OBJECTIVE: Acute symptoms are used to diagnose and manage acute otitis media (AOM). We studied whether AOM could be predicted by the reason for parental suspicion of AOM or by the occurrence, duration, and/or severity of symptoms. We also compared scores including or excluding tympanic-membrane examination of children with and without AOM.
PATIENTS AND METHODS: Children aged 6 to 35 months with parental suspicion of AOM were eligible. Before tympanic-membrane examination, we registered on a structured questionnaire the reason for parental suspicion of AOM, symptoms, and score components.
RESULTS: Of 469 children studied, 237 had AOM and 232 had respiratory tract infection without AOM. The most common reason for parental suspicion of AOM, restless sleep, was not predictive for AOM (RR: 1.0 [95% CI: 0.8–1.2]), nor was ear-rubbing (relative risk [RR]: 0.7 [95% confidence interval (CI): 0.5–1.0]). Neither the occurrence of fever (RR: 1.2 [95% CI: 1.0–1.4]) nor the highest mean temperature within 24 hours predicted AOM, nor did the occurrences of ear-related, nonspecific, respiratory, or gastrointestinal symptoms. The duration and severity of symptoms were not predictive for AOM, although rhinitis lasted longer and conjunctivitis was more severe in children with AOM. The clinical/otologic score (median: 4.0 vs 2.0; P = .000) and the AOM total-severity index (11.0 vs 6.0; P = .000), both including symptoms and tympanic-membrane examination, were higher in those with AOM. The AOM severity-of-symptom scale, based solely on symptoms, was equal in children with and without AOM (6.0 vs 6.0; P = .917).
CONCLUSIONS: AOM cannot be predicted by the occurrence, duration, or severity of symptoms at otitis-prone age. Likewise, solely symptom-based scores do not differentiate between respiratory tract infections with or without AOM. Thus, tympanic-membrane examination is crucial in the diagnosis and severity classification of AOM in clinical practice and research settings.
WHAT'S KNOWN ON THIS SUBJECT:
Acute symptoms and scores are used as tools in the diagnosis and management of AOM. However, their predictive value for AOM is not known for young children whose parents suspect AOM.
WHAT THIS STUDY ADDS:
The occurrence, duration, and severity of symptoms are not predictive for AOM at otitis-prone age. Symptom-based scores cannot differentiate AOM from respiratory tract infection. Tympanic-membrane examination is crucial for the diagnosis and scoring of AOM.
Acute symptoms play a crucial role in the diagnosis of acute otitis media (AOM). These symptoms, such as ear pain and fever, are included in the diagnostic criteria of AOM.1 Parents base their suspicion of AOM on the symptoms of their child. The guidelines advise physicians to use the severity of symptoms to choose the most appropriate treatment (antimicrobial therapy versus observation option) of AOM and to follow-up the episode of AOM.2 In clinical trials of AOM, symptoms have been used for severity scoring to assign children to different treatment groups,3 to study only 1 severity grade of AOM,4 and as the primary outcome in randomized trials that assess the effect of antimicrobial treatment.3,5,–,8
Clinical experience gives the impression that symptoms occurring at the time of AOM are variable. Only a few studies have focused on symptoms of AOM in the outpatient setting.9,–,11 All of the studies have included children with verbal skills, and none have examined the severity of symptoms. Thus, there is a lack of data on the predictive value of the occurrence, duration, and severity of symptoms in children aged 6 to 35 months, which is the age group with the highest incidence of AOM.
We studied the symptoms of children in this otitis-prone age group when their parents sought medical attention because AOM was suspected. The aim of this study was to find out if AOM could be predicted by the occurrence, duration, and/or severity of symptoms. We also compared scores from recent literature that included or excluded tympanic-membrane examination of children with and without AOM.
PATIENTS AND METHODS
This study was conducted between November 2006 and December 2008 and was part of a project examining the diagnosis, microbiology, and treatment of AOM at the primary care level. After telephone contact by parents, children aged 6 to 35 months were brought for an outpatient visit because of parental suspicion of AOM based on suggestive symptoms.
Written informed consent was obtained from a parent of each child before any study procedure was performed. All visits were free of charge, and no compensation for participation was given. The study protocol was approved by the ethical committee of the Hospital District of Southwest Finland.
Before tympanic-membrane examination, the study physician asked the reason for parental suspicion of AOM and the occurrence, duration, and severity of 17 symptoms by using a standardized, structured questionnaire. The highest measured temperature (≥38°C) within 24 hours was recorded. For the occurrence and duration of fever, we also accepted the parents' assessment of fever with no temperature measurement. Severity was classified as mild, moderate, or severe for the following symptoms: ear pain reported by parents and the child's verbal expression of ear pain; ear-rubbing; and irritability. In the following symptoms, the severity was classified as mild or severe: excessive crying; restless sleep; less playful or active; poor appetite; rhinitis; nasal congestion; cough; hoarse voice; conjunctivitis; mucus vomiting (retching and throwing up swallowed mucus); vomiting (throwing up partially digested foods and drinks); and diarrhea. The duration of symptoms was measured in days (with an accuracy of 0.5 days). Parents evaluated their child's overall condition with the AOM-faces scale,12 scoring 1 to 7, at worst within 24 hours and at the time of the visit (used with the kind permission of David P. McCormick, MD).
The study physicians first performed tympanometry (MicroTymp2 [Welch Allyn, Skaneateles Falls, NY]). After careful cerumen removal, we systematically assessed the middle-ear status (position, translucency, color, hyperemia, light reflex, mobility of the tympanic membrane, and possible air-fluid interfaces) by using pneumatic otoscopy (Macroview otoscope model 23810 [Welch Allyn]); the otoscopy score (8 grades) (OS-8)13; and which ear had the worse status. Digital pneumatic video otoscopy (Jedmed, St Louis, MO) was used to document the findings.
The diagnosis of AOM required 3 criteria: (1) middle-ear fluid detected by pneumatic otoscopy (at least 2 of the following signs on the tympanic membrane: bulging position; decreased or absent mobility; abnormal color or opacity not caused by scarring; or air-fluid interfaces); (2) at least 1 acute inflammatory sign of the tympanic membrane (distinct erythematous patches/streaks or increased vascularity over full/bulging/yellow convexity); and (3) symptoms and signs of acute infection.
Before examining the study population, all study physicians were validated to assess the tympanic-membrane findings and the OS-8. Of the 5 study physicians, 3 of us (Drs Laine, Tähtinen, and Ruohola) made more than 90% of the diagnoses and had an excellent agreement (κ values from 0.80 to 0.92). An ear-nose-throat specialist assessed the images and videos of 150 children without knowing their symptoms and/or our diagnosis. He agreed with 95% of our AOM diagnoses.
On the basis of our detailed symptom questionnaire and tympanic-membrane examination, we calculated 3 scores used in previous literature. First, we used the clinical/otologic score14 primarily developed by Dagan et al15 to determine the severity of AOM and to measure the treatment outcome of AOM (temperature, irritability, redness of tympanic membrane, and bulging position were scored from 0 to 3, for a total range of 0–12). Second, we used the modified AOM total-severity index (AOM-Si) as suggested by McCormick et al4 who used this score to determine the severity of AOM for studying the treatment of nonsevere AOM only. The AOM-Si score (range: 1–14) was calculated by including the highest OS-8 in pneumatic otoscopy (measuring the severity of tympanic-membrane inflammation [range: 0–7]) and the highest AOM-faces scale (measuring parental perception of their child's worst overall condition within 24 hours [range: 1–7]). Third, we used the AOM severity-of-symptom scale (AOM-SOS) (version 3.0) created by Shaikh et al16 to measure the outcome in clinical studies of AOM. The AOM-SOS score consisted of ear-rubbing, excessive crying, irritability, restless sleep, less playful or active, poor appetite, and fever, scored as 0 (none), 1 (a little, including our categories mild and moderate), or 2 (a lot, including our category severe). We classified temperature of <38°C as 0 (none), 38.0°C to 38.9°C as 1 (a little), and ≥39°C as 2 (a lot). A score range of 0 to 14 was the result.
In addition, we assessed illness severity according to the American Academy of Pediatrics (AAP) 2004 guidelines for the diagnosis and management of AOM.2 The child had severe illness if ear pain (parentally reported and/or reported by the child verbally) was moderate or severe and/or the highest temperature within 24 hours was >39°C. Otherwise, the child had nonsevere illness.
We did not analyze these 3 scores or the AAP's definition for illness severity if 1 or more components of a score (except measured temperature) were missing. If the child had had fever and the temperature had not been measured within 24 hours, we used the highest measured mean temperature of his or her study group.
The proportions were compared with χ2 test or Fisher's test as applicable. The means were compared with the t test and the medians with the Mann-Whitney U test. The likelihoods were estimated by calculating the relative risk (RR) with respective 95% confidence intervals (CIs). The relationships between the scores were assessed by Spearman correlation coefficients. The statistical analyses were performed by using the SPSS 16.0 statistical package (SPSS Inc, Chicago, IL).
The study population comprised 469 children: 237 had AOM (AOM group) and 232 had respiratory tract infection (RTI) without AOM (non-AOM group). Accordingly, parental suspicion of AOM proved to be correct for 51% of all children, for 48% (68 of 141) of children with no previous AOM, and for 52% (169 of 325) of children with previous AOM. Characteristics of children in the AOM and non-AOM groups are shown in Table 1. In the non-AOM group, 61 children (26%) had unilateral or bilateral middle-ear fluid. The symptoms of these 61 children did not differ from those with completely healthy ears; rather, the latter tended to have more severe symptoms (data not shown).
Reasons for Parental Suspicion of AOM
The most common reason for parental suspicion of AOM was restless sleep, listed for 134 of 468 (29%) children (data missing for 1 child in the AOM group). However, restless sleep could not predict AOM (RR: 1.0 [95% CI: 0.8–1.2]) (Fig 1), nor could irritability that had evoked parents' suspicion of AOM in 84 children (18%). Parents suspected AOM because of ear-rubbing in 64 children (14%), but when ear-rubbing was a reason for suspicion, AOM was improbable (RR: 0.7 [95% CI: 0.5–1.0]). Ear pain (parentally reported or reported by the child verbally), fever, and severe or prolonged rhinitis/cough were rare reasons for parental suspicion and could not predict AOM.
Occurrence, Duration, and Severity of Symptoms
The occurrence of ear-related symptoms could not predict which children had AOM (Table 2). Almost all parents reported that their child had ear pain, although children themselves had rarely expressed it verbally. Ear-rubbing tended to be more common in children who did not have AOM (70% [AOM group] vs 78% [non-AOM group]; P = .050). Occurrence of fever could not predict AOM (RR: 1.2 [95% CI: 1.0–1.4]). Furthermore, the highest mean temperature within 24 hours (38.7°C [AOM group] vs 38.6°C [non-AOM group]; P = .508) (Fig 2) and the duration of fever (2.1 days [AOM group] vs 1.8 days [non-AOM group]; P = .234) did not differ between the groups. The occurrences of nonspecific symptoms (irritability, excessive crying, restless sleep, less playful or active, and poor appetite) were not predictive for AOM. In fact, irritability suggested that the child did not have AOM (RR: 0.7 [95% CI: 0.6–0.9]). Respiratory symptoms, conjunctivitis, mucus vomiting, vomiting, and diarrhea could not predict which children had AOM.
The duration of symptoms had no predictive value for AOM in children having RTI with and without AOM except for rhinitis that had lasted ∼1 day longer in those with AOM compared with children with no AOM (Table 2).
The severity of parentally reported ear pain, child's verbal expression of ear pain, ear-rubbing, and irritability did not differ between children with and without AOM (Fig 3). Furthermore, nonspecific, respiratory, and gastrointestinal symptoms were equally severe in both groups (Table 3). Only conjunctivitis was more severe in those children who had AOM compared with those with only RTI.
The clinical/otologic score was significantly higher in children who had AOM than in children who had only RTI (median: 4.0 vs 2.0; P = .000) (Fig 4). The AOM-Si score was likewise higher in the AOM group than in the non-AOM group (11.0 vs 6.0; P = .000). However, the AOM-SOS score based solely on symptoms was equal between the AOM and non-AOM groups (6.0 vs 6.0; P = .917). The same applied to the scoring of the overall condition of the child by the AOM-faces scale. According to the AAP's definition for illness severity, 68% of children with AOM and 60% of children without AOM would have been categorized as having severe illness (P = .087).
Our main finding was that the occurrence, duration, and severity of symptoms did not predict AOM in children at otitis-prone age when their parents suspected AOM. Furthermore, solely symptom-based scores could not differentiate children with AOM from those without AOM. These results are important, because symptoms are used in the diagnosis and management of AOM.
For a new perspective, we analyzed the reasons parents had when they suspected AOM in their child. The most common reason given was restless sleep, the symptom that most disturbs the parents' own life, although it should be noted that it was not predictive for AOM. Parents of almost half of the children suspected AOM on the basis of a nonspecific symptom, and none of the reasons for suspicion could predict AOM.
Ear pain has been the symptom commonly associated with AOM.9,–,11 In our study, the occurrence of parentally reported or the child's verbal expression of ear pain could not predict AOM. Similarly, the severity of ear pain did not differentiate children with and without AOM. Our study design requiring AOM suspicion might explain why almost all parents reported ear pain in their child. In contrast, fewer than one-fifth of these young children verbally expressed ear pain, a finding that agrees with previous study results.9,11,17,18 Shaikh et al16 excluded ear pain from the AOM-SOS score, which, as in our results, points to parental difficulties with assessing ear pain in children at this preverbal and otitis-prone age.
We found it surprising that ear-rubbing tended to be more common in children without AOM than in children with AOM. This finding disagrees with the results of Niemelä et al,9 who included children with any kind of acute illness. Ear-rubbing has been used almost as a synonym for ear pain.2,–,4,9 However, infants may rub their ear because of a blocked ear or merely when becoming acquainted with their body. Furthermore, Baker19 reported that if ear-rubbing was a child's only complaint, then no child had AOM. On the basis of these results, we recommend against using ear-rubbing as evidence of ear pain or as a sign of AOM.
Consistent with the results of previous studies,9,–,11 the occurrence of fever did not predict AOM in young children, although pediatric textbooks describe fever as indicative of AOM.1,20,21 As in our study, fever usually has occurred in less than half of the AOM population.4,5,9,11,22,–,24 The distribution of the highest mean temperature in children with and without AOM was surprisingly similar. For this reason, fever might be associated with the viruses causing RTI rather than specifically with AOM.17,25
The occurrence, duration, and severity of nonspecific symptoms were equal in children with and without AOM. Nonspecific symptoms occur during viral infections, and frequently with no infection, in young children. As others have shown, nonspecific symptoms are not predictive for AOM at otitis-prone age.9,11
Respiratory symptoms could not predict AOM, with the exception of conjunctivitis, which was more severe in children with AOM than in children with only RTI. The so-called conjunctivitis-otitis syndrome is well known26,27 but has a limited role in predicting AOM because of its rare occurrence.9,11
Gastrointestinal symptoms were not predictive for AOM. In previous studies, the occurrence of vomiting was more than 10 times higher than in our study.9,28,29 However, we asked separately about actual vomiting and mucus vomiting and found that mucus vomiting was reported 10 times more frequently than actual vomiting. On the basis of this new finding, it would be important to report these 2 symptoms separately, because mucus vomiting is a respiratory symptom rather than a gastrointestinal symptom, and antimicrobial agents may cause mainly actual vomiting.
Applying the scores from recent literature to our data raised interesting and new considerations. The clinical/otologic score and the AOM-Si score, both based on symptoms and ear-related signs, were significantly higher in children with AOM than in children with only RTI. It is important to note that this was because of the ear-related signs of the scores; the symptom-based components did not differ. Similar to the AOM-faces scale, the AOM-SOS score based solely on symptoms gave equal scores to children having RTI with and without AOM. It should be noted that Shaikh et al16,30 developed the AOM-SOS score not to determine the severity of AOM but, rather, specifically to measure outcome in clinical studies of AOM. The AAP's criteria for severe illness is designed to determine the optimal treatment for AOM.2 In our study population, illness severity could not predict the probability of AOM. Because solely symptom-based scores and the AAP's definition of illness severity lead to similar results in young symptomatic children with and without AOM, we recommend validating scoring systems that would include tympanic-membrane examination.
This study has several strengths. We used a standardized, structured questionnaire for the symptom survey. Data were collected before tympanic-membrane examination. Although our AOM diagnosis was based on pneumatic otoscopy, which is somewhat subjective, our physicians also used tympanometry and had excellent interobserver agreement. We had a prespecified definition for AOM, even stricter than that of the AAP.2 Our AOM group had typical findings of positive bacterial culture: full/bulging position of tympanic membrane; purulent middle-ear fluid; and tympanometric B curve.24,31,32 In contrast, our non-AOM group had typical findings of negative bacterial culture or nonexistence of middle-ear fluid: retracted position of tympanic membrane and tympanometric C curve.24,33 Our approach was based on what is daily encountered in an outpatient setting: an acutely ill, young child with anxious parents who suspect AOM.
The limitations also must be elucidated. Our results cannot be generalized to older children with verbal skills. Because we focused on children with parental suspicion of AOM, our results are not applicable to the entire age group. Children without AOM in our study may have had more severe symptoms than did children with RTI in general. In addition, children with AOM in our study may have had a different profile of symptoms than did children with accidentally diagnosed AOM. Finally, the methodology could have been improved by asking the parents to complete the questionnaire.
The message of our study for clinicians and parents is that symptoms cannot predict AOM at otitis-prone age. Therefore, the diagnosis and management of AOM cannot be made by telephone contact. All children with symptoms that cause parental anxiety deserve careful clinical examination, including cerumen removal followed by pneumatic otoscopy.
The significance of this study for the otitis media authorities, guideline-makers, and researchers is that the symptom-based scores can poorly differentiate young children with AOM from those with only RTI. Even if children have the same diagnosis, each one has an individual spectrum of symptoms. Furthermore, each child has a different spectrum of symptoms in different episodes of infections caused by different microbes. Accordingly, if the severity scores are based solely on symptoms, the scoring actually depends more on the occurrence of the symptoms than on the severity of the symptoms. We propose that the scores also include evaluation of the tympanic membrane.
AOM cannot be predicted by the occurrence, duration, or severity of symptoms at otitis-prone age. Likewise, solely symptom-based scores do not differentiate RTI with and without AOM. Therefore, treatment recommendations should not be entirely based on symptom severity, and the role of tympanic-membrane examination should be emphasized in the diagnosis and scoring of AOM in clinical practice and research settings.
This work was supported by grants from the Foundation for Paediatric Research; the Research Funds from Specified Government Transfers; the Finnish Cultural Foundation, Varsinais-Suomi Regional Fund; Turku University Hospital Research Foundation; the Maud Kuistila Memorial Foundation; the Jenny and Antti Wihuri Foundation; and the Paulo Foundation. We acknowledge the European Society for Paediatric Infectious Diseases for the fellowship to Dr Ruohola (2006–2007).
We thank Raakel Luoto, MD, and Elina Lahti, MD (residents of pediatrics at Turku University Hospital) for their contribution to data collection, and Tuomo Puhakka, MD (ear-nose-throat specialist at Turku University Hospital) for assessing digital images and videos of tympanic-membrane findings of the study patients.
- Accepted December 4, 2009.
- Address correspondence to Aino Ruohola, MD, PhD, Department of Pediatrics, Turku University Hospital, PL 52, 20521 Turku, Finland. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
- AOM =
- acute otitis media •
- OS-8 =
- otoscopy score, 8 grades •
- AOM-Si =
- AOM total-severity index •
- AOM-SOS =
- AOM severity-of-symptom scale •
- AAP =
- American Academy of Pediatrics •
- RR =
- relative risk •
- CI =
- confidence interval •
- RTI =
- respiratory tract infection
- Bluestone CD,
- Klein JO
- 2.↵American Academy of Pediatrics, Subcommittee on Management of Acute Otitis Media. Diagnosis and management of acute otitis media. Pediatrics. 2004;113(5):1451–1465
- Kaleida PH,
- Casselbrant ML,
- Rockette HE,
- et al
- McCormick DP,
- Chonmaitree T,
- Pittman C,
- et al
- Appelman CL,
- Claessen JQ,
- Touw-Otten FW,
- Hordijk GJ,
- de Melker RA
- Damoiseaux RA,
- van Balen FA,
- Hoes AW,
- Verheij TJ,
- de Melker RA
- Le Saux N,
- Gaboury I,
- Baird M,
- et al
- Little P,
- Gould C,
- Williamson I,
- Moore M,
- Warner G,
- Dunleavey J
- McCormick DP,
- Saeed KA,
- Pittman C,
- et al
- Arola M,
- Ruuskanen O,
- Ziegler T,
- et al
- Baker RB
- Kliegman RM,
- Behrman RE,
- Jenson HB,
- Stanton BF
- McMillan JA,
- Feigin RD,
- DeAngelis CD,
- Jones MD
- Palmu AA,
- Herva E,
- Savolainen H,
- Karma P,
- Mäkelä PH,
- Kilpi TM
- Bodor FF
- Bodor FF,
- Marchant CD,
- Shurin PA,
- Barenkamp SJ
- Little P,
- Gould C,
- Moore M,
- Warner G,
- Dunleavey J,
- Williamson I
- Schwartz RH,
- Stool SE,
- Rodriguez WJ,
- Grundfast KM
- Copyright © 2010 by the American Academy of Pediatrics