Dental Erosion: Effect of Diet Drink Consumption on Permanent Dentition

Abstract

Objective:

The aim of this study is to examine the effect of diet drinks on dental erosion among a representative sample of US adults.

Methods:

Adult dietary and dental data were analyzed from the 2003–2004 cycle of the National Health and Nutrition Examination Survey. Erosion was measured with a modified tooth wear index and was analyzed as a dichotomous variable. Cluster analysis was performed, and the cluster number was based on having a separate diet drink cluster and the R² values. Survey procedure and sample weights were used.

Results:

Most of the population (80%) had some form of dental erosive lesions. When compared with the total sample, people with erosion were more likely to be male (52.5%) and older. People with no erosive lesions were younger (42.3%) and non-Hispanic Black (21.2%). Cluster analysis resulted in 4 distinct clusters: high water, high diet drinks, high coffee/tea, and high soda. The respective percentage of individuals in each cluster who had erosion was 78.9%, 85%, 83.9%, and 76.2%, where the “high diet drinks” cluster showed the highest erosion (P = 0.28). Logistic regression modeling showed that the “high diet drinks” cluster had increased odds of erosion (odds ratio = 1.27; 95% CI = 0.58 to 2.77) when compared with the “high water” cluster, but the relationship was not statistically significant.

Conclusion:

High diet drinks consumption slightly increased the odds of dental erosion among US adults, although this relationship was not statistically significant. It is thus not yet clear that dentists should recommend diet drinks, as they might be linked to systemic diseases. Further research is needed to explore more about risk factors of erosion.

Knowledge Transfer Statement:

The findings of this study are suggestive, though not significantly, that diet drinks may increase risk for dental erosion. While further research is needed, it is not clear that dentists should recommend these drinks as healthy substitutes for sugary drinks.

Keywords

beverages tooth wear sweetening agents artificially sweetened beverages carbonated beverages cluster analysis

Introduction

An erosive lesion is a noncarious lesion, defined as the loss of dental hard tissues, which is due to acidic substance exposure (Ganss 2006; Struzycka et al. 2014). These acids might be of intrinsic origin, such as gastroesophageal reflux, or extrinsic origin, such as fruit juices and carbonated beverages (Ganss 2006; Struzycka et al. 2014). If erosion is not treated, it can lead to tooth discoloration and sensitivity (Okunseri et al. 2011).

While there are many researchers studying erosion among children (Alvarez Loureiro et al. 2015; McGuire et al. 2009; Murakami et al. 2011; Salas, Nascimento, Huysmans, et al. 2015; Salas, Nascimento, Vargas-Ferreira, et al. 2015), few studies have researched erosion in adults, and these were mainly conducted in Europe and in low-caries countries. The prevalence of dental erosion among adults showed a wide range among countries, populations, and samples (Bartlett et al. 2013). When erosion prevalence was compared in 7 European countries (Estonia, Latvia, Finland, France, Italy, Spain, and United Kingdom), 57% of the adults aged 18 to 35 y had some form of erosion, with UK participants having the highest prevalence of erosive lesions (Bartlett et al. 2013).

Dietary habits and beverage consumption, especially acidic beverages, are considered important factors in the process of developing dental erosive lesions. Natural fruit juice/drinks and carbonated beverages such as soda are the most studied acidic beverages, and both demonstrate a positive relationship with dental erosion (Salas, Nascimento, Vargas-Ferreira, et al. 2015). A systematic review aimed to assess the influence of diet on dental erosion among children and adolescents concluded that natural fruit juices and soft drinks increase the odds for dental erosion (OR = 1.20, P = 0.03; OR = 1.61, P = 0.001, respectively), while milk showed a protective effect (OR = 0.96, P = 0.008; Salas, Nascimento, Vargas-Ferreira, et al. 2015).

Most of the current studies examined the effect of single beverages on dental erosion. This relationship might be affected by consuming other beverages during the day, resulting in a different outcome than expected. For example, one might consume soda (increase erosion) and milk (protective against erosion) in the same day, and the resulting effect on dental erosion is still unknown.

Another limitation of these studies is that they did not examine the effects of regular and diet drinks separately, although these beverages have different pH levels and in vitro studies showed inconsistent results when comparing the effects of sugar-sweetened and diet drinks on dental erosion (Barac et al. 2015; Ehlen et al. 2008; Gravelle et al. 2015; Lussi and Carvalho 2015; Jameel et al. 2016). When low calorie sweetener (LCS) is added to diet drinks, citric acid is also added to compensate for the decreased sugar content, which might result in different levels of erosion.

The aim of our study is to examine the effect of diet drinks on dental erosion among a representative sample of US adults via cluster analysis.

Methods

This study has complied with the protocol of the STROBE checklist.

Data Source and Sample

Data from the 2003–2004 cycle of the National Health and Nutrition Examination Survey (NHANES) were used. NHANES is a stratified multistage probability survey that is conducted by the National Center for Health Statistics. NHANES includes the civilian noninstitutionalized population residing in the 50 states and the District of Columbia (Centers for Disease Control and Prevention 2011). We used this cycle because it is the only one where erosion measurement is available. The study sample consisted of 2,368 adults aged between 21 and 60 y who completed the dental examination and dietary interview. Individuals aged >60 y were excluded because their beverage consumption was mainly tea, coffee, and water.

Dietary Assessment

Beverage consumption data were obtained from two 24-h recall dietary interviews. Beverages were extracted and grouped manually, and the mean consumption was calculated (g/d). The percentage of each beverage was calculated and used for analysis, instead of absolute amounts, as it is easier to compare and group individuals. Beverages with low percentages were grouped with similar beverages—for example, plain water was grouped whether it was bottled or tap, and drinks with LCS were grouped. The main beverage categories that were used in the cluster analysis were defined as follows:

Milk: Plain nonsweetened milk, buttermilk

Water: Plain nonsweetened, nonflavored tap and bottle water

100% juice: Fresh fruit and vegetable juices with no added sugars

Juice drink: Drinks containing <100% fresh juice; bottled, canned, or powder mix fruit-based drinks

Soda: Regular sugared carbonated soda

Diet drinks: Reduced sugar and diet soda; fruit drinks with reduced sugar or LCS; water, coffee, or tea with LCS; low-calorie sports/energy drinks

Coffee and tea: Nonsweetened coffee and tea

Sports/energy drinks: Regular sugared sports/energy drinks

Sugared drinks: Flavored and sugared milk, sweetened water, sugared and flavored coffee, sugared tea, and nutritional and protein shakes

Other drinks: Soy, almond, and malted milk; coconut and oats beverages

Erosion Measurement

The index used to assess erosion by NHANES was the tooth wear index (TWI) developed by Smith and Knight in 1984 and modified by Millward et al. in 1994. This index was modified by Millward et al. to measure erosion and has been used for measuring and studying erosion (Dye et al. 2008; McGuire et al. 2009). The TWI protocol used in the 2003–2004 NHANES was derived from the 1998 UK Adult Dental Health Survey, which excluded cervical surfaces as they are mainly related to abrasion (Millward et al. 1994; Dye et al. 2008).

NHANES dental examinations were performed by dentists and recorded by a dental recorder. Dentists used a surface-reflecting mirror and an examining light and dried the tooth surface before examination. Only natural permanent tooth surfaces were examined. The TWI was used to measure erosion for the following teeth: central, lateral, canine, and first molar. It also was used for the following surfaces: lingual, facial, and incisal/occlusal. For each surface measured, erosion was classified into 4 stages:

Stage 0: Sound (including any wear restricted to enamel)

Stage 1: Loss of enamel just exposing dentin

Stage 2: Loss of enamel exposing dentin for more than one-third of the surface for buccal/lingual surfaces; on incisal/occlusal surfaces, exposed dentin facets ≥2 mm in dimension buccolingually

Stage 3: Complete loss of enamel and pulp exposure or exposure of secondary dentin

More information regarding the examination and scoring is included in the NHANES dental examiner manual (National Center for Health Statistics 2004).

In this analysis, erosion was dichotomized into no erosion versus at least 1 erosive lesion, as a very small number of the sample had erosion stages 2 and 3.

Covariates

Sociodemographic variables were as follows: gender (male/female), age (≥21 y, defined as 9-y intervals), race/ethnicity, ratio of family income to poverty (categorized per the Medicare eligibility criteria into 4 categories: ≤1.38, >1.38 to ≤2.5, >2.5 to ≤4.0, and >4.0; ObamaCareFacts.com 2014), and education (categorized into high school or less, some college, and college graduate or above). Marital status was used as several researchers suggested a link between marital status and dietary behaviors; specifically, social relationships can influence the variety and types of foods consumed and therefore affect health including oral health (Conklin et al 2014; Vinther et al. 2016). The categorizations are based on partnership and cohabitation. Married and living with partner were combined as 1 group; widowed, divorced, and separated were combined into another group; and singles were the third group. In addition, the percentages of wine and alcoholic beverage intake were used as possible confounders.

Data Analysis

Our research utilized cluster analysis to account for the complexity of beverage consumption and to better understand the effect of different consumption patterns on the dentition. Cluster analysis is a statistical method for grouping data, where objects in the same cluster are more similar to one another than objects in the other clusters. This methodology was previously used to study the effects of beverage consumption patterns on dental caries (Sohn et al. 2006; Samman et al. 2020).

SAS PROC FASTCLUS was used to perform a disjoint cluster analysis, which groups observations according to the Euclidean distance, resulting in every observation being assigned to only 1 cluster (Wirfält and Jeffery 1997; Sohn et al. 2006; Samman et al. 2020).

Cluster analysis was performed by using the percentage of each of the previously defined beverages. Since the literature did not identify the appropriate number of clusters, we evaluated different cluster numbers (from 2 to 10). The final cluster number was based on whether it served the purpose of the study of having a separate diet drink cluster. This number was confirmed by comparing the R² values across different cluster numbers.

Chi-square tests were used to examine the crude association between erosion and 1) beverage consumption clusters and 2) sociodemographic characteristics. Multiple logistic regression was used to predict erosion from the beverage consumption clusters and the sociodemographic variables. Statistical analysis was performed with SAS Software version 9.4 (SAS Institute Inc.). Survey procedures were used to account for the complex NHANES sampling design. In addition, sample weights were used to generate national level estimates of consumption in the US population.

Results

Eighty percent of the sample had some form of dental erosive lesions. When compared with the total sample, people with erosion were more likely to be male (52.5%) and older. People without erosive lesions were younger (42.3%), non-Hispanic Black (21.2%), and never married (27.8%; Table 1).

Table 1.

Study Sample Characteristics.

	Study Sample^a
Variable	All, n (%)		No Erosion (n = 519), %		Erosion (n = 1,849), %		P Value
Participants	2,368		19.7		80.3
Gender							<0.0001^b
Male	1,116 (49.1)		35.4		52.5
Female	1,252 (50.9)		64.6		47.5
Age, y							<0.0001^b
21 to 30	670 (26.1)		42.3		22.1
31 to 40	627 (25.9)		27.5		25.6
41 to 50	598 (26.1)		16.8		28.4
51 to 60	473 (21.9)		13.4		24.0
Race/ethnicity							0.003^b
Mexican American + Hispanic	548 (12.0)		7.9		13.0
Non-Hispanic White	1,204 (71.1)		67.2		72.1
Non-Hispanic Black	511 (12.0)		21.2		9.8
Other	105 (4.9)		3.8		5.2
Education							0.37
High school or less	1,074 (37.8)		33.5		38.8
Some college	750 (34.4)		38.7		33.3
College graduate or above	544 (27.9)		27.8		27.9
Marital status							0.03^b
Married/living with partner	1,531 (65.8)		62.2		66.7
Widowed/divorced/separated	337 (13.2)		10.1		14.0
Never married	499 (21.0)		27.8		19.3
Ratio of family income to poverty							0.71
≤1.38	750 (24.8)		26.7		24.3
>1.38 to ≤2.5	470 (17.2)		17.6		17.0
>2.5 to ≤4.0	434 (19.6)		20.7		19.3
>4.0	714 (38.5)		35.0		39.3
	Beverage Consumption^c
	mg/d	%	mg/d	%	mg/d	%
Beverage consumption							<0.0001^b
Total	2,531.4	100	2,651.8	100	2,743.7	100
Water	1,177.1	43.3	1,218.2	43.0	1,167.0	40.3
Soda	362.9	15.8	387.8	16.6	356.8	14.3
Coffee/tea	367.2	15.0	295.1	11.1	384.9	14.6
Juice drink	167.4	7.5	183.4	8.1	163.4	6.7
Milk	142.2	5.8	126.2	5.1	146.1	5.6
Diet drinks	183.2	7.2	154.8	5.6	190.2	7.2
Alcoholic drinks	175.7	10.4	137.1	4.2	185.1	5.7
Sugared drinks	104.4	4.2	101.6	4.3	105.1	3.9
100% Juice	23.1	1.1	31.1	1.3	21.2	0.9
Wine	18.5	0.9	14.2	0.5	19.5	0.7
Other drinks	3.6	0.1	2.5	0.1	3.9	0.1

Values are presented as n (%) and weighted percentage.

Significant difference at the 0.05 level.

Values are presented as mean and weighted percentage.

Beverage Consumption Clusters

Total beverage consumption among adults comprised mostly water (43.3%), followed by soda (15.8%). Coffee and tea composed 15%, while juice and diet drinks constituted around 7% each. Although wine is considered an important risk factor for erosion, it constituted <1% of the study sample’s total beverage consumption; thus, it was removed from the cluster analysis and controlled for as a separate confounder.

Cluster analysis resulted in 4 distinct clusters, with each cluster named according to the highest percentage beverage. For example, the first cluster was named “high water” because water constituted of 65.2% of the total beverages consumed in that cluster. Similarly, other groups were named as follows: high diet drinks, high coffee/tea, and high soda. The “high water” cluster had 1,153 adults in it; “high coffee/tea,” 594; “high soda,” 460; and “high diet drinks,” 156 (Table 2).

Table 2.

Beverage Profile and Cluster Distributions.

	High Water (n = 1,153)		High Diet Drinks (n = 156)		High Coffee/Tea (n = 594)		High Soda (n = 460)
Beverage	%	g/d	%	g/d	%	g/d	%	g/d
Milk	5.0	135.1	5.4	141.6	7.8	176.3	5.2	108.2
Water	65.2	1,832.5	27.6	733.2	26.5	669.1	20.9	500.4
100% juice	1.2	28.0	0.2	7.3	1.3	24.4	0.8	16.5
Juice drink	6.1	161.2	1.7	47.6	9.9	204.7	9.9	182.4
Soda	8.0	224.4	2.2	48.5	9.4	212.1	52.0	1,101.9
Diet drinks	3.8	104.3	50.3	1,263.6	3.9	93.0	0.8	16.5
Coffee/tea	7.1	193.7	9.7	238.3	35.6	846.2	5.7	129.4
Sugared drinks	3.5	93.2	2.7	59.8	5.5	140.4	4.7	98.6
Other drinks	0.1	3.6	0.1	5.5	0.1	2.0	0.2	4.9
Total	100	2,775.9	100	2,545.9	100	2,368.2	100	2,158.9

Values are presented as weighted percentage and weighted mean. Bold indicates maximum values per cluster.

Sociodemographic Characteristics of Beverage Clusters

The “high diet drinks” cluster had the highest percentage of females (64%), while the “high soda” and “high coffee/tea” clusters had high percentages of males (58.8% and 56.3% respectivley; Table 3). Young adults were mostly in the “high soda” cluster (44.6%), while “high diet drinks” consisted mostly of older adults (35.6%). “High soda” consumers were mostly non-Hispanic Black (19%), less educated, and never married and had a lower ratio of family income to poverty. In contrast, those in the “high diet drinks” cluster were wealthier and college graduates (41.6%) as compared with other beverage patterns.

Table 3.

Sociodemographic Characteristics of Beverage Clusters.

	Beverage, % (SE)^a
Variable	High Water	High Diet Drinks	High Coffee/Tea	High Soda	P Value
Gender					<0.0001^b
Male	43.5 (1.8)	36.0 (3.4)	56.3 (3.1)	58.8 (3.7)
Female	56.5 (1.8)	64.0 (3.4)	43.7 (3.1)	41.2 (3.7)
Age, y					<0.0001^b
21 to 30	30.4 (2.4)	11.0 (3.0)	11.6 (2.2)	44.6 (3.0)
31 to 40	26.2 (2.1)	31.7 (4.6)	24.1 (3.8)	25.4 (2.2)
41 to 50	24.3 (2.0)	21.7 (4.5)	34.0 (2.1)	20.7 (2.4)
51 to 60	19.2 (1.7)	35.6 (7.3)	30.3 (3.6)	9.4 (2.2)
Race/ethnicity					<0.0001^b
Mexican American + Hispanic	13.8 (2.9)	5.8 (1.7)	9.0 (2.7)	14.5 (3.2)
Non-Hispanic White	66.9 (3.7)	79.9 (5.7)	80.7 (4.0)	63.0 (6.5)
Non-Hispanic Black	13.2 (2.1)	5.1 (1.7)	7.7 (1.9)	19.0 (4.3)
Other	6.1 (1.3)	9.2 (5.4)	2.5 (0.6)	3.5 (1.4)
Education					<0.0001^b
High school or less	34.2 (2.6)	25.3 (4.6)	38.2 (2.9)	52.1 (3.2)
Some college	33.9 (2.2)	33.1 (6.6)	33.8 (2.5)	37.1 (2.8)
College graduate or above	32.0 (2.8)	41.6 (6.5)	27.9 (2.8)	10.8 (2.2)
Marital status					0.003^b
Married/living with partner	64.8 (2.2)	70.8 (3.8)	70.8 (3.0)	58.5 (3.8)
Widowed/divorced/separated	12.4 (1.1)	18.3 (3.8)	12.8 (1.5)	13.4 (3.1)
Never married	22.8 (2.7)	10.9 (2.5)	16.4 (3.1)	28.1 (4.4)
Ratio of family income to poverty					<0.0001^b
≤1.38	23.6 (2.0)	20.5 (2.4)	19.4 (2.3)	38.0 (3.1)
>1.38 to ≤2.5	19.0 (1.7)	15.3 (3.9)	12.2 (1.5)	21.0 (3.2)
>2.5 to ≤4.0	18.0 (2.2)	12.7 (3.2)	24.9 (2.4)	18.7 (2.4)
>4.0	39.5 (2.3)	51.4 (3.9)	43.5 (3.4)	22.3 (3.7)
Erosion					0.28
No erosion	21.1 (4.1)	15.0 (5.5)	16.1 (4.3)	23.8 (7.2)
Any form of erosion	78.9 (4.1)	85.0 (5.5)	83.9 (4.2)	76.2 (7.2)

Weighted column percentage.

Significant difference at the 0.05 level.

Erosion and Beverage Clusters

The crude association between erosion and beverage clusters showed no statistically significant differences (P = 0.28), although the “high diet drinks” cluster showed the highest percentage of people with erosion (85%; Table 3). The multivariable logistic regression showed similar results after adjusting for confounders: while the “high diet drinks” cluster increased the odds of erosion to 1.27 when compared with the “high water” cluster, the results were not statistically significant (95% CI = 0.58 to 2.77). The other clusters, “high soda” and “high coffee/tea,” showed virtually no differences versus the “high water” cluster (Table 4).

Table 4.

Adjusted Multiple Logistic Regression Predicting Erosion.

	Point Estimate	95% CI
Beverage clusters
High diet drinks	1.27	0.58 to 2.77
High coffee/tea	0.99	0.54 to 1.81
High soda	0.94	0.55 to 1.60
High water	Reference
Gender
Male	2.19	1.66 to 2.89^a
Female	Reference
Age, y
31 to 40	1.88	1.21 to 2.94^a
41 to 50	3.57	2.35 to 5.43^a
51 to 60	3.88	2.06 to 7.31^a
21 to 30	Reference
Race/ethnicity
Mexican American + Hispanic	1.78	0.83 to 3.79
Non-Hispanic Black	0.41	0.22 to 0.75^a
Others	1.34	0.49 to 3.69
Non-Hispanic White	Reference
Education
Some college	0.79	0.48 to 1.30
College graduate or above	0.84	0.60 to 1.19
High school or less	Reference
Marital status
Widow/divorced	1.34	0.94 to 1.91
Single	1.12	0.62 to 2.04
Married	Reference
Ratio of family income to poverty
>1.38 to ≤2.5	1.01	0.61 to 1.66
>2.5 to ≤4.0	0.86	0.44 to 1.69
>4.0	0.93	0.57 to 1.54
≤1.38	Reference
Wine	6.24	0.20 to 196.12
Alcohol	0.96	0.88 to 1.04

The association between beverage clusters and erosion, adjusted for gender, age, race/ethnicity, education, marital status, ratio of family income to poverty, and wine and alcohol.

Significant difference at the 0.05 level.

Discussion

This study aimed to examine the effect of diet drinks on dental erosion among US adults. Our results revealed that having a “high diet drink” pattern increased the odds of erosion, although this relationship was not statistically significant. This observation is consistent with a previous study examining the association between erosion and dietary factors among adolescents (Sovik et al. 2015). It found that while diet soft drinks increased the odds of developing erosive lesions, they were not statistically associated with erosion (OR = 1.3, 95% CI = 0.6 to 2.7) after adjusting for gender, toothbrushing frequency, sour sweets consumption, and other beverage intake (regular soda, sports drinks, and juice drinks). That study reported that the only beverage significantly affecting erosion was sports drinks (OR = 4.8, 95% CI = 1.3 to 18.4). When these drinks were combined by using the “total intake of erosive drinks” index (based on the combined daily consumption of these drinks in liters), it showed no statistically significant relationship (OR = 2.3, 95% CI = 1.5 to 3.6; Sovik et al. 2015).

In vitro studies that examined the effect of different drinks on erosion in a sample of human teeth had inconsistent results. Some studies concluded that regular sugar-sweetened soda, such as Coca-Cola, was the most erosive agent when compared with other drinks, such as fruit juice, coffee, or tea. Other studies showed the opposite effect and concluded that the low pH of the consumed beverage was the major risk factor and not the sugar contained in it (Ehlen et al. 2008; Khamverdi et al. 2013; Barac et al. 2015; Gravelle et al. 2015; Lussi and Carvalho 2015; Jameel et al. 2016).

Being in a “high diet drinks” group might suggest an attempt to follow a healthy lifestyle and overall diet by avoiding sugary soda drinks. This pattern is seen in Table 2, which shows that harmful sugary drinks, such as soda, juice drinks, and sugared drinks, constitute a very low percentage of that cluster. This suggests that other factors may play a role in the relationship between erosion and diet, such as good oral hygiene practices and frequent dental visits. Another possible explanation for the nonsignificant relationship between diet drink and erosion is the small sample size of the “high diet drinks” cluster, which resulted in a wide confidence interval. In addition, NHANES used the modified TWI as its erosion measurement. To diagnose erosion, the etiology and individual risk factors need to be identified, although this is difficult to accomplish in epidemiologic studies. Because NHANES data measured only the TWI without collecting information on the etiology of the lesion, there may be misclassification, therefore biasing the result toward the null.

Our logistic model showed that being male and middle or older age increased the odds of erosion (not to be interpreted as total effects). These findings agree with some of the previous literature, which found that erosion increases with age among children (Nunn et al. 2003; Ganss 2006; McGuire et al. 2009; Murakami et al. 2011) and adults (Lussi and Schaffner 2000; Bartlett et al. 2013). Males have more erosive lesions, probably due to their different lifestyle and dietary habits, including increased consumption of acidic foods and drinks (McGuire et al. 2009; Mulic et al. 2012; Struzycka et al. 2014; Alvarez Loureiro et al. 2015). One study showed a slight increase in erosive lesion prevalence in females versus males, but this relationship was not statistically significant (OR = 1.65, 95% CI = 0.78 to 3.48, P = 0.58; Kirthiga et al. 2015). Our results also showed an increased odds of erosion among wine consumers, but this relationship had a very wide confidence interval, perhaps due to the very small percentage of individuals who drink wine (<1%). However, regular soda had a neutral effect on erosion. This finding is consistent with a study conducted in Europe, which noted that soft drinks were not highly associated with erosion; nevertheless, it revealed a significant association between erosion and fruit juice and mentioned that this could be due to the low titratable acidity of soft drinks (Bartlett et al. 2013).

There are several strengths in this study. To our knowledge, this is the first study to examine the prevalence of erosion among US adults and its relationship with diet drinks via cluster analysis. Using cluster analysis to analyze the data accounted for the complexity of beverage consumption behavior, as an individual does not exclusively consume single beverages during one’s daily routine. Beverages should be considered in combination and studied as a whole as they may add or counteract one another’s effects on dentition and dental disease. Another strength is analyzing NHANES data, which allowed us to study a large representative sample of the US population. Combining the 2- to 24-h recall dietary assessments is a strength and is considered a reliable measure to use in nationwide data (Lee and Nieman 1996; Ahluwalia et al. 2016).

Our study has some limitations, as it is a cross-sectional study; hence, the reported relationships do not indicate causality. The dietary interview might not capture an individual’s total diet and is subject to recall bias. In an attempt to decrease this bias, interviewers asked about foods consumed the day before the interview, including the portion, the timing, and whether there was any food that the interviewee had forgotten to mention. This dietary interview is a reliable measure for group analysis (Lee and Nieman 1996; Ahluwalia et al. 2016). The data used in this analysis were from the 2003–2004 NHANES, which unfortunately was the only cycle to measure surface-level erosion. Future studies should explore the effects of other possible confounders, such as drinking technique (holding the drink in the mouth), drinking pattern (e.g., with or between meals), type of toothbrush, and type of toothpaste. Another limitation is that NHANES uses the modified TWI as its erosion measurement. As mentioned earlier, the measure could not clearly distinguish erosion from other tooth wear (Ganss and Lussi 2008).

Our study used data collected in 2003 and 2004; hence, the result would be incongruent with the current prevalence of erosion. Since 2004, substantive diet changes have occurred, with more people are becoming vegetarians, vegans, or fruitarians, and these highly acidic diets might contribute to an increased risk of dental erosion. In addition, well-designed longitudinal studies are recommended to confirm the relationship between diet drink consumption and dental erosion. While our results showed no effect of diet drinks on increased erosion, dentists should recommend consuming water and nonacidic beverages instead of diet drinks as a substitute for people who use sugary drinks, as there is a possibility for diet drinks to cause dental erosion. Diet drinks may have additional health consequences since they are also reported to be associated with systemic diseases: dementia, diabetes, vascular disease, metabolic disease, and stroke (Nettleton et al. 2009; Gardener et al. 2012; Suez et al. 2014; Pase et al. 2017). Hence, dentists should continue to educate patients about the potential harmful effects of the excessive consumption of diet drinks as well as sugared beverages.

Conclusion

High consumption of diet drinks seems to increase the odds of dental erosion among US adults, although this relationship was not statistically significant after adjusting for sociodemographic factors. Further research is needed to explore more about risk factors of erosion.

Author Contributions

M. Samman, contributed to conception, design, data analysis, and interpretation, drafted and critically revised the manuscript; E. Kaye, H. Cabral, T. Scott, W. Sohn, contributed to design, data analysis, and interpretation, critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

M. Samman

References

Ahluwalia

Dwyer

Terry

Moshfegh

Johnson

2016. Update on NHANES dietary data: focus on collection, release, analytical considerations, and uses to inform public policy. Adv Nutr. 7(1):121–134.

Alvarez Loureiro

Fabruccini Fager

Alves

Alvarez Vaz

Maltz

2015. Erosive tooth wear among 12-year-old schoolchildren: a population-based cross-sectional study in Montevideo, Uruguay. Caries Res. 49(3):216–225.

Barac

Gasic

Trutic

Sunaric

Popovic

Djekic

Radenkovic

Mitic

2015. Erosive effect of different soft drinks on enamel surface in vitro: application of stylus profilometry. Med Princ Pract. 24(5):451–457.

Bartlett

Lussi

West

Bouchard

Sanz

Bourgeois

2013. Prevalence of tooth wear on buccal and lingual surfaces and possible risk factors in young European adults. J Dent. 41(11):1007–1013.

Centers for Disease Control and Prevention. 2011. National Health and Nutrition Examination Survey: 2011–2012 overview. Hyattsville (MD): US Department of Health and Human Services [accessed 2021 Sep 7]. https://wwwn.cdc.gov/nchs/nhanes/ContinuousNhanes/overview.aspx?BeginYear=2011

Conklin

Forouhi

Surtees

Khaw

Wareham

Monsivais

2014. Social relationships and healthful dietary behaviour: evidence from over-50s in the EPIC cohort, UK. Soc Sci Med. 100(100):167–175.

Dye

Nowjack-Raymer

Barker

Nunn

Steele

Tan

Lewis

Beltran-Aguilar

2008. Overview and quality assurance for the oral health component of the National Health and Nutrition Examination Survey (NHANES), 2003–04.J Public Health Dent. 68(4):218–226.

Ehlen

Marshall

Qian

Wefel

Warren

. 2008. Acidic beverages increase the risk of in vitro tooth erosion. Nutr Res (New York, NY). 28(5):299–303.

Ganss

. 2006. Definition of erosion and links to tooth wear. Monogr Oral Sci. 20:9–16.

10.

Ganss

Lussi

2008. Current erosion indices—flawed or valid? Summary. Clin Oral Investig. 12 Suppl 1:S1–S3.

11.

Gardener

Rundek

Markert

Wright

Elkind

MSV

Sacco

. 2012. Diet soft drink consumption is associated with an increased risk of vascular events in the Northern Manhattan Study. J Gen Intern Med. 27(9):1120–1126.

12.

Gravelle

Hagen Ii

Mayhew

Crumpton

Sanders

Horne

2015. Soft drinks and in vitro dental erosion. Gen Dent. 63(4):33–38.

13.

Jameel

Khan

Abdul Rahim

Bakri

Siddiqui

2016. Analysis of dental erosion induced by different beverages and validity of equipment for identifying early dental erosion, in vitro study. J Pak Med Assoc. 66(7):843–848.

14.

Khamverdi

Vahedi

Abdollahzadeh

Ghambari

. 2013. Effect of a common diet and regular beverage on enamel erosion in various temperatures: an in-vitro study.J Dent (Tehran). 10(5):411–416.

15.

Kirthiga

Poornima

Praveen

Sakeena

Disha

2015. Dental erosion and its associated factors in 11-16-year old school children. J Clin Pediatr Dent. 39(4):336–342.

16.

Lee

Nieman

. 1996. Nutritional assessment. 2nd ed. Cairo (Egypt): FAO. Lussi

Carvalho

. 2015. Analyses of the erosive effect of dietary substances and medications on deciduous teeth. PloS One. 10(12):e0143957.

17.

Lussi

Schaffner

2000. Progression of and risk factors for dental erosion and wedge-shaped defects over a 6-year period. Caries Res. 34(2):182–187.

18.

McGuire

Szabo

Jackson

Bradley

Okunseri

2009. Erosive tooth wear among children in the united states: relationship to race/ethnicity and obesity. Int J Paediatr Dent. 19(2):91–98.

19.

Millward

Shaw

Smith

Rippin

Harrington

1994. The distribution and severity of tooth wear and the relationship between erosion and dietary constituents in a group of children. Int J Paediatr Dent. 4(3):151–157.

20.

Mulic

Skudutyte-Rysstad

Tveit

Skaare

. 2012. Risk indicators for dental erosive wear among 18-yr-old subjects in oslo, norway. Eur J Oral Sci. 120(6):531–538.

21.

Murakami

Oliveira

Sheiham

Nahas

Pires

Correa

Haddad

Bonecker

2011. Risk indicators for erosive tooth wear in Brazilian preschool children. Caries Res. 45(2):121–129.

22.

National Center for Health Statistics. 2004. National Health and Nutrition Examination Survey: dental examiners procedures manual. Hyattsville (MD): US Department of Health and Human Services.

23.

Nettleton

Lutsey

Wang

Lima

Michos

Jacobs

2009. Diet soda intake and risk of incident metabolic syndrome and type 2 diabetes in the Multi-ethnic Study of Atherosclerosis (MESA). Diabetes Care. 32(4):688–694.

24.

Nunn

Gordon

Morris

Pine

Walker

2003. Dental erosion—changing prevalence? A review of British national childrens’ surveys. Int J Paediatr Dent. 13(2):98–105.

25.

ObamaCareFacts.com . 2014. ObamaCare facts: federal poverty level guidelines [accessed 2021 Sep 7]. http://obamacarefacts.com/federal-poverty-level/

26.

Okunseri

Gonzalez

Visotcky

Szabo

2011. Erosive tooth wear and consumption of beverages among children in the United States. Caries Res. 45(2):130–135.

27.

Pase

Himali

Beiser

Aparicio

Satizabal

Vasan

Seshadri

Jacques

. 2017. Sugar- and artificially sweetened beverages and the risks of incident stroke and dementia: a prospective cohort study. Stroke. 48(5):1139–1146.

28.

Salas

Nascimento

Huysmans

Demarco

. 2015. Estimated prevalence of erosive tooth wear in permanent teeth of children and adolescents: an epidemiological systematic review and meta-regression analysis. J Dent. 43(1):42–50.

29.

Salas

Nascimento

Vargas-Ferreira

Tarquinio

Huysmans

Demarco

. 2015. Diet influenced tooth erosion prevalence in children and adolescents: results of a meta-analysis and meta-regression. J Dent. 43(8):865–875.

30.

Samman

Kaye

Cabral

Scott

Sohn

2020. The effect of diet drinks on caries among us children: cluster analysis. J Am Dent Assoc. 151(7):502–509.

31.

Sohn

Burt

Sowers

. 2006. Carbonated soft drinks and dental caries in the primary dentition. J Dent Res. 85(3):262–266.

32.

Sovik

Skudutyte-Rysstad

Tveit

Sandvik

Mulic

2015. Sour sweets and acidic beverage consumption are risk indicators for dental erosion. Caries Res. 49(3):243–250.

33.

Struzycka

Rusyan

Boguslawska-Kapala

2014. Prevalence of dental erosion in young adults aged 18 years in Poland. Przegl Epidemiol. 68(4):689–693.

34.

Suez

Korem

Zeevi

Zilberman-Schapira

Thaiss

Maza

Israeli

Zmora

Gilad

Weinberger

, et al. 2014. Artificial sweeteners induce glucose intolerance by altering the gut microbiota. Nature. 514(7521):181–186.

35.

Vinther

Conklin

Wareham

Monsivais

2016. Marital transitions and associated changes in fruit and vegetable intake: findings from the population-based prospective EPIC-Norfolk cohort, UK. Soc Sci Med. 157:120–126.

36.

Wirfält

Jeffery

. 1997. Using cluster analysis to examine dietary patterns: nutrient intakes, gender, and weight status differ across food pattern clusters. J Am Diet Assoc. 97(3):272–279.