Nalidixic acid was accidentally discovered as a chloroquine byproduct in 1962 and several additional quinolones, including ciprofloxacin, levofloxacin, and moxifloxacin, have since been marketed. These are widely prescribed for the treatment of infectious diseases in both inpatient and outpatient adults because of their broad-spectrum antimicrobial activity, excellent oral absorption, high tissue penetration, and long elimination half-life which permits once-daily dosing (1). While fluoroquinolones (FQs) also have favorable pharmacokinetic properties in children, the response to these drugs differs between children and adults (2). In addition, the efficacy of FQs in children is similar to other antibiotics for the treatment of respiratory infections, cystic fibrosis, febrile neutropenia, and urinary tract infections (3–7). Some FQs have FDA approval for the treatment of complicated urinary tract infections, inhalation anthrax, and pyelonephritis in children (8). As the prevalence of antibiotic-resistant pathogens such as Pseudomonas aeruginosa, Tuberculosis, and Mycoplasma, has increased, FQs are being increasingly prescribed to children. In the United States, 372,357 oral FQs were dispensed to children from 2006 to 2016 (9).

Quinolone safety is an area of focus in pediatrics. Early toxicology studies in juvenile animals found that quinolones can induce dose-related arthrotoxicity (10, 11), leading to a ban on their use among children aged <18 years in China. In addition, a clinical trial about the safety of ciprofloxacin in children demonstrated that the ciprofloxacin group had a small but statistically significant increase in musculoskeletal adverse events (AEs) in comparison to the control group (9.3% and 6%, respectively) 6 weeks after treatment (8). Another randomized controlled study of pediatric patients with complicated intra-abdominal infections found that QT prolongation, a heart rhythm disorder, was higher in the moxifloxacin group than in the control group (7% and 1.3%, respectively) (2). However, other studies found no AEs associated with FQ use in children (8). Human studies have shown that there is a very low risk of FQ-associated musculoskeletal disorders. A 5-year follow-up study indicated that levofloxacin-treated children reported slightly lower musculoskeletal AEs than comparator children, and indicated that the risk of cartilage injury was rare (12). Moreover, no growth abnormalities were associated with FQ treatment in children (12). In 2018, a multicenter study (NCT01069900) examined the efficacy and safety of moxifloxacin in children with complicated intra-abdominal infections (2). The study found that after 5–14 days of treatment, the overall incidence of AEs was similar between the moxifloxacin and comparator groups and most AEs were mild or moderate. The proportion of children who experienced musculoskeletal AEs was comparable between the two treatment arms (moxifloxacin group: 4.3%; comparator group: 3.3%) (2). A recent systematic review including 154,638 children found that the most common AEs associated with systemic quinolones were gastrointestinal reactions such as diarrhea, vomiting, nausea, and abdominal pain (13). This study also found that the quinolone-treated children had a higher risk of musculoskeletal AEs than those in the comparator group after 2–12 months of follow-up (2–6 months: RR 2.56, 95% CI, 2.26–2.89; 7–12 months: RR 1.35, 95% CI, 0.98–1.86); however, the incidence of musculoskeletal AEs was uncommon and reversible (13). Considering the inadequacy of long-term safety studies, the American Academy of Pediatrics (AAP) has recommended FQ use in children only when no effective and safe alternative is available (7).

While a few studies have identified other AEs associated with quinolone use among pediatric patients, such as cardiotoxicity and neurotoxicity, most have focused on musculoskeletal side effects. In addition, except for levofloxacin, few large and long-term studies of FQ safety have been conducted in children. To gain further insight into quinolone-associated AEs in children and better protect pediatric patients, continued surveillance of safety data is essential. Data mining is a useful method for monitoring the safety of new drugs and regularly supervising the effects of established drugs. To our knowledge, no studies have mined the FDA adverse events reporting system (FAERS) database for AEs and adverse outcomes associated with the clinical use of systemic or topical quinolones in children. To address this, the current study sought to characterize and map the safety profiles of quinolone-associated AEs in children using FAERS data.

The ROR and BCPNN methods were used to assess the association between AEs and quinolone treatment in pediatrics department and further to evaluate its medicine safety in children. Our findings were generally consistent with the AEs in the manufacturer’s instructions, demonstrating the validity of the data mining methods.

Our study found that the most prevalent AEs associated with quinolone use (oral or intravenous administration) in children were gastrointestinal reactions. We found three significant (ciprofloxacin, levofloxacin, and moxifloxacin) signals for vomiting, and two significant signals (ciprofloxacin and levofloxacin) for nausea, abdominal pain, and diarrhoea. Our analysis found a significant signal for abdominal distension with moxifloxacin, which were consistent with a meta-analysis by Li et al. (13) and a similar studies in adults (24, 25). A strong signal of arthralgia was elicited from ciprofloxacin and levofloxacin, which were similar to the reported research by Adefurin et al. (26).

In terms of age, 12–18-year-old patients were the most reported age group in our study, which might be related to the higher frequency of quinolone use in this population (13). FQ-induced tendon disorders were reported to occur tendon ruptures within 1 week of administration and tendinopathies within the first month (27), Similarly, we observed that the onset of the most frequent FQ-musculoskeletal ADRs was 7 days for the median treatment with ciprofloxacin, 8 days for levofloxacin, 5 days for moxifloxacin and ofloxacin, and 41 days for ozenoxacin. Peripheral neuropathy caused by quinolones in adults has been reported in previous studies (28), but has not been observed in children., two signals of CNS disorders, dyskinesia (ROR = 7.01, IC = 2.44) and peripheral neuropathy (ROR = 27.02, IC = 3.56) were detected in this study, which require further study.

Our analysis found significant signals of ciprofloxacin and levofloxacin in musculoskeletal adverse events, which is showed in the instructions. An estimated risk of one musculoskeletal AE for every 62.5 patients and a 57% increased risk of arthropathy associated with ciprofloxacin was reported by Adefurin et al. (26). Our results identified that musculoskeletal pain were associated with levofloxacin and ciprofloxacin, which are concordant with the previous findings (2, 29). Although quinolone use in children may be lead to musculoskeletal AEs, long-term follow-up studies suggest that they are reversible and do not contribute to growth problems (12). The exact pathophysiology of tendon injury induced by FQ is still mysterious (27). FQ-induced tendon damage is thought to be associated with oxidative stress and mitochondrial toxicity (28). FQs interact with regulatory proteins of the tenant and further damage the tendon structure. Besides, the final event in the pathogenetic mechanism has been suggested to be apoptosis (30).

Cardiac arrest was uncommon, only linked to moxifloxacin and ozenoxacin. An RCT conducted by Stefan et al. (2) found a drug-related QT prolongation of 9.3% and a recent systematic review and meta-analysis identified a prolonged electrocardiogram QT interval of 0.03% (13). It might be a fatal arrhythmia. Prolongation of the QTc interval increases the risk of Torsade de Pointes (TdP) arrhythmia, especially when the QTc interval exceeds 500 ms or the extension is larger than 60 ms compared to the pre-processing value. There are many risk factors of prolonged electrocardiogram QT interval, including female sex, bradycardia, hypokalemia, hypocalcemia, hypomagnesemia, history of cardiac disease and treatment with more than one QTc-prolonging medication. Therefore, any link between moxifloxacin-induced QTc extension and risk factors was hardly established (2).

For drug resistance, our findings showed significant signals in moxifloxacin and ofloxacin.

Therefore, the occurrence of drug resistance should also be closely monitored in children. The use of FQs elevates in this population, and drug resistance will also consequentially increase. One study reported (31) the correlation between FQs use and the emergence of ciprofloxacin and levofloxacin resistance in gram-negative bacilli in hospitalized children.The current study identified drug resistance signals induced by moxifloxacin and ofloxacin.

Quinolones have also been associated with seizures in children. We only found significant signals in ozenoxacin. An increase in seizures (0.63%) was found among children without CNS disorders (0.02%) after using quinolones such as ciprofloxacin, levofloxacin, and gatifloxacin (32). Ozenoxacin is a novel, non-fluoroquinated quinolone, and approved for topical use by the FDA in 2017. This drug is bactericidal against gram-positive pathogens, including methicillin-resistant and methicillin-sensitive Staphylococcus aureus. Ozenoxacin lacks a fluorine atom and have fewer side effects than other fluoroquinolones (33). Ofloxacin produced typical quinolone-induced lesions in the articular cartilage of three of 10 juvenile rats, while chondrotoxicity was unassociated with ozenoxacin (34). Oral administration of ozenoxacin in juvenile dogs demonstrated no chondrotoxicity or toxicologically in select target organs. Ozenoxacin-associated rhinitis, and rare AEs, such as cardiac arrest and epilepsy, were identified by Savion Gropper et al. (35). Only rosacea and seborrheic dermatitis are listed as AEs on the ozenoxacin package insert, while the current study identified strong signals for coma, somnolence, suicide and attempted suicide. Although the number of AEs was low, these side effects should be closely monitored during ozenoxacin use in children. Further studies are needed to assess its safety profile in this population.

Our study has several limitations. First, AEs are often underreported in spontaneous reporting systems, which is caused by various factors, such as the underreporting, absence, and exclusion of healthy individuals, the lack of a denominator, and the presence of potential confounders (36). The FAERS database is not appropriate for estimating incidence rates, due to the absence of a denominator (37). Second, most reporters are from the United States, European and South American countries; therefore, it is uncertain whether the findings are applicable to individuals of other races and ethnicities. Third, although most of the reports (75.4%) are reported by health professionals (physician, pharmacist and other health-professional), and the rest of the cases submitted by consumers have not been verified, and the reliability of the reported information may be affected by these reasons. Last, the FDA does not require a proof of the cause-and-effect relationship between AE and drug, so the reports often lack detailed information about the AE.

In summary, this study provides an objective reference for pharmacovigilance work by mining the safety signals of quinolone use in children. Focus should be placed on AEs with strong real-world signals, such as the cardiac arrest and suicide death associated with ozefloxacin that are not included in the label. In addition, both the risk and benefit should be appropriately weighed when quinolones are prescribed for children. Pharmaceutical care should be strengthened for patients with risk factors for heart disease, skeletal muscle disease, and hematological conditions. Close attention should be paid to disease progression and timely intervention measures should be taken when AEs occur to reduce the risk of poor outcomes.