Early Human Development 90S1 (2014) S71–S74
Infection in late preterm infants Simonetta Picone *, Roberto Auﬁeri, Piermichele Paolillo Division of Neonatology and Neonatal Intensive Care, Casilino General Hospital, Roma, Italy
A R T I C L E Keywords: Late preterm Infection Neonatal sepsis Biomarkers Cultures
I N F O
A B S T R A C T Background: Late preterm (LP) are at higher risk than term infants to develop infections due to their more immature immune system. Little data about the risks and incidence of infection and sepsis in LP are present in literature. Aims: To evaluate treated infection rates and risk factors for infection in moderate and late preterm infants (gestational age = 32–36 weeks). Study design: We retrospectively studied a population of 771 moderate and late preterm infants consecutively admitted to our unit from June 2008 to November 2013. Results: Treated infections were 128, with an incidence of 16.6%; the 90% (n = 115) occurred during the ﬁrst 72 hours of life. Blood cultures were positive in 22% of cases, umbilical venous catheter cultures were positive in 26% of cases; Coagulase-negative staphylococci were the most frequently isolated pathogens. Patients of the sepsis group had a C-reactive protein (CRP) mean value of 28.27 mg/L and a procalcitonin mean value of 25.3 μg/L. Risk factors for infections were umbilical venous catheter (UVC) insertion (χ2 = 15.9; p ≤ 0.05), prophylaxis with antenatal corticosteroids (χ2 = 16.7; p ≤ 0.05) and birth by cesarean section, with observed values very similar to the expected values (χ2 = 15.9; p = 0.1). Respiratory symptoms were found in 47 of the 60 patients in the sepsis group (78.3%). Conclusions: Late and moderate preterm infants have an increased signiﬁcant risk of infection compared to term infants. Infections, given the high frequency of negative cultures in neonates, should be often suspected and treated on the basis of clinical features and inﬂammatory markers, trying always to avoid a possible overtreatment. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Late preterm (LP) infants are deﬁned as those born at 34 0/7 to 36 6/7 weeks of gestational age (GA); they account for about 74% of all preterm births and 8–9% of total births in the United States . These infants experience the interruption of the normal fetal development during the last six weeks of gestation, a critical period for growth of brain, lungs and other systems development. Though most LP infants (∼80%) will have a neonatal course with no signiﬁcant complications , LP compared with term infants, are at increased risk for resuscitation at birth, feeding diﬃculties, jaundice, hypoglycemia, temperature instability, apnea and respiratory distress with a threefold higher mortality rate (7.7 vs 2.5 per 1000 live births) . Meanwhile several authors described and investigated these complications in LP infants; still little data about the risks and incidence of infection and sepsis in LP are present in literature. Preterm infants are at higher risk than term infants to develop infectious diseases due to their more immature immune system. Several risk factors frequently associated with the late preterm * Corresponding author: Simonetta Picone, Policlinico Casilino. Via Casilina, 1049, 00169 Roma, Italy. Tel.: +39 0623188 260/261; fax: +39 0623188 393. E-mail address: [email protected]
(S. Picone). 0378-3782/$ – see front matter © 2014 Elsevier Ireland Ltd. All rights reserved.
birth are also associated with an increased incidence of sepsis: premature rupture of membranes (PROM), prolonged time of premature rupture of membranes, maternal peripartum infections, chorioamnionitis, cesarean section or instrumental delivery and Apgar scores less than 7 at 5 minutes. The incidence of sepsis in the neonatal period is greater than at any other period of life with values widely depending on population studied: ranging from 1–5 cases per 1000 live births in the developed countries to 49–170 cases per 1000 live births in developing countries . Signs and symptoms of neonatal sepsis are often non-speciﬁc, especially in preterm infants and diagnosis is complicated by the frequent presence of non-infectious conditions that resemble those of sepsis. Although the presence of growth of an organism from a sterile site is the “gold standard” for deﬁnitive diagnosis, up to 25% of neonates with bacterial sepsis have a low organism burden that would not be detected with culture of smaller blood volumes . Therefore, to identify neonates with suspected infection it must be evaluated the presence of risk factors in combination with clinical history and diagnostic tests results. Risk factors depend on the time of onset of signs or symptoms, reﬂecting the differing etiologies. Early onset sepsis (EOS) is deﬁned by the onset of signs or symptoms and an associated positive culture at or before 72 hours of life; late onset sepsis (LOS) is characterized by the onset of symptoms at greater than 72 hours of life.
S. Picone et al. / Early Human Development 90S1 (2014) S71–S74
Evaluation of the cerebrospinal ﬂuid is recommended for infants with suspected sepsis or with blood culture proven bacteremia. Organisms are not isolated from blood in up to 38% infants who have a pathogen isolated from cerebrospinal ﬂuid (CSF) . Urine culture should be used in the evaluation for LOS. Tracheal aspirate culture reﬂects mainly colonization rather than infection. The median incidence of blood-culture conﬁrmed sepsis was 16 per 1000 live births (from 2.2 to 9.8 per 1000 live births for EOS and from 20.7 to 50 per 1000 for LOS). The rate of mortality is now 5–20% for EOS, but over 30% in very preterm infants . In addition to blood, CSF and urine culture, laboratory tests may be beneﬁcial in identifying infected neonates and in deciding on duration of antimicrobial therapy. Among these, neutrophil count, immature to total neutrophil ratio, C-reactive protein (CRP) and procalcitonin are the most used. Neutrophil indices have proven most useful for excluding infants without infection rather than for identifying infants with infection . CRP is an acute-phase protein synthesized and secreted by the liver in response to inﬂammatory cytokines and increases within 6 to 8 hours of infection, with a peak from 24 to 26 hours following infection . Thus has a low sensitivity if measured during the ﬁrst hours of life and its normal values variates with gestational age . Non infectious processes can also have an elevated CRP up to 10 times the normal concentration. Procalcitonin is a peptide precursor of the hormone calcitonin whose values appear to increase 2 hours following pathogen exposure, peak within 12 hours and return to baseline levels within 48 to 72 hours. It has better sensitivity but less speciﬁcity than CRP for identifying infants with neonatal sepsis . Procalcitonin levels increase physiologically after birth and in others conditions, such as respiratory distress syndrome.
number of patients, therefore not included in our analysis. According to the normal values for CRP and procalcitonin for preterm infants published in 2011 by Chiesa et al. , have been considered as elevated values of CRP >10 mg/L and values of procalcitonin >10 μg/L. In order to reduce the number of false positivity have been taken into account only procalcitonin levels determined after 24 hours of life. Patients were classiﬁed into four groups based on the clinical, laboratory and bacteriologic ﬁndings: (a) Sepsis: presence of clinical signs/symptoms, elevated CRP or procalcitonin levels, positive or negative blood culture (or culture from a deep site); (b) Speciﬁc infection: absence of clinical signs/symptoms, presence of elevated CRP or procalcitonin levels, positive blood culture (or positive deep culture or Group B streptococci (GBS) positive swab); (c) Non-speciﬁc infection or inﬂammation: absence of clinical signs/ symptoms, elevated CRP or procalcitonin levels and negative cultures; (d) Contamination: absence of clinical signs/symptoms, normal CRP and procalcitonin levels, a positive culture from any site. Intravenous antibiotics were administered in 86 infants (67% of cases). Ampicillin/sulbactam was the antibiotic used in all patients, alone or in association with amikacin, except in two cases of LOS where vancomycin and metronidazole were added to the ﬁrst two antibiotics. In two cases of non-bacterial infections, a case of herpes simplex virus (HSV) infection and a case of Candida systemic infection, speciﬁc treatment was administered (respectively acyclovir and ﬂuconazole).
2. Materials and methods
In our population of 771 late to moderate preterm infants, 137 had at least one clinical, laboratory and bacteriologic ﬁnding suggestive for infection. Demographic features and risk factors of the 137 infants are shown in Table 1. Of the 137 cases of suspected infection: 60 (44.2%) were classiﬁed as sepsis; 22 (16.2%) as speciﬁc infections; 46 (33%) as non-speciﬁc infections or inﬂammation and 9 cases (6.6%) were considered as a contamination, as reported in Table 2. Therefore, excluding the 9 contaminations, the treated infections were 128, with an incidence of 16.6%. The 90% of infections cases (n = 115) occurred during the ﬁrst 72 hours of life.
We retrospectively studied a population of 771 infants, both inborn and outborn, consecutively admitted to our unit from June 2008 to November 2013, with a GA between 32 and 36 weeks (GA 32 0/7–36 6/7 weeks). Moderate term infants, with GA of 32 and 33 weeks, were also included along with the LP (34–36 weeks of GA), being a population often not taken into account in many studies focused only on preterms with less than 32 weeks of GA. For all infants who showed at least one clinical, laboratory or bacteriologic ﬁnding suggestive for infection the following data were collected and analysed: history of maternal diabetes, use of antenatal corticosteroids, type of delivery, birth weight, Apgar score at 5 minutes, presence of respiratory disease or other symptoms during the hospital stay, insertion of an UVC, values of CRP and procalcitonin at different hours of life, results of cultures, antibiotic treatment. For each patient results of cultures performed (blood culture for aerobic and anaerobic bacteria, urine culture, bronchoalveolar ﬂuid culture, tip of oro-tracheal tube culture, tip of UVC culture, ear canal swab and pharyngeal swab cultures) have been recorded. Lumbar puncture for CSF culture was not routinely performed in preterm neonates in the period studied. Rectal and vaginal maternal swabs cultures, being taken usually at 35–37 weeks of gestation, were available only for a limited
Table 2 Treated infections in moderate and late preterm infants (GA = 32–36 weeks), classiﬁed on the basis of the clinical, laboratory and bacteriologic ﬁndings. GA
Speciﬁc infection (n)
Aspeciﬁc infection, inﬂammation (n)
32 33 34 35 36 Total
4 22 31 34 46 137
3 7 18 10 22 60
1 4 5 7 5 22
0 10 7 13 16 46
0 1 1 4 3 9
Table 1 Demographic features and risk factors of the 137 moderate and late preterm infants (GA = 32–36 weeks) studied. GA
Mean birth weight (g)
Mean Apgar at 5 min
Maternal diabetes (n)
Antenatal corticosteroids (n)
Cesarean section (n)
32 33 34 35 36
4 22 31 34 46
1710 1916 2085 2409 2623
8 8.6 7.8 8.7 8.9
2 1 7 2 1
2 12 16 10 4
0 4 2 5 1
4 18 23 21 32
1 5 3 2 5
4 14 14 8 16
S. Picone et al. / Early Human Development 90S1 (2014) S71–S74
Table 3 Number and type of specimens sampled, positive and negative cultures and microorganisms isolated. Type of specimen
Positive cultures n (%)
Negative cultures n (%)
Microorganisms isolated (n)
Coagulase-negative Staphylococcus (11) GBS (3) Citrobacter koseri (2) Streptococcus bovis (1) Streptococcus mitis (1) Streptococcus viridans (1) Escherichia coli (1) Staphylococcus aureus (1) Klebsiella pneumoniae (1) Anaerobic bacteria (1) HSV (1)
Escherichia coli (2) Klebsiella pneumoniae (2) GBS (1) Candida (1)
Coagulase-negative Staphylococcus (1)
Coagulase-negative Staphylococcus (10) Pseudomonas aeuginosa (2) Enterococcus faecalis (2) Candida (1)
GBS (8) Staphylococcus aureus (1) Candida (1)
The majority of cultures performed resulted sterile. Blood cultures were positive in 22% of cases, UVC cultures were positive in 26% of cases. Number and type of specimens sampled, positive and negative cultures and microorganisms isolated are shown in Table 3. The CRP mean value in patients of the sepsis group was 28.27 mg/L, much higher than the cut-off value considered. Also procalcitonin values of patients of the sepsis group were elevated, with a mean value of 25.3 μg/L. Among the possible risk factors for infection (UVC insertion, PROM, maternal diabetes, IUGR, cesarean section, antenatal corticosteroids) those with a signiﬁcant statistically association found were UVC insertion (χ2 = 15.9; p ≤ 0.05) and prophylaxis with antenatal corticosteroids (χ2 = 16.7; p ≤ 0.05). Birth by cesarean section seems to be the risk factor most frequently associated to infections in our population, with observed values very similar to the expected values (χ2 = 15.9; p = 0.1). Lower gestational ages instead were not correlated to an increased risk for infections (χ2 = 12.7; p = 0.5) Gram-positive organisms were those most frequently responsible for infections in our population of infants: Coagulase-negative staphylococci were the most frequently isolated pathogens (in 12 blood cultures and 10 UVC tip cultures), followed by GBS and Gramnegatives such as E. Coli, K. pneumoniae and Citrobacter koseri, the latter responsible for two cases of LOS with associated necrotizing enterocolitis. Pathogens have been also identiﬁed in the two cases of non-bacterial systemic infections occurred; in the HSV infection patient, detection of HSV DNA by Polymerase Chain Reaction was necessary. Respiratory symptoms were found in 47 of the 60 patients in the sepsis group (78.3%): 36 infants developed a severe respiratory distress who required endotracheal intubation (29 with surfactant administration), complicated by frank pneumonia in 6 cases; pneumothorax in 3 cases and pulmonary hypertension in 3 cases (one of which required nitric oxide); other 11 infants had instead a milder distress not requiring intubation. Two infants of the sepsis group developed a NEC with severe acute course that required surgery. In three infants sepsis was associated with disseminated intravascular coagulation. In the remaining 11 patients of the sepsis group the symptoms were mild or absent.
Jaundice that required phototherapy has occurred in 28 of the 128 infants with a treated infection (21.8% of cases). The short-term outcome of the treated infections in late and moderate preterm infants in our unit was good. Except for two infants with NEC who required to be transferred for surgical treatment, all other cases have been successfully treated and there has been no death due to sepsis. To compare the incidence of treated infections between late and moderate preterm infants and term infants we reviewed records of term infants (GA >37 weeks) born at our department over a period of six months (June 2013 to November 2013). Of the 1080 infants born at term, 58 (5.37%) had a treated infection, diagnosed with values of CRP >15 mg/l (mean CRP 45.33 mg/L; range: 15.54– 204 mg/L). Of the 52 blood cultures performed, 10 (19.2%) were positive (8 staphylococci, 1 GBS, 1 Escherichia coli), the remaining 42 resulted negative. CSF culture was positive in 1 of the 8 performed. Clinical symptoms were present in 18 of the 58 term infants with treated infections: 11 had respiratory symptoms; 7 other symptoms (septic shock, septic appearance, tremors, hypertonicity); 40 were asymptomatic. To compare the incidence of treated infections in late and moderate preterm infants with that of preterm infants with