Otitis Media in the Neonatal Intensive Care Unit

(1)

Otitis

Media

in the

Neonatal

Intensive

Care

Unit

Steven A. Berman, M.D., Thomas J. Balkany, M.D., and Michael A. Simmons, M.D.

From the Departnaents of Pediatrics and Otokiryngology, (Inicersity of Colorado Medical Center, Denver

ABSTRACT. Thirty-eight of 125 premature infants who

were hospitalized in a neonatal intensive care unit (NICU)

had abnormal tympanic membrane mobility compatible

with otitis media. Twenty-five of these 38 had received

antibiotics within one week of otoscopic examination and

were considered to have either serous otitis or partially

treated bacterial otitis media; tyinpanocentesis was not

performed in them. Tympanocentesis was performed in the

remaining 13 infants who had not received antibiotics.

Bacterial otitis media was confirmed in ten of the 13. Either

staphylococcal (six cases) or Gram-negative enteric

orga-nisms (four cases) were isolated in cultures obtained by

tympanocentesis in these cases. The four cases of

Gram-negative infections occurred in infants within six weeks of

birth. Nasotracheal intubation for more than seven days was

significantly correlated with impaired tympanic membrane

mobility compatible with otitis media. Otitis media occurs

frequently among premature infants who are hospitalized in

an NICU, and it should be looked for in any infant in whom

sepsis is clinically suspected. Pediatrics 62:198-201, 1978,

newborn, otitis media, nasotracheal intubation, tympanocen-tesis.

The occurrence of acute bacterial otitis media

in infants under 6 weeks of age has been

docu-mented in both clinical’5 and autopsy studies.67

Failure to accurately diagnose this condition has

important clinical implications, because

unrecog-nized acute bacterial otitis media may act as a

focus for the dissemination of the organism into

the bloodstream and/or CNS.’28

Although suppurative otitis media in older

children is usually caused by Streptococcus

pneu-moniae, Haemophilus influenzae, and S. pyogenes

group A,8 12 studies in infants within six weeks of

birth have revealed a high incidence of

Gram-negative enteric and staphylococcal

orga-nisms.’’5 These neonatal studies, however, have

not distinguished between infants living at home

and those hospitalized in the neonatal intensive

care unit (NICU), where the risk of nosocomial

Gram-negative enteric infection is greater.

The present study was undertaken to

deter-mine prospectively the clinical presentation,

bacteriologic features, and clinical evolution of

acute otitis media in the NICU.

MATERIALS

AND METHODS

One hundred twenty-five infants were

exam-med in the NICUs at the Colorado General

Hospital and the Denver Children’s Hospital

between December 1975 and May 1976. The

infants’ postnatal ages ranged from 1 day to 4

months and their gestational ages ranged from 25

weeks to term. A pediatrician (S.A.B.) examined

the infants with a hand-held otoscope with a

pneumatic attachment. From this group, 50

infants whose ages ranged from 1 day to 4 months

and who weighed from 700 gm to 3.2 kg were

independently evaluated by an otolaryngologist

(

T.J.B.)

using

a

binocular operating microscope.

An adequate view of the tympanic membrane was

obtained by inserting the speculum while

retract-ing the pinna posteriorly. The speculum was

gently advanced while pneumatic pressure was

used

to distend the collapsing external auditory

canal skin. The ear canals were cleaned by

curettage or suction when debris prevented an

adequate view of the tympanic membrane. Care

was taken to distinguish ear canal motion from

tympanic membrane mobility.

Each of the 38 infants with abnormal tympanic

membrane mobility was retrospectively matched

for gestational age

(

± 2 weeks) and postnatal age

(

± 2 weeks) with an infant with normal otoscopic

findings. The presence of abnormal mobility

compatible with otitis media was correlated with

the duration of nasotracheal intubation. In

addi-tion, the number, indications for, and percent of

Received October 26, 1976; revision accepted for

publica-tion February 3, 1978.

Dr. Simmons is an Established Investigator of the American Heart Association.

ADDRESS FOR REPRINTS: (M.A.S.) Department of

Pedi-atrics, The Johns Hopkins Hospital, 601 North Broadway,

(2)

ARTICLES 199

positive yield from cultures of blood, CSF, and

urine (septic workup) in the 38 infants with

abnormal otoscopic features were compared to

the 38 age-matched controls.

RESULTS

Thirty-eight (30.5%) of the 125 infants

exam-med during a four-month period had abnormal

tympanic membrane mobility. Of these, 25 were

considered to have either serous otitis media or

partially treated suppurative otitis media since

they were receiving antibiotics. Results of

tympa-nocentesis performed in the 13 infants with

abnormal tympanic membrane mobility who had

not received antibiotics within the preceding

week were considered positive when Gram stain

of a middle ear aspirate showed many bacteria

and WBCs or when culture of the aspirate was

positive with a negative culture of the external

auditory canal.

Sex. There was no influence of the infant’s sex

on the incidence of otitis documented by positive

findings on tympanocentesis or by abnormal

mobility without tympanocentesis.

Postnatal Age. Of the ten cases of documented

bacterial otitis media, four occurred in infants

within 3 weeks of birth, three between 3 and 6

weeks, one between 6 and 9 weeks, and two

between 9 and 12 weeks.

Birth Weight. Birth weight, which ranged from

700 to 3,200 gm, did not affect the incidence of

cases

with abnormal tympanic membrane mobili-ty or documented otitis.

Clinical Manifestations. The most common

symptoms in infants with bacterial otitis media

were fever (60%), abdominal distention (50%),

vomiting (40%), diarrhea (40%), irritability (30%),

poor feeding (30%), and congestion (20%).

Seventy-six percent of the nursery infants had two

or more of the indicated symptoms, and

approxi-mately one half of the infants had at least three of the symptoms.

Septic Workups. Septic workups in the NICU

consisted of two blood cultures, CSF culture,

suprapubic urine culture, complete blood cell

count, platelet count, and chest roentgenogram.

Such workups were performed on those patients

presumed to have significant bacterial infection

based on clinical findings. A retrospective

evalu-ation of infants who had undergone septic

workups (their physicians were unaware of the

presence or absence of suppurative otitis media

because of failure to perform otoscopic

examina-tions) demonstrated that a large number (84%) of

all neonates later shown to have abnormal

tympanic membrane mobility compatible with

otitis were thought by their physicians to have

systemic infection, but in only five (6%) of these

was a cause demonstrated by cultures of throat,

blood,

sputum, stool, CSF, or urine. Fewer (37%)

of the age-matched non-otitis group were thought

to be septic on clinical grounds, but a large

number (36%) of those who were had positive

cultures.

Relationship Between Culture and Grain Stain Results. Gram stain and culture were both

posi-live in eight of the 13 specimens obtained by

tympanocentesis; Gram stain was positive and

culture was negative in one; Gram stain was

negative and culture was positive in one; and

Gram stain and culture were both negative in

three. In one case Gram stains demonstrated

polymorphonuclear leukocytes and

Gram-posi-tive organisms but the culture was negative. A

culture from a repeat tympanocentesis three

weeks later yielded Staphylococcus epiderniidis. In another case a positive culture with a negative

Gram stain was found. The percentage of positive

findings on tympanocentesis in patients clinically

thought to have bacterial otitis media was 77%.

Bacteriology. The following organisms were

found

in the ten patients: Sta. epidermidis in four

(40%), Sta. aureus in two (20%), Klebsiella

pneu-noniae in two (20%), Eschericliia coli in one

(10%), and Enterobacter in one (10%).

Simulta-neous

negative ear canal cultures documented the middle ear origin of the isolated organisms.

Relationship of Nasotracheal Inttibation and Otitis Media. Of the 38 infants with abnormal

tympanic membranes in this study, 24 had been

intubated for seven days or longer. Only six of the

38 age-matched infants with normal tympanic

membranes had been nasally intubated for seven

or more days. All six of these infants had received

at least two weeks of antibiotic therapy prior to

otoscopy, so the absence of otitis might have been

obscured by treatment. This association of otitis

media with nasotracheal intubation for at least

one week is statistically significant (P < .001). Associated Infections. The two cases of Kleb-siella otitis media in nursery patients were asso-ciated with Klebsiella septicemia.

Accuracy of Hand-Held Pneumatic Otoscopy.

One hundred tympanic membranes were

inde-pendently evaluated by a pediatrician (S.A.B.)

and otolaryngologist (T.J.B.) in order to compare

the accuracy of a hand-held otoscope (with a

pneumatic attachment) with an operating

iiiicro-scope. The observers agreed on mobility in 93 of

100 tympanic membranes examined (55 normal,

38 abnormal). If one assumes that the operating

microscope is a superior diagnostic tool in the

by guest on September 8, 2020

www.aappublications.org/news

(3)

determination of mobility, there were two

false-positives and five false-negatives with the

hand-held otoscope. Normal color, usually a dull pink,

was found in 19 of the 37 tympanic membranes

with abnormal mobility. Abnormal mobility was

noted in 23 of 25 tympanic membranes noted to

have abnormal color.

We have found that tympanometry does not

correlate well with otoscopic findings in the

premature infant within six weeks of birth. This is

probably related to the inability of the technique

to distinguish ear canal from tympanic membrane

mobility in these patients.

DISCUSSION

Bacterial otitis media is a common finding in

the NICU. While 30% of the 125 infants examined

were noted to have abnormal tympanic

mem-branes, most of these infants were receiving

antibiotics and tympanocentesis was not thought

to be justified. Among the 13 infants in whom

suppurative otitis media was diagnosed, ten had

the diagnosis confirmed by tympanocentesis.

All ten cases of bacterial otitis media had either

Gram-negative enteric (four) or staphylococcal

(six)

organisms isolated from cultures. A Gram-negative enteric infection was identified in four of

the seven infants with otitis media who were less

than 6 weeks of age. The three infants aged 6

weeks or older with otitis had staphylococcal

infections.

Other studies in children within six weeks of

birth in non-NICU settings suggest that postnatal

age affects the incidence of Gram-negative

enter-ic otitis media. In a study by Bland,1 72% of

positive cultures obtained by tympanocentesis

yielded Gram-negative enteric organisms. In

studies by Shurin et al.2 and Tetzlaff et al.,5 the

Gram-negative enteric organisms accounted for

10% and 18% of bacterial otitis in infants within

six weeks of birth. No Gram-negative enteric

infections were identified by Tetzlaff et al. in

infants older than 6 weeks. Our own findings

suggest that in addition to the effect of postnatal

age there may be a higher incidence of otitis

media caused by Gram-negative enteric

orga-nisms among infants hospitalized in an NICU.

Staphylococcus epidermidis, which was isolated

in four cases, is not usually considered a

patho-genic organism. It was considered pathogenic in

our study when associated with a positive Gram

stain, positive ear aspirate culture, and negative

ear canal culture. Other studies have also

identi-fled Sta. epidermidis as a causative agent of

bacterial otitis media.3 ‘‘

The demonstration of an association between

prolonged nasotracheal intubation and abnormal

tympanic mobility was an unexpected finding. It

is not clear whether the abnormal mobility

asso-ciated with nasotracheal intubation is caused by

bacterial otitis media or is secondary to serous

otitis. It would be difficult to establish a causal

relationship between prolonged nasotracheal

intubation and bacterial otitis since infants

requiring prolonged intubation for respiratory

support

have other underlying conditions that

might predispose to the development of bacterial

otitis media. Since we use the nasal route for

intubation almost exclusively, we cannot predict

any association of otitis with oral intubation.

Since 80% of the infants who had been

intu-bated for at least seven days had abnormal

tympanic membrane mobility, it is very

impor-tant to determine the frequency of bacterial

infection associated with this impaired mobility.

We recommend periodic otoscopy for these

patients and careful examination when sepsis is

suspected. Failure to perform such otoscopy in

the NICU might increase the prevalence of

chronic partially treated bacterial otitis media.

The comparison between the number of septic

workups among the infants with otoscopic

find-ings compatible with otitis media and among the

retrospectively age-matched controls suggests

that infants with otitis media have a greater

number of negative septic workups, and therefore

often receive multiple short courses of antibiotics

while awaiting culture results. A short course of

antibiotic therapy is probably inadequate

treat-ment for neonatal otitis media and, we suspect,

might lead to persistent, partially treated otitis.

In five of the ten cases of documented bacterial

otitis media, abnormal otoscopic findings

per-sisted and were associated with recurrent

symp-toms compatible with sepsis (such as fever and

vomiting). Cultures of blood, CSF, and urine were

consistently negative in these patients. All five of

these infants had chronic respiratory disease

requiring continual respiratory support with

nasotracheal intubation. Cultures from repeat

tympanocentesis after each had received three to

four weeks of oral antibiotic therapy based on

culture and sensitivity results yielded

staphylo-coccal organisms (three with Sta. epidermidis and

two with Sta. a areas).

The management of bacterial otitis media in

the NICU must be individualized. When results

of otoscopy are compatible with middle ear

infection, a tympanocentesis should be

per-formed. If a bacterial infection is documented by

culture or Gram stain, a complete septic workup

(4)

ARTICLES 201

CSF, and urine. Bacterial otitis media in the

NICU can be associated with sepsis and

meningi-tis.7 Two patients with Kiebsiella otitis media in

this study also had Klebsiella septicemia.

Antibiotic therapy of bacterial otitis media in

the NICU must be based on sensitivities to the

organism isolated by tympanocentesis and the

clinical condition of the infant. In our experience,

staphylococcal organisms were most often

responsible for unsuccessful treatment. In such

refractory cases of otitis, repeat tympanocentesis should be performed.

It should be emphasized that bacterial otitis

media occurs frequently in the NICU and

pre-sents with signs indistinguishable from other

infections. A search for otitis is essential in all

infants suspected of having any infection

regard-less of their birth weight or age.

REFERENCES

1. Bland R: Otitis media in the first six weeks of life:

Diagnosis, bacteriology, and nianagement.

Pediat-ncs 49:187, 1972.

2. Shurin PA, Pelton SI, Klein JO: Otitis media in the

newborn infant. Ann Otol Rhino! Laryngol 85

(suppl 25):216, 1976.

3. Warren WS, Stool SE: Otitis media in low birth weight

infants. I Pediatr 79:740, 1971.

4. Jaffe JF, Hurtado F, Hurtado E: Tympanic membrane

mobility in the newborn (with seven months

follow-up). Laryngoscope 30:36, 1970.

5. Tetzlaff TR, Ashworth C, Nelson JD: Otitis media in

children less than 12 weeks of age. Pediatrics

59:827, 1977.

6. deSa DJ: Infection and amniotic aspiration of middle ear

in stillbirths and neonatal deaths. Arch Di.s Child

48:872, 1973.

7. McLellan MS, Strong JP, Johnson QR, Dent JH: Otitis

media in premature infants: A histopathologic

study. I Pediatr 61:53, 1962.

8. Ermocilla R, Cassady G, Ceballos R: Otitis media in the

pathogenesis of neonatal meningitis with group B

beta-hemolytic streptococcus. Pediatrics 54:643,

1974.

9. Feingold M, Klein JO, Haslam GF, et al: Acute otitis

media in children. Am I Dis Child 1 1 1:361, 1966.

10. Bluestone CD, Shurin PA: Middle ear disease in

chil-dren: Pathogenesis, diagnosis, and management.

Pediatr Clin North An, 21:379, 1974.

11. Howie VM, Plussard JH, Lecter RL: Otitis media: A

clinical and bacteriological correlation. Pediatrics

45:29, 1970.

12. Nelson BW, Poland RL, Thompson RS, et al: Acute

otitis media: Treatment results in relation to

bacte-rial etiology. Pediatrics 43:351, 1968.

13. Feigin RD, Shackelford PC, Campbell J, et al:

Assess-ment of the role of Staphylococcus epiderinidis as a

cause of otitis media. Pediatrics 52:569, 1973.

ACKNOWLEDGMENT

We thank Frederick C. Battaglia, M.D., Bonnie W. Camp,

M.D., and J. K. Todd, M.D., for their careful review of the

manuscript and helpful criticism. We appreciate the aid

given by Fred Bruhn, M.D., in identifying six cases of

outpatient otitis seen at the Denver General Hospital.

(5)

1978;62;198

Pediatrics

Steven A. Berman, Thomas J. Balkany and Michael A. Simmons