Safety signals for drug-related respiratory depression: a disproportionality analysis of the FDA Adverse Event Reporting System (FAERS)
Original Article

Safety signals for drug-related respiratory depression: a disproportionality analysis of the FDA Adverse Event Reporting System (FAERS)

Yang Rui1#, Zhe Chen1# ORCID logo, Jun Zhou2

1Laboratory of Cough, Affiliated Kunshan Hospital of Jiangsu University, Kunshan Key Laboratory of Chronic Cough, Suzhou, China; 2Department of Pharmacy, Affiliated Kunshan Hospital of Jiangsu University, Suzhou, China

Contributions: (I) Conception and design: Y Rui, Z Chen; (II) Administrative support: J Zhou; (III) Provision of study materials or patients: None; (IV) Collection and assembly of data: All authors; (V) Data analysis and interpretation: All authors; (VI) Manuscript writing: All authors; (VII) Final approval of manuscript: All authors.

#These authors contributed equally to this work.

Correspondence to: Jun Zhou, MD. Department of Pharmacy, Affiliated Kunshan Hospital of Jiangsu University, No. 566 Qianjin East Rd., Suzhou 215300, China. Email: 13862620262@163.com.

Background: Drug-induced respiratory depression is a severe and potentially life-threatening adverse event (AE); however, its systematic risk profile based on real-world data remains incomplete. Using real-world data from the U.S. Food and Drug Administration (FDA) Adverse Event Reporting System (FAERS), this study systematically identified drugs associated with respiratory depression and generated reporting safety signals to provide exploratory evidence for optimizing clinical medication safety.

Methods: The reporting odds ratio (ROR) was used to evaluate reports of drug-induced respiratory depression from the first quarter of 2015 to the third quarter of 2025. Single-factor, least absolute shrinkage and selection operator (LASSO), and multi-factor regression models were employed to analyze the association between various medications and respiratory depression events. Odds ratios (ORs) and 95% confidence intervals (CIs) were used to assess the associated risks.

Results: A total of 869 drugs associated with respiratory depression were identified, with nervous system drugs being the predominant category. Multivariate analysis identified 16 of 17 LASSO-screened drugs and male patients under 41 years of age as significant safety signals for respiratory depression, exhibiting strong disproportionate reporting signals. The median duration of drug-related respiratory depression was 1 day [interquartile range (IQR), 23 days], with approximately 75% of AEs occurring within this 23-day period.

Conclusions: These findings assist clinicians in the early identification of the risk of drug-related respiratory depression, underscoring the necessity for proactive assessment and management of this risk in both analgesia and disease treatment.

Keywords: Respiratory depression; FDA Adverse Event Reporting System (FAERS); acute respiratory distress syndrome (ARDS); adverse event (AE)

Submitted May 01, 2026. Accepted for publication Jun 18, 2026. Published online Jun 29, 2026.

doi: 10.21037/jtd-2026-1239

Highlight box

Key findings

• A total of 869 drugs associated with respiratory depression were identified, with nervous system drugs being the predominant category. 16 drugs, including morphine, baclofen, and monoclonal antibodies, showed disproportionately high reporting signals for respiratory depression. Males aged <41 years also represented a subgroup with strong safety signals. The median duration of drug-related respiratory depression was 1 day, with approximately 75% of adverse events (AEs) occurring within this 23-day period.

What is known and what is new?

• Opioids and benzodiazepines are well-recognized causes of respiratory depression, primarily through direct suppression of the central nervous system.

• This study systematically characterizes safety signals for respiratory depression across a broad range of drug classes, including muscle relaxants, gabapentinoids, antipsychotics, and monoclonal antibodies. It also identifies young males as a high-signal subgroup, a finding that may be largely attributable to behavioral factors.

What is the implication, and what should change now?

• Clinicians should avoid the non-indicated concomitant use of multiple high-risk drugs, such as opioids, benzodiazepines, and gabapentinoids. Enhanced respiratory monitoring should be considered for young male patients receiving these medications. Medication safety education should also be updated to include non-traditional high-risk drugs, such as baclofen and monoclonal antibodies.

Introduction

Respiratory depression refers to impaired ventilation function that leads to insufficient gas exchange, specifically manifested as a decreased respiratory rate and reduced tidal volume, which subsequently results in hypoxemia and hypercapnia (1). In clinical practice, respiratory depression is a common and life-threatening complication, particularly prominent in perioperative and intensive care settings. The causes of respiratory depression are complex and diverse, encompassing central nervous system disorders, neuromuscular diseases, metabolic disturbances, and severe pulmonary diseases, among others. However, among the preventable iatrogenic factors, drug-induced respiratory depression is the most prevalent (2). Opioids, such as morphine, fentanyl, and oxycodone, are widely used in clinical practice due to their potent analgesic effects; however, these drugs are also a key cause of respiratory depression (3). According to a systematic review of postoperative patients, the incidence of opioid-induced respiratory depression is approximately 0.5% (4). This risk significantly increases in patient populations with high safety signals, such as cardiopulmonary disease or obstructive sleep apnea. Findings from another multicenter prospective trial (PRODIGY trial) indicate that approximately 14% of patients experienced at least one event of respiratory depression after receiving opioid treatment (5). In addition to opioids, other central nervous system depressants can also cause respiratory depression. Studies have shown that both benzodiazepines and gabapentinoids may induce varying degrees of respiratory depression (6). More importantly, when these drugs are used in combination with opioids, the risk of respiratory depression increases exponentially and significantly (7).

The U.S. Food and Drug Administration (FDA) Adverse Event Reporting System (FAERS) represents one of the largest global voluntary reporting systems for adverse drug events, documenting reports submitted by consumers, healthcare professionals, and pharmaceutical companies (8). Previous studies have utilized the FAERS database to investigate adverse drug reactions associated with tramadol, fentanyl, gabapentin, and other medications (9-12). However, current research on drug-induced respiratory depression is predominantly limited to single-drug or small-sample observational studies, lacking a systematic analysis of drug-related respiratory depression. This study aims to systematically analyze and identify independent signals associated with drug-induced respiratory depression using the FAERS database. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1239/rc).

Methods

Source of data

This study primarily utilized the FAERS database. We extracted drug-related data spanning from the first quarter of 2015 to the third quarter of 2025. Specifically, we focused on adverse event (AE) reports associated with drug-induced respiratory depression, employing the preferred terms (PTs) “HYPOPNOEA”, “RESPIRATORY DEPRESSION”, “APNOEA”, and “HYPOVENTILATION”. The dataset obtained comprises seven key tables: demographic information (DEMO), drug usage records (DRUG), drug therapy duration (THER), drug indications (INDI), adverse event records (REAC), sources of adverse events (RPSR), and patient outcomes (OUTC) (8).

Data processing

During the data processing phase, we primarily employed the FDA-recommended method for the removal of duplicate data, effectively eliminating problematic records. We began by sorting the data according to CASEID, FDA_DT, and PRIMARYID. When duplicate CASEIDs were identified, the record with the maximum FDA_DT value was retained; if both the CASEID and FDA_DT were identical, the report with the largest PRIMARYID value was preserved. In this study, the relevant AE data were coded using PTs from the Medical Dictionary for Regulatory Activities (MedDRA) (13). Given the diversity of data sources, we performed standardized processing on all medications. Additionally, we adopted the World Health Organization’s Anatomical Therapeutic Chemical (ATC) classification system to systematically categorize various types of medications.

Statistical analysis

In this study, we primarily employed the reporting odds ratio (ROR) along with the 95% confidence interval (CI) to conduct a disproportionality analysis aimed at evaluating suspected drugs that induce respiratory depression. This method is fundamentally based on the 2×2 contingency table utilized in the disproportionality method (Table 1). The criteria for signal detection using ROR were established as follows: the number of reports (a) must be ≥3, and the lower limit of the 95% CI for ROR must exceed 1, which indicates a positive signal. The calculation formulas are detailed in Table 2.

Table 1

Four-fold table of disproportionality measures

Type of drug Targeted adverse event Other adverse events Total
Target drug a b a+b
Other drugs c d c+d
Total a+c b+d a+b+c+d

a, number of reports with the target drug and the target adverse event; b, number of reports with the target drug and all other adverse events; c, number of reports with all other drugs and the target adverse event; d, number of reports with all other drugs and all other adverse events.

Table 2

Formulas and threshold values of ROR

Methods Calculation formula Algorithmic signal generation conditions
ROR ROR=adbcSE(lnROR)=(1a+1b+1c+1d)95% CI=eln(ROR)±1.96×1a+1b+1c+1d 95% CI (lower limit) >1, a≥3

a, number of reports with the target drug and the target adverse event; b, number of reports with the target drug and all other adverse events; c, number of reports with all other drugs and the target adverse event; d, number of reports with all other drugs and all other adverse events. CI, confidence interval; ROR, reporting odds ratio; SE, standard error.

The P value resulting from Fisher’s exact test and the Bonferroni correction was termed P adjust. A volcano plot was created using −log (P-adjust) on the y-axis and logROR on the x-axis.

A single-factor analysis was performed on potentially suspicious drugs, employing a 95% CI with a lower limit of ROR greater than 1, a report count of above 100, and a P-adjustment value lower than 0.01. Drugs identified with a significance level of P<0.01 in the single-factor analysis were employed in least absolute shrinkage and selection operator (LASSO) regression. Subsequently, multi-factor logistic regression was performed to evaluate the presence of respiratory depression safety signals associated with drugs, utilizing the drugs selected by LASSO along with foundational patient information as independent variables. Statistical analysis was conducted using R software (version 4.3.2).

It should be noted that the FAERS database used in this study is a spontaneous reporting system, and its data-generating process is inherently subject to selection bias, such as preferential reporting for newly approved or highly scrutinized drugs, as well as reporting bias, whereby serious AEs are more likely to be submitted. These biases cannot be fully corrected by any statistical model. In this context, the use of LASSO regression and multivariable logistic regression in the present study was not intended for causal inference or for developing a clinical prediction model generalizable to the overall patient population. Rather, these methods were applied as exploratory analytical tools. Specifically, LASSO regression was used to screen signals from high-dimensional variables and to reduce the risk of false-positive findings arising from multiple comparisons. The receiver operating characteristic-area under the curve (ROC-AUC) was used only to assess the model’s internal fitting and discriminative performance for reported events within the current dataset; therefore, it should not be interpreted as evidence of clinical predictive accuracy. Accordingly, all odds ratios (ORs) should be interpreted as measures of relative signal strength rather than unbiased estimates of absolute clinical risk.

Ethical statement

Because the FAERS database is publicly accessible and patient records are anonymized, this study does not require ethical approval or informed consent. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Results

Baseline characteristics of drug-induced respiratory depression

In this study, we retrieved data from the FAERS database covering the period from the first quarter of 2015 to the third quarter of 2025. After excluding duplicate and questionable entries, a total of 14,782,616 complete data reports were included. Through data mining, we identified 869 drugs associated with AEs related to respiratory depression, affecting 10,075 patients. Based on the collected AE data concerning respiratory depression, we conducted a baseline analysis. Among the reporters, there were 4,564 females (45.3%), 4,319 males (42.9%), and 1,192 individuals with unspecified gender (11.8%). Regarding age distribution, after excluding missing values, the majority of cases were concentrated in the age group below 36 years (27.2%). A total of 5,487 cases were reported by healthcare professionals. Concerning treatment outcomes for patients experiencing respiratory depression, 27.2% required hospitalization. The reports were predominantly sourced from the United States (4,278, 42.5%) and Canada (1,006, 10.0%). Specific data are presented in Table 3. In terms of the annual distribution of reports, the average number of reported AEs has been 916 cases per year since 2015 (Figure 1).

Table 3

Basic patient information

Characteristics Drug-related respiratory depression (N=10,075), n (%)
Gender
   Female 4,564 (45.3)
   Male 4,319 (42.9)
   Unknown 1,192 (11.8)
Age, years
   <36 2,739 (27.2)
   36–52 1,423 (14.1)
   53–65 1,465 (14.5)
   ≥66 1,185 (11.8)
   Unknown 3,263 (32.4)
Reported person
   Consumer 2,360 (23.4)
   Health professional 1,778 (17.6)
   Pharmacist 1,052 (10.4)
   Physician 2,657 (26.4)
   Missing 2,228 (22.1)
Outcome
   Congenital anomaly 43 (0.4)
   Death 2,024 (20.1)
   Disability 62 (0.6)
   Hospitalization 2,744 (27.2)
   Life-threatening 1,693 (16.8)
   Other 3,509 (34.8)
Reported country (top 3)
   United States 4,278 (42.5)
   Canada 1,006 (10.0)
   France 731 (7.3)
Figure 1 Drug-related respiratory depression reports by year.

Drugs associated with respiratory depression

We utilized volcano plots to illustrate the relationship between respiratory depression and suspected drugs (Figure 2). In this figure, the x-axis represents the logarithm of the ROR. A positive x-axis indicates a higher frequency of drug-related adverse reactions compared to other adverse reactions. The y-axis represents the negative logarithm of the P value following Fisher’s exact test and Bonferroni correction, where a positive y-axis indicates significant differences. The color of the dots corresponds to the logarithm of the number of reported cases; the redder the dot, the higher the number of reports. Consequently, drugs positioned in the upper right corner of the graph not only demonstrate significant signal strength but also exhibit significant differences.

Figure 2 Drug-related respiratory depression volcano plots. CI, confidence interval; ROR, reporting odds ratio.

Furthermore, we classified the top 50 most frequently identified drugs in this study according to the ATC system, which included: nervous system (35/50), antineoplastic and immunomodulating agents (3/50), alimentary tract and metabolism (3/50), respiratory system (2/50), musculoskeletal system (2/50), cardiovascular system (2/50), genitourinary system and sex hormones (1/50), antiinfectives for systemic use (1/50), and sensory organs (1/50) (Table 4).

Table 4

Distribution of the top 50 drug-related respiratory depression

ATC category Drug Number of cases ROR 95% CI P value P adjust
Nervous system Fentanyl† 417 14.13 12.81−15.6 <0.001 <0.001
Oxycodone† 364 3.2 2.88−3.55 1.8514E−116 1.6089E−113
Oxybate sodium† 296 6.04 5.38−6.79 2.2296E−262 1.9376E−259
Tramadol† 274 13.93 12.34−15.72 <0.001 <0.001
Diazepam† 214 24.1 21.01−27.64 <0.001 <0.001
Paracetamol† 202 3.54 3.08−4.08 8.64814E−80 7.51523E−77
Quetiapine† 177 7.02 6.05−8.15 1.2283E−194 1.0674E−191
Olanzapine† 167 7.39 6.34−8.62 1.3165E−196 1.144E−193
Gabapentin† 165 5.16 4.42−6.02 9.6413E−119 8.3783E−116
Pregabalin† 146 2.97 2.52−3.5 3.11647E−42 2.70822E−39
Buprenorphine† 143 6.22 5.27−7.34 2.9015E−134 2.5214E−131
Alprazolam† 127 8.46 7.09−10.08 3.7355E−178 3.2462E−175
Methadone† 110 28.37 23.44−34.32 5.8846E−115 5.1137E−112
Levetiracetam† 99 3.2 2.62−3.9 1.70571E−33 1.48226E−30
Lorazepam† 90 11.92 9.67−14.69 2.2455E−191 1.9513E−188
Hydromorphone† 88 16.72 13.53−20.67 2.3791E−274 2.0674E−271
Sertraline† 70 2.71 2.14−3.43 1.10457E−17 9.59869E−15
Risperidone 63 1.24 0.97−1.59 0.102939536 >0.99
Bupropion† 63 4.47 3.48−5.72 1.39703E−37 1.21402E−34
Ibuprofen† 62 1.51 1.18−1.94 0.001506288 >0.99
Venlafaxine† 61 3.37 2.62−4.34 3.52081E−23 3.05958E−20
Zolpidem† 59 5.8 4.48−7.49 3.08684E−51 2.68247E−48
Sevoflurane† 56 37.63 28.79−49.18 4.98133E−66 4.32877E−63
Dexmedetomidine† 56 40.55 31.02−53.02 9.09833E−68 7.90645E−65
Propofol† 55 17.17 13.14−22.43 3.87061E−47 3.36356E−44
Lamotrigine† 52 2.68 2.04−3.53 3.57774E−13 3.10905E−10
Clonazepam† 50 5.21 3.94−6.88 1.52117E−37 1.3219E−34
Carbidopa levodopa† 49 1.35 1.02−1.79 0.042943557 >0.99
Midazolam† 48 24.97 18.74−33.28 1.0213E−48 8.87507E−46
Aripiprazole† 47 1.41 1.06−1.88 0.022878387 >0.99
Paroxetine† 46 5.35 4−7.15 5.87448E−36 5.10492E−33
Valproic acid† 46 3.32 2.48−4.43 2.72168E−17 2.36514E−14
Fluoxetine† 46 3.78 2.83−5.05 1.89591E−21 1.64754E−18
Clozapine 41 0.75 0.55−1.02 0.079420422 >0.99
Duloxetine 40 1.31 0.96−1.78 0.1078394 >0.99
Antineoplastic and immunomodulating agents Adalimumab 123 0.39 0.33−0.47 1.45133E−26 1.26121E−23
Infliximab 43 0.55 0.41−0.74 7.99541E−05 0.069480121
Methotrexate 43 0.55 0.41−0.74 8.12193E−05 0.070579529
Alimentary tract and metabolism Metformin† 79 2.14 1.71−2.67 1.12551E−11 9.78067E−09
Loperamide† 74 8.01 6.37−10.08 1.05037E−97 9.12768E−95
Ondansetron† 49 6.92 5.22−9.17 5.69556E−54 4.94944E−51
Respiratory system Omalizumab† 173 4.36 3.75−5.06 1.79783E−96 1.56231E−93
Salbutamol 53 1.12 0.85−1.46 0.465206079 >0.99
Musculo-skeletal system Baclofen† 165 13.55 11.61−15.82 <0.001 <0.001
Rocuronium† 59 30.42 23.46−39.44 2.86282E−64 2.48779E−61
Cardiovascular system Lidocaine† 61 6.07 4.71−7.81 2.70891E−56 2.35404E−53
Amlodipine† 55 2.25 1.72−2.93 1.55754E−09 1.3535E−06
Genito urinary system and sex hormones Morphine† 386 25.79 23.27−28.59 <0.001 <0.001
Antiinfectives for systemic use Palivizumab† 104 13.81 11.37−16.78 1.6839E−261 1.4633E−258
Sensory organs Clonidine† 41 11.71 8.6−15.94 2.72297E−29 2.36626E−26

P adjust, P value after Bonferroni correction; P adjust <0.01, statistically significant; †, the drug was positive for ROR. ATC, anatomical therapeutic chemical; CI, confidence interval; ROR, reporting odds ratio.

Safety signals for drug-related respiratory depression

We conducted a univariate analysis on suspected drugs that had reported case numbers exceeding 100, with lower limits of the 95% CI for ROR greater than 1, and a P adjusted value of less than 0.01. Drugs that exhibited P values below 0.01 in the univariate analysis were subjected to LASSO regression analysis, resulting in the identification of 17 drugs (Figure 3). A multivariate logistic regression analysis was subsequently performed on these drugs, incorporating patient age, gender, and weight (Figure 4). The ROC-AUC, which indicates the prediction accuracy of the model, was 0.683 (Figure 5). Among the drugs included in the analysis, a total of 16 were found to be positively associated with the risk of respiratory depression, suggesting that they may serve as independent safety signals for this condition. Notably, morphine (OR =38.586, 95% CI: 30.214–48.536) and oxycodone (OR =27.934, 95% CI: 21.921–35.049) demonstrated the strongest associations. Other significant medications with an OR greater than 7 included diazepam, methadone, baclofen, tramadol, fentanyl, palivizumab, alprazolam, omalizumab, and sodium oxybate. These drugs can be classified according to the ATC classification system as follows: nervous system (12/16), genitourinary system and sex hormones (1/16), respiratory system (1/16), musculoskeletal system (1/16), and antiinfectives for systemic use (1/16). Furthermore, when using the 41–54-year age group as a reference, individuals aged less than 41 years (OR =1.580, 95% CI: 1.390–1.799) exhibited a significantly increased risk of respiratory depression (P<0.001). In comparison to females, males were found to have a higher risk of respiratory depression (OR =1.450, 95% CI: 1.338–1.571, P<0.001). Weight gain was associated with a slightly reduced risk of respiratory depression (OR =0.991, 95% CI: 0.989–0.992).

Figure 3 Results of the LASSO regression analysis. (A) AUC of drug-induced respiratory depression; (B) drug-induced respiratory depression coefficients diagram. AUC, area under curve; LASSO, least absolute shrinkage and selection operator.
Figure 4 Results of the multivariate logistic regression analysis. P adjust, P value after Bonferroni correction; P adjust <0.01, statistically significant. CI, confidence interval; OR, odds ratio.
Figure 5 The ROC curves of drug-related respiratory depression safety signals. AUC, area under curve; ROC, receiver operating characteristic.

Time interval between drug use and the onset of respiratory depression

The time from drug administration to the onset of respiratory depression was evaluated (Figure 6). The median time to drug-induced respiratory depression was found to be 1 day [interquartile range (IQR), 23 days], with approximately 75% of reported cases occurring within 23 days following medication use.

Figure 6 The time interval between drug administration and drug-related respiratory depression episodes. (A) Violin plot of time to drug-related respiratory depression occurrence; (B) cumulative incidence of drug-related respiratory depression.

Discussion

This study found that, after adjusting for factors such as age, gender, and body weight, opioids, benzodiazepines, certain antiepileptic and antipsychotic drugs, and some biological agents emerged as significant independent safety signals for respiratory depression. Meanwhile, male gender and age under 41 years were identified as independent demographic safety signals, while weight gain exhibited a weak protective effect.

In the present study, age younger than 41 years (OR =1.580) and male sex (OR =1.450) were both positively associated with reports of respiratory depression. Younger patients typically exhibit higher activity of hepatic microsomal enzymes, such as the CYP450 system, which may accelerate the conversion of certain prodrugs into more potent metabolites, consequently leading to increased blood drug concentrations (14,15). However, these pharmacokinetic differences alone are unlikely to fully explain the observed association. Women may be at greater risk of adverse drug reactions because of lower average body weight, whereas older adults are generally more vulnerable owing to age-related declines in hepatic and renal function (16). Therefore, behavioral factors may have contributed substantially to the elevated signal observed in younger male patients. Young men represent a high-risk population for nonmedical use of prescription opioids, recreational drug misuse, alcohol abuse, and self-directed dose escalation. When these exposures occur concomitantly with prescribed medications, they may exert synergistic respiratory depressant effects. However, such concomitant non-prescription exposures are often incompletely captured in FAERS reports, in which the recorded suspect medications are typically limited to prescribed drugs. This may lead to an overestimation of the apparent risk attributed to the reported prescription medications. In addition, clinicians may adopt more aggressive opioid-based treatment strategies in younger male patients, which could further increase cumulative drug exposure. From the perspective of pharmacovigilance reporting bias, the higher reporting frequency of respiratory depression among younger patients may also reflect reporter expectation bias. Respiratory depression occurring in older adults or in patients with multiple comorbidities may be perceived as a relatively expected event and therefore may be less likely to be reported. By contrast, when the same serious AE occurs in a younger and previously healthy individual, its unexpected nature may substantially increase the likelihood of reporting by clinicians. Thus, the observed signal is likely to reflect the combined influence of biological factors, clinical monitoring practices, and reporting bias. It is also noteworthy that the hospitalization rate associated with respiratory depression in this study was 27.2%. This relatively high proportion may be partly attributable to the inherent characteristics of the FAERS spontaneous reporting system. Reporters are more likely to submit serious AEs that result in hospitalization, whereas mild events or events not requiring hospitalization are more likely to be underreported. Therefore, the hospitalization rate of 27.2% should be interpreted as the proportion of hospitalization among reported FAERS cases, rather than the absolute hospitalization rate of respiratory depression in the target population.

In this study, we found that strong opioid receptor agonists, such as morphine (OR =38.586), oxycodone (OR =27.934), and methadone (OR =16.363), pose extremely high risks, which is entirely consistent with the known pharmacological mechanisms (17,18). Its mechanism of action involves direct inhibition of the preBötzinger complex (preBötC) in the ventrolateral medulla, with the inhibitory effect being dose-dependent (19). The association between tramadol and respiratory depression (OR =11.837) may be directly related to its active metabolite, O-desmethyltramadol (M1) (20). In patients with increased CYP2D6 activity, the formation of M1 may be enhanced, thereby further increasing the risk of opioid-related respiratory depression (21). Benzodiazepines, including diazepam (OR =19.088) and alprazolam (OR =8.253), broadly suppress central neuronal activity by promoting GABA-A receptor-mediated chloride influx (22). When used alone, benzodiazepines generally cause only mild respiratory depression; however, when combined with opioids, they may produce a marked synergistic depressant effect on respiration and constitute one of the most common drug combinations associated with fatal respiratory depression (7,23).

It should be emphasized that the extremely high reporting ORs observed for strong opioids such as morphine and oxycodone in FAERS may reflect not only their genuine pharmacological effects but also systematic amplification driven by clinical expectations, public safety warnings, and targeted prescription monitoring. In medical education and clinical training, opioids are commonly associated with respiratory depression (24). Clinicians may be particularly vigilant when respiratory abnormalities occur in patients receiving opioid therapy and may be more likely to report such events, whereas similar events related to drugs without a well-established association may be overlooked. FDA has also issued multiple public safety communications regarding opioid overdose and respiratory risks (25). Such warnings can substantially increase the reporting rate of AEs for the implicated drugs; therefore, even in the absence of a true increase in risk, the number of reports may rise sharply over a short period. In addition, patients receiving strong opioids for pain management or postoperative analgesia are often subject to more intensive monitoring and adverse-event documentation within routine clinical pathways. This increases the probability that episodes of respiratory depression will be detected and reported compared with patients who are not under similar clinical scrutiny. Therefore, the very high OR observed in the present study, such as that for morphine (OR =38.586), should be interpreted as the combined result of a true pharmacological risk and reporting bias.

In addition to the medications discussed above, our analysis also identified independent safety signals for skeletal muscle relaxants, γ-hydroxybutyrate analogues, gabapentinoids, and sedating antipsychotics, all of which remained statistically significant (all P<0.001). Baclofen, a GABA-B receptor agonist, showed a strong association with respiratory depression (OR =14.043). This effect may be partly explained by its ability to inhibit acetylcholine release from anterior horn motor neurons in the spinal cord, thereby reducing respiratory muscle tone (26). The risk appears to be particularly pronounced in cases of overdose or intrathecal administration (27). Oxybate sodium was also associated with an increased risk (OR =7.030). As a potent sedative-hypnotic agent, sodium oxybate can induce respiratory depression in a dose-dependent manner, and this effect has been linked to the development of respiratory acidosis (28). Gabapentin, which selectively binds to the α2δ-1 subunit of voltage-gated calcium channels, may suppress the release of excitatory neurotransmitters such as glutamate, providing a plausible neuropharmacological basis for its respiratory effects (OR =6.235) (29). Olanzapine, which exerts prominent sedative effects mainly through antagonism of 5-HT2A and H1 receptors, was likewise associated with respiratory depression (OR =5.636) (30). Although these agents are generally considered to carry a relatively low risk of respiratory depression when used alone, their concomitant use with opioids may substantially amplify this risk. Clinically, non-essential co-prescribing of these medications with opioids should therefore be avoided whenever possible.

This study also found that the monoclonal antibodies palivizumab (OR =8.413) and omalizumab (OR =7.948) were significantly and positively associated with respiratory depression, with all associations reaching statistical significance (all P<0.001). Unlike centrally acting depressant drugs, the respiratory risk associated with these biologics is more likely to arise from immune-mediated acute hypersensitivity reactions, including bronchospasm, vasodilation, and increased capillary permeability (31,32). In addition, during the initial administration or high-dose infusion of certain monoclonal antibodies, rapid and excessive release of pro-inflammatory cytokines, such as IL-6, TNF-α, and IFN-γ, may occur within a short period, leading to cytokine release syndrome (33,34). From a respiratory perspective, this syndrome may present with manifestations resembling acute respiratory distress syndrome (ARDS), characterized by diffuse lung injury and severe hypoxemia. These findings extend the risk spectrum of respiratory depression beyond conventional central nervous system depressants to include biologic agents.

Limitations

First, FAERS is a spontaneous reporting system, and its data-generating process is inherently susceptible to selection bias and reporting bias. For example, newly approved drugs, highly scrutinized medications, and drugs with previous safety warnings may be systematically over-reported, whereas serious events or events with positive outcomes are more likely to be submitted than mild or expected AEs, which are often under-reported. Although LASSO and multivariable regression analyses were applied in the present study, these methods cannot eliminate the intrinsic biases of FAERS data. Therefore, the ROC-AUC obtained in this study should be interpreted only as a measure of model fit within the FAERS reporting dataset, rather than as evidence of clinical predictive performance. Similarly, variables selected by LASSO may also be influenced by underlying reporting biases and should be regarded as exploratory safety signals rather than definitive safety signals.

Second, as a spontaneous reporting system, FAERS often contains incomplete or missing information on patients’ underlying medical conditions, such as chronic obstructive pulmonary disease (COPD), sleep apnea, other pulmonary disorders, epilepsy, a history of spasticity or convulsions, and neuromuscular diseases. These conditions may serve both as indications for certain medications, including muscle relaxants, antiepileptic drugs, and sedatives, and as independent safety signals for respiratory depression. For instance, patients with COPD may be more susceptible to drugs with respiratory depressant effects; patients with epilepsy may receive sedating antiepileptic agents; and patients with spasticity may be treated with muscle relaxants. Because these variables are largely unavailable or highly incomplete in FAERS, they could not be incorporated into the multivariable regression model for adjustment, leaving residual confounding unaddressed. Consequently, some of the observed drug-respiratory depression associations may partly reflect confounding by indication, which could have inflated the apparent strength of the drug-related signals. This represents an important limitation of spontaneous reporting systems. Therefore, the findings of this study should be interpreted as exploratory safety signals rather than unbiased estimates of causal effects.

Finally, FAERS does not systematically collect variables related to patient behavior, including but not limited to alcohol use, illicit drug use, adherence to prescribed medications, self-directed dose escalation, or urine toxicology screening results. These behavioral factors may be particularly relevant among younger male patients and are well-recognized predictors of increased risk for respiratory depression. The absence of urine toxicology data further means that this study could not determine whether unreported illicit substances, non-prescribed opioids, or other undocumented concomitant medications were present at the time of respiratory depression. Because these variables were unavailable, behavioral confounding could not be statistically adjusted for in the present analysis. Therefore, the risk signals associated with male sex and younger age observed in this study may reflect the combined effects of prescribed drug exposure and underlying behavioral safety signals. On the one hand, the stronger reporting signals observed among younger male patients may be partly attributable to concomitant recreational substance use, such as illicit opioids, benzodiazepines, or alcohol. On the other hand, in some severe cases of respiratory depression, urine toxicology screening could have helped confirm or exclude exposure to specific substances; however, such confirmatory information is not available within the current FAERS data structure.

Conclusions

Using the large-scale FAERS database, this study systematically characterized the patterns and features of disproportionate reporting of drug- and patient-related respiratory depression, thereby constructing a multilayered spectrum of safety signals. It not only confirms the core high-risk status of opioids and benzodiazepines but also reveals the widespread potential risks associated with antiepileptic drugs, muscle relaxants, antipsychotics, and biologics, with a particular emphasis on young males as a frequently overlooked high-risk population. Future prospective studies or in-depth analyses are necessary to clarify the dose-response relationship and investigate the specific mechanisms underlying the heightened risk in younger patients, thereby providing a scientific basis for developing age-specific analgesic and sedation guidelines.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1239/rc

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1239/prf

Funding: This work was funded by National Natural Science Foundation of China (No. 82370036), Key R&D Program of Jiangsu Province (No. BE2022761) and Kunshan Key Laboratory of Chronic Cough (No. KZ202401).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1239/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

  1. Dahan A, Aarts L, Smith TW. Incidence, Reversal, and Prevention of Opioid-induced Respiratory Depression. Anesthesiology 2010;112:226-38. [Crossref] [PubMed]
  2. Baldo BA. Opioid-induced respiratory depression: clinical aspects and pathophysiology of the respiratory network effects. Am J Physiol Lung Cell Mol Physiol 2025;328:L267-89. [Crossref] [PubMed]
  3. Boom M, Niesters M, Sarton E, et al. Non-analgesic effects of opioids: opioid-induced respiratory depression. Curr Pharm Des 2012;18:5994-6004. [Crossref] [PubMed]
  4. Gupta K, Nagappa M, Prasad A, et al. Risk factors for opioid-induced respiratory depression in surgical patients: a systematic review and meta-analyses. BMJ Open 2018;8:e024086. [Crossref] [PubMed]
  5. Kor JJ, Sprung J, Khanna AK, et al. Continuous Monitoring Detected Respiratory Depressive Episodes in Proximity to Adverse Respiratory Events During the PRODIGY Trial. J Patient Saf 2022;18:738-41. [Crossref] [PubMed]
  6. Evoy KE, Morrison MD, Saklad SR. Abuse and Misuse of Pregabalin and Gabapentin. Drugs 2017;77:403-26. [Crossref] [PubMed]
  7. Baillargeon J, Singh G, Kuo YF, et al. Association of Opioid and Benzodiazepine Use with Adverse Respiratory Events in Older Adults with Chronic Obstructive Pulmonary Disease. Ann Am Thorac Soc 2019;16:1245-51. [Crossref] [PubMed]
  8. Sakaeda T, Tamon A, Kadoyama K, et al. Data mining of the public version of the FDA Adverse Event Reporting System. Int J Med Sci 2013;10:796-803. [Crossref] [PubMed]
  9. Park S, Lee GH, Kim S, et al. Risk Factors for Respiratory Depression Associated with Tramadol Based on the Global Pharmacovigilance Database (VigiBase). Pharmaceuticals (Basel) 2024;17:205. [Crossref] [PubMed]
  10. Zhang J, Wang W, Lu Z, et al. Integrated pharmacovigilance of fentanyl: multinational analysis reveals novel safety signals and demographic-specific risk profiles. Front Pharmacol 2025;16:1700112. [Crossref] [PubMed]
  11. Le H, Hong H, Ge W, et al. A systematic analysis and data mining of opioid-related adverse events submitted to the FAERS database. Exp Biol Med (Maywood) 2023;248:1944-51. [Crossref] [PubMed]
  12. Vickers-Smith R, Sun J, Charnigo RJ, et al. Gabapentin drug misuse signals: A pharmacovigilance assessment using the FDA adverse event reporting system. Drug Alcohol Depend 2020;206:107709. [Crossref] [PubMed]
  13. Tieu C, Breder CD. A Critical Evaluation of Safety Signal Analysis Using Algorithmic Standardised MedDRA Queries. Drug Saf 2018;41:1375-85. [Crossref] [PubMed]
  14. Subash S, Prasad B. Age-Dependent Changes in Cytochrome P450 Abundance and Composition in Human Liver. Drug Metab Dispos 2024;52:1363-72. [Crossref] [PubMed]
  15. Jin J, Zhong XB. Epigenetic Mechanisms Contribute to Intraindividual Variations of Drug Metabolism Mediated by Cytochrome P450 Enzymes. Drug Metab Dispos 2023;51:672-84. [Crossref] [PubMed]
  16. Zucker I, Prendergast BJ. Sex differences in pharmacokinetics predict adverse drug reactions in women. Biol Sex Differ 2020;11:32. [Crossref] [PubMed]
  17. Bateman JT, Saunders SE, Levitt ES. Understanding and countering opioid-induced respiratory depression. Br J Pharmacol 2023;180:813-28. [Crossref] [PubMed]
  18. Davis MP, DiScala S, Davis A. Respiratory Depression Associated with Opioids: A Narrative Review. Curr Treat Options Oncol 2024;25:1438-50. [Crossref] [PubMed]
  19. Baldo BA, Rose MA. Mechanisms of opioid-induced respiratory depression. Arch Toxicol 2022;96:2247-60. [Crossref] [PubMed]
  20. Rodieux F, Vutskits L, Posfay-Barbe KM, et al. When the Safe Alternative Is Not That Safe: Tramadol Prescribing in Children. Front Pharmacol 2018;9:148. [Crossref] [PubMed]
  21. Ingelman-Sundberg M, Sim SC, Gomez A, et al. Influence of cytochrome P450 polymorphisms on drug therapies: pharmacogenetic, pharmacoepigenetic and clinical aspects. Pharmacol Ther 2007;116:496-526. [Crossref] [PubMed]
  22. Goldschen-Ohm MP. Benzodiazepine Modulation of GABA(A) Receptors: A Mechanistic Perspective. Biomolecules 2022;12:1784. [Crossref] [PubMed]
  23. McCarthy M. US drug labels to warn of risks of combining opioids and benzodiazepines. BMJ 2016;354:i4784. [Crossref] [PubMed]
  24. Palkovic B, Marchenko V, Zuperku EJ, et al. Multi-Level Regulation of Opioid-Induced Respiratory Depression. Physiology (Bethesda) 2020;35:391-404. [Crossref] [PubMed]
  25. Schneider L. FDA Requires Stronger Opioid Pain Medication Warning Labels. JAMA 2025;334:1135. [Crossref] [PubMed]
  26. Javaheri S, Randerath WJ, Safwan Badr M, et al. Medication-induced central sleep apnea: a unifying concept. Sleep 2024;47:zsae038. [Crossref] [PubMed]
  27. Olivier PY, Joyeux-Faure M, Gentina T, et al. Severe Central Sleep Apnea Associated With Chronic Baclofen Therapy: A Case Series. Chest 2016;149:e127-31. [Crossref] [PubMed]
  28. Felmlee MA, Morse BL, Morris ME. γ-Hydroxybutyric Acid: Pharmacokinetics, Pharmacodynamics, and Toxicology. AAPS J 2021;23:22. [Crossref] [PubMed]
  29. Revol B, Jullian-Desayes I, Cracowski JL, et al. Gabapentinoids and sleep apnea syndrome: a safety signal from the WHO pharmacovigilance database. Sleep 2019; [Crossref]
  30. Park YS, Kim GM, Sung HJ, et al. Atypical antipsychotic drug olanzapine inhibits 5-HT(3) receptor-mediated currents by allosteric and non-competitive mechanisms. Korean J Physiol Pharmacol 2025;29:431-42. [Crossref] [PubMed]
  31. Cutroneo PM, Arzenton E, Furci F, et al. Safety of Biological Therapies for Severe Asthma: An Analysis of Suspected Adverse Reactions Reported in the WHO Pharmacovigilance Database. BioDrugs 2024;38:425-48. [Crossref] [PubMed]
  32. Galvão VR, Castells MC. Hypersensitivity to biological agents-updated diagnosis, management, and treatment. J Allergy Clin Immunol Pract 2015;3:175-85; quiz 186. [Crossref] [PubMed]
  33. Xiao X, Huang S, Chen S, et al. Mechanisms of cytokine release syndrome and neurotoxicity of CAR T-cell therapy and associated prevention and management strategies. J Exp Clin Cancer Res 2021;40:367. [Crossref] [PubMed]
  34. Cosenza M, Sacchi S, Pozzi S. Cytokine Release Syndrome Associated with T-Cell-Based Therapies for Hematological Malignancies: Pathophysiology, Clinical Presentation, and Treatment. Int J Mol Sci 2021;22:7652. [Crossref] [PubMed]
Cite this article as: Rui Y, Chen Z, Zhou J. Safety signals for drug-related respiratory depression: a disproportionality analysis of the FDA Adverse Event Reporting System (FAERS). J Thorac Dis 2026;18(7):785. doi: 10.21037/jtd-2026-1239

Article Options

Download Citation

Share