Otitis Media Prone Children with Cystic Fibrosis: A New Normal

Abstract

Objectives:

To determine if children with cystic fibrosis (CF) who are otitis media prone and treated with tympanostomy tube placement (TTP) follow the natural course of non-CF children regarding the incidence of tympanostomy tube otorrhea (TTO) (21-34%).

Methods:

All CF patients seen at a large tertiary pediatric hospital were retrospectively reviewed from 2010-2019. A total of 483 patients were identified and seventeen met the inclusion criteria and were included in the analysis. Data collected included demographics, CF diagnosis history including date of diagnosis and genotype, TTP notes, and otorrhea found in otolaryngology clinic and pediatrician clinic notes for up to 18 months post-TTP.

Results:

CF was diagnosed at a median age of 13 days (0 days to 6 years). In terms of surgical frequency, 14/17 (82.4%) patients had one TTP, 2/17 (11.8%) had two TTPs, and 1/17 (5.9%) had five TTPs. The median (range) age at first TTP was 2 years (3 months to 13 years). After the first TTP, TTO occurred in 5 (29.4%) patients at 3 months, 6 (35.3%) at 6 and 9 months, and 7 (41.2%) at 12 and 18 months at median (range)=1 (0-5)) otolaryngology appointments and median (range)=0 (0–8)) pediatrician appointments.

Conclusion:

To our knowledge this is the first study to report that CF children are more likely to be severely affected with recurrent acute otitis media (RAOM), to require TTP, and to exhibit a natural history of TTO commensurate with the non-CF population.

1. Introduction

Cystic fibrosis (CF) has traditionally been considered protective of the middle ear from otitis media and rarely were such patients “severely affected” to require myringotomy and tympanostomy tube placement (TTP). Historically, retrospective and prospective designed studies have been the backbone of this notion [1,2] with only 3.4% of children presenting with recurrent acute otitis media (RAOM) [3]. Numerous studies have described the low prevalence of middle ear disease in patients with CF, suggesting that the poor mucociliary transport in cystic fibrosis does not necessarily induce otitis media. The prevalence of inflammatory ear disease in children with CF has been reported to be between 3 and 43% [2-7], while it is generally reported that 80% of children will have at least one episode of acute otitis media before age two [8]. In 1979, Bak-Petersen et al. described a cohort of 111 patients with CF and found 35% had a history of acute otitis media [6]. Conversely, in a retrospective review of 450 cases, otologic disease occurred in only 3% of patients with only five patients (1%) receiving ear tubes [2]. A smaller study of 23 patients found no patients had hearing loss [9]. Temporal bone histology and mastoid pneumatization in patients with CF has also been studied. In general, it has been shown that patients with CF have well pneumatized mastoids [10]. Todd et al. studied 20 patients and found that patients with cystic fibrosis had significantly less otitis media and better pneumatized mastoids than the population without CF [10]. In a large study of 186 temporal bone CT scans, Seifert et al. found that temporal bone pneumatization did not differ between patients with CF and patients without CF, but they did find that DeltaF508 homozygosity, the gene mutation which most commonly causes CF, positively correlates with greater pneumatization [11].

Studies comparing patients with CF to age-matched controls have found no greater incidence of middle ear disease. In fact, one study found a lower prevalence of otitis media in patients with CF than controls [1]. In a retrospective review of 450 cases, otologic disease occurred in only 3% of patients with only five patients (1%) receiving ear tubes [2]. Anatomical studies reveal comparable pneumatization and mucosal histology of middle ear between temporal bones of patients with and without CF [10,11]. One theory that explains these findings is that the CFTR protein might play a more limited role in anion transport in middle ear mucosa compared to nasal epithelium [12]. It is also possible that children with more severe CF are physiologically protected from inflammatory processes in the middle ear. It has been hypothesized that the distribution of goblet cells in the respiratory tract may differ between patients with and without CF. Yildirim et al. found a statistically significant reduction in the density of goblet cells in the middle ear mucosa and Eustachian tube of patients with CF compared with control temporal bones, which may serve a protective role against middle ear infections [13].

Curiously, Kreicher, et al. suggest that children with chronic sinusitis might be at lower risk for inflammatory middle ear disease and subsequent hearing loss than children with less severe sinus disease [14].

As the idiom states, "where there is smoke there is fire" and despite a reasonable body of published evidence, both from clinical and basic science studies, suggesting a potential "protective role" of the middle ear in some or most CF children, the senior author of this study has been astonished by the general lack of consensus or even awareness of these data. Yet, in nearly three decades of practice in a tertiary children's hospital featuring one of the largest and most active CF centers in the world, the senior author has very rarely inserted tympanostomy tubes in CF children.

Our group recently reported an increase in middle ear disease in cystic fibrosis (CF) children with a significant number of these patients requiring TTP [15]. Otorrhea is the most common complication of TTP and is considered a more valid and objective surrogate of otitis media [16,17]. To further substantiate our recent finding that otitis media is more common in CF children than once believed, we employed this more valid and objective surrogate of otitis media, otherwise referred to as acute otitis media with a tympanostomy tube (AOMT), to study the prevalence of post-tympanostomy tube otorrhea (TTO) in the CF population.

The natural history of post-operative complications following bilateral myringotomy with tympanostomy tube placement (BMT) has never been reported in cystic fibrosis (CF). The objective of this study was to determine if CF children, classified as “severely otitis media prone” and treated with TTP follow the natural course of non-CF children for the reported incidence of otorrhea (21-34%) [18] post-operatively and 83% incidence in a controlled environment [19].

2. Methods

An IRB protocol (STUDY19080188) was approved by the University of Pittsburgh Human Research Protection Office to retrospectively review cystic fibrosis patient records in the setting of one of the largest tertiary pediatric CF centers in the world. All CF patients that were seen inpatient and/or outpatient at UPMC Children’s Hospital of Pittsburgh from 2010 to 2019 were reviewed with a total of 483 patients. This list of patients was derived from searching ICD 9 and 10 codes (E84.0, 277.00, 277.01, 277.02, 277.03, 277.09). Exclusion criteria are listed in Figure 1. All patients were included if they were diagnosed with cystic fibrosis by the Cystic Fibrosis Center at UPMC Children’s Hospital of Pittsburgh supported by confirmed CFTR genetic mutations. Patients born in 2019 were excluded due to insufficient follow-up. Those that were a carrier for CF, had atypical CF, a CFTR disorder or inconclusive testing, myringotomy without tubes, without myringotomy and tubes, and having TTP at an outside hospital (OSH) or provider for TTP were excluded. Interestingly, four out of the eight atypical CF/CFTR related disorder patients underwent TTP. All four of these patients had an abnormal sweat test but normal CFTR genetic testing or a history of one non disease-causing mutation.

Figure 1.

Inclusion and exclusion with cystic fibrosis and tympanostomy tube placement

The duration from each patient’s first TTP to 18 months postop was assessed. Data collected included demographics, CF diagnosis history including date of diagnosis and genotype, TTP notes, and AOMT found in otolaryngology (ENT) clinic and pediatrician notes for up to 18 months post-TTP. Well-established risk factors for RAOM were also collected, including breastmilk exposure, secondhand smoke exposure, and daycare attendance. Only available and in-network pediatrician appointments were included. An otorrhea episode was considered if found at the ENT or pediatrician appointment or reported at the appointment. The number of TTP for each patient was included as well as the median age at first TTP. The frequency of otorrhea episodes were assessed at 3, 6, 9, 12, and 18 months after each TTP. Findings of middle ear effusion were reviewed at the preoperative ENT appointment and during surgery.

SPSS version 24 was used for statistical analysis with p<.05 considered significant. Statistics were performed with Fisher’s Exact Test and Mann-Whitney U test. GraphPad Prism 8 was used for figure configuration.

3. Results

A total of 21/483 (4.3%) patients underwent TTP. Four patients were excluded from the analysis due to outside hospital or provider performing the TTP. Seventeen patients with CF met the inclusion criteria and were included in the analysis. A majority, 10 (58.8%), were male. As expected from the known racial demographics of CF, 16 (94.1%) patients were white and 1 (5.9%) black. Patients were diagnosed with CF at a median age of 13 days (0 days to 6 years). Genetic testing revealed over half (10/17, 58.8%) of the patients were homozygous for the most common CF genotype, ΔF08. Six of 17 (6/17) patients were heterozygous ΔF08 and the remaining patient had two other genetic mutations on the CFTR gene. Twelve (70.5%) patients had no breastmilk exposure, 7 (41.2%) had secondhand smoke exposure, and 3 (17.3%) were in daycare.

In terms of surgical frequency, 14/17 (82.4%) patients had one TTP, 2/17 (11.8%) had two TTPs, and 1/17 (5.9%) had five TTPs. The median (range) age at first TTP was 2 years (3 months to 13 years). The patient with TTP at 3 months had flat tympanograms bilaterally with hearing loss and four subsequent TTP for recurrent acute otitis media (RAOM) and chronic otitis media with effusion (COME). The median (range) age at 18 months after last TTP was 4 years (2 – 16 years). One patient has since deceased.

Table 1 shows otorrhea found at the preoperative ENT appointment, effusion found at the preoperative ENT appointment, effusion found in surgery, and COME for the reason for TTP surgery for each frequency of TTP. Over half of patients with their first TTP had effusion found at the time of surgery and was found at all subsequent surgeries. COME was the reason for TTP in a third of all surgeries combined. After the first TTP, otorrhea was found in 5 (29.4%) patients at 3 months, 6 (35.3%) at 6 and 9 months, and 7 (41.2%) at 12 and 18 months at median (range) = 1 (0 -5)) ENT appointments and median (range) = 0 (0 – 8)) pediatrician appointments. Table 2 shows the median and range for ENT and pediatrician appointments during the 18-month study period after each TPP, otorrhea episodes at 3, 6, 9, 12, and 18 months post-TTP, total reported otorrhea episodes, and total number of CF patients with otorrhea for each frequency of TTP. Figure 2 displays the frequency of new otorrhea episodes after the first TTP at 3, 6, 9, 12, and 18 months.

Table 1.

Predictors for postoperative otorrhea in pediatric cystic fibrosis patients

	Otorrhea found preop ENT appt n (%)	Effusion found preop ENT appt n (%)	Effusion found in surgery n (%)	COME reason for surgery n (%)
1st TTP, n=17	2 (11.8%)	12 (70.6%)	9 (52.9%)	4 (23.5%)
2nd TTP, n=3	1 (33.3%)	1 (33.3%)	3 (100%)	2 (66.7%)
3rd TTP, n=1	0 (0%)	1 (100%)	1 (100%)	1 (100%)
4th TTP, n=1	0 (0%)	0 (0%)	1 (100%)	0 (0%)
5th TTP, n=1	0 (0%)	0 (0%)	1 (100%)	0 (0%)

Abbreviations: TTP, Tympanostomy tube placement; COME, chronic otitis media with effusion

Table 2.

Postoperative tympanostomy tube placement course over the 18-month study period

	ENT appts 18 post-TTP mdn (range)	Pediatrician appts 18 post-TTP mdn (range)	Otorrhea at 3 months n (%)	Otorrhea at 6 months n (%)	Otorrhea at 9 months n (%)	Otorrhea at 12 months n (%)	Otorrhea at 18 months n (%)	Total reported otorrhea episodes	CF patients with otorrhea n (%)
1st TTP, n=17	1 (0-5)	0 (0-8)	5 (29.4%)	1 (5.9%)	0 (0%)	2 (11.8%)	1 (5.9%)	9	7/17 (41.2%)
2nd TTP, n=3	2 (1-2)	0 (0-3)	1 (33.3%)	1 (33.3%)	0 (0%)	0 (0%)	0 (0%)	2	1/3 (33.3%)
3rd TTP, n=1	3 (3-3)	3 (3-3)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0	0/1 (0%)
4th TTP, n=1	4 (4-4)	3 (3-3)	1 (100%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	1	1/1 (100%)
5th TTP, n=1	3 (3-3)	8 (8-8)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0	0/1 (0%)

Abbreviations: TTP, Tympanostomy tube placement; Mdn, median

Figure 2.

Frequency of new otorrhea episodes after first tympanostomy tube placement at 3, 6, 9, 12, and 18 months (n=17)

There were no differences in age in months between those with and without otorrhea 18 months post-operatively (with median (range) = 48 (21-141) months and without median (range) = 41.5 (30-180) months), p=.962. AOMT was correlated with middle ear effusion found during surgery (p=.036) but not with effusion found at the preoperative ENT appointment or otorrhea found at the preoperative ENT appointment, p=.314 and p=.154, respectively.

4. Discussion

This study serves as another proof of principal supporting the changing face of otolaryngology manifestations of cystic fibrosis children, as they are not protected against middle ear disease as traditionally thought. To the contrary, CF children exhibit an incidence of middle ear disease commensurate with that reported for the general population. Using a much more objective and valid surrogate of otitis media, we measured an incidence of AOMT in CF children that is comparable to the reported incidence in non-CF children. To our knowledge this is the first study to report that CF children may be severely affected with RAOM, to require TTP, and to exhibit a natural history of post-tympanostomy tube otorrhea commensurate with the non-CF population.

Our conclusion that the natural history of post-tympanostomy tube otorrhea is commensurate with the non-CF population is conservative. In reality, our incidence of post-TTO may be lower than an age-matched cohort of children who do not have CF. Because of our retrospective study design lacking a control group, we had to base this comparative statement on the historical reported incidence of post-TTO. The range of the reported incidence is wide. Our statement is a simple result of our reported incidence falling within the broad reported range in the literature (21%-34%) [18]. A meta-analysis of 9 studies looking at post-operative prophylactic antibiotic drops reported an even broader range of 3.4% to 74% otorrhea post-operatively [20], however this range includes studies with randomization of both ears and patients. These 9 studies reported incidence rates from short term follow-ups appointments post-operatively: 1 week [21], 2 weeks [22-25], 1 month [26,27], 12 weeks [28], and initial post-op appointment [29]. Ah-Tye characterized the incidence of otorrhea after TTP in a very rigorous long-term, prospective study of child development in relation to early-life otitis media [19]. During the first 18 months after TTP, the proportion of children who had tubes in place and who developed 1 or more episodes of otorrhea increased progressively, reaching 74.8% after 12 months and 83.0% after 18 months. Ah-Tye’s incidence nearly doubles the rates we measured (41.2% at 12 and 18 months, respectively) and if both were to be confirmed in a prospective controlled study design, this would support the long-held tenant that CF does indeed protect the middle ear against otitis media. Using Ah-Tye’s incidence versus our reported incidence, those without CF were 7 times more likely to have otorrhea post-operatively compared to CF children, OR: 6.7, 95%CI: 2.1-23.3, p=.002.

A critical missing data set from our study is sub-analysis of the microbiology of otorrhea in our patient population. The most likely explanation for the historical contention that CF children are not otitis prone relates to the Eustachian tube [10]. Children who develop post-TTO due to Eustachian tube dysfunction generally develop infection due to the three most common nasopharyngeal pathogens S. pneumoniae, non-typeable H. influenzae and M. catarrhalis [30]. Our group [31-33] and others reported that those who develop infection from external auditory canal pathogens gaining access to the middle ear via the tube lumen are S. aureus and P. aeruginosa [34]. Future studies must include such microbiology data since finding a preponderance of external auditory canal pathogens would favor traditional teaching versus a significant prevalence of nasopharyngeal pathogens which would support a pathophysiologic change. Additionally, the identification of pathogens unique to CF patients such as Burkholderia cepacia is also paramount given that these complex bacteria are resistant to many common antibiotics and able to acquire resistance against many more. Given the current dearth of published AOMT microbiology in the CF population, as a result of this study, the senior author has decided to culture otorrhea from all CF patients not only for treatment guidance using culture directed antimicrobial therapy but also for microbiologic surveillance understanding that the nasopharynx and by extension of the Eustachian tube, the middle ear, are major ecological reservoirs of potential respiratory pathogens.

One interesting finding of note was the increased incidence of severely otitis prone children requiring TTP in the sub-population of atypical CF/CFTR related disorder. Unfortunately, the total number of his genotype was relatively small (8 patients), however, it potentially provides some insight into the overall trend demonstrating that CF does not afford the same protection against middle ear infections as once believed. Future studies examining patients with these atypical CF/CFTR related disorders in terms of their susceptibility to otitis media are crucial in order to either confirm or refute what was suggested in our small sample size and, if confirmed, to better understand the pathophysiology of today’s CF population. By definition, because autosomal recessive inheritance (i.e., cystic fibrosis as an example) presents with the locus on an autosomal chromosome, both inherited alleles must be mutant to express the phenotype. For a person to have CF, there must be a mutation in both copies of the CFTR gene [35]. The ΔF508 mutation being the most common supporting our justification to focus on it scientifically and clinically [36]. Additionally, the implementation of universal newborn screening (NBS) in the US has enhanced our ability to identify these patients earlier [37] and those with less severe phenotypes [35]. With increased detection of mutations, one can justifiably hypothesize more variability in the clinical phenotype and manifestations of the disease which includes middle ear disease.

The limitations of this study largely result from its retrospective design. The retrospective case-series study design obviated the inclusion of a control group. The determination of preoperative middle ear effusion was done by clinicians who were not validated pneumatic otoscopists. Furthermore, documentation of intra-operative middle ear effusion was often done by a member of the surgical team rather than by the attending surgeon of record. Long term post-operative otorrhea reported could have inadvertently been captured by a related upper respiratory infection or water contamination. Phone triage of otorrhea handled by a clinic nurse were not statistically analyzed. These limitations would all be eliminated by a prospective study which is important to conduct not only to confirm these findings but also to investigate the pathophysiology of this ostensible change in the otolaryngology manifestations of CF.

5. Conclusion

Severe middle ear disease in CF pediatric patients follows the same post-operative course as non-CF patients despite conventional teaching that CF protects against otitis media. Otolaryngologists must prioritize middle ear disease along with sino-nasal in the CF population. This study serves as another proof of principal supporting the changing face of otolaryngology manifestations of CF children as they are not protected against middle ear disease as traditionally thought. To the contrary, CF children exhibit an incidence of middle ear disease commensurate with that reported for the general population. Using a much more objective and valid surrogate of otitis media, we measured an incidence of AOMT in CF children that is comparable to the reported incidence in non-CF children. To our knowledge this is the first study to report that CF children are more likely to be severely affected with RAOM, to require TTP, and to exhibit a natural history of post-tympanostomy tube otorrhea commensurate with the non-CF population. Further prospective study not only confirming these findings but also investigating the pathophysiology of this ostensible change in the otolaryngology manifestations of CF is needed. Until then, otolaryngologists must now understand that ENT CF care can no longer primarily focus on sino-nasal disease but must equally include middle ear disease as well. Indeed, the data now suggests that this may be “a new normal!”