Use of medicines in pediatric patients entails more difficulties than in adults as physiological development brings about significant changes in the bioavailability of the drugs in the patients’ body1,2. As a result, special attention must be paid not only to the dosing required by each particular individual but also to any potential drug-drug interactions (DDIs) that could interfere with the effect of the therapy, negatively impacting its effectiveness and/or safety profile3,4.

When DDIs appear as a clinical manifestation in the patient they are called real DDIs (rDDIs). Such clinical manifestations, regardless of their specific mechanisms, may involve the appearance of adverse drug reactions (ADRs) or a modification of the drug's therapeutic effect. DDIs without clinical manifestations are called potential DDIs (pDDIs)5.

There are certain factors that are specific to pediatric patients and which could increase the risk of pDDIs. These include sudden changes in hepatic and renal function parameters, in the physiology of the stomach, the patency of the cells, and the ratio of total body water to fat-free body mass, all of which tend to affect the drugs’ pharmacokinetic profile. Such changes usually affect neonates more than other patient groups given that the drug clearance rate in those patients is lower, which increases the medicines’ half-life6. Moreover, other factors such as polypharmacy7-9, especially during adolescence10,11; long hospital stays7,8; certain conditions such as neoplasms, autoimmune diseases, diseases of the central nervous system and congenital diseases7-9; admission to the intensive care unit7; and use of antineoplastic, antiepileptic, gastrointestinal, diuretic or adrenergic drugs, or opioids, antibiotics, or inmunosuppressants8,9,11,12, are also associated with a higher incidence of pDDIs.

Although multiple studies have in the last few years analyzed the occurrence of DDIs in pediatric patients8-15, the information available is still scarce and almost exclusively restricted to pDDIs or their risk factors. The estimated prevalence of pDDIs among pediatric patients on two or more drugs is highly variable, ranging from 1.4%11 to 83.5%14. Such variability reflects a significant methodological diversity and often results from the use of multiple data bases as DDI identification tools, which may differ in terms of predictive value, classification of clinical relevance, access, and ease of management16.

In addition, evidence on the occurrence of DDIs in pediatric patients is still scarce. A study of acute patients who required hospitalization showed a very low prevalence of rDDIs (0.009%)17. However, the study was not specifically geared toward evaluating DDIs bur rather to an analysis of medication-related problems in general, which means that DDIs were not examined or described in great detail.

Another limitation is the lack of information on patients admitted to pediatric burn units. The challenging clinical situation of these patients, which usually requires complex pharmacotherapeutic management and the continuous use of drugs with high DDI potential, would deserve a more specific description18, along the lines of the reports on pediatric patients admitted to critical care units or on those with a hemato-oncologic diagnosis7,9. As a result, study of DDIs in these patient groups is essential, distinguishing the specificities of such interactions for clinical practice and their implications for the patients’ safety. The purpose of this study was therefore to explore the prevalence and characteristics of pDDIs and rDDIs in two pediatric hospitals in Mexico City.

A total of 111 patients were included in the study. Their characteristics are shown in table 1. The main causes for admission were second degree burns, bacterial pneumonia, and febrile neutropenia, which accounted for 59.4% of admissions. A total of 36 patients (32.4%) were admitted to the critical care unit. The main diagnoses were infection-related and included bacterial pneumonia (14.4%) and septic shock (4.5%). In addition, five patients (4.5%) presented with burns and were admitted to the UQTI. Median hospital stay was 11 days (range: 1-302 days), while the median number of concomitant medications was 5 (range 2-33). Variables such as sex, age group and nutritional status did not show themselves to be associated to the presence of pDDIs (Table 1).

A total of 332 pDDIs (142 different pairs) were identified in 68 patients, which represented an overall prevalence of 61.3% (95% CI 52.2-70.4%). At the same time, five rDDIs were observed in four patients, with a prevalence of 3.6% (95% CI 0.1-7.1%), among children who received two or more drugs while hospitalized.

The mean number of pDDIs per patient in this study was 4.8. A positive association was observed between the total number of pDDIs and the number of drugs administered (rS = 0.80; p < 0.001) across all the hospital units under analysis (Figure 1). The length of hospital stay was not correlated with the number of pDDIs or with being on polypharmacy. Some patients admitted to non-critical units with moderate polypharmacy and short hospital stays showed similar pDDI rates. A total of 52.9% of patients presented with neutropenic fever or minor burns affecting 10% of their body surface. Other patients with de novo neoplasms had highly variable hospital stays and a low pDDI rates. On the other hand, 51.2% of patients admitted to critical care units such as PICU or ICBU presented with pneumonia or burns affecting 20-60% of their body surface on admission and displayed highly variable lengths of hospital stay, exposure to polypharmacy and pDDIs (Figure 2).

Figure 1.

Distribution of the number of potential drug-drug interactions, number of medicines and length of hospital stay per department.

(0,08MB).

Figure 2.

Distribution of potential drug-drug interactions, medications and lengths of hospital stay per diagnosis on admission.

(0,09MB).

Of the 332 pDDIs, 17 (5.1%) were classified as contraindicated, 201 (60.5%) as major, 97 (29.2%) as moderate, and 17 (5.1%) as minor. The level of evidence was considered excellent in 20 cases, (6.0%), good in 81 (24.4%), and sufficient in 231 (69.6%). Only one of the five rDDIs observed resulted in an extension of the patients’ hospital stay. The remainder required treatment and closer monitoring than usual (Table 2).

Antibiotics were the most usually prescribed drugs (23.4%), followed by analgesics (19.3%), electrolytes (6.5%) and systemic corticosteroids (5.6%). The drugs most frequently involved in pDDIs were analgesics (22.6%, opioids 18.2%), sedative hypnotics (13.3%), antibiotics (13.1%), diuretics (8.6%), histamine type-2 receptor antagonists (6.8%) and systemic corticosteroids (5.1%).

This is the first study to estimate the prevalence of pDDIs and rDDIs in pediatric patients with burns, cancer, and other medical conditions. A pDDI prevalence of 61.3% was found among patients on two or more drugs; 60.5% of those pDDIs were considered major. Moreover, a prevalence of 3.6% was observed in rDDIs that caused temporary damage and required a medical intervention or resulted in an extension of the patients’ stay in hospital.

The observed prevalence of pDDIs is aligned with the findings of other studies of hospitalized pediatric patients, which reported rates between 41.7% and 67.1%8-10,12,25. Direct comparisons between the estimated pDDI prevalence rates should be avoided given the diversity of sources of information, the different sampling methods used, the wide range of medical specialties involved, and the heterogeneity of age groups in the pediatric patient universe. Moreover, prescription patterns could vary across different countries. For example, the proportion of patients with pDDIs reported in studies including pediatric cancer patients ranges from 56.7%9 to 95%26 in inpatients and 83.5%14 in outpatients whereas in this study the proportion was 32.8%. This could be attributable to the non-probabilistic sampling method used, the low proportion of cancer patients on polypharmacy, and to the fact that only three patients with neoplasms were admitted to critical care.

This is also one of the few studies to look into the occurrence of rDDIs. The previously reported prevalence of rDDIs in pediatric patients requiring hospitalization was 0.009%17, which is significantly lower than the rate found in this analysis. This could be due to the selection bias arising from the sampling method used, whereby patients requiring critical care, especially children with burns who required more drugs to be properly managed and stabilized7,27, were overrepresented18.

Furthermore, rDDIs between paracetamol and cholestyramine, considered minor on the Micromedex® severity scale, required the analgesic regimen to be modified as a result of an objective reduction in effectiveness. In a large-scale population-based study, the majority of pDDIs observed involved reductions in therapeutic effectiveness13. Although this suggests that theoretically minor pDDIs should not be underestimated in clinical practice, it must be borne in mind that the severity categories on the Micromedex® scale were not compared with those of other databases, which could lead to an under- or overestimation of the available evidence on pDDIs.

Like other studies, this analysis demonstrated the existence of a correlation between degree of polypharmacy and number of pDDIs3,5, but not between degree of polypharmacy and length of hospital stay. The small sample size and the wide length-of-stay variability in patients with the same diagnosis made it impossible to establish such a correlation, which had however been found by other authors7,8. Diagnoses showing the widest length-of-stay and length-of-drug-exposure variability were unspecific bacterial pneumonia, de novo neoplasms, septic shock, and congenital malformations. Some of them, such as pneumonia and burns involving over 20% of body surface, were more common in critical units. However, the overrepresentation and grouping together of several single diagnoses led to a wider dispersion between drug exposure and length of hospital stay. These diagnoses included optical neuritis, pancreatitis and intraventricular hemorrhage, whose considerable severity and clinical complexity diluted the correlation between length of hospital stay and pDDI.

Another aspect found to be relevant for the type of pDDI experienced by patients was the observation in this study and in those of other authors that opioid analgesics, antibiotics and CNS drugs were among the three groups of drugs most commonly involved in pDDIs7,10,11,26,28. These groups of drugs are widely used to manage patients in the critical care setting7, including children with large burns (> 20% of their body surface), where opioids, benzodiazepines and sedative hypnotics are used for the management of pain, and antibiotics are used to treat complex infections18,29.

Given its exploratory nature, the present pilot study includes certain methodological aspects, such as the small sample size, the non-probabilistic process used for patient selection, and the cross-sectional design of the analysis, which cannot exclude potential confounding and selection biases, thereby restricting the external validity of the results obtained. Another important limitation was a failure to compare the pDDI results with a second source of information, as is usually done in other studies9,14,30. Although the Micromedex® database is considered a very comprehensive, reliable, and user-friendly source of information21, use of an additional data base to compare the results would have allowed higher levels of sensitivity and accuracy in the identification of pDDIs2, allowing for more robust prevalence estimations. For these reasons, the results of the study must be taken with caution and be set within the context of an exploratory study, i.e., the results are not intended to establish a definitive parameter but to serve as a starting point for further research into this area. Although the participation of a pharmacist was important in this study, hospital pharmacy as a profession is in a teething phase in Mexico and faces many obstacles to effective implementation, which on many occasions could involve risks for the patients’ health.

In a nutshell, the prevalence of pDDIs as identified by means of the Micromedex® database by a resident physician was 61.3% among acute pediatric patients who received two or more drugs during their hospital stay. Only 3.6% of patients at risk presented with clinical manifestations resulting from rDDIs that caused temporary damage. However, up to 65.6% of pDDIs identified were considered severe or contraindicated, which suggests the need to place a greater emphasis on research into preventing risks to patient safety.

No funding.

No conflict of interests.