We conducted a retrospective cohort study developed at a Portuguese private hospital in Lisbon, Portugal. It included a population of children between 2 and 10 years old, admitted for bilateral myringotomy with tube placement and adenoidectomy, between May 1st, 2011, and May 1st, 2012.

All children were operated on the same hospital, and the eligibility criteria included: (1) age<10; (2) indication for surgery following diagnosis of chronic Otitis media with effusion (OME) (ICD-10 code H65.2, H65.3, H65.4) in accordance with American Academy of Otolaryngology guidelines for OME; (3) Bilateral tube placement and adenoidectomy; Sheppard tube; and (4) follow-up longer than one year.

Patients with a diagnosis other than OME (e.g. recurrent otitis media), who had previous surgery (myringotomy and/or tube placement surgery and/or adenoidectomy), or received a tube other than Sheppard (e.g. Goode) were excluded. Other exclusion criteria included a follow-up period less than a year after surgery (or lost for follow-up), lack of records on personal clinical file or drop out.

The participants included in this study where operated by 9 (nine) different surgeons, who prescribed homogeneous post-operative care, except for ear protection when subjected to water exposure. All of them were allowed water exposure, but 7 (seven) surgeons always prescribed ear canal protection when water exposure was expected (not exposed to the intervention), while the remaining two did not (exposed to the intervention). The prescriptions reflect surgeon's personal preferences at the time of surgery, and this study did not interfere with their choice. This difference allowed the constitution of two cohorts, the exposure being defined as not prescribing ear canal protection from water exposure after surgery.

The level of prescribed care was checked on the medical record and confirmed by the parents. As we also wanted to understand how compliant the participants were regarding the level of care prescribed, they were asked how thorough they were about protecting the ears (plugs, headband), and how deep they were diving (e.g. just surface swimming or diving deep too).

The primary outcome variable was otorrhea episodes. It was used as a surrogate for middle ear inflammation. The question was “how many episodes of ear drainage during the first year” (0, 1, 2–4, >5) (primary outcome variable). Patients did not have a specific log book for these events, but were all instructed to write in their health log book and to inform the surgeon whenever any relevant event happened, which in turn was logged on the patients’ medical record.

The secondary outcome variable was impact on quality of life; we studied the covariable “level of impact on quality of life caused by precautions taken” and “level of impact on quality of life by prescribed precautions (if different)”.

The study was authorized by the Hospital Ethics Committee (doc. CEHCD.ORL1-26092012), and the parents gave consent following a telephonic interview. A trained investigator conducted the interviews and questionnaires. Data was obtained from medical records and telephonic questionnaires, with multiple choice questions. Parents completing the questionnaire were asked to check the child's health log book for clinical events (in Portugal every child has a health log book, kept by the parents, where both parents and doctors record relevant clinical events).

Each question addressed one covariable of the study. The covariables were: “age at time of surgery”, “gender” and “upper airway tract infection episode”.

The covariable “upper airway tract infection episode” addressed the coexistence of upper airway infection symptoms with acute otorrhea episode. These included blocked nose, rhinorrhoea, sore throat, or any other symptom in this area suggestive of at least one of the following: rhinitis, sinusitis, adenoiditis, tonsillitis, pharyngitis, laryngitis, epiglottitis or tracheitis. With this question, we want to query a possible association of acute otorrhea episode with upper airway infection.

To assess otorrhea episodes following exposure to water, we studied the covariable “episodes related to water passing into the ear”.

To characterize the otorrhea episodes as to their duration and severity, we asked respectively “how long did it last most of the times” (<1d, 1d, 2–3d, 4–7d, 7d); and “treatment” (e.g. oral antibiotics, drops, paracetamol/acetaminophen, other anti-inflammatory treatment, other treatment).

Statistical treatment was carried out with STATA 13 software. Descriptive analysis included absolute and relative frequencies, Chi-square (x2) to compare proportions (categorical variables), and t-Student to compare averages. Inferential analysis used logistic regression, with non-adjusted and adjusted odds ratio. The level of significance was p<0.05.

After enrolling 193 children, only 143 met the inclusion criteria and were not excluded (Fig. 1). All 143 children included had a follow-up appointment at least 12 months after the surgery. For 116 patients, surgeons prescribed ear plugs. We included these 116 patients in Group B designated ‘Not exposed’. The remaining 27 patients were advised not to wear ear plugs and were included in Group A, designated ‘Exposed’ to the intervention.

Figure 1.

Inclusion and exclusion criteria.

(0.26MB).

The average age at the time of surgery was 3.8 years in the group without protection (A) and 4.2 in the group with protection (B). This was considered not to have significant statistical difference (p=0.07) between the groups. The gender in the exposed and non-exposed group also did not differ significantly (p=0.5). The two cohorts were considered homogeneous regarding “age at time of surgery”, “gender” and “upper airway tract infection episode”, as shown in Table 1.

In Group A (exposed/without protection), 16 patients had no episodes of otorrhea; the rest of the group had at least one episode in the study period. In Group B (not exposed/with protection), 74 patients did not have any episodes of otorrhea, with the remaining patients having had at least one episode. Almost all episodes lasted less than one day. Treatment comprised oral and topical antibiotics in 8 (72.7%) affected subjects on Group A, and in 35 (83.3%) patients in Group B.

Interestingly, when we split patients in Group A into two additional categories, comparing the ones who of were just swimming but not diving deep, with the ones who swam and dived deep, we found that only 2 out of a total of 11 not diving participants (18.2%) of the first group had at least one episode, versus 9 out of 16 (56.3%) of the diving ones. This has no statistical significance because of the small number of participants, but suggests a difference when not diving deep.

The odds ratio of having at least one episode of otorrhea in Group A patients (‘Exposed’/Not using ear protection) was 1.21, with a 95% confidence interval of 0.51–2.85 (Table 2). This indicates that not wearing protection poses a slightly higher risk, but the difference between both groups was deemed not statistically significant, with p=0.6.

Otorrhea episodes were related to upper respiratory tract infection on 4 (36.4%) subjects in Group A, and 21 (48.8%) in Group B (totalling 46.3% in both groups).

When adjusted (Table 3), the odds of having otorrhea in Group A were 1.59 (with 95% CI 0.50–5.04) and remained non-significant. Interestingly, age and male gender had an odds ratio bellow 1, and so, although not significant, it suggests that being older and male is a protective factor.

On the other hand, having an upper airway infection (UAI) was associated with an odds ratio of having otorrhea of 136.1 (with a 95% CI 16.23–1141.6 and p<0.001). This odds ratio is not surprising considering that otorrhea is surrogate of UAI.

As to the secondary outcome, the level of impact on quality of life, on the exposed Group A, 16 patients felt no impact and 11 felt a negative impact (1 negligible impact, 2 slight impact, 1 felt moderate, 2 felt significant, and 5 severe). In Group B (not exposed/with protection) 60 patients felt no impact and 56 felt a negative impact on their quality of life (7 negligible impact, 8 slight impact, 18 felt moderate, 9 felt significant, and 14 severe). The odds ratio of feeling any negative impact on quality of life when exposed to not having been prescribed water protection was 0.74 (with a 95% CI 0.31–1.72) (Table 2), indicating that not wearing protection would cause less impact, though not reaching significance, as the difference between both groups showed p=0.5 significance, even after adjustment for age and gender (Table 3).