Although all Italian level III NICUs were invited to join the survey, it is likely that those that participated are more active in research and/or are more interested in self-audit and improvement of a unanimously acknowledged clinical care. Nevertheless, a third of all the Italian NICUs participated and represent the north, center, and south of Italy. Besides the geographic distribution, the estimated rate of admission of newborns <32 weeks’ gestational age in participating NICUs (0.85%) is close to that reported in Italy (0.9%)  and also supports the representativeness of collected data at the national level. The wide variation in dosages used does not seem to depend on area of the country and, in any case, highlights the need for national, or better international, shared guidelines incorporating current knowledge on the treatment of neonatal sepsis.
Fluconazole prophylaxis has been found to reduce the incidence of invasive fungal infection in VLBW infants, and recent data on long term outcomes and antifungal resistance seem positive so far [26, 27]. This study’s data show that, in Italy, a majority of NICUs (79%) use fluconazole prophylaxis, a rate much higher than that found in all the European NICUs participating in the survey  and in the UK/Ireland and US studies [22, 23]. Rates of prophylaxis in different studies may be based on incidence, but rates in infants at the lower gestational ages are high in all studies. No correlation was found between reported prevalence of fungal infections in the NICUs and their choice to use fluconazole prophylaxis. The reported incidence in the survey fits within the range reported by other studies , although methodologies in data collection differ and direct comparison with incidence data is not possible. Incidence rates of fungal infection differ between studies, from 2,6 to 9% [29, 30]. A large part of the variation in rates is due to the number of extremely low birth weight infants admitted at each NICU [31, 32], but may also reflect differences in clinical practices .
There was variation between the centers concerning the indications for initiating fluconazole prophylaxis, but birth weight, presence of a central venous catheter, and gestational age were considered the most important and these data are similar to those found by Burwell et al . The guidelines addressing prophylaxis [12, 13] only take into account birth weight and rate of disease in hospital units as the factors on which to base prophylaxis initiation, while other factors should also be considered . Guidelines use birth weight as an indicator, for example, but gestational age has a more linear relationship with invasive candida infection . Furthermore, the guidelines should be updated to reflect the finding that fluconazole prophylaxis can reduce incidence even in NICUs with a low rate of invasive candida infection . It has been shown that implementing a guideline that includes specific criteria for identifying high-risk VLBW infants reduced invasive fungal sepsis rates without evidence of fluconazole resistance emerging . The possibility of resistance emerging, as well as insufficient safety data, make the creation of standard, more accurate guidelines identifying newborns at the highest risk of acquiring fungal infection, and therefore candidates for prophylaxis, a high priority.
The main guidelines on candidiasis  suggest considering fluconazole prophylaxis in neonates with birth weights <1000 g, and at a dosage of 3 mg/kg iv every 72 hours for the first 2 weeks of life, every 48 hours during the third and fourth weeks, and daily in the fifth and sixth weeks of life. The other cited guidelines  offer limited information on prophylaxis, but the initial dose is similar. Dosage data from the NICUs refers to a single dose only, so a comparison with the dosage change over time, as specified in the main guidelines, is not possible, however, 38% of the NICUs report using the same dosage and route as the guidelines, i.e. 3 mg/kg every 72 hours intravenously. Only one center followed the guidelines completely, considering also duration of prophylaxis. The US survey did not assess dose, and data from the UK/Ireland study, which is not directly comparable because it involved individual physicians, had almost half the responses reporting the use of 3 mg/kg and almost one fourth 6 mg/kg, compared to three fourths using the lower dose in the Italian data. Studies have been performed to test less intensive dose regimens, including twice weekly dosing [17, 37], and their results were encouraging, but additional trials are needed to identify the most appropriate dosing regimen .
Significant concern about safety in using fluconazole for prophylaxis resulted from the data. This differs from the US data , which revealed concern especially about resistance, and the UK/Ireland data , which emphasized the importance or the perceived level of invasive candidiasis in the unit. Even when compared with the data from all the European NICUs surveyed , showing that those that don’t use prophylaxis were more likely to be concerned about both resistance and the lack of a statement by pediatric societies promoting routine use in a subset of newborns, concerns seem to differ. To reduce the possibility of resistance in cases of fungal infection in infants following fluconazole prophylaxis, treatment with a different antifungal than fluconazole should be initiated . Only 19/30 of the NICUs reported using a drug other than fluconazole in cases of confirmed fungal infection in patients under fluconazole prophylaxis. The guidelines should address this issue specifically.
Half of the NICUs use fluconazole for treatment of fungal sepsis, and only about one third follow the guideline recommendations on dosage, while the rest use lower dosages. Concerning the indications for starting treatment with fluconazole, the majority of NICUs that use the drug follow the guidelines in that they initiate treatment when sterile body fluids test positive for Candida spp. There did not seem to be general concern over safety, but some interest was expressed for both the need for an official statement supporting the choice of fluconazole and for more efficacy studies. These, however, did not seem to be valid indicators in the choice to not use the drug. The emerging picture concerning fluconazole treatment is that there is a need for more guidance for clinicians, through greater acknowledgement of the existing guidelines and acquisition of additional data on the drug.
One of the major limits of this study was the sample size. The Italian data originated from 38 NICUs and this limited data comparison.
Researchers in Italy have attempted to prioritize shared information on the care of extremely low birth weight babies [38, 39], but the situation needs further developments. The collection of data on rates of infection by a centralized, international public body would permit transparency and homogeneous, accurate data on incidence, and this would, in turn, motivate centers to implement the most effective, evidence based practices.
The TINN project has already begun to improve the availability of information on fluconazole and ciprofloxacin, as also documented in published articles [20, 40, 41], and will provide additional clinical evidence in the early future.