal and urinary disorders, musculoskeletal and connective tissue disorders and nervous system disorders. A greater proportion of AEs in the neoplasms benign, malignant, and unspecified (including cysts and polyps) was also exhibited by leuprolide, goserelin, and triptorelin. In addition to the reproductive system and breast disorders mentioned above, buserelin also displayed a higher percentage of AEs in the gastrointestinal disorders. While histrelin displayed a higher percentage of AEs in psychiatric disorders and skin and subcutaneous tissue disorders, which was the most different case compared to the other four products. Serious adverse events/Important medical events were significant factors that deserve our special attention because of restrictions on the widespread use of drugs. An Important Medical Event (IME) is a clinical event that does not meet the stringent criteria for a serious adverse event (SAE) but has the potential to be clinically significant and requires proactive medical intervention to avoid a deterioration or serious outcome (such as hospitalization, disability, or death). IME is also a document in the form of a list of updated versions regularly published on EMA’s official website ( https://www.ema.europa.eu/en/human-regulatory/research-development/pharmacovigilance/eudravigilance/eudravigilance-system-overview ). Univariate and multivariate logistic regression analyses were conducted to determine the odds ratio (OR) of IMEs associated with long-acting GnRHa across various exposure factors. The results presented in Table 1 indicated that “drug, gender, off-label use and age” were independent factors that significantly influence IMEs associated with long-acting GnRHa. As shown in Table 1 , compared with patients receiving leuprolide, the risk of IMEs was 27.94 times higher in patients with buserelin, 6.78 times higher with goserelin, 1.22 times higher in histrelin, and 3.58 times higher in triptorelin. Furthermore, females and males exhibited comparable risks of IMEs. Patients using long-acting GnRHa off-label use had a higher risk of IMEs with 1.40 times compared to those on-label use. Compared to the reference group (aged 19–45 years), the risk of IMEs with long-acting GnRHa was significantly lower in patients under 18 years (risk ratio: 0.29), while no significant differences were observed in older age groups (46–60, 61–80, and >80 years). Ref: Reference; OR: Odds Ratio; CI: Confidential Interval. Statistically significant values are marked in boldface.

In this study, we found that the reported adverse events were basically consistent with the drug label. In our study, up to September 2023, the following number of reports were submitted to FAERS for each drug: 43,683 reports for leuprolide, 5,138 for goserelin, 3,143 for triptorelin, 501 for histrelin, and 52 for buserelin. Ultimately, a comprehensive analysis and comparative assessment were conducted, revealing 402 significant signals for leuprolide, 285 for goserelin, 152 for triptorelin, 35 for histrelin, and 19 for buserelin. Furthermore, the effects of various exposure parameters were examined by us on important medical events and contrasted the proportion of significant AEs among the five long-acting GnRHa across the nine systems. However, our research indicated that there were adverse event profile differences in the major systems of the five drugs. What makes us think deeply is that the therapeutic effects of long-acting GnRHa in psychiatric system and reproductive and breast system coexist with the adverse effects. The results of our analysis better illustrate the importance of actual pharmacovigilance studies. In summary, significant disparities exist between the adverse event profiles of long-acting GnRHa, affected by drug routes and pharmacological actions. As a result, drug safety management should consider the disease, the drug itself (e.g., route of administration), age and other factors, so as to monitor the drug according to these characteristics. Collaborative efforts among healthcare providers, society, and patients through enhanced surveillance, education, regulations, and feedback systems are expected to mitigate adverse drug reactions. In the future, we will focus on the relationship between route of administration, dose and AEs related with long-acting GnRHa.

FAERS serves as a post-market safety surveillance program for all medications that have received approval from FDA. We extracted the data utilized in this study from FAERS, which consists of spontaneous reports submitted by consumers, healthcare professionals, and manufacturers worldwide. This retrospective pharmacovigilance study employed a disproportionality analysis based on adverse drug reactions reported in FAERS, one of the prominent public spontaneous reporting systems encompassing reports from over 100 countries [ 22 ]. Serving as a comprehensive post-marketing safety surveillance initiative for FDA-approved drugs and therapeutic biological products globally, FAERS collects spontaneous safety reports and post-marketing clinical studies related to drug use within and outside the United States. The FAERS database utilizes the Medical Dictionary for Regulatory Activities (MedDRA) coding system. In this study, the international standardization of adverse drug events (ADEs) was achieved through the application of system organ class (SOC) and preferred terms (PTs) from MedDRA [ 23 ]. In this study, the original data is sourced from the FAERS database, extracted and queried using the OpenVigil 2.1 website ( http://openvigil.sourceforge.net/ ). The time range for data extraction is limited from the establishment of FAERS database to September 2023. As primary or secondary suspected drugs, we conducted a search in the FAERS database, spanning from January 2004 to September 2023, using the keywords “buserelin,” “goserelin,” “histrelin,” “leuprolide,” and “triptorelin” to identify all relevant indications. The duplicate case is considered if the case ID, reporting date, and patient characteristics are all the same. The Reporting Odds Ratio (ROR) analysis was conducted to quantify the statistical linkage between the drug of interest and specific AEs recorded in the FAERS database. The key algorithms leveraged for signal detection are outlined in S1 Table . AE would be deemed highly correlated with the treatment involving the drug of interest if all criteria outlined in S1 Table were concurrently satisfied, with elevated values suggesting a stronger statistical association. Each validated AE was assigned a preferred term (PT) and subsequently categorized into its respective System Organ Class (SOC) utilizing the Medical Dictionary for Regulatory Activities (MedDRA, version 24.1) as a reference. In this study, the disproportionality method was used for signal mining of ADEs. The disproportionality method is a widely used method in ADR monitoring, which is based on a 2 × 2 league table (four-cell table) to calculate the ratio of observed to expected drug-event values, and an imbalance is defined if it exceeds a prespecified threshold. The reporting odds ratio (ROR) in the disproportionality method is one of the most commonly used frequency methods, which are characterized by simple calculation and good consistency of results [ 24 ]. The corresponding algorithms, equations and criteria listed in S1 Table , which were used as the judgment conditions for the signal generation, and the fulfillment of the above conditions suggested that the drug was statistically associated with ADEs. Excel 2019 software was used for data cleaning, mapping between system organ classification (SOC) and preferred term (PT) and statistical analysis. A descriptive analysis of the demographic characteristics was conducted, and the proportion difference of AE reports among long-acting GnRHa was analyzed using either the chi-square test or Fisher’s exact test, both implemented in SPSS 26.0. We executed Disproportionality analyses utilizing ROR. Following the removal of duplicates, we compared the proportion of valid signals within each System Organ Class (SOC) using the chi-square test. The entire process of data processing and analysis is shown in Fig 2 .

(DOCX) (XLSX) (XLSX) (XLSX)