The work was performed at the neonatal intensive care unit of the Pediatric Department, Medical University of Graz, Austria.
Neonates presenting with temperature symptoms: Role in the diagnosis of early onset sepsis
Article first published online: 9 APR 2012
© 2012 The Authors. Pediatrics International © 2012 Japan Pediatric Society
Volume 54, Issue 4, pages 486–490, August 2012
How to Cite
Hofer, N., Müller, W. and Resch, B. (2012), Neonates presenting with temperature symptoms: Role in the diagnosis of early onset sepsis. Pediatrics International, 54: 486–490. doi: 10.1111/j.1442-200X.2012.03570.x
- Issue published online: 26 JUL 2012
- Article first published online: 9 APR 2012
- Accepted manuscript online: 2 FEB 2012 06:58AM EST
- Received 17 July 2011; revised 18 January 2012; accepted 23 January 2012.
- bacterial infection;
- neonatal sepsis;
- newborn infant
Background: In this study, we aimed to evaluate the role of fever, hypothermia, and temperature instability in term and preterm newborns during the first 3 days of life and to identify risk factors for early onset sepsis (EOS) among newborns presenting with these temperature symptoms.
Methods: In this retrospective cohort study set in our level III neonatal intensive care unit, we included all newborns hospitalized within the first 24 h of life from 2004 to 2007.
Results: Of 851 newborns, 127 presented with temperature symptoms during the first 3 days of life (15%): 69 had fever, 69 had hypothermia, and 55 had temperature instability (8%, 8%, and 6%, respectively). Of 127 newborns presenting with temperature symptoms, 14 had culture-proven EOS/pneumonia (33% of all 42 newborns with culture-proven EOS/pneumonia), 67 had clinical EOS (30% of all 209 newborns with clinical EOS) and 46 were EOS-negative (8% of all 600 EOS-negatives). Factors associated with culture-proven EOS/pneumonia in newborns presenting with temperature symptoms were maternal fever (P= 0.009), chorioamnionitis (P < 0.001), antibiotic therapy of the mother (P= 0.04), poor skin color (P= 0.001) and syndrome of persistent fetal circulation (P= 0.01).
Conclusions: Every seventh newborn hospitalized at our neonatal intensive care unit developed fever, hypothermia and/or temperature instability during the first 3 days of life. Two-thirds of them had culture-proven or clinical sepsis. Temperature symptoms were rarely observed in EOS-negative newborns (8%) but despite low sensitivity, were highly specific for bacterial infection in preterm and term newborns.
Febrile neonates <1 month of life are at risk for bacterial infection, including meningitis, bacteremia, and urinary tract infections. For neonates presenting with fever to the emergency department or the practitioner, the rates of serious bacterial infections range from 4 to 28%.1–10 The available criteria for identifying young febrile infants at low risk for sepsis are of limited reliability when applied to neonates.1–4 The presenting symptoms are often subtle and may clinically be indistinguishable from those occurring in non-infectious conditions, including almost every sign of neonatal distress.11
Voora et al. reported on about 1% of term newborns from a transitional nursery who had developed fever during the first 4 days of life.12 Although common causes of fever during the first days of life include dehydration, breast-feeding, high birthweight, and cesarean section, the possible presence of an early onset bacterial sepsis (EOS) implies careful observation, sepsis work-up, and/or antibiotic therapy.12–15 Especially in an intensive care setting, newborns are a priori at a higher risk for serious bacterial infections and require a thorough septic work-up.
The aim of this study was: (i) to evaluate the prevalence of fever, hypothermia, and temperature instability in term and preterm newborns in the setting of a tertiary care neonatal intensive care unit (NICU) during the first 3 days of life; (ii) to determine risk factors for EOS among newborns presenting with temperature symptoms; and (iii) to identify diagnoses associated with presence of temperature symptoms in EOS-negative newborns.
The study site was the NICU of the Paediatric Department, Medical University of Graz, a tertiary care centre with a catchment area of about 8000 births per year. Data collection was done retrospectively by analyzing medical reports, case histories and the electronic patient filing system (“MEDOCS”). The system includes all laboratory results, perinatal data, information on medication, and detailed results of all clinical examinations done during hospitalization. Examinations were performed every 2–4 h; heart and respiratory rate and peripheral oxygen saturation were recorded continuously. All clinical, perinatal and laboratory data were crosschecked to insure the accuracy of the data.
Inclusion criteria were hospitalization within the first 24 h of life with documentation of clinical and laboratory data in order to classify newborns regarding EOS as well as accurate documentation of body and environmental temperature.
Unit policies regarding sepsis work-up algorithms and antibiotic treatment remained unchanged during the study period.
Definition of fever, hypothermia, and temperature instability
Temperature was measured rectally every 2–4 h during hospitalization or more frequently when needed and was documented together with incubator or bed temperature. Fever was defined as a rectal temperature >38.5°C, hypothermia as a rectal temperature <36.0°C.16 Temperature instability was defined as an increase or decrease of rectal temperature of >1.5°C within 3 h regardless of whether or not upper and lower temperature limits were exceeded. Presence of fever, hypothermia, or instable temperature was not regarded as being true if associated with changes of incubator or bed temperature or with other environmental causes, such as prolonged insufficient warming during resuscitation. Infants consecutively presenting with two or more temperature symptoms were counted as positive in each category. The term “temperature symptoms” was used for fever, hypothermia, and temperature instability.
The practice of temperature control in term and preterm newborns in our NICU is as follows: stable term neonates are cared for in a standard care bed without external temperature regulation devices. Instable or ill term neonates are cared for in an open bed with radiant warmer. All preterm neonates are cared for in an incubator. Neonates <28 weeks gestational age have incubator temperature set to 37°C and neonates from 28 to 37 weeks to 35°C. During the study period, therapeutic body cooling was not performed at our department.
Definition of EOS
Group 1: Culture-proven EOS/pneumonia
Proven EOS was defined as a positive culture from blood, cerebrospinal fluid, or tracheal aspirate in cases of clinically suspected sepsis during the first 72 h of life. Blood cultures were performed on every newborn with a suspected infection from a peripheral vein or, if available, from a central venous catheter at the time of catheter insertion before starting antibiotic treatment. Cerebrospinal fluid cultures were obtained from every patient with suspected meningitis. Cultures of tracheal aspirates were obtained from every intubated patient with suspected sepsis or pneumonia. Positive cultures from tracheal aspirates were regarded as true pneumonia only with one or more abnormal laboratory markers (C-reactive protein >8 mg/L, white blood cell count >34000/µl or <9000/µl, absolute neutrophil count >14400/µL or <7000/µL [<2000/µl during the first 24 h of life], immature to total neutrophil ratio >0.2).17–19 Other bacteria recovered in cultures were considered to be pathogenic unless they consisted of normal skin or upper respiratory flora.
Group 2: Clinical EOS
Clinical EOS was defined as culture-negative clinical sepsis with three or more clinical signs: (i) respiratory symptoms (apnoea, tachypnoea >60/min, retractions, cyanosis, respiratory distress); (ii) cardiocirculatory symptoms (tachycardia >180/min or bradycardia <100/min, arterial hypotonia); (iii) neurological symptoms (lethargy, irritability, seizures); (iv) poor skin color or prolonged capillary refilling time >2 s with either positive maternal risk factors (preterm rupture of membranes [>18 h in term newborns], chorioamnionitis and fever during labor) or at least one abnormal laboratory marker (see above).20,21
Group 3: EOS-negative
A control group was formed by all remaining sepsis-negative newborns hospitalized within the first 24 h of life during the study period.
Statistical analysis was done with spss version 17 (spss, Chicago, IL, USA) and Microsoft Excel 2007 (Microsoft, Redmond, WA, USA). We determined prevalence of temperature symptoms and EOS as well as sensitivity, specificity, positive and negative predictive values, and odds ratio with 95% confidence intervals for temperature symptoms, fever, hypothermia, and temperature instability in diagnosis of culture-proven EOS/pneumonia. Risk factors for presence of culture-proven EOS/pneumonia among newborns with temperature symptoms were determined using the χ2-test and the Mann–Whitney U-test. Binary logistic regression analysis with stepwise forward selection was performed using all significant variables from the univariate analysis as predictor and culture-proven EOS/pneumonia as outcome variable.
Risk factors for temperature symptoms among EOS-negative newborns were determined using the χ2-test and the Mann–Whitney U-test. Binary logistic regression analysis with stepwise forward selection was performed using all significant variables from the univariate analysis as predictor and presence of temperature symptoms as outcome variable. Statistical significance was set at P < 0.05.
Of 863 newborns hospitalized during the study period, 12 were excluded because of incomplete clinical information. Thus, the study population included 851 newborns with a median gestational age of 34 weeks and a median birthweight of 2120 g. Perinatal data are presented in detail in Table 1.
|Total study population||851(100)|
|Gestational age <28 weeks||89(10)|
|Gestational age 28–32 weeks||233(27)|
|Gestational age 33–37 weeks||253(30)|
|Gestational age >37 weeks||276(32)|
|Sex(male : female)||460:391(54:46)|
|Small for gestational age||87(10)|
|Positive maternal risk factors(premature rupture of membranes, chorioamnionitis, or fever during labor)||221(26)|
|Antibiotic therapy of the mother prior to birth||148(17)|
A total of 127 newborns presented with temperature symptoms (15%): 69 with fever (8%), 69 with hypothermia (8%), and 55 with temperature instability (6%). Fifty-one infants presented with more than one temperature symptom. Culture-proven EOS/pneumonia was diagnosed in 42/851 newborns (5%) with group B Streptococci (50%) being the most frequently isolated pathogens. Cultures from tracheal aspirates were positive in 11 cases (26% of all positive culture results) with Ureaplasma urealyticum (91%) being the most frequently isolated pathogen. A total of 209 newborns (25%) had diagnosis of clinical EOS and 600 were diagnosed as being EOS-negative (71%).
Overall, the prevalence of temperature symptoms did not significantly differ between term and preterm newborns (14.5% vs 15.1%, respectively). Presence of fever, hypothermia, and temperature instability did not correlate with gestational age, though prevalence of hypothermia was significantly higher in extremely low gestational age (<28 weeks) infants compared to other preterm and term newborns (18.0% vs 7.0%, P < 0.001).
Temperature symptoms and EOS
Of 69, 69, and 55 newborns presenting with fever, hypothermia, and temperature instability, culture-proven EOS/pneumonia was diagnosed in 8/69 (12%), 7/69 (10%), and 5/55 (9%) cases, respectively; clinical EOS was diagnosed in 37/69 (54%), 39/69 (57%), and 30/55 (55%) cases; and 24/69 (35%), 23/69 (33%), and 20/55 (36%) were EOS-negative. Sensitivity, specificity, positive and negative predictive value, and odds ratio of temperature symptoms in diagnosis of culture-proven EOS/pneumonia were 33%, 92%, 23%, 95%, and 6.0 (3.9–12.2), respectively (see Table 2). In EOS-positive term newborns, fever was the most frequent temperature symptom (27%) whereas in preterm newborns hypothermia was the most common symptom (19%).
|Temperature symptoms||Fever||Hypothermia||Temperature instability|
In 127 newborns presenting with temperature symptoms, risk factors associated with presence of culture-proven EOS/pneumonia in the univariate analysis were antibiotic therapy of the mother (7/14 EOS-positive newborns presenting with temperature symptoms, P= 0.04), maternal fever (2/14 newborns, P= 0.009), and chorioamnionitis (4/14 newborns, P < 0.001). Clinical findings associated with culture-proven EOS/pneumonia were poor skin color (8/14 newborns, P= 0.001), and syndrome of persistent fetal circulation (3/14 newborns, P= 0.01). A total of 12/14 newborns (86%) with culture-proven EOS/pneumonia who presented with temperature symptoms had at least one of these perinatal or clinical risk factors (7/14 had perinatal risk factors and 10/14 had clinical risk factors). Significant variables from the univariate analysis were inserted in a binary logistic regression model. Chorioamnionitis (OR 52.2 [3.7–725.5], P= 0.003) and poor skin color (OR 11.2 [2.1–60.0], P= 0.005) were associated with presence of culture-proven EOS/pneumonia.
Temperature symptoms in EOS-negative newborns
Of 127 newborns presenting with temperature symptoms, 46 were EOS-negative (36%). Twenty-four newborns had fever, 23 had hypothermia, and 20 had temperature instability (4%, 4%, and 3% of all EOS-negative newborns, respectively). By univariate analysis, risk factors associated with the presence of temperature symptoms in EOS-negative newborns were severe asphyxia (Apgar 1 ≤ 3; 7/46 EOS-negative newborns presenting with temperature symptoms, P= 0.009) and meconium staining of the amniotic fluid (6/46 newborns, P= 0.028). Clinical findings associated with presence of temperature symptoms in EOS-negative newborns were arterial hypotension (13/46 newborns, P < 0.001), tachycardia (16/46 newborns, P < 0.001), lethargy (30/46 newborns, P < 0.001), and syndrome of persistent fetal circulation (1/46 newborns, P= 0.024). A total of 44/46 EOS-negative newborns presenting with temperature symptoms had at least one of these conditions. By regression analysis symptoms associated with presence of temperature symptoms in EOS-negative newborns included tachycardia (5.7 [2.7–12.0], P < 0.001) and lethargy (5.8 [2.3–15.2], P < 0.001).
Every seventh newborn hospitalized at our neonatal intensive care unit developed fever, hypothermia and/or temperature instability during the first 3 days of life. Two-thirds of them had culture-proven or clinical sepsis. Temperature symptoms were rarely observed in EOS-negative newborns (8%) but despite low sensitivity, were highly specific for bacterial infection in preterm and term newborns.
The clinical problem of interpreting fever in the newborn is that it might always be the only clinical sign for neonatal bacterial infection.22 Febrile and hypothermic newborns are evaluated by pediatricians or neonatologists and the decision to perform septic work-up and/or to start antibiotic therapy depends on the newborn's appearance and its perinatal history.
Voora et al. reported on a prevalence of fever of 1% in term newborns with 10% of these febrile newborns having sepsis.12 Term newborns were described as being more likely to react to a bacterial infection with fever while preterm newborns were more likely to react with hypothermia.23 In our study we found comparable results. The differences in patterns of response to infection were most evident in the group of extremely low gestational age newborns <28 weeks with sensitivities of fever and hypothermia of 9% and 36%, respectively. Interestingly, preterm newborns ≥28 weeks gestational age had similar values compared to term infants. Early postnatal hypothermia was primarily considered to be associated with insufficient warming or heat conservation and not possible EOS. Therefore our prevalence was lower compared to previous analyses reporting on hypothermia being present in up to 40% of preterm infants <26 weeks gestational age at the time of first temperature measurement24 and in 10% of near-term newborns.25
Schwartz et al. described a decrease of the prevalence of serious bacterial infections in febrile neonates presenting to the emergency department with increasing postnatal age declining from 22% (3–7 days of life) to 12% (22–28 days of life).1 We conducted this study in the setting of an NICU, where infants present with a vast variety of clinical signs and symptoms. Additionally, a high percentage of newborns carry maternal risk factors associated with bacterial infection, and hospitalization is a result to rule out sepsis. We report a higher prevalence of sepsis of 65% among febrile newborns and of 64% among newborns presenting with any temperature symptom, though the majority of these cases were not culture-proven. The predictive values for sepsis did not differ between fever, hypothermia, and temperature instability ranging from 64% to 67%.
Neonatal fever may occur when secretion of pyretic inflammatory cytokines raises the hypothalamic temperature set point, leading to heat-conservation and heat-generation physiologic responses. However, mechanisms besides inflammatory patterns may as well lead to elevated body temperature.14,22 Non-infectious conditions being associated with neonatal fever include dehydration, breast-feeding, high birthweight, and cesarean section.12–14 Excessive bundling may also elevate newborn temperature.26,27 Overall rates of temperature symptoms in EOS-negative infants were very low, making these symptoms a highly specific diagnostic tool in the diagnosis of bacterial infection. The vast majority of infants presenting with temperature symptoms suffered severe illnesses and/or had other organ systems involved. Temperature symptoms were extremely rare in otherwise healthy infants without any other sign of illness.
Fever was defined as a rectal temperature >38.5°C. This is higher compared to some previous studies, where limits were set from 37.5 upwards.28 The higher cut-off value was found to be of improved specificity in diagnosis of EOS and was routinely used as the standard definition during the study period.
As far as blood cultures lack sensitivity, the number of culture-positive newborns might be underestimated in our study. Newborns with clinical sepsis, not confirmed by positive culture results, were not classified as true sepsis cases according to international definitions and therefore excluded from analysis.
In conclusion, temperature symptoms were a common symptom in our NICU with a prevalence of 15%. Fever, hypothermia, and temperature instability had almost identical predictive values for sepsis with two-thirds of the affected infants having culture-proven or clinical sepsis. Temperature symptoms were rarely observed in EOS-negative newborns (8%). Despite its low sensitivity, temperature symptoms were highly specific for bacterial infection in preterm and term newborns.
- 14Bacterial sepsis and meningitis. In: Remington JS, Klein JO, Wilson CB (eds). Infectious Diseases of the Fetus and Newborn Infant, 6th edn. Elsevier Saunders, Philadelphia, PA, 2006; 247–95., , .
- 15American College of Emergency Physicians Clinical Policies Committee; American College of Emergency Physicians Clinical Policies Subcommittee on Paediatric Fever. Clinical policy for children younger than three years presenting to the emergency department with fever. Ann. Emerg. Med. 2003; 42: 530–45.
- 17Haematological values in the newborn. In: Rennie JM, Roberton NRC (eds). Textbook of Neonatology, 3rd edn. Churchill Livingstone, Edinburgh, 1999; 1399..
- 22Temperature regulation of the premature infant. In: Taeusch HW, Ballard RA, Gleason CA (eds). Avery's Diseases of the Newborn, 8th edn. Elsevier Saunders, Philadelphia, PA, 2004; 364–71., .