Risk of drug eruption associated with the use of tetracyclines, macrolides, and fluoroquinolones: real-world evidence from a pharmacovigilance study utilizing the FDA adverse event reporting system

Abstract

Recently, concerns regarding the risk of drug eruption associated with the use of tetracyclines were raised by the United States Food and Drug Administration (U.S. FDA). Therefore, it is necessary to reevaluate this risk. This study aimed to quantify the signal of drug eruption associated with the use of tetracyclines. Also, the signal of drug eruption with the use of two other antimicrobial drugs—macrolides and fluoroquinolones—was investigated for comparison. A pharmacovigilance case/noncase study was conducted using the U.S. FDA Adverse Event Reporting System (FAERS). Reports of drug eruptions using the preferred term were retrieved from the database from 2020-2025 quarter 1. The associations between the risk of drug eruption and the three antibiotic classes were quantified using disproportionality analysis utilizing both Bayesian and traditional statistical analyses, including the reporting odds ratio (ROR), proportional reporting ratio (PRR), empirical Bayes geometric mean (EBGM) and information component (IC). All three antibiotic classes were significantly associated with the risk of drug eruption. However, compared with fluoroquinolones and macrolides, tetracyclines were more strongly associated with drug eruption. Disproportionality analysis revealed that the ROR for tetracycline was 6.65 (95% confidence interval (CI), 4.00–11.09) in 2020; in 2024, it was 30.75 (95% CI, 24.46–38.67). For macrolides, the RORs were 2.16 (95% CI, 1.20–3.92) and 10.12 (95% CI, 7.16–14.31) in 2020 and 2024, respectively. The RORs for fluoroquinolones were 4.57 (95% CI, 3.20–6.52) in 2020 and 2.27 (95% CI, 1.31–3.92) in 2024. Males were predominant in the tetracycline group (74%), and females were predominant in the macrolide group (56%). Approximately 41% and 40% of patients in the fluoroquinolone and macrolide classes, respectively, were hospitalized. All three antibiotic classes were associated with the signal of drug eruption. The signal of drug eruption occurred across all age groups with all three antibiotics. Also, the signal has affected both males and females across the years studied. However, there is no specific pattern for the signal.

Introduction

Adverse drug reactions (ADRs) can impact medication adherence and lifestyle factors among patients. The frequency of ADRs ranges from very rare to common on the basis of several factors, including the drug’s mechanism of action, the patient’s risk factors, and potential drug interactions. Furthermore, ADRs can affect any organ system in the human body, including the skin (Rolfes et al. 2016; AlShammari and Almoslem 2018; Kommu et al. 2024).

Drug eruption is an adverse skin reaction that occurs as a result of drug administration, where the drug is deemed the primary cause (Svensson 2008). Drug eruption is not uncommon, as it accounts for approximately 30% of adverse skin-related drug reactions, and is considered to occur frequently in the fields of pharmacovigilance and drug safety (Anderson and Lee 2021; Zhang et al. 2021). Furthermore, drug eruption can be a severe reaction and can lead to fatal or life-threatening situations such as Stevens‒Johnson syndrome (SJS), toxic epidermal necrolysis (TEN), generalized bullous fixed drug eruption (GBFDE), acute generalized exanthematous pustulosis (AGEP), and drug reactions with eosinophilia and systemic symptoms (DRESS). Also, Fixed drug eruption (FDE) is a type of drug eruption in which a skin-related adverse reaction characterized by the appearance of a rash appears at a specific location on the body each time a particular medication is taken. Although these severe forms are uncommon, the annual incidence rates of SJS and TEN range from approximately 27 cases per million people. However, the mortality rate of these events could reach as high as 50%. Therefore, these cases must be taken seriously when patients are treated with medications that may lead to such complications (Weyers and Metze 2011; Zhang et al. 2021).

From a pathophysiological perspective, drug eruption involves genetic associations with HLA and non-HLA genes, drug-specific T-cell-mediated cytotoxicity, T-cell receptor (TCR) restriction, and various cytotoxic mechanisms. Furthermore, it is believed to be triggered by the activation of a latent infection caused by viruses, such as human herpesvirus 6, cytomegalovirus, and Epstein–Barr virus. Drug eruptions that are associated with antimicrobial agents are delayed type IV hypersensitivity (Weyers and Metze 2011; Aushal and Rakesh 2020; Zhang et al. 2021; Kommu et al. 2024).

As for the medications that were reported to be associated with the development of drug eruption, there are several medications include, Nonsteroidal anti-inflammatory drugs (NSAIDs), anticonvulsants, and some antimicrobial agents. Additionally, antibiotic medications, including sulfa-containing drugs (e.g., cotrimoxazole), vancomycin, penicillin, and tetracyclines, are potentially associated with drug eruption. The main concern with drug eruption is that there is no specific treatment for it; instead, all therapeutic approaches are based on supportive care, including antihistamines, fever-lowering agents, and stopping the causative drug (Anderson and Lee 2021; Sokolewicz et al. 2021; Zhang et al. 2021; McClatchy et al. 2022).

There are several case reports of drug eruptions associated with different medications, especially antimicrobial agents. Examples of these medications include doxycycline, cotrimoxazole, azithromycin, and fluoroquinolone (An et al. 2017; Iliyas et al. 2018; Takenaka and Nishizawa 2020; Adhimoolam and Akili 2022; Brehon et al. 2025; Talasila et al. 2025).

A study conducted in the dermatology ward of a university hospital revealed that amoxicillin, lincosamides, and cephalosporins were among the antibiotics that are most commonly associated with drug-induced skin reactions. Furthermore, there were also cases of tetracyclines, fluoroquinolones, and macrolides; however, the number of cases was very low (Sokolewicz et al. 2021). A recently published nested case‒control study examined the risk of serious cutaneous adverse drug reactions (cADRs) associated with various antibiotics. The results revealed that sulfonamide and cephalosporins have the highest risk associated with the development of cADRs (Lee et al. 2024).

Recently, the United States Food and Drug Administration (U.S. FDA) investigated the signal and potential serious risk of drug eruption associated with tetracyclines. (FDA 2024) Therefore, the aim of this study was to quantify the association between the signal of drug eruption and the use of tetracyclines. Furthermore, in this study, the signal of drug eruption associated with the use of two other types of antimicrobial drugs—macrolides and fluoroquinolones—was investigated.

Methods

Study design and settings

This observational pharmacovigilance study was conducted between January 2020 and March 2025 and utilized the FDA Adverse Event Reporting System (FAERS). The FAERS database is a freely accessible database that includes seven datasets, which include demographic and administration information (DEMO), drug information (DRUG), indication for use (INDI), patient outcomes (OUTC), adverse reactions (REAC), report sources (RPSR), and therapy date information (THER). All these datasets include a unique identifier (i.e., primaryid) that is used to link them during data management if a linkage between some or all of these datasets is needed. These files or datasets are released every quarter (i.e., every three months) on the FDA website (https://fis.fda.gov/extensions/FPD-QDE-FAERS/FPD-QDE-FAERS.html) (FDA 2025). The data were accessed on May 14, 2025. This study followed the REporting of a Disproportionality Analysis for DrUg Safety Signal Detection Using Individual Case Safety Reports in PharmacoVigilance (READUS-PV) guidelines (Fusaroli et al. 2024a, b).

The U.S. FDA receives various types of reports from pharmaceutical manufacturers, healthcare professionals (HCPs), patients/consumers, and the public. These reports encompass multiple types, including adverse events, medication errors, and product quality reports. Notably, all these reports are sent voluntarily, except for reports from pharmaceutical manufacturing companies, which are required to submit any safety reports related to their medications to the U.S. FDA. Although the U.S. FAERS is an American database that is run and maintained by the U.S. FDA, it can still accept safety reports from any country worldwide (Alenzi et al. 2024; FDA 2025; Thaibah et al. 2025a, b).

Data management

This study examined the risk of drug eruption in three different antibiotic groups. To find all reports with adverse events of interest, which are “drug eruption” and “fixed drug eruption”, the preferred term (PT), which is part of the Medical Dictionary for Regulatory Activities (MedDRA) hierarchy, was used. This was done in two phases. 1) In the first phase, some of these terms were used to capture all possible cases. These terms were “Erupt”, “Fixed drug”, and “drug erupt”. 2) In the second phase, all the extracted terms were reviewed and further checked to include all the abovementioned PTs, which were “drug eruption” and “fixed drug eruption”.

To define the exposure of interest, both the “drugname” and “prod_ai” fields within the FAERS databases were used. Since this study focused primarily on tetracyclines, tetracyclines approved by the FDA were included. The included medications were tetracycline, doxycycline, minocycline, oxytetracycline, sarecycline, demeclocycline, and eravacycline. For macrolides, the following medications were included in the analysis: azithromycin, clarithromycin, and erythromycin. Furthermore, ciprofloxacin, levofloxacin, and moxifloxacin were included as fluoroquinolones.

The associations between drugs and adverse events within the FAERS database are classified into 4 categories: primary suspected “PS”, secondary suspected “SS”, concomitant “C”, and interacting “I”. To minimize bias from other categories, only the PS category was included in the analysis. Furthermore, to remove duplicate reports, the primary identification number “primaryid” along with the event dates “edven_dt” and PT were used for this purpose. Additionally, the PS code was used to confirm that there were no duplicate reports.

Statistical analyses

Pharmacovigilance analyses using the case/noncase method were disproportionately employed to examine the risk of drug eruption associated with the exposures of interest (i.e., tetracyclines, macrolides, and fluoroquinolones). The cases in the case/noncase method use a contingency table, also known as a 2 × 2 table (Table 1), which has four components: a, b, c, and d. These components are defined as follows: a represents the number of reports of cases (outcome of interest) for the studied drugs; b represents the number of reports of noncases (no outcome of interest, adverse events) for those drugs; c represents the number of reported cases for other medications; and d represents the number of reports of noncases for all other medications (Poluzzi et al. 2012, Thaibah et al. 2025a, b).

Table 1.

Contingency (2X2) table for disproportionality analyses

	Event of interest	Other events in the database
Exposure of interest	a	b
Other exposures in the database	c	d

Descriptive analyses were conducted to compute the demographic data (including age, sex, reporter information, reporting country, and other data) of the report information. Furthermore, traditional and Bayesian statistical analyses were performed to examine and quantify the risk of drug eruption associated with each of the aforementioned exposures of interest. Traditional statistical analyses include reporting odds ratio (ROR) and proportional reporting ratio (PRR) metrics, whereas Bayesian statistical analyses include empirical Bayes geometric mean (EBGM) and information component (IC) metrics. Additionally, their confidence intervals were computed. For a signal to be a potentially significant safety signal, it must have a positive value (i.e., more than zero) for the IC, two for the EBGM, above two for the PRR, and above one for the ROR (Table 2). All the statistical analyses were conducted using the statistical software R (version 4.5.0) and R Studio (Version 2025.05.0 + 496) (Poluzzi et al. 2012, Thaibah et al. 2025a, b).

Table 2.

Traditional and Bayesian disproportionality analyses metrics, and their significant criteria (van Puijenbroek et al. 2002, Poluzzi et al. 2012, Noren et al. 2013, Zhou et al. 2021)

Measures	Formula	Criteria
ROR	$ROR=_{}^{[a\ast d}\left./,_{b\ast c]}^{}\right)$ ${95CI\%=e}^{ln\left(R,O,R\right)\pm 1.96\sqrt{\frac{1}{a}+\frac{1}{b}+\frac{1}{c}+\frac{1}{b}}}$	95%CI/> 1,N ≥ 2
PRR	$PRR=_{}^{[\frac{a}{a+b}}\left./,_{\frac{c}{c+d]}}^{}\right)$ ${95CI\%=e}^{ln\left(P,R,R\right)\pm 1.96\sqrt{\frac{1}{a}+\frac{1}{b}+\frac{1}{c}+\frac{1}{b}}}$	PRR ≥ 2, χ2 ≥ 4,N ≥ 3
EBGM	$\left[\frac{a\left(a,+,b,+,c,+,d\right)}{\left(a,+,c\right)\left(a,+,b\right)}\right]$ ${95CI\%=e}^{ln\left(E,B,G,M\right)\pm 1.96\sqrt{\frac{1}{a}+\frac{1}{b}+\frac{1}{c}+\frac{1}{b}}}$	EBGM $\ge$ 2
IC IC025 IC975	Log2 $\left[\frac{Nobserved+0.5}{N\ast expected+0.5}\right]$ IC-3.3*(O + 0.5)−1/2 −2*(O + 0.5)−3/2 IC + 2.4*(O + 0.5)−1/2 −0.5*(O + 0.5)−3/2	Lower limit of 95% CI $\ge 0$

CI, confidence interval; EBGM, Empirical Bayes Geometric Mean; IC, information component; PRR, proportional reporting ratio; ROR, reporting odds ratio

N^*_expected = (N_drug × N_reaction)/N_total

Results

During the study period, 378 cases of drug eruption were suspected to be associated with the use of the studied medications. One hundred forty-nine (39.4%) cases were linked to tetracycline group use, 119 cases were associated with fluoroquinolone use (31.5%), and 110 cases were associated with macrolide use (29.1%). Most of the patients in the tetracycline group were between the ages of 18 and 39 years (64 patients, 43%); however, the patients aged between 60 and 89 years (52 patients, 44%) were the most common age cohort in the fluoroquinolone group, whereas very young patients (i.e., less than 18 years; 32 patients, 29%) were the most commonly affected age cohort in the macrolide group (Table 3). Notably, there were no duplicate reports.

Table 3.

Characteristics of drug eruption reports associated with the three antibiotic classes

Medication Class	Tetracyclines		Macrolides		Fluoroquinolones
Characteristics	n	%	n	%	n	%
Patient's Sex
Male	110	74	46	42	54	45
Female	29	19.4	62	56	52	44
Unknown	10	6.6	2	2	13	11
Patient's Age (years)
0–17	2	1.3	32	29	0	0
18–39	64	43	26	23.7	17	14.2
40–59	44	29.5	19	17.3	26	22
60–89	27	18.1	22	20	52	44
≥ 90	1	0.7	3	2.7	1	0.8
Unknown	11	7.4	8	7.3	23	19
Reporter's Occupation
Physician	40	26.8	17	15.5	42	35.5
Pharmacist	11	7.4	3	2.7	9	7.5
Other Health Professional	93	62.5	67	60.9	55	46
Lawyer	2	1.3	0	0	3	2.5
Consumer/Patient	3	2	23	20.9	10	8.5
Report date
2020	15	10	11	10	31	26
2021	10	6.7	21	19	16	13.5
2022	11	7.4	21	19	27	22.6
2023	18	12.1	17	15.5	28	23.5
2024	79	53	33	30	13	11
2025*	16	10.8	7	6.5	4	3.4
Outcome
Hospitalization—Initial or Prolonged	16	10.8	44	40	49	41
Disability	2	1.3	2	2	0	0
Life-Threatening	0	0	0	0	2	1.6
Other Serious (important medical event)	121	81.2	57	51.5	51	43
Death	3	2	0	0	4	3.4
Unknown	7	4.7	7	6.5	13	11
Total	149		110		119

*Includes Quarter 1 (January to March 2025)

Male sex was more common in the tetracycline group (110 males, 74%), whereas the sex ratio was more balanced in the fluoroquinolone group (54 males, 45%). However, female sex was more common in the macrolide group (62 females, 56%). Approximately 41% and 40% of the patients were hospitalized because of drug eruptions associated with the use of fluoroquinolones and macrolides, respectively. Moreover, approximately 11% of the patients were hospitalized because of the use of tetracyclines. However, three patients died from the use of tetracyclines, and four died from the use of fluoroquinolones (Table 3).

Healthcare professionals were the most common reporters for all three groups; however, consumer reports were more common in the macrolide group (23, 21%) than in the other two groups (Table 3). Interestingly, most of the reports for all three groups were from outside the U.S. For tetracyclines, 130 (87%) reports were from outside the U.S.; the majority of these reports came from France (100, 67%). There were 97 fluoroquinolone reports from outside the U.S. (81%); the majority of these reports were from Japan and France, each accounting for 15 reports (13%). However, more reports were from the U.S. (22, 18%) than from Japan and France. Finally, for macrolides, 99 (90%) of the reports were from outside the U.S., and almost half of these reports were from China (56, 51%) (Fig. 1).

Fig. 1.

Top reporting countries categorized by antibiotic class. CN, China; FR, France; GB, Great Britain; JP, Japan; US, United States

There were differences in the year of reporting among the three groups, as the peak number of reports for tetracyclines and macrolides occurred in 2024, with 79 (53%) and 33 (30%) reports, respectively. In comparison, the peak number of reports for fluoroquinolones occurred in 2020, with a total of 31 reports (26%) (Table 3).

Doxycycline had the highest number of reports in the tetracycline group (145, 97%), whereas azithromycin (78, 71%) was the most commonly reported drug in the macrolide group. However, both ciprofloxacin (52, 44%) and levofloxacin (44, 37%) had comparable numbers of reports (Fig. 2).

Fig. 2.

Number and percentages of FAERS drug eruption reports retrieved per antibiotic class

A statistically significant association between the risk of drug eruption and all three drug groups was found. However, the magnitude of the association with tetracycline use was greater than that for the other two drug groups. There was a slight but persistent increase in the risk of drug eruption with the use of tetracyclines during the first four years (i.e., 2020–2023), with an ROR of 6.65 (95% confidence interval (CI) 4.00–11.09) in 2020 and an ROR of 7.44 (95% CI 4.67–11.85) in 2023. However, the magnitude of the risk sharply increased in 2024 and the first quarter of 2025, with RORs of 30.75 (95% CI, 24.46–38.67) and 26.80 (95% CI, 16.16–44.46), respectively (Table 4).

Table 4.

Disproportionality analyses of the signals of drug eruption with the use of the three antibiotic classes classified by year

Year	Antibiotic class	Number of drug eruption ADEs	Reported ADEs with the drug of interest	ROR (95% CI)	PRR (95% CI)	EBGM (95% CI)	IC (IC025-IC975)
2020	Tetracyclines	15	2,151	6.65 (4.00–11.08)	6.61 (4.00–11.02)	5.57 (3.35–9.28)	2.47 (2.37–2.56)
	Macrolides	11	4,801	2.16 (1.20–3.92)	2.16 (1.20–3.92)	2.05 (1.14–3.72)	1.04 (0.90–1.14)
	Fluoroquinolones	31	6,503	4.57 (3.20–6.52)	4.55 (3.19–6.49)	4.25 (2.98–6.08)	2.10 (2.04–2.13)
2021	Tetracyclines	10	1,697	6.81 (3.65–12.71)	6.78 (3.63–12.65)	5.29 (2.84–9.88)	2.40 (2.24–2.52)
	Macrolides	21	4,133	5.90 (3.83–9.10)	5.87 (3.81–9.10)	5.23 (3.40–8.05)	2.38 (2.31–2.44)
	Fluoroquinolones	16	5,002	3.69 (2.25–6.10)	3.68 (2.25–6.03)	3.38 (2.07–5.55)	1.76 (1.66–1.83)
2022	Tetracyclines	11	1,740	6.61 (3.65–12.00)	6.58 (3.63–11.93)	5.27 (2.91–9.57)	2.39 (2.26–2.50)
	Macrolides	21	3.012	7.34 (4.76–11.31)	7.30 (4.74–11.24)	6.31 (4.10–9.72)	2.65 (2.58–2.71)
	Fluoroquinolones	27	4,029	7.07 (4.83–10.36)	7.03 (4.80–10.30)	6.26 (4.28–9.18)	2.64 (2.59–2.69)
2023	Tetracyclines	18	2,203	7.43 (4.66–11.85)	7.38 (4.63–11.77)	6.25 (3.93–9.97)	2.64 (2.56–2.71)
	Macrolides	17	3,100	4.97 (3.08–8.02)	4.95 (3.06–7.99)	4.42 (2.74–7.14)	2.14 (2.05–2.21)
	Fluoroquinolones	28	4,926	5.18 (3.56–7.53)	5.15 (3.54–7.50)	4.75 (3.27–6.91)	2.25 (2.19–2.29)
2024	Tetracyclines	79	2,333	30.75 (24.46–38.66)	29.75 (23.66–37.40)	24.35 (19.36–30.60)	4.60 (4.24–4.87)
	Macrolides	33	2,822	10.12 (7.16–14.31)	10.01 (7.08–14.16)	8.71 (6.16–12.31)	3.12 (3.07–3.16)
	Fluoroquinolones	13	4,847	2.27 (1.31–3.92)	2.27 (1.31–3.92)	2.16 (1.25–3.73)	1.11 (1.00–1.20)
2025	Tetracyclines	16	552	26.80 (16.15–44.45)	26.05 (15.70–43.20)	14.53 (8.76–24.10)	3.86 (3.76–4.45)
	Macrolides	7	782	7.94 (3.75–16.81)	7.88 (3.72–16.68)	5.35 (2.53–11.33)	2.42 (2.20–2.58)
	Fluoroquinolones	4	950	3.69 (1.37–9.90)	3.68 (1.37–9.87)	2.82 (1.05–7.57)	1.49 (1.12–1.71)

Tetracyclines (italic), Macrolides (underlined), Fluoroquinolones (bold)

ROR, reporting odds ratio; PRR, proportional reporting ratio; EBGM, empirical Bayes geometric mean; IC, information component

The risk of drug eruption associated with macrolides was greater than the risk associated with fluoroquinolones; however, fluctuations in the magnitude of the risk were detected during the study period (i.e., 2020–2025). The EBGM was also significant for the risk of drug eruption with macrolides. The magnitudes were 2.05, 6.31, 4.42, and 5.35 for 2020, 2022, 2023, and 2025, respectively. Similarly, for fluoroquinolones, there was a fluctuation in the magnitude of the risk over the study period. The PRRs were 3.69 (95% CI 2.25–6.04), 7.03 (95% CI 4.80–10.30), 5.15 (95% CI 3.55–7.50), and 2.27 (95% CI 1.31–3.92) in 2021, 2022, 2023, and 2024, respectively. All the metrics, including ICs, revealed statistically significant associations between the risk of drug eruption and the use of the three types of drugs (Table 4).

Discussion

Drug eruption is not an uncommon ADR and can lead to serious conditions, such as SJS, TEN, AGEP, GBFDE, and DRESS, which can be life-threatening. Medications are common reasons for drug eruption; however, the risk of drug eruption varies across medications (Zhang et al. 2021; Zhu et al. 2022; Shrestha et al. 2025). Recently (2024), the U.S. FDA has been investigating the risk of drug eruption associated with tetracyclines to determine whether regulatory action is warranted (FDA 2024).

Therefore, this study examined the signal of drug eruption associated with the use of tetracyclines. Furthermore, the study included two additional antibiotic classes (macrolides and fluoroquinolones) for comparison purposes and to minimize the bias associated with the disease. Both macrolides and fluoroquinolones share indications similar to those of tetracyclines “e.g., urinary tract infections, respiratory tract infections, pneumonia, and acne” (Regunath and Oba 2024; Reynolds et al. 2024, Sabih and Leslie 2024). Macrolides were selected because their mechanism of action is similar to that of protein synthesis inhibitors. Often, medications with similar mechanisms of action are used for similar indications, and some adverse events are due to these mechanisms. However, one of the suggested pathophysiological mechanisms of drug eruption is a type of hypersensitivity that is considered a type B adverse drug reaction (Kommu et al. 2024). These reactions are characterized as unpredictable, not dose-related, and often serious. However, fluoroquinolones were selected because both tetracyclines and fluoroquinolones are associated with the risk of photosensitivity, which is a skin reaction (Zhu et al. 2022). This study followed the READUS-PV guidelines in several parts, including the background, methods, results, and discussion. For example, the databases that are used in this study were mentioned and explained, and the disproportionality analyses were explained in detail. With respect to the results, both the descriptive findings and the results of the disproportionate analyses were mentioned. Other studies were compared within the discussion analyses, and the limitations of the study were also mentioned (Fusaroli et al. 2024a, b).

This study revealed a signal associated with all three groups; however, the signal associated with tetracyclines is greater than that associated with the other groups, particularly in 2024 and the first quarter of 2025. In 2024, the EBGM and IC were 24.35 and 4.60, respectively, whereas for macrolides, these parameters were 8.71 and 3.12, respectively. Nonetheless, the signal was higher in 2022 for macrolides than for the other two drug groups. This difference may be attributed to the extensive use of macrolides, particularly azithromycin, during the COVID-19 pandemic (Herrera-Lasso Regas et al. 2020; Bednarcuk et al. 2023; Barisic et al. 2025). The signal increased sharply by more than double during 2021 and continued to grow in 2022 before declining in 2023.

The sharp increase in the signal of drug eruption with the use of tetracyclines, particularly doxycycline, could be due to the 50–70% increase in the utilization of the drug, especially for postexposure prophylaxis for bacterial sexually transmitted diseases (STDs) (Bednarcuk et al. 2023; Sunagawa et al. 2025). In 2023, a clinical trial revealed the efficacy of doxycycline in reducing the incidence of bacterial STDs by two-thirds (Luetkemeyer et al. 2023). Additionally, the Centers for Disease Control and Prevention (CDC) released a recent guideline for 2024 regarding the use of doxycycline as a postexposure prophylaxis (PEP) for bacterial STDs (Bachmann et al. 2024). The signal associated with tetracyclines, as well as macrolides and fluoroquinolones, was also high in all the study years, including the initial years (i.e., during the pandemic). This finding could also be due to several assumptions and opinions regarding the need for these medications and their use in patients with COVID-19 to limit complications or treat pneumonia associated with the disease. However, it is essential to note that the use of tetracyclines as PEP may exceed their use during the COVID-19 pandemic. (Karampela and Dalamaga 2020; Al-Kuraishy et al. 2021; Chedid et al. 2021; Mosquera-Sulbaran and Hernandez-Fonseca 2021; Garrido-Mesa et al. 2022; Kournoutou and Dinos 2022; Hrovat et al. 2024).

Considering the literature on a similar topic, few studies have quantified the risk of drug eruption associated with these three antibiotic classes; instead, previous studies have focused mainly on the frequency of drug eruption and other skin reactions. A study by EM Sokolewicz et al. reported that compared with the other two drug groups, tetracyclines were more strongly associated with the risk of drug eruption. The results revealed that 6% of all cutaneous drug reactions were attributed to tetracycline use, with five leading causes of this reaction. Furthermore, 2.4% of reactions were reported for both macrolides and fluoroquinolones (Sokolewicz et al. 2021). The results of this study align with those of our study, as both studies revealed that tetracyclines are associated with drug eruptions.

Nevertheless, the study by Sokolewicz et al. was conducted in one ward in a university hospital, whereas our study included databases that cover different populations and perhaps different practices. Additionally, we have a relatively larger sample size than Sokolewicz et al.’s study does. Additionally, our study is updated, as we have included data up to the first quarter of 2025.

Furthermore, van der Linden et al. reported a higher rate of cutaneous reactions with fluoroquinolones (1.6%) than with tetracyclines (0.5%) and macrolides (0.3%). However, the study was not recent, and the utilization of these medications may have changed, possibly because of recent regulatory actions on fluoroquinolones in particular. Additionally, the sample size was relatively small compared with that in our study and was limited to the Netherlands (van der Linden et al. 1998; Barberan et al. 2024). Another study by Alshareef H et al. investigated the pharmacovigilance of different antibiotics on a small scale. However, this study revealed that the frequency of dermatological reactions was greater in the fluoroquinolone group (4.8%), followed by the macrolide group (1.8%) and then the tetracycline group (0.4%) (Alshareef et al. 2025).

The sex distributions differed between the three antibiotic classes in this study, with males being predominant in the tetracycline and fluoroquinolone groups and females being predominant in the macrolide group. These sex differences could be due to female children being more likely to develop skin reactions than their male counterparts are; accordingly, both tetracyclines and fluoroquinolones are not recommended or indicated in pediatric populations (Zhang et al. 2021). Additionally, macrolides are commonly used in pediatric populations for multiple indications, particularly since the pediatric population (i.e., those < 17 years old) has the highest incidence of drug eruptions and is often prescribed macrolides. Additionally, a study by Jadhav, Anuja et al. reported similar results; they reported that most of the patients in their study were female (58%). However, they included nonantibiotic medications in their study (Jadhav et al. 2021).

The study results were similar to those of EM Sokolewicz et al. as the risk associated with tetracyclines and fluoroquinolones was greater among males than among females (Sokolewicz et al. 2021). Since neither tetracyclines nor fluoroquinolones are recommended during pregnancy, this could also be a factor in the higher prevalence of males among these two antibiotic classes (Chae et al. 2024; Nakitanda et al. 2024).

Furthermore, a study by Yeh and Liu, which characterized patients with fixed drug eruption, reported results similar to those of the current study, with males accounting for 63%; however, the study included both antibiotic and nonantibiotic medications (Yeh and Liu 2025). It is generally believed that the risk of adverse drug reactions, including skin reactions, is greater in females than in males. Therefore, in general, it is important to focus on all patients using the three antibiotic classes in this study, particularly for tetracyclines and fluoroquinolones (Rademaker 2001).

Most age groups were affected by drug eruption. The age group of 18–59 years is the most commonly affected by tetracyclines among the other drug groups. Furthermore, elderly patients (i.e., 60–89 years old) were more commonly affected by fluoroquinolones, and pediatric patients (i.e., 0–17 years old) were most commonly affected by macrolides. Notably, all three drug groups had a considerably high frequency among adult populations (i.e., greater than 15%). Approximately 43% of the patients who were on tetracyclines were aged 18–39 years, possibly because those patients are probably sexually active, and this drug class is commonly used to treat bacterial STDs (Beutel et al. 2002; Luetkemeyer et al. 2023; Bachmann et al. 2024; Sunagawa et al. 2025). Furthermore, acne is common among this age group, and doxycycline is among the treatment options for those patients (Baldwin 2020). On the other hand, patients with drug eruption who were on fluoroquinolones were more likely to be adults or elderly individuals. This may be due to the types of diseases for which these medications are used, such as urinary tract infections (UTIs) and respiratory tract infections (Armstrong 2020; Sabih and Leslie 2024). There were no patients aged less than 18 years among the fluoroquinolone group because these medications are not indicated for pediatric populations because of the risk of tendon rupture (Bradley et al. 2011; Choi et al. 2013). Patients on macrolides were more prevalent in pediatric populations; however, they were also distributed across all age groups, except those over 90 years old. This might be because macrolides are commonly used for different indications, such as upper and lower respiratory tract infections, bacterial STDs, acne, and Helicobacter pylori (H. pylori) infections (Armstrong 2020; Baldwin 2020; Patel and Hashmi 2025).

A study by Yeh and Liu reported results similar to those of this study, as the majority of the patients in their study (53.4%) were aged 19–64 years (Yeh And Liu 2025). Additionally, EM Sokolewicz et al. reported similar age distributions in drug eruption cases. Similarly, studies have reported a similar trend among all age groups, as observed in the current study (Jadhav et al. 2021; Rajendran et al. 2021). Nevertheless, it is necessary to consider the interaction between age and sex, which might impact the prevalence of drug eruption among those who were on the three antibiotic classes among different age groups for different sexes.

Most of the reported events were from healthcare professionals (HCPs), including physicians, pharmacists, and other HCPs. Additionally, reports from consumers or patients are rare. This is expected because patients usually go directly to primary care or hospitals when they experience adverse skin reactions, especially if they are moderate or severe. Additionally, reactions related to organ systems in the human body are often identified and diagnosed by skilled healthcare professionals, particularly physicians, unlike mental or behavior-related adverse reactions, which are usually reported more by consumers/patients than by HCPs (Aydınkarahaliloğlu et al. 2018; Thaibah et al. 2025a, b). There are few reports by lawyers, who are believed to have a very minimal impact on safety signals within spontaneous reporting systems, such as FAERS (Rogers et al. 2019).

The majority of the outcomes reported among the three antibiotic classes were other serious medical events, followed by hospitalizations or prolonged hospitalizations, which also occurred at a high frequency. However, compared with those among tetracyclines, hospitalizations were almost triple the rate among fluoroquinolones and macrolides. This might be because patients taking fluoroquinolones and macrolides are older than those taking tetracyclines are (44%, 20%, and 18.1%, respectively). Additionally, since these two antibiotic classes are more commonly used among patients with different bacterial respiratory tract infections, they may also be admitted to the hospital and stay longer than expected (Armstrong 2020; Abdelsalam Elshenawy et al. 2024).

On the other hand, serious events occurred more frequently with tetracyclines (81.2%) than with fluoroquinolones (43%) and macrolides (51.5%). The results of this study revealed that the signal of drug eruption is greater with tetracyclines than with the other two antibiotic classes, which may explain why the majority of outcomes are serious. Additionally, adverse skin reactions usually lead to serious conditions or complications (Zhang et al. 2021; Al Aboud et al. 2025). Deaths occurred in 2% and 3.4% of patients with drug eruption who were taking tetracyclines and fluoroquinolones, respectively. However, these findings need to be interpreted with caution, as they cannot be directly linked to the use of these medications. Nevertheless, monitoring patients on these medications is worthwhile (Marzano et al. 2016; Zhang et al. 2021; Al Aboud et al. 2025).

Doxycycline was the drug that was most commonly associated with drug eruption. This finding may be because doxycycline is the most commonly prescribed tetracycline, as it is listed in various guidelines for conditions such as Lyme disease, postexposure STD prophylaxis, acne, UTIs, and pneumonia; furthermore, this drug is affordable (Zaenglein et al. 2016; Armstrong 2020; Dorobisz et al. 2021; Lantos et al. 2021; Sabih and Leslie 2024; Sunagawa et al. 2025). It is almost the same as azithromycin (a macrolide), as it is the most widely utilized drug. This is mainly because it is included in most guidelines for treating various diseases, and it is also easy for patients to use, as it is administered once daily for only a few days. Clarithromycin is usually prescribed and used to treat peptic ulcers associated with H. pylori (Armstrong 2020; Patel and Hashmi 2025; Sabih and Leslie 2024). The effects of fluoroquinolones differ slightly from those of previous antibiotic medications because there were no significant differences among the three drugs within the group, especially between ciprofloxacin and levofloxacin, as they have similar indications. Additionally, these two drugs were approved earlier than the newest drug (i.e., moxifloxacin). Moreover, moxifloxacin has narrower indications than ciprofloxacin and levofloxacin do (Armstrong 2020; Sabih and Leslie 2024).

In general, non-U.S. countries have more cases of drug eruption than the U.S., especially France, Japan, and Great Britain. In addition to the previous high utilization of these medications, this might also be because people from these countries have greater susceptibility to developing skin reactions (Tang et al. 2025). Interestingly, China has the greatest number of cases of drug eruptions associated with macrolides. This could be because macrolides are widely used in China (Yang et al. 2022; Li et al. 2023a, b). These results are consistent with those of D Li et al. (Li et al. 2023a, b). In addition, the distribution of different HLAs among different countries and the linkage of these HLAs to specific antibiotic-induced drug eruptions might also impact the particular occurrence of these antibiotic-induced drug eruptions in various countries. For instance, the suggested HLAs for macrolide-induced drug eruptions are HLA-B*51:01, HLA-A*02:07,and HLA-B*15:27, which are common HLAs in Asian populations, including Chinese and Japanese individuals. (Zhu et al. 2022; Wung et al. 2023; Mahmud et al. 2025). With respect to fluoroquinolones, few reports exist concerning the genetic predisposition; however, some reports link this reaction to HLA-B*22, HLA-B*57:01, HLA-B*13:01, and HLA-B*13:02, which are common in Europeans and Asians, including Japanese and Chinese individuals (Jain and Jain 2013; Zhu et al. 2022; Jiang et al. 2023; Mahmud et al. 2025).

On the other hand, no specific HLA has been identified as responsible for tetracycline-induced drug eruption, especially in European or North American populations, as most cases associated with tetracyclines have been reported from these regions. Instead, these higher case numbers might be due to phototoxicity and photoallergic reactions, which lead to the generation of reactive oxygen species (ROS) and ultimately damage keratinocytes (Hofmann and Weber 2021). Therefore, patients who are prescribed tetracyclines or even fluoroquinolones are advised to avoid sun exposure (Di Bartolomeo et al. 2022). Nevertheless, as mentioned earlier, immune-mediated hypersensitivity involving HLA is one of the mechanisms that is associated with such a reaction (Hamilton and Guarascio 2019). Therefore, national pharmacovigilance systems in different countries must consider this information when future cases are evaluated.

This study has the advantages of investigating three commonly used antibiotic classes (tetracyclines, macrolides, and fluoroquinolones) in relation to the signal of drug eruption and quantifying this signal using both Bayesian and traditional disproportionality analyses. Additionally, a vast database such as FAERS, which receives reports from all countries worldwide, and different reporting sources such as consumers and HCPs are utilized. Additionally, the database reflects real-world patient data and has the advantage of the ability to measure signal over multiple years. The study also utilized all four metrics of Bayesian and traditional disproportionality analyses (i.e., the ROR, PRR, EBGM, and IC) to enhance the robustness and validity of the study’s findings.

However, this study has limitations; as a result of the nature of the spontaneous reporting system, the information may not be complete and may contain missing data such as age, sex, and comedications. For example, a lack of ethnicity information might impact the risk of developing the event of interest. Also, not all tetracycline medications were included in this study (e.g., tigecycline). However, tigecycline was not included because the drug has a black-box warning, as studies have found that all-cause mortality was higher with this drug. It is used only when other alternatives are not suitable (FDA, TYGACIL®, 2025). Therefore, we did not include it to avoid any potential bias that could affect the results. Additionally, bias due to channeling bias and confounding indications might be an issue, as doxycycline and azithromycin are used for acne; however, this may not significantly impact the results, since fluoroquinolones also carry a risk. We limited the report categories to “PS” and excluded “SS”, “C”, and “I” codes to reduce biases; however, this could also omit potentially relevant signals. Although this study used the drug eruption PT term to retrieve the data, it does not necessarily cover all types of drug eruptions, such as SJS/TEN, DRESS, and AGEP. Underreporting is also a major concern in pharmacovigilance and spontaneous reporting systems; therefore, not all real cases of drug eruption have been reported (Fusaroli et al. 2024a, b). Finally, owing to the nature of this observational study, which uses spontaneous reporting systems, we cannot conclude causality from this study; instead, we use the term 'association'.

Recent concerns about drug eruptions linked to tetracyclines, raised by the U.S. FDA, could lead to notoriety bias. However, this issue may not be significant, as the risk had already existed prior to the FDA's investigation. Additionally, the risk was observed to increase with macrolides during the year when the FDA’s concerns were publicized. Furthermore, a study has indicated that such bias is not always present in this database (Neha et al. 2021). Furthermore, as the nature of this study was observational, it was unable to adjust for multiple confounders, such as concomitant medications and comorbidities, owing to the lack of this information.

Conclusion

All quantitative study results, including ROR, PRR, EBGM, and IC, demonstrated a signal between drug eruption and all three antibiotic classes studied. The signal of drug eruption occurred across all age groups with all three antibiotics. Also, the signal has affected both males and females across the years studied. Therefore, healthcare professionals should be aware of the risk and its magnitude across various groups during their practice, especially since it can be serious, as evidenced by the results. Furthermore, consumers and patients should be educated about these risks to increase their knowledge and facilitate reporting. A further well-controlled pharmacoepidemiological study might be required to confirm this signal of drug eruption with these antibiotic groups.

Author contribution

Thamir M. Alshammari contributed to all aspects of the research, including conceptualization, methodology, project administration, writing—original draft, writing—review and editing, and approval of the final manuscript version.

Funding

This research received no external funding.

Data availability

The FDA Adverse Event Reporting System (FAERS) quarterly data extract files are publicly available and can be obtained from https://fis.fda.gov/extensions/FPD-QDE-FAERS/FPD-QDE-FAERS.html (accessed on 4 October 2024).

Declarations

Institutional review board statement

Not applicable.

Informed consent statement

Not applicable.

Competing interests

The author declares that he has no competing interests.