Background

A cephaelhaematoma is defined as a subperiosteal haemorrhage in a newborn secondary to birth trauma. It occurs in 1–2 % of spontaneous vaginal deliveries and 3–4 % of forceps or vacuum-assisted deliveries [1]. Usually, it is a benign condition which resolves spontaneously over weeks without any treatment. Possible complications are hypotension, anaemia, jaundice or exostosis [2]. Moreover, there is also a potential risk of either primary infection through skin lesions or secondary infection through bacteraemia.

We present an illustrative case of an infant with an infected cephalhaematoma which was initially misdiagnosed. We added a literature review summarizing clinical presentation, risk factors, investigations results and treatment of previously reported cases.

Case presentation

A 5-week-old girl presented with decreased feeding and a fever of 38.5 °C. The infant was delivered at term by vacuum-extraction after an uneventful pregnancy. There was no protracted labour or use of a scalp electrode. A cephalhaematoma on the left parieto-temporal region was noted on the first day of life. At presentation on the 35th day of life, the only clinical symptom, apart from fever, was a non-tender cephalhaematoma of 12 × 16 cm without local signs of inflammation (Fig. 1). Laboratory tests revealed a haemoglobin level of 79 g/l, a total white blood cell count of 13.2 × 109/l, a platelet count of 233 × 109/l, a C-reactive protein level of 212 mg/l and an erythrocyte sedimentation rate above 110 mm/h. The girl was started on intravenous treatment with ceftriaxone 95 mg/kg/d and amoxicillin/clavulanate 150 mg/kg/d. Urine, blood and cerebrospinal fluid cultures remained sterile. A magnetic resonance imaging of the whole body, which was done in order to search for an infectious focus, revealed no other pathology apart from the cephalhaematoma (Fig. 1). The elevated inflammatory markers did not decrease under antibiotic treatment and were thought to be a reaction to the resorption process of the haematoma. The antibiotic therapy was stopped after 10 days and the girl was discharged.

Fig. 1
figure 1

Clinical presentation and magnetic resonance imaging findings of the 5-week-old girl with infected cephalhaematoma

Full size image

On follow-up one week later, a 2 × 2 cm sized, fluctuant, reddened superimposed area with blisters had formed on the cephalhaematoma. A diagnostic tap was done and the girl was started on intravenous cefuroxime 150 mg/kg/d. On the next day a debridement and evacuation of the haematoma was performed. The culture of the aspirate grew Escherichia coli. According to the antibiotic suspectibilities the treatment was changed to amoxicillin 150 mg/kg/d for total duration of 11 days. On follow-up four months later, the girl was well. There were no residual local findings and the neuropsychological development was normal.

Conclusion

A literature review was performed using Medline and Web of Science (1950 to January 2016) with the search terms: (cephalhematoma OR cephalhaematoma OR cephalohematoma) AND (infected OR infection) AND (treatment OR management). References were hand-searched for additional articles. Criteria for inclusion of publications were: (i) infants less than 4 months of age, (ii) case reports supplying sufficiently detailed clinical, diagnostic and therapeutic data, (iii) publications in English, Spanish, Italian, French or German. A total of 36 publications containing 42 case reports plus our case were included in the review [1, 3–37]. Descriptive statistical analysis was performed analysing demographic data, risk factors, clinical presentation, investigation results and treatment. Neonates were classified as preterm babies when born at a gestational age of less than 37 weeks. Prolonged rupture of membranes was defined as rupture ≥ 18 h before delivery. Fever was defined as a body temperature ≥ 38 °C and leucocytosis as a total white blood cell count ≥ 15 × 109/l. Infants were categorised as non-traumatic cases when delivered spontaneously without the use of a scalp electrode and in absence of skin abrasions. The time of diagnosis was defined as the time when the percutaneous tap was performed. The diagnosis of sepsis was made by positive blood cultures and of meningitis by positive cultures from cerebrospinal fluid. Osteomyelitis was either diagnosed when osteolytic lesions were present on skull X-rays or lytic bone changes on computed tomography images.

The clinical characteristics of the 43 infants with infected cephalhaematomas are summarised in (Table 1). The median age at the time of the first medical presentation was 10 (range 2–98, mean 15) days. The main reasons for presentation were an increase in size of the cephalhaematoma (n = 21, 49 %) or fever (n = 20, 47 %). Other reasons were poor feeding (n = 7, 16 %), lethargy (n = 8, 19 %), icteric skin colour (n = 8, 19 %) or pallor (n = 3, 7 %). One patient presented with seizures.

Table 1 Demographic and clinical characteristics of the 43 infants with infected cephalhaematomas
Full size table

Local findings indicating infection were secondary enlargement (n = 25, 58 %), erythema (n = 25, 58 %), tenderness (n = 17, 40 %), fluctuance (n = 13, 30 %) or skin abrasions (n = 9, 21 %). Systemic findings included fever (n = 28, 65 %), leucocytosis (n = 19, 44 %) or irritability (n = 11, 26 %). The white blood cell count was available from 23 patients, the median was 18 (range 5.5–34.5, mean 17) × 109/l. The C-reactive protein levels were available from 14 infants, the median was 166 (range 40–280, mean 168) mg/l.

Twenty-six percent (n = 11) of infants were diagnosed with meningitis. The median age at time of diagnosis was 17 (range 7–98, mean 23) days, which is seven days later than the median age at presentation. Patients presenting with sepsis and meningitis were diagnosed at a younger median age than children with osteomyelitis at 12, 18 and 22 days, respectively.

The main associations with infection of cephalhaematomas were instrumental delivery assistance in 44 % (n = 19) (forceps in 21 % (n = 9), vacuum-extraction in 21 % (n = 9) and both methods in one child) and sepsis in 42 % (n = 18). Further associated factors included the use of a scalp electrodes during delivery in 23 % (n = 10), skin abrasions in 21 % (n = 9) and prolonged rupture of membranes for more than 18 h in 7 % (n = 3) of infants. Maternal sepsis was not reported in any of the reviewed cases. Maternal administration of antibiotics prior or during delivery was not specified in the reviewed cases.

The used imaging methods and results are summarised in (Table 2). Osteomyelitis occurred in 38 % (n = 18) of infants. In eight patients the diagnosis was made because of osteolytic lesions on plain skull radiographs and in ten patients because of lytic bone changes on computed tomography images. In three patients a probable abscess was diagnosed with imaging (two via ultrasound and one via magnetic resonance imaging).

Table 2 Used imaging technics and results of the infants with infected cephalhaemtomas
Full size table

The pathogens isolated from the cephalhaematomas are summarised in (Table 3). E. coli was isolated from 67 % (n = 29) of haematomas, followed by other bacteria in much lower numbers. In patients without trauma E. coli was isolated in 75 % (n = 12/16), while in patients with skin abrasions it was isolated in 44 % (n = 4/9). In all of the 18 infants with sepsis, Gram negative rods were isolated from blood cultures, 17 of them were identified as E. coli. In all of the 11 infants with meningitis, Gram negative rods were isolated from cerebrospinal fluid; eight of them were E. coli. Only in 1 of the 28 children, who had a urine analysis, a urinary tract infection was found. However, while the pathogen isolated from urine was enterococcus, a culture from the cephalhaematoma of the same child grew E. coli. One child did not have an aspirate of the cephalhaematoma, but because the infant was found to have E. coli bacteraemia, meningitis and osteomyelitis of the parietal bone as well as skin lesions above the cephalhaematoma, it was presumed that the cephalhaematoma was infected.

Table 3 Pathogens isolated from infected cephalhaematomas
Full size table

All children received antibiotic treatment. The median duration was 21 (range 5–67, mean 27, data available from n = 41) days. The choice of antibiotics and the duration of therapy were heterogeneous. All but one infant were initially treated with intravenous antibiotics. Eleven patients were switched to oral antibiotics after a median duration of 21 (range 4–46, mean 19) days.

Surgical management was most commonly aspiration followed by incision and drainage (42 %, n = 18) or incision and drainage only (30 %, n = 13). In two infants aspiration was the only surgical intervention. Scalp bone was partially removed in 12 % (n = 5) of infants. Two patients died before a surgical intervention was possible and one child healed without a surgical intervention.

The median duration of hospitalisation was 23 days (range 5–47, mean 38, data available from n = 38). Six children were hospitalised twice. Of the two children who died, one suffered from unmanageable sepsis and meningitis, while the other one was found death in his bed 13 days after being treated for E. coli meningitis. Post-mortem examination revealed a grossly infected cephalhaematoma with growth of E. coli from an aspirate.

An infected cephalhaematoma is a rare, but potentially life-threating condition. Half of the infants with infected cephalhaemtoma present with non-specific signs of sepsis, such as fever, reduced feeding or lethargy and the other half because of changes of the haematoma, most frequently secondary enlargement or erythema of the overlying skin. Sepsis, instrumental assistance during delivery, the use of scalp electrodes and skin abrasions are the most important associated risk factors. Plain radiographs, computed tomography or magnetic resonance imaging have limited power to determine if a cephalhaematoma is infected, but can help in identifying associated osteomyelitis. Inflammatory markers in blood are often elevated, but this does not necessary mean that there is an infection. In two infants an infection of a cephalhaemtoma was suspected because of raised inflammatory markers, but no organisms were isolated from the haematomas or various other body fluids [2, 38]. At our clinic, we also looked after an infant who presented with a cephalhaemtoma and remarkably elevated inflammatory markers. The infant healed without any antibiotics or other interventions. Therefore, we suggest that elevated inflammatory markers can be part of the reabsorption process of haematomas and do not necessarily indicate that there is an infection, which can cause further difficulties in diagnosis.

Although, aspiration is contraindicated as a treatment option in cephalhaematomas, because of the potential risk of inoculating microbes, a percutaneous tap is necessary for the diagnosis of an infection.

Overall, E. coli is by far the most frequent pathogen responsible for infecting cephalhaematomas. Given that the identified associated risk factors for infection of cephalhaematomas are skin abrasions, instrumental assistance and the use of scalp electrodes during delivery, it is not surprising that entry for organisms acquired from the birth canal is facilitated. However, E. coli is more frequently isolated in non-traumatic cases than it is in patients with skin lesions. A further interesting finding is that even though skin abrasions are a risk factor for infection, no case of infection with Staphylococcus aureus as causative organism was reported.

Apart from primary invasion of pathogens, secondary infection associated with sepsis or meningitis is a further pathogenetic mechanism. Here, the spectrum of microbes is expected to be the same than in neonates with sepsis and/or meningitis without cephalhaematomas. However, studies including similar case numbers of neonates with sepsis or meningitis with comparable demographic and clinical characteristics (community-acquired late onset sepsis in term babies from developed countries before the implementation of intrapartum antibiotic prophylaxis) identified much lower numbers of Gram negative rods and E. coli as causative pathogens [39]. The proportion of E. coli in isolated late-onset meningitis is also much lower (32 %) than in infants with infected cephalhaematomas and meningitis (82 %) [40]. A further interesting finding is the rarity of Group B streptococci (GBS) as causative organisms in infected cephalhaematomas. Since the peripartal antibiotic prophylaxis for GBS does not influence the incidence of late-onset sepsis and most children with infected cephalhaematomas present at an age of 3 weeks or more, one would expect higher rates of GBS.

When an infant with a cephalhaematoma shows a decline in general well-being, fever or local signs of inflammation, infection of the haematoma should be suspected. Parents need to be informed about the potential risk of infection in cephalhaematomas when leaving the birth clinic.

Antibiotic treatment should cover the typical causative organism of neonatal sepsis/meningitis, particularly E. coli. Surgical intervention might be necessary.