Evaluation of phenoxymethylpenicillin treatment of acute otitis media in children aged 2–16

Abstract

Objective. To study the clinical recovery from acute otitis media (AOM) in children, 2–16 years of age, managed with or without treatment with phenoxymethylpenicillin (PcV). Design. An open, prospective randomized trial. Children aged between 2 and 16 years, presenting with one- or double-sided AOM (without perforation) with symptom duration of less than four days, were included. The children were randomized to PcV for five days or to no primary antibiotic treatment. A health score and compliance were registered on a daily basis for seven days. Setting. A total of 32 health centres and 72 GPs in south-east Sweden. Subjects. Children aged 2–16 presenting with earache. Main outcome measures. Recovery time, symptom duration, frequency of complications (up to three months) and consumption of healthcare services independent of treatment with or without antibiotics. Results. A total of 179 patients carried out the trial; 92 were randomized to PcV, 87 to no primary antibiotic treatment. The median recovery time was four days in both groups. Patients who received PcV had less pain (p <0.001) and used fewer analgesics. There were no significant differences in the number of middle-ear effusions or perforations at the final control after three months. Children randomized to PcV treatment consulted less (p <0.001) during the first seven days. Conclusions. Our investigation supports that PcV treatment of AOM does not affect the recovery time or complication rates. PcV provided some symptomatic benefit in the treatment of AOM in otherwise healthy children, aged 2–16 years.

• Acute otitis media
• antibiotics
• AOM
• children
• family practice
• general practice
• PcV
• phenoxymethylpenicillin
• primary healthcare

Acute otitis media (AOM) is one of the most common reasons for visiting primary health services in childhood. In most cases AOM is preceded by a viral respiratory tract infection. Nocturnal onset of earache, most often moderately increased fever, and general distress may be present.

Placebo-controlled studies of treatment with and without antibiotics [1–6] have shown limited differences in outcome between antibiotic-treated and non-treated groups. Antibiotics gave less contralateral AOM and a shortened period of pain, especially when Streptococcus pneumoniae was present [1], [7], [8].

Spread of multiresistant pneumococcal micro-organisms, especially among preschool children in the 1990s [9], [10], led to a re-evaluation of the guidelines for the treatment of acute otitis media in the year 2000 [11]. Until 2000, the recommended treatment strategy for all purulent AOM was phenoxymethylpenicillin (PcV): 25 mg/kg×2 for 5 days in Sweden. The new national Swedish guidelines still recommend empirical treatment with PcV for children below the age of two years, to those with perforated acute otitis media, general distress, or underlying disorders. For otherwise healthy children, 2–16 years of age, with sporadic cases of AOM (less than three episodes of AOM in six months), the guideline suggested a “wait and see” treatment alternative. If symptoms persisted for three days from symptom onset, the child should be examined again, and antibiotics should be prescribed if the clinical condition persisted. The guidelines concluded that few studies had been performed in daily practice in primary care and requested prospective evaluations of the usefulness of PcV, regarding the effect both on patient recovery and on the workload of primary care.

The aim of this investigation was to evaluate the possible benefits of PcV treatment as compared with an active “wait and see” policy in the treatment of uncomplicated acute otitis media in children aged 2–16 years in primary care. Symptom duration, frequency of complications (up to three months), and consumption of healthcare services were monitored.

Bacterial resistance in respiratory tract infections is a growing problem in Sweden and worldwide.

• The benefit of antibiotic treatment of AOM in children aged 2–16 is limited.

• Children who received PcV had less pain, used fewer analgesics and consulted less during the first seven days.

• PcV treatment of AOM did not affect the recovery time or complication rate

Material and methods

In 2002, 32 health centres and 72 general practitioners in Kalmar, Jönköping and Östergötland County agreed to participate in the study. All participating centres were equipped with aural microscopes.

Written introduction to the trial was given to all children aged 2–16 and/or their guardian when consulting due to earache. After clinical examination and confirmation of the diagnosis, the children were randomized by an Internet-based random number generator to treatment with PcV 25 mg/kg×2 for five days or no antibiotic treatment. The guardians also received written information about how to act if the condition did not improve or got worse within three days after randomization.

Exclusion criteria

Patients with perforation of the eardrum, chronic ear conditions or impaired hearing, previous adverse reactions to penicillin, concurrent disease that should be treated with antibiotics, recurrent AOM (three ore more episodes of AOM during the past six months), children with immunosuppressive conditions, genetic disorders, and mental disease or retardation were not included.

Additional treatment

Symptomatic treatment with paracetamol or NSAIDs, drugs reducing the swelling of the nasal mucosa (e.g. xylomethazolin), and nasal steroids was allowed.

Patients’ diary

All participants were asked to register in a semi-structured way on a daily basis symptoms such as pain (0 = no pain; 1 = some, 2 = moderate, 3 = severe), number of doses of given medications, fever, sleeping disturbance, rash, vomiting, diarrhoea, absence from day-care/school/sick-leave, and the day they appraised that the condition of their child was back to normal (recovery day). The diary was returned to the primary care centre after one week. A nurse telephoned all participants after approximately 14 days to supplement the information in the diary and to register all acute contacts that had occurred during the first week of treatment.

The final follow-up was performed after three months and perforations and serous otitis media were registered. This consultation also included information on healthcare contacts during the past three months. Patients who did not show up at the three-month follow up were interviewed by phone.

Definitions

AOM

The diagnosis should be based on direct inspection of the eardrum by a pneumatic otoscope or preferably an aural microscope. Findings should include a bulging, red eardrum displaying reduced mobility.

Treatment failure

Absence of improvement or increasing symptoms after at least three days of treatment.

Ethics

The study was approved by the Ethical Committee in Linköping and the Swedish Medical Products Agency.

Statistics

Differences between the three groups (see Table I) were analysed using a chi-squared test for proportions if not otherwise stated (followed by Fisher's exact test in the case of significance) and non-parametric tests for continuous variables (Kruskal–Wallis test followed by Mann–Whitney U-test in the case of significance). In the case of two groups (see Tables II and III) Fisher's exact test was used for proportions and a Mann–Whitney U-test for continuous variables.

Table I.  Registered data on 92 children randomized to PcV treatment (1), 87 children randomized to wait and see (2) and 82 excluded children (3). Differences between any of the three groups (indicated by the p-value) were analysed by chi-squared test for proportions (followed by Fisher's exact test in the case of significance) and non-parametric tests for continuous variables (Kruskal–Wallis test followed by Mann–Whitney U-test in the case of significance).

Parameters	PcV (1)	no PcV (2)	Excluded (3)	p-value
n	92	87	82
Mean age (years)	5.8 (2–13)	5.9 (2–16)	6.2 (2–16)	0.687
Boys/Girls (n)	45/47	45/42	44/38	0.819
Children in day care (n;%)	68 (77%)	66 (84%)	64 (82%)	0.561
Bilateral otitis (n;%)	10 (11%)	7 (8%)	7 (9%)	0.787
No of days of illness prior to first visit (median (range))	1.0 (0–20)	1.0 (0–30)	2.0 (0–20)	0.022^1
No of days with temp.>38°C prior to first visit (median (range))	0.0 (0–6)	0.0 (0–5)	1.0 (0–7)	0.042^2
Nasal discharge (n;%)	71 (79%)	54 (68%)	67 (86%)	0.02^3
Fever > 38°C at day of first visit (n;%)	18 (20%)	13 (16%)	11 (14%)	0.586
Pain severity 0–1/2–3 (n)	43/48	34/46	27/51	0.246
Presence of cough (n;%)	56 (62%)	45 (56%)	48 (62%)	0.695
Sleep disturbance (n;%)	60 (67%)	63 (79%)	67 (86%)	0.011^4
Appetite (n;%)	51 (57%)	52 (65%)	54 (69%)	0.225
Diarrhoea (n;%)	5 (6%)	3 (4%)	4 (5%)	n.s.
Vomiting (n;%)	6 (7%)	7 (9%)	10 (13%)	0.391
Analgesics (no of doses; median (range))	1.0 (0–4)	2.0 (0–9)	1.0 (0–6)	0.002^5

¹Group 3 differed from 1 (p = 0.015) and 2 (p = 0.035). ²Group 3 differed from 1 (p = 0.043) and 2 (p = 0.032). ³Group 2 differed from 1 (p = 0.034) and 3 (p = 0.006). ⁴Group 1 differed from 3 (p = 0.017). ⁵Group 1 differed from 2 (p = 0.002) and 3 (p = 0.003).

Table II.  Follow-up variables during the first week in terms of pain severity, analgesics used and fever.

Parameters	PcV	no PcV	p-value
Pain severity 2–3 (n;%)
Day 0	48 (53%)	46 (58%)	0.542
Day 1	5 (5%)	22 (28%)	<0.001
Day 2	3 (3%)	8 (10%)	0.116
Days 3–7 (max)	2 (2%)	4 (5%)
Mean no. of days (range) with pain severity 2–3	0.1 (0–6)	0.5 (0–5)	<0.001
Analgesics (Yes; n;%)
Day 0	57 (63%)	61 (76%)	0.095
Day 1	20 (22%)	43 (54%)	<0.001
Day 2	9 (10%)	27 (34%)	<0.001
Day 3	4 (4%)	10 (13%)	0.091
Day 4–7 (max)	3 (3%)	8 (10%)
Median (range) no. of doses day 1–7	0.0 (0–8)	1.0 (0–17)	<0.001
Fever > 38° C (n;%)
Day 0	18 (20%)	13 (16%)	0.557
Day 1	3 (3%)	9 (11%)	0.069
Day 2	1 (1%)	5 (6%)	0.101
Day 3–7 (max)	3 (3%)	5 (6%)
Mean no. of days (range) of fever 1–7	0.7 (0–5)	0.8 (0–8)	0.585

Day 0 is the day of first visit. Days with ≤10% of the patients in every group are not shown separately but as maximum number per day (%) during the remaining days.

Table III.  Follow-up variables in terms of phone contacts, new appointments, risk of long-term complications and economic implications.

Parameters	PcV	no PcV	p-value
Phone contacts day 1–7
Total phone contacts (mean (range))	0.1 (0–1)	0.2 (0–3)	0.603
New urgent appointments
Disease-specific; days 1–7 (n;%)	4 (4%)	13 (15%)	0.021
All; days 1–7 (n;%)	4 (4%)	16 (18%)	0.004
Treatment failure/perforation; days 2–7 (n;%)	0 (0%)	4 (5%)	0.054
Disease specific; 0.5–3 months	9 (10%)	10 (13%)	0.808
Planned follow-up (14 days and 3 months)
Recovered day 14 (n;%)	71 (82%)	70 (85%)	0.541
Perforation at 3 months	None	None	–
Serous otitis media at 3 months	10 (12%)	8 (11%)	1.00
Subjective recovery 3 months	73 (85%)	63 (84%)	1.00
Economic aspects
Parents at home (n;% of all)	49 (56%)	42 (53%)	0.755
Days at home from work (median (range))	1.2 (0–7)	1.2 (0–7)	0.904

Results

A total of 268 patients met the criteria for acute otitis media, but 82 patients were excluded since they refused the randomization procedure. In addition, due to non-compliance with the study protocol, seven patients were excluded: two patients were under-age for the study, one received amoxicillin and one received erythromycin instead of PcV, one had a tympanostomy tube in the non-affected ear, and two patients had by mistake not been registered in the database. Thus, a total of 179 patients (67%) completed the investigation; 92 were randomized to PcV and 87 to no primary antibiotic treatment.

Baseline data for the 179 randomized participants and 82 of the excluded patients are presented in Table I. In the PcV group, 83% had completed their medication as prescribed according to the patient diaries. The cumulative number of recoveries day by day was similar in the two randomized groups (p = 0.606) (Figure 1). The median recovery day was day four in both groups and approximately 80% had recovered on day seven independent of treatment choice. Children randomized to PcV treatment had less pain, from moderate to severe pain (pain score 2–3) to no or some pain (0–1), on average 0.4 days earlier compared with the other children (p < 0.001), but after two days there was no significant difference in the occurrence of pain between the groups (Table II). The use of analgesics during days 1–3 (expressed as number of children given analgesics day by day) was significantly higher among children randomized to no antibiotics than among children randomized to PcV treatment. The median number of doses of analgesics during day 1–7 was significantly higher (p < 0.001) among children receiving no antibiotics compared with those on antibiotic treatment.

Figure 1.  The cumulative number of patients by parents reported as recovered according to parent's diary records.

Number of phone contacts with the primary health centre and number of days a parent stayed home from work to take care of their child were similar when comparing the groups (Table III). The number of new consultations for symptoms related to AOM during the first week, such as perforations at days 0 and 1, ear discomfort, and hearing disturbance, was significantly higher in children randomized to no antibiotics (15%) than among children randomized to PcV (4%) (p < 0.021). There were no significant differences in the occurrence of treatment failures or perforations between the groups. The prevalence of serous otitis media was similar after three months.

Discussion

This was an open randomized trial, which may affect perception of symptoms by parents and medical staff. The lack of placebo or blinding of the study may lead to a more careful registration of symptoms in patients without antibiotic treatment, which could overestimate the occurrence of symptoms and lead to more phone contacts or new visits. The study showed a higher incidence of new visits among those not treated with PcV, but not on median recovery time, of perforations or therapeutic failures.

All participating staff members received careful instructions and education on AOM by study nurses and the project leader. The doctors are all familiar with the aural microscope. The very strict inclusion criteria would ensure that all included patients had a purulent AOM since our primary aim was to evaluate those patients who, according to the new guidelines, could be managed without antibiotic treatment. However, the relatively small number of patients (n = 179) is a limitation of the study and this number is substantially less than was initially intended in order to reach a desirable power (>80%) for most of the clinically important differences. After two years, the allocation of patients asymptotically reached a plateau and the study, therefore, had to be closed. The reasons for this decline may be that the recommendation of antibiotic restriction was spreading in the community and that people therefore adopted a wait-and-see policy. The limited number of patients resulted in relatively low power, especially as concerns small counts (low proportions). That means that it is not possible to separate statistically the almost threefold number of children having moderate or severe pain at day 2 in the non-pc group compared with the pc group (8 and 3 children, respectively, corresponding to 10% and 3%; see Table II) (p = 0.12), or the difference between 0 and 4 children presenting with treatment failure (ear-drum perforation) at days 2–7 (see Table III).

Our results were in concordance with most placebo-controlled studies. Nicole Le Saux et al. [12] enrolled 512 children with moderate severity of AOM, and showed that children who received amoxicillin had less pain and fever during the first two days. Their study could not show any statistical differences in recurrence rates or middle-ear effusion. Mygind et al. [1], in a double-blind, placebo-controlled study of 149 children, found that PcV decreased the number of days with pain but that there was no major difference in outcome or complications between the PcV and the placebo group. The study concluded that the attitude of “masterly inactivity” with regard to the treatment of AOM was justifiable, provided sufficient analgesic treatment was given and the patient could be closely followed up.

In daily practice, all treatment choices are open and placebos are not allowed. We found that the total number of new consultations was higher in the no antibiotics group, 18% as compared with 4% in the PcV group. However, the cumulative number of patients recovering in all treatment groups did not differ (see Figure 1), and it can be argued that the observed difference might be due to the open design. In the open study by Little et al. [13], children with AOM were randomized to either immediate treatment with antibiotics or to wait and see for at least 72 hours before considering collecting the prescribed antibiotic. In the wait-and-see group, 24% of children did collect their antibiotic prescription which can be compared with the 18% of children randomized to no antibiotics in our study who revisited, of whom 5% received antibiotics because of treatment failure. It seems probable that the number of revisits would decline once not prescribing antibiotics is the established strategy and generally accepted both among staff and in the general population

It is a well-known fact that high usage of antibiotics is a major risk factor for development and spread of antimicrobial resistance [14–17]. There are also indications that PcV might have less negative ecological effects than most other antibiotic choices [15], [17–19]. Although antibiotic use by children has decreased by approximately 50% in Sweden during the last decade, resistance rates in pneumococci have not declined [20]. Thus, when evaluating the positive effects of antibiotic treatment of AOM, the risk of increased spread of antimicrobial resistance and increased individual side effects must be taken into account.

The most feared complication of AOM, acute mastoiditis, is rare and the low incidence rate demands long-term surveys on a regional or national basis in order to discover changes in incidence rate due to changes in treatment policy in the treatment of AOM [[22],23].

Conclusions

This open primary care-based study confirms the results found in comparable placebo controlled studies that the benefit of antibiotic treatment of AOM is limited [1–6], [12]. Despite the small number of patients, the study supports an active “wait and see” policy in the treatment of uncomplicated acute otitis media in children aged 2–16 years in primary care.

Acknowledgements

The authors gratefully acknowledge the valuable assistance of participating general practitioners and the other staff members at the participating primary care centres. The study was financially supported by FORSS – Medical Research Council of Southeast Sweden, Department of Research and Development, Kalmar County Council, The Swedish Society of Medicine, and STRAMA – The Swedish Strategic Programme for the Rational Use of Antimicrobial Agents and Surveillance of Resistance.