Background

Antimicrobials are the most commonly prescribed drugs in the community and hospital setting, especially among paediatric patients [1]. However, antibiotics are often unnecessarily used both in the community, where too many children receive broad-spectrum antibiotics for viral infections, and in the hospital, where long courses of broad-spectrum antibiotics are frequently prescribed [2]. Recent studies have found that up to 50% of antimicrobial prescriptions are inappropriate [3,4].

The emergence of multi-drug resistant (MDR) pathogens and their rapid global spread, strictly associated with an inappropriate use of antimicrobials, are important global public health threats with a substantial impact on patient outcomes such as hospital length of stay and mortality, as well as on healthcare costs [5–8]. The European Antimicrobial Resistance Surveillance Network (EARS-Net) system has reported a dangerous rise in MDR bacteria in the last years showing that some countries such as Italy are strongly contributing to this worrying increase [9].

Many studies have evaluated antibiotic prescriptions in the paediatric outpatient population highlighting the problem that Italian prescribing habits that differ from those of other European countries. An Italian child is more likely to be exposed to antibiotics than children are in North Europe [10] and, in particular, the prevalence of antibiotic prescriptions in childhood have been reported to be 4 times higher than in the UK and 6 times higher than in the Netherlands [11,12]. Moreover, Italy reported the highest prescription rate (1.3 per infants per year) in a study comparing antibiotic use in the first year of life in five European countries [13]. In fact, data from the Gagliotti et al study in 2006 show that the 55% of Italian infants in the community have already received at least one course of antibiotics at 1 year of age and 84% at 2 years of age [14].

Although a positive correlation between outpatient and inpatient antibiotic use has been noted [15], few studies describing the real antibiotic consumption in Italian children’s hospitals have been published. A single centre study was carried out in Rome in 2008 [16] confirming the abuse of antibiotics observed in the outpatient population. A more recent paper evaluating the trend of antibiotic use in all the paediatric wards of Emilia-Romagna Region over an 8-year-period [17] indicated a slight increase of antibiotic consumption over time, an inadequate tendency to prefer penicillin plus inhibitors to plain penicillins, an over-use of third generation cephalosporins and a worrisome increase in linezolid prescriptions.

In this paper, we present the results of a point-prevalence survey (PPS) on antibiotic prescriptions carried out in seven large Italian paediatric institutions in 2012. The aims of our study were: i) to describe prevalence rates of antibiotic prescriptions for prophylaxis and treatment of infections for neonatal (<30 days) and paediatric (age ≥30 days) patients in seven Italian centers; ii) to evaluate antibiotic prescriptions, indications, number and type of antibiotic agents and administration route in the same age sub-groups both for prophylaxis and treatment of infections; iii) to describe over-all consumption and off-label use of particular classes of antibiotics, such as carbapenems and quinolones, in our cohort; and iiii) to identify targets for improving the quality of antimicrobial prescribing in these centers.

Methods

This research has been conducted according to the principles expressed in the Declaration of Helsinki. Ethical approval has been obtained for the coordinating centre. No consent was given, because data were collected by reviewing medical charts and were analyzed anonymously. Every patient record was given a unique non-identifiable survey number, which was automatically generated by a computer program specifically designed for anonymous data entry.

A 1-day PPS on antibiotic use in hospitalized children was performed in Italy between October and December 2012 as part of the Antibiotic Resistance and Prescribing in European Children project (ARPEC). Seven paediatric or mixed adult-paediatric hospitals in seven Italian cities were involved (Genoa, Milan, Padua, Florence, Viareggio, Rome and Naples). The survey included all admitted patients less than 18 years of age present in the ward at 8:00 am on the day of the survey who had at least one on-going antibiotic prescription. The wards of admission were: medical (general neonatal and maternal wards, and general paediatric wards), special medical (cardiology, nephrology, onco-hematology, neuromuscolary, neurology, bronchopneumology, infectious diseases unit), neonatal and paediatric intensive care (NICUs and PICUs), surgical (neonatal surgery, paediatric surgery, orthopedics, neurosurgery). For feasibility reasons, one hospital provided data from randomly selected wards, maintaining the patient distribution among medical, special medical, surgical and intensive care units, in agreement with the coordinating centre. Full details of the ARPEC methodology are described elsewhere [18].

Results

The PPS was performed in 61 wards within seven Italian institutions. Characteristics of the centres involved are shown in Table 1. A total of 899 patients was present in the hospitals at 8:00 am on the day of the survey and 349 (38.9%) of these had an on-going prescription for one or more antibiotics. However, this rate was variable among the hospitals ranging from 25.7% to 53.8% (Table 1). Combination therapies were variably used among the institutions (21.7–60.3%) with a ratio between number of prescribed antibiotics and treated patients ranging from 1.25 to 1.76 (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of the 7 Italian institutions involved in the ARPEC project.

https://doi.org/10.1371/journal.pone.0154662.t001

A wide variability also existed in the proportions of patients treated with at least one antibiotic stratified by ward type. In particular, special medical wards and intensive care units accounted for higher proportions of patients receiving antibiotics compared to surgical and medical wards (“Fig 1”).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Proportion of paediatric patients treated with at least one antibiotic by ward type.

https://doi.org/10.1371/journal.pone.0154662.g001

Characteristics of all patients enrolled are summarized in Table 2. Median patient age was 24 months and 12.3% were less than 30 days old. Overall, 24.6% of patients were affected by a medical/surgical underlying condition and the most frequent was an oncologic/hematologic disease. Noteworthy is that the rate of oncologic/hematologic patients admitted to the hospital at the time of the survey was 5% (45/899), but this rate increased to 22.3% (78/349) looking at the group of patients receiving antibiotics. These data reflect the fact that oncologic and hematologic patients were responsible for a large proportion of antibiotic consumption in our survey.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Characteristics of the 349 Italian patients enrolled in the 24-hour ARPEC PPS.

https://doi.org/10.1371/journal.pone.0154662.t002

We analyzed antibiotic prescriptions separately in neonates and children (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Demographic characteristics and antibiotic prescription patterns of the neonates and children enrolled in the study.

https://doi.org/10.1371/journal.pone.0154662.t003

Neonates

At the time of the survey, 248 neonates were admitted in participating hospitals and 43 were receiving antibiotics (17.3%). As shown in Table 3, neonates treated with antibiotics were mostly admitted in the NICUs (83.7%, 36/43). Moreover, 62.8% (27/43) of newborns were receiving antibiotics for prophylaxis and only 37.2% (16/43) were being treated for infection. The top two active antibiotic prescriptions were penicillins (69.8%, 30/43) and aminoglycosides (58.1%, 25/43). Details about indications to receive antibiotics are summarized in Table 3. All antibiotic classes prescribed are listed in “Fig 2”.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Antibiotic prescriptions among the neonates and children.

https://doi.org/10.1371/journal.pone.0154662.g002

Discussion

The 1-day ARPEC PPS provided very useful data on hospital antibiotic prescriptions for paediatric and neonatal patients in Italy. According to data collected in seven large Italian institutions, 38.9% of inpatients received at least one antibiotic prescription during hospitalization. This rate is similar to the mean rate reported from the worldwide ARPEC PPS (36.7%) [18].

To better analyze antibiotic prescription patterns and their appropriateness, we assessed antibiotic prescriptions for prophylaxis and treatment of infection separately.

Our results show that overall 39% of patients were prescribed antibiotics for prophylaxis with the highest rate observed in the neonatal population (63% of neonatal prescriptions were for this indication). The main indication for neonatal prophylaxis was the presence of perinatal conditions (e.g. prematurity, maternal fever during labor, prolonged rupture of the membranes). Prophylactic monotherapy was prescribed just in 37% of neonates and penicillin was the preferred agent, while combination prophylactic therapies including penicillin plus aminoglycosides or glycopeptides plus cephalosporins/aminoglycosides were widely used in neonatal patients. This approach is not in-line with the international literature. Although neonates represent a high risk population due to their immature immune system and the invasive procedure they are likely to undergo in NICU (e.g. indwelling catheters, invasive mechanical ventilation), recent reviews reject the routine use of antibiotic prophylaxis due to lack of efficacy in many conditions [19–22]. Moreover, in 2010 the Center for Disease Control and Prevention revised their guidelines regarding the prevention of perinatal Group B streptococcal disease in healthy neonates, restricting the need for prophylaxis only to well-defined subgroups of patients [23]. Prolonged courses of antibiotics have also been associated with increased risk of necrotizing enterocolitis or death in low birth weight infants [24].

In the paediatric group, the rate of antibiotic prescriptions for prophylaxis was 35.5% of all the prescriptions. Approximately, half of these patients were receiving antibiotics for surgical prophylaxis in accordance with previous European reports in which the proportion of children receiving surgical prophylaxis ranged from 13 to 42% [25, 26]. Third generation cephalosporins ranked first in prescription frequency in this scenario, used often in monotherapy but combined with metronidazole in some cases, confirming their alarming overuse for this indication. This problem in fact was already raised by Ciofi et al in 2008 [16], but a recent paper published by Buccellato et al in 2015 shows that a limitation on the prescriptions of these drugs has not yet been reached [17]. However, it is worth noting that this finding was very variable among the seven centers, since some hospitals preferred the first generation cephalosporins for surgical prophylaxis, as suggested by international guidelines [27].

Cotrimoxazole was the most prescribed antibiotic for medical prophylaxis, used alone in most cases. As explanation, most of the treated children were affected by onco-hematological diseases and cotrimoxazole is the best treatment to prevent Pneumocystis jirovecii pneumonia in immunocompromised patients [28].

Regarding the prescription patterns for treatment of infection, the 37.2% of our neonatal cohort was prescribed at least one antibiotic for treating an infection, the main reason was sepsis and the most common antibiotic class was penicillins, combined with aminoglycosides in a large number of patients, in line with international literature [29, 30]. It is hard to compare prescription habits in our centres with other NICUs because of a wide variability of the rate of neonates prescribed antibiotics across hospitals, as shown by a recent multicenter study involving 127 NICUs in the US [31]. The 40-fold variations in prescription frequencies noted in this study did not appear to be related to higher infection burden, necrotizing enterocolitis incidence, surgical volume or mortality rate [31]. They have instead been attributed to frequent inappropriate courses of antibiotics in inpatient neonates, more commonly owing to an unnecessary antibiotic continuation than starting of a non-required therapy [32].

Focusing instead on the paediatric group, we noticed an excessive use of third generation cephalosporins for treatment of infection similar to that seen for surgical prophylactic use, as underlined before. In children with LRTIs, ceftriaxone was the most prescribed antibiotic, used as monotherapy or often combined with macrolides, with a wide total daily dose variability ranging from 12.1 mg/kg/day to 153.8 mg/kg/day. The frequent choice of ceftriaxone as first line therapy for treatment of uncomplicated LRTIs and, in some cases, the high dosage prescribed, are reasons of concern because they are not supported by current guidelines [33]. In fact, other European countries, as the UK and France, seem to have different prescribing patterns for LRTI, preferring amoxicillin/clavulanic acid as first line therapy [34].

An abuse of parenteral cephalosporins in Italian hospitalized children was already denounced in a study conducted by Esposito in 2001 [35] and is a well-known problem also in the adult population [36].

Noteworthy was also the widespread use of carbapenems and quinolones. Indeed, in our study population, among the 349 patients receiving antibiotics, 8.9% were being treated with carbapenems, whereas proportion of carbapenems for therapeutic use reported in the literature in European paediatric units is 4,2% [37]. Though carbapenems were prescribed in most cases for fever in cancer patients, which often requires aggressive antibiotic treatment considering the patients’ immunological status and predisposition to severe infections, we are concerned about the increasingly popular usage of these agents for community acquired-infections, empiric treatment and combination therapy. This is a very alarming finding, considering the doubling of carbapenem resistance rates in invasive isolates of Klebsiella pneumoniae reported by the European Antimicrobial Resistance Surveillance Network (EARS-Net)’s report from 2010 to 2013 [9] for Italy. We also found 18.7% off-label use in patients below 3 months of life, but this could be explained by the involvement of many of our centers in the European NEOMERO study, which aimed to evaluate pharmacokinetics, safety and efficacy of meropenem in neonatal sepsis and meningitis [38].

Similar problems were noticed also for quinolone prescription. Quinolones were widely used in our cohort, even if the license for the use of this antibiotic class below 18 years of age is restricted to few rare indications such as cystic fibrosis with pulmonary exacerbations, complicated urinary tract infections, post exposure prophylaxis against inhalational anthrax and severe infections with allergies to other antibiotics [39,40]. Among our patients, the main indication for treatment with quinolones was LRTI. In this group of patients, quinolones were often prescribed empirically and combined with other drugs, though current guidelines do not suggest quinolones as a first-line treatment considering that infections caused by pneumococci or atypical bacteria can still be successfully treated with high doses of β-lactams [41]. Furthermore, the Scottish Antimicrobial Prescribing Group (SAGP) in 2008 and the National Institute for Health and Care Excellence (NICE) in 2015 [42] recommended to avoid the use of quinolones as first line agents for empirical treatment of most commonly infections in primary care, because the overuse of these broad-spectrum antibiotics is associated with a significantly increased risk of Clostridium difficile infection [43,44]. The wide use of quinolones in our cohort could be explained by the finding that most of our patients receiving this treatment had underlying chronic pulmonary diseases, such as cystic fibrosis or secondary to immunodeficiencies, but the lack of data about their microbiological status did not allow us to evaluate the appropriateness of these prescriptions.

Our study highlights many feasible targets that need a prompt intervention with appropriate antimicrobial stewardship programs. International guidelines for stewardship identify a wide set of interventions including: disease-specific clinical pathways, audit with feedback and formulary restriction with preauthorization of select agents. The best type of interventions must be tailored according to local practices, resistance trends, and available resources [4]. While the most effective antimicrobial stewardship programs are built on proactive interventions, in settings where a robust antimicrobial stewardship team is hard to establish, clinical pathways tool represent a reasonable and feasible first step for implementation standardizing care without adversely affecting patient safety or outcomes [45, 46]. Moreover, annual PPS could be a useful tool to measure the impact of these interventions on antibiotic prescribing practices [18]. Thus, implementation of clinical pathways in Italian paediatric hospitals associated with annual PPSs could be a good start to reduce the abuse and misuse of antibiotics.

Our study has some limitations. First, data about microbiological isolates and antibiotic susceptibility tests, length of therapies and prophylaxis and previous antibiotic courses could have been useful to better define the appropriateness of antimicrobial prescriptions. Moreover, the characteristics of involved institutions may have affected at least in part the reliability of some results. Most institutions are in fact tertiary care hospitals that usually manage more complicated and severe cases, which might significantly impact antibiotic prescriptions. In addition, the heterogeneity among institutions, in terms of presence/absence of onco-hematology departments or intensive care units, may strongly affect antimicrobial prescription patterns. Finally, the survey was conducted in 2012 and since then a greater awareness of antibiotic stewardship programs has spread among Italian hospitals [47–49].

Conclusions

Our study took a picture of the Italian situation in terms of antibiotic prescriptions in hospitalized neonates and children and identified many feasible targets that require a prompt intervention to reduce the abuse and misuse of antibiotics. Antibiotic stewardship programs should immediately introduce measures to control prescription patterns in particular for prophylaxis, both in neonatal and paediatric populations, and to limit the over-use of third generation cephalosporins, that seems to persist over-time. Surveillance and educational programs are also needed to restrict the use of carbapenems to more severe conditions. The implementation of disease-specific clinical pathways associated with annual PPSs could be a good way to monitor and ameliorate antibiotic prescription patterns in neonatal and paediatric inpatients over time, in order to reduce as much as possible the worrisome emergence of MDR bacteria in this vulnerable population.

Supporting Information