The Role of Diet in Tympanostomy Tube Otorrhea

Abstract

Objective

The objective of this study was to evaluate the role of diet quality in children with tympanostomy tube placement (TTP) complicated by tympanostomy tube otorrhea (TTO).

Study Design

Three‐day 24‐hour diet recall.

Setting

Tertiary care medical center.

Methods

Children between the ages of 2 to 6 years old with TTP performed 6 months to 2 years prior to enrollment were included. Children with a history of Down syndrome, cleft palate, craniofacial syndromes, known immunodeficiency, g‐tube dependent, or a non‐English speaking family were excluded. The primary outcome variable was TTO. The primary predictor was total caloric intake measured by percent estimated energy rate (%EER).

Results

A total of 120 families completed the 3‐day diet recall. The median age was 27 months (interquartile range: 7.9‐68.5), with 57% male sex. Most children reported dietary intake within the recommended range percent intake for carbohydrates and fat and less than recommended range for percent vitamin D. Within this cohort 63 (52.5%) participants had >1 TTO episode and 57 (47.5%) $\le$1 TTO episode. Children with an EER% that was average or high were at higher odds of >1 TTO episodes compared to participants with a low EER% with ORs of 4.6 (95% confidence interval [CI]: 1.4, 15.6) and 5.7 (95% CI: 1.5, 22.1) respectively.

Conclusion

Children with a typical or high total daily caloric intake are approximately 5 to 6 times more likely to have multiple TTO episodes compared to those with low intake.

The most frequently performed ambulatory surgical procedure in the United States is tympanostomy tube placement (TTP). This procedure is indicated in children with either frequent ear infections (3 episodes in 6 months or 4 in 12 months) or chronic ear fluid present for >3 months. Despite TTP, some children continue to have episodes of ear infections which present as ear drainage, known as tympanostomy tube otorrhea (TTO). Children with TTO are initially treated with ear drops either as antibiotics only or antibiotics and steroids. However, some cases are persistent and necessitate oral antibiotics, and in some cases removal and replacement of ear tubes. Children with persistent drainage can cause undue stress and burden to children and their families.

Prior work has identified a preliminary association between diet and TTO, specifically an increase in TTO when children consume higher levels of dessert. The role of diet has been better evaluated in children with otitis media (OM). ¹ , ² , ³ Choi et al identified the correlation between high‐fat diet and OM in a population‐based cross‐sectional survey in Korea. ⁴ Peracchio et al performed a cross‐sectional study identifying an association between frequent OM episodes and increased BMI in preschoolers (3‐5 years old) from 2 urban centers in Connecticut. ⁵ The study also suggested a relationship between frequent episodes of OM and a higher preference for foods with a higher content of fat and/or sugar. ⁵

Though these studies alluded to a relationship between TTO and OM, methods for obtaining associated diet data were not rigorous. This study sought to evaluate the association between total caloric intake and TTO using the gold standard method of a 3‐day 24‐hour diet recall by a research nutritionist. The hypothesis for this study was that children with multiple episodes of TTO had worse diet quality than those with at most 1 episode.

Methods

Study Design and Sampling

We conducted a 3‐day 24‐hour diet recall in children between the age of 2 to 6 years old with TTP performed at the Children's Hospital of Philadelphia 6 months to 2 years prior to enrollment. Those with craniofacial syndromes, Down syndrome, cleft palate, known immunodeficiency, G‐tube dependent, or from a non‐English speaking family were excluded. The study was approved by the Children's Hospital of Philadelphia Institutional Review Board (IRB), IRB# 20‐018387.

Survey Design and Protocol

Prior to performing the diet recall, informed consent was obtained, and the family was counseled not to change their normal food habits for the study. A research nutritionist, who is part of the Children's Hospital of Philadelphia Bio‐nutrition Unit, collected the diet intake data via phone. Both the caregiver (parent or legal guardian) and child were able to partake in the diet recall, which occurred over the course of 2 weeks on 3 separate days, 2 weekdays, and 1 weekend day. Additionally, a medical chart review was performed on all participants and caregivers filled out a questionnaire assessing otorrhea episodes and potential OM risk factors.

Diet Record Analysis

Estimated energy requirement (EER) was calculated for each participant and the percent EER consumed for total calories for each day was calculated. Percent dietary reference intake for micronutrients for each day and individual diet records if estimated energy requirement percentage (%EER) was <50% or >150% to screen for plausibility/typicality of the parent records was examined.

Z Scores were calculated for participants (weight for age and length or height for age) using the Centers for Disease Control and Prevention reference database and examined diet records if the z score did not seem to corroborate the %EER consumed.

Measures

The primary outcome variable was TTO within 6 to 24 months of TTP which we dichotomized to either “$\le 1$” or “>1”. This question was both asked on the survey and extracted via chart review to allow for validation of survey data. In cases where this question was answered, “do not recall,” data from the medical chart review were used, in all other cases survey data were used.

The primary predictor variable was total daily caloric intake (kcal/d). The EER% was the indicator of diet quality in this study. The EER was calculated using sex, age, physical activity level (PAL), weight (kg), and height (cm). For this analysis, the PAL was estimated to be 1.2, which is within the sedentary category, for all participants. Within this cohort, the average caloric intake from the 3‐day, 24‐hour diet recall was utilized. The reported average caloric intake was divided by the calculated EER to obtain the EER%. The cohort was divided into 3 categories:$\mathrm{EER}\%\le 80,$ lower caloric intake; $80<\mathrm{EER}\%<120$, typical caloric intake; and $\mathrm{EER}\%\ge 120$, higher caloric intake.

Covariates included prior known reported risk factors of OM. ¹ , ² , ³ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ Age at the time of TTP, was a continuous variable. Sex was defined as either male or female. A total number of siblings in the household was recorded$.$ Race was categorized as American Indian/Alaska Native, Asian, black/African American, white, other, and not reported. Ethnicity was categorized as Hispanic or non‐Hispanic. Socioeconomic status was indicated by type of insurance (public/private) and household education and employment. For participants with 2 caregivers, education and employment were ascertained for each caregiver separately and then reported as the highest level of education/employment reported between the 2 caregivers. Additional survey clinical and environmental factors are as follows: breastfeeding 6 months, pacifier use and/or bottle use within the past year, immunizations, family history of ear infections and TTP, daycare/school attendance, household smoke, and vaping exposure. Dietary data included total caloric intake, average percent intake over the 3 days recall for fat, carbohydrate, added sugar, and vitamin D (Table 1).

Table 1.

Cohort Characteristics

	Total	>1 TTO episode
	N = 120	No N = 57	Yes N = 63	P value
Age, months at time of TTI [mean (SD)]	27 (13)	29 (14)	24 (11)	.041
Sex				.93
Male	69 (57%)	33 (58%)	36 (57%)
Female	51 (42%)	24 (42%)	27 (43%)
Race				1
Asian	2 (2%)	1 (2%)	1 (2%)
Black or African American	17 (14%)	8 (14%)	9 (14%)
White	95 (79%)	45 (79%)	50 (79%)
Other	3 (2%)	1 (2%)	2 (3%)
Ethnicity				.212
Hispanic	6 (5%)	1 (2%)	5 (8%)
Non‐Hispanic	113 (94%)	55 (96%)	58 (92%)
Insurance type				.79
Public	33 (28%)	15 (26%)	18 (29%)
Private	85 (71%)	41 (72%)	44 (70%)
Highest combined employment				.747
Full‐time	111 (92%)	54 (95%)	57 (90%)
Part‐time	2 (2%)	0 (0%)	2 (3%)
Unemployed/disabled/retired	2 (2%)	1 (2%)	1 (2%)
Highest combined education				.19
Some colleges of less	18 (15%)	8 (14%)	10 (16%)
College degree	42 (35%)	25 (44%)	17 (27%)
Graduate degree	58 (48%)	24 (42%)	34 (54%)
Breastfeed > 6 months				.87
No	45 (38%)	22 (39%)	23 (37%)
Yes	74 (62%)	35 (61%)	39 (62%)
Smoke exposure				.279
No	111 (92%)	54 (95%)	57 (90%)
Yes	8 (7%)	2 (4%)	6 (10%)
Daycare/school				.51
No	15 (12%)	6 (11%)	9 (14%)
Yes	104 (87%)	51 (89%)	53 (84%)
Pacifier				.72
No	88 (73%)	43 (75%)	45 (71%)
Yes	31 (26%)	14 (25%)	17 (27%)
Bottle				.47
No	97 (81%)	48 (84%)	49 (78%)
Yes	22 (18%)	9 (16%)	13 (21%)
Siblings at home (mean, SD)	1 (1)	1 (1)	1 (1)	.17
Average % carbohydrates				1
<Recommended	9 (8)	4 (7)	5 (8)
Recommended	105 (81)	50 (88)	55 (87)
>Recommended	6 (5)	3 (5)	3 (5)
Average % vitamin D level				.78
<Recommended	70 (58)	34 (60)	36 (57)
Recommended	50 (42)	23 (40)	27 (43)
Average % fat				.41
<Recommended	4 (3)	3 (5)	1 (2)
Recommended	85 (71)	38 (67)	47 (75)
>Recommended	31 (26)	16 (28)	15 (24)
Average % added sugar (mean, SD)	32 (17)	31 (20)	33 (15)	.54
Percent EER				.01
Low	19 (16)	15 (26)	4 (6)
Average	72 (60)	31 (54)	41 (65)
High	29 (24)	11 (19)	18 (29)
Immunizations				1
No	1 (1%)	0 (0%)	1 (2%)
Yes	118 (98%)	56 (98%)	62 (98%)
Premature birth				.912
<28 wk	3 (2%)	1 (2%)	2 (3%)
28‐37 wk	17 (14%)	7 (12%)	10 (16%)
>37 wk	64 (53%)	29 (51%)	35 (56%)
Missing	36 (30%)	20 (35%)	16 (25%)
FMHx OM				.26
No	50 (42%)	27 (47%)	23 (37%)
Yes	69 (57%)	30 (53%)	39 (62%)
Missing	1 (1%)	0 (0%)	1 (2%)
FMHx TTI				.12
No	75 (62%)	40 (70%)	35 (56%)
Yes	44 (37%)	17 (30%)	27 (43%)
Missing	1 (1%)	0 (0%)	1 (2%)

Abbreviations: EER, estimated energy rate; FMHx, family history; OM, otitis media; TTI, time to treatment initiation; TTO, tympanostomy tube otorrhea.

Statistical Analysis

The sample size was calculated using an effect size of 25%, β = .8, and α = .05, which estimated that 60 subjects would be required in each group ($\le 1$TTO vs >1 TTO).

All data were analyzed in STATA 16™. A κ statistic was calculated to address the level of agreement between chart review and survey responses addressing the presence of otorrhea. Pearson χ ² and Fischer's exact tests were used to compare both demographics and predictors between the 2 groups. Frequency histograms were created to illustrate the percent intake of carbohydrates, fats, added sugar, and vitamin D. To test the hypothesis that children with multiple TTO episodes have poorer diet quality than those with single no TTO episodes, we utilized a multivariable logistic regression. This model adjusted for confounders which were significant on the univariate analysis: EER% and age at the time of tube placement. This regression was repeated to include all sociodemographic risk factors: age, sex, race, ethnicity, insurance, combined highest education, and combined highest employment.

Results

A total of 137 families completed the survey, of which 120 completed the 3‐day diet recall. Sociodemographic characteristics were similar between participants who did and did not complete the diet recall, except for age at the time of tube placement. The 17 children who did not complete the diet recall were older in age (median age 34 months [interquartile range, IQR: 9.1‐68.6]) compared to those who did (median age 27 months [IQR: 7.9‐68.5]).

The 120 participants who completed the recall were mostly male (57%) with 42% female sex. Seventy‐nine percent were white, 14% were African American or black, 2% were Asian, and 2% reported “other race.” Most (94%) were of non‐Hispanic ethnicity. Among participants, 71% had private insurance, 92% reported having at least 1 household member with full‐time employment, and 48% reported at least 1 household member having a college degree. Most children in this cohort had no smoke exposure, attended daycare/school, did not use pacifiers or drink from bottles in the past year, and were up to date on immunizations (Table 1). In this study, most children reported dietary intake within the recommended percent intake for carbohydrates and fat and less than recommended for percent vitamin D. The mean average 3‐day percent intake for added sugar was 32% (SD 17%) (Figure 1A‐D).

Figure 1.

Three day 24‐hour diet recall nutritional markers. (A) Three‐day diet average % carbohydrate intake. The recommended average % carbohydrate intake is between 45% and 65%. (B) Three‐day diet average % total fat intake. The recommended average % fat intake is between 20% and 35%. (C) Three‐day diet average % vitamin D reference intake. The recommended % dietary reference intake is >100%. (D) Three‐day diet average % energy from added sugar intake.

Within this cohort 63 (52.5%), participants had >1 TTO episode and 57 (47.5%) $\le$1 TTO episode. The 2 groups differed in age at the time of TTP and %EER. The >1 TTO group were younger with a mean age of 24 months (SD 11 months) compared to a mean age of 29 months (SD 14 months) in children with at most 1 TTO episode. Children who had $\le$1 TTO episode also had children with low EER% compared to those with >1 episode. There were no other sociodemographic differences between the groups. Family history of ear infections or tube placement and prematurity did not significantly impact rates of TTO (Table 1).

The multivariate regression model compared children with and without >1 TTO episode, with the primary predictor being EER%. In this model age at the time of TTI was adjusted for, as this was the only significant factor in the univariate analysis. Children with an EER% that was average or high were at higher odds of >1 TTO episode compared to participants with a low EER% with ORs of 4.6 (95% confidence interval [CI]: 1.4, 15.6) and 5.7 (95% CI: 1.5, 22.1), respectively. The null hypothesis was rejected, the results reveal that children with multiple TTO episodes have higher caloric intake compared to those with at most 1 TTO episode. An additional model adjusting for sociodemographic factors (sex, race, ethnicity, insurance, combined education and employment) did not impact the above findings (Table 2).

Table 2.

Adjusted Odds Ratio for the Impact of Percent EER on >1 TTO Episode

>1 TTO	Odds Ratio	Standard error	P value	Lower 95% CI	Upper 95% CI
Percent EER (reference low)
Average	4.65	2.87	.013	1.39	15.60
High	5.74	3.95	.011	1.49	22.14
Age at the time of TTI	0.72	0.13	.078	0.50	1.04

Abbreviations: CI, confidence interval; EER, estimated energy rate; TTI, time to treatment initiation; TTO, tympanostomy tube otorrhea.

Discussion

Children with >1 TTO episode accounted for 52.5% of this cohort. The present study identifies that children who have ingested the average or higher than average amount of calories as measured by EER% have approximately a 5 to 6 times higher risk of multiple TTO episodes compared to low caloric intake. When looking at specific components of diet (carbohydrates, fats, added sugar, vitamin D), there was no difference between children with and without multiple TTO episodes.

Only 1 prior study has evaluated the role of diet in children with TTO, using a short food questionnaire in almost 300 children. The study evaluated the role of fresh fruits, fresh vegetables, fruit juice and drinks, sweet tea, diet and regular soda, and dessert intake in TTO. They identified that children with a higher dessert intake (a proxy for sugar and fat) were at higher risk of TTO, compared to those with a lower intake. ¹¹ These findings differ from the present study, as the prior study evaluates children with and without TTO, while this study only evaluated the presence and absence of multiple TTO episodes.

More studies have been performed evaluating the relationship between diet and OM, using multiple methods. Peracchio et al performed a cross‐sectional survey on dietary behaviors, food preferences and OM history in US preschoolers (3‐5 years old) from low‐income families. They reported that children with higher OM rates had a lower preference for vegetables and fruit and with higher preference for foods high in fats and sugar compared to those with lower OM rates. However, they only assessed food preferences and not intake. ⁵ Another study evaluated the effect of diet on nasopharyngeal pneumococcal carriage in Finnish children in daycare. Each family filled out a food frequency questionnaire and the child underwent a nasopharyngeal swab to test for bacteria. They concluded that children with a diet higher in sugar had an increased rate of pneumococcal carriage and those with a higher fruit and berry intake had a lower rate of acute otitis. ¹² Moreover, Choi et al used a 1‐day 24‐hour diet recall method within the Korean National Health and Nutritional Exam Survey database to evaluate the association between OM and a higher fat diet in children 4 to 13 years old. They found that a 10% increase in fat intake in healthy‐weight children was significantly associated with a 1.4 higher odds of OM. ⁴

All children in this study had a history of OM that required TTP. Though the present study did not identify any clinically meaningful difference between fat, carbohydrates, or added sugar, children with a lower caloric intake were noted to have less risk of multiple TTO episodes. Additionally, the diet information in this study was robust and utilized three 24‐hour diet recalls to better characterize the child's total intake. This is a preferred method to evaluate diet, as it is the more comprehensive and thorough method for obtaining dietary data, especially when administered over 3 days. ¹³ The EER% was the quality of diet measure, as this compares the actual calories reported by the respondent over the calculated required calories for that respondent taking into account the participant's age, sex, and PAL. Several factors can lead to higher caloric intake, including knowledge regarding nutrition, food availability, food choices, portion size, resources to obtain nutritious food, and food security.

Furthermore, vitamin D deficiency in children with OM has been associated with increased OM risk. Kim and colleagues evaluated the influence of vitamin D deficiency on OM progression in rats. Rats with vitamin D deficiency had a significant increase in pro‐inflammatory and decreased anti‐inflammatory cytokines compared to those within the vitamin reference range, which may exacerbate OM. ¹⁴ A case‐control study reported vitamin D deficiency in children with OM and a significant improvement in disease frequency among children who were treated with vitamin D supplementation. ¹⁵ In this study, most children had a lower than recommended vitamin D intake; however, no difference was identified between those with and without >1 TTO episode.

The shortcomings of this study include nonresponder/recall bias and small sample size. We compared sociodemographic characteristics of those who did and did not complete the 3‐day diet recall. Aside from the children of nonresponders caregivers being of older age, there were no other significant differences between groups. To address the impact of recall bias, we conducted a validation study to compare survey responses of TTO episodes to that obtained by chart review, finding 80% agreement, consistent with substantial agreement (κ statistic 0.61). Our sample size calculation was based on identifying a difference of at least 25% between groups. Though we did identify an association between total daily caloric intake and TTO episodes, a larger sample may help identify differences in specific diet intake variables. Nonetheless, this is the only study in children with TTP using high‐quality diet intake methods, and therefore allowed for the exploration of associations between diet and TTO. ¹⁶

Conclusion

This study identifies that children with an average or high total daily caloric intake are approximately 5 to 6 times more likely to have multiple episodes of TTO compared to those with low intake. For a majority of children in this cohort, the average percent 3‐day dietary intake was within recommended levels for carbohydrates and fat, and below for vitamin D. This study adds to the growing body of evidence illustrating the role of diet in children with OM, suggesting the utility of addressing diet intake in children with OM, specifically those with multiple episodes of TTO.

Author Contributions

Kavita Dedhia, study conceptualization and design, data acquisition, data analyses and interpretation, drafting the initial manuscript, review and revision of the manuscript, and approval of the final version to be published; Alyssa Tindall, study conceptualization and design, data analyses and interpretation, review and revision of the manuscript, and approval of the final version to be published; Jill Karpink, data acquisition, interpretation of data, critically revising for important intellectual content, and final approval of the version to be published; Ashley Williams, data acquisition, interpretation of data, critically revising for important intellectual content, and final approval of the version to be published; Terri Giordano, data acquisition, interpretation of data, critically revising for important intellectual content, and final approval of the version to be published; Virginia Stallings, study conceptualization and design, data analyses and interpretation, review and revision of the manuscript, and approval of the final version to be published.

Disclosures

Competing interests

The authors have no conflicts of interest to disclose.

Funding

This work was supported by the Department of Otolaryngology Pilot Grant (K.D.) and the National Institutes of Deafness and Other Communications Disorders (1K23DC021217‐01A1 [K.D.]).