Nosocomial infections are one of the main challenges we face in the management of very preterm newborns in neonatal intensive care units (NICUs), and it is one of the leading causes of morbidity and mortality in this population. The World Health Organization estimates that there are 4 million neonatal deaths a year, of which one third are due to severe infection.1

The incidence of nosocomial infection is higher in patients with birth weights of less than 1500g, and mortality is higher in this population.2 On the other hand, the indiscriminate use of antibiotics can promote infection by drug-resistant pathogens that are associated with higher mortality rates.3 Effective strategies to prevent nosocomial infections in NICUs need to be established.4

The risk factors for the development of nosocomial infections include those intrinsic to the patient, such as gestational age and weight at birth, genetic predisposition, the permeability of skin and mucosal barriers, male sex and immunosuppression,5 and extrinsic factors related to the interventions implemented in the NICU.

Chief among the extrinsic factors is the use of external devices, such as ventilatory support devices and especially central venous catheters5,6 and the administration of parenteral nutrition.7 The other key factor concerns human resources. Poor adherence with hand hygiene measures and inappropriate nurse-to-patient ratios are associated with an increase in the incidence of nosocomial infections and outbreaks of infection by multidrug-resistant pathogens.8

Other extrinsic factors associated with increases in the incidence of nosocomial infection are delayed initiation of enteral feeding, administration of ranitidine, prolonged antibiotic treatment and postnatal steroids use.5,9

The literature on the prevention of nosocomial infection is extensive. Some of the proposed measures are hand hygiene,10 aseptic technique in the handling of intravenous catheters,11 promoting breastfeeding,12 limiting the use of antibiotics13 and monitoring colonisation by multidrug-resistance microbes.8

We present the findings of a study based on a protocol that introduced 5 measures to decrease the incidence of nosocomial infection in very preterm and very low birth weight infants.

The primary objective of the study was to assess the impact of a bundle of 5 interventions, implemented over a period of 12 months, on the incidence of nosocomial infections in the population of very preterm and very-low-birth-weight infants.

A total of 33 patients were included in each of the periods under study. Fig. 1 presents the flow chart of patient inclusion. Patients in both groups were comparable in terms of the various demographic variables under analysis (Table 3).

Figure 1.

Flow chart of patient inclusion in the study.

(0.07MB).

A total of 18 nosocomial infection episodes were diagnosed, 14 in the pre-intervention period and 4 in the post-intervention period. The distribution by type of infection was: CRS, 7 (50%) pre-intervention and 3 (75.0%) post-intervention; non-catheter related sepsis, 3 (28.6%) in the pre-period and 0 in the post-period; NEC, 2 (14.3%) in the pre-period and 1 (25.0%) in the post-period; VAP, 1 (7.1%) in the pre-period and 0 in the post-period. The results of blood culture were positive in 15 out of the 18 episodes (83.3%). The isolated bacteria were Staphylococcus epidermidis (S. epidermidis) in 8 cases (53.3%), Klebsiella pneumoniae (K. pneumoniae) in 3 (20.0%), Enterococcus faecalis in 2 (13.3%), Candida albicans in 3 (20.0%) and Enterobacter cloacae in 1 (6.6%). More than one bacterium was isolated from blood culture in 2 episodes. The 3 episodes in which no bacteria were recovered from culture corresponded to 2 episodes of NEC and an episode of suspected non-catheter related sepsis.

We observed a statistically significant reduction in the incidence rate of nosocomial infection between the two periods, with a rate of 8.7 nosocomial infections/1000 patient days pre-intervention and a rate of 2.7 nosocomial infections/1000 patient days post-intervention (P<.05). The number of patients with more than 1 episode of nosocomial infection during their stay was 10 (30.3%) in the pre-intervention period and 4 (12.1%) in the post-intervention period, a result that tended towards statistical significance (P=.07).

There was a statistically significant reduction in the number of central catheter days per patient, with a median of 13.0 (IQR, 8.0–17.5) in the pre-intervention period and of 9.0 (IQR, 0–15.0) in the post-intervention period (P<.05). In 6 out of the 10 episodes of CRS (54.5%), the same organism was recovered from the blood culture and the catheter tip culture (S. epidermidis in 3 cases, C. parapsilosis and K. pneumoniae). In 4 episodes (45.5%), sepsis was attributed to the indwelling catheter but the same organism was not recovered from the tip culture. We observed a reduction in the number of CRS episodes/1000 central catheter days from 14.4 in the pre-intervention period to 9.6 in the post-intervention period, although the difference was not statistically significant (P=.54).

Table 4 summarises the findings for the rest of the secondary variables under analysis. The most salient findings were statistically significant decreases in the proportion of days of mechanical ventilation during the stay (both invasive and noninvasive) and in the proportion of patients that received vasoactive drugs and blood product transfusions. We found decreases in length of stay, days of parenteral nutrition and the proportion of patients that died or had severe lesions in neuroimaging that approximated statistical significance.

We observed a statistically significant reduction in the number of days of antibiotherapy relative to patient days (Table 4). This reduction was sustained in the separate analysis of the days of empirical antibiotherapy. The proportion of patients that received human milk between days 14 and 28 of life was higher in the post-intervention period, and the difference was statistically significant for day 14 post birth.

During the pre-intervention period, there were 3 outbreaks by multidrug-resistant organisms (OXA1-type beta-lactamase-producing K. pneumoniae and Serratia marcescens in 2 occasions). These outbreaks caused 3 cases of 3 sepsis and 10 local infections (9 cases of conjunctivitis and 1 case of surgical site infection). There were no cases of systemic infection by multidrug-resistant organisms in the post-intervention period. The screening conducted in February 2016 identified 2 patients that were asymptomatic carriers of broad-spectrum beta-lactamase-producing K. pneumoniae. These patients were isolated, and colonisation by this organism was not identified in any further patients. The rest of the screens in the post-intervention period were negative.

Despite the numerous interventions that have been proven useful in the prevention of nosocomial infections, the latter continues to be one of the main problems in neonatal care units.7,21 There is wide variability in the incidence of nosocomial infection between hospitals,5 and every facility needs to develop preventive strategies suitable to its own circumstances.21

In January 2013, our unit became a IIIB level unit (from being a IIIA unit). After the first 12 months of offering care for very preterm infants, we reviewed different quality assessment indicators in our unit. The difference in the incidence of nosocomial infection compared to the one reported in the literature22 (8.7 vs 6.4 nosocomial infections/1000 patient days) motivated us to design a control cycle to improve care. The intervention included 5 evidence-based strategies adapted to our needs. The implementation of these measures achieved an incidence rate within the established quality standards (2.7 nosocomial infections/1000 patient days).

Catheter-related sepsis was the most frequent type of sepsis in both periods under study. The introduction of specific protocols for the insertion and maintenance of central catheters has been shown to reduce the incidence of CRS.11 Different institutions and authors have published reviews and clinical practice guidelines based on current scientific evidence to achieve this goal.4 Our outcomes corresponded to an incidence rate that exceeded those reported in other series22 (14.4 vs 11.1 cases of CRS/1000 catheter days). The new protocol for the insertion and handling of central catheters succeeded in reducing this rate by nearly 25% (to 9.6 CRS cases/1000 catheter days).

Many studies have demonstrated the protective effect of breastfeeding against NEC and late-onset neonatal sepsis.12 There is evidence of a dose-response relationship in this effect between days 1 and 28 post birth.23,24 Breast milk feeding has also been proposed as a protective factor against colonisation by multidrug-resistant bacteria.25 We believed that we could improve the percentage of patients that were fed breast milk, and the efforts devoted to the purpose increased this percentage by more than 50%. In the only case of NEC diagnosed during the post-intervention period, the patient could not be fed breast milk because the mother had a disease that could be transmitted through breast milk and we could not provide donor milk, as our hospital did not have a human milk bank or cooperate with one.

Among the strategies described in the literature, hand hygiene constitutes another cornerstone in the prevention of nosocomial infections, as it prevents colonisation of the skin, airways and gastrointestinal tract by potentially pathogenic organisms. These organisms are disseminated from patient to patient through the hands of health care staff.4,8 Training and awareness campaigns on correct handwashing and adherence to the use of alcohol-based antimicrobial gels aimed at health care staff are effective in reducing colonisation by multidrug-resistant organisms.4,26 Correct hand hygiene prevents outbreaks of nosocomial infection by this type of pathogen, which have even led to the closure of some neonatal care units.16 During the pre-intervention period, there were 3 outbreaks by multidrug-resistant bacteria, a situation that did not recur once the preventive measures were implemented.

The use of antibiotics is widespread in NICUs, and 47% of patients admitted to NICUs do receive them.20 Preterm infants that have long courses of empirical antibiotherapy (>5 days) with a sterile blood culture are at higher risk of sepsis, enterocolitis and death, as well as increased neurologic sequelae.13 Thanks to the antimicrobial stewardship protocol, the percentage of days of empirical antibiotherapy in the post-intervention period decreased by 30%. This result probably contributed to the in nosocomial infection and colonisation by multidrug-resistant organisms.

During the second period, the reduction in the incidence of nosocomial infection was accompanied by a decreased mortality that tended towards statistical significance. We also observed a decrease in other illnesses (assessed by the use of blood products, vasoactive drugs and days of mechanical ventilation). Furthermore, the proportion of patients with lesions detected by neuroimaging decreased by a factor of 5. Therefore, these strategies and their results may have benefits in psychomotor development at later stages in life.

The reduction in the number of nosocomial infections and other associated comorbidities entails a reduction in health care costs. It has been estimated that episodes of LONS in very preterm infants result in a 15% increase in the cost of their care.27 In the post-intervention period, we observed a median reduction in length of stay of 25%, which a probability that neared statistically significance.

The weaknesses of our study derive from the small sample size. The number of patients included per period allowed us to identify statistically significant differences in the primary outcome variable, however, the analysis of secondary outcomes was limited due to this reason.

In conclusion, the strategies introduced in our hospital were efficacious in reducing the incidence of nosocomial infection. Our experience may be useful for units which, as was our case, have rates that exceed the limit established by quality standards.

The authors have no conflict of interests to declare.