Association of Adult Atopic Dermatitis with Impaired Oral Health and Oral Dysbiosis: A Case-Control Study

Abstract

Background

Systemic alterations in the oral cavity can be reflected in skin disorders like psoriasis. However, data about oral health factors that are affected and controlled mainly by oral microbiota in atopic dermatitis (AD) are sparse. This study compared the oral status and oral microbiota of AD patients and healthy controls.

Methods

This was a prospective sex- and age-matched case-control study comparing adult participants with and without dermatologist-verified AD. A dentist assessed oral health status, and oral flora samples were collected and subjected to 16S rRNA sequencing for microbiome analysis.

Results

Forty-five AD participants and 41 non-AD controls were recruited. Compared to the participants in the control group, those with AD had significantly higher plaque levels (P = .04), poorer oral hygiene indices (P = .04), and higher gingival index trends (P = .05). The oral microbiome in the AD group showed significantly higher diversity, both in α and β diversities (P = .001, P = .0007, respectively). Furthermore, AD patients had a significantly increased abundance of taxa correlated with oral diseases and a decreased abundance of bacteria associated with a healthy oral status.

Conclusion

AD appears to be associated with poor oral health and oral dysbiosis. There is a need to increase both patients’ and physicians’ awareness of oral health.

Introduction

Atopic dermatitis (AD) is a chronic multifactorial inflammatory skin disorder with an incidence of 10% to 20% in children and 1% to 3% in adults. Dysbiosis of the skin is considered a hallmark of AD.¹ However, the causal relationship between dysbiosis and AD remains unclear.¹ Furthermore, much less is known about microbiota in different habitats of AD patients, such as the gut or the oral cavity.

Gingivitis and periodontitis are common chronic oral diseases affecting the tissues supporting the teeth – the gums, the periodontal ligament, and the alveolar bone.^2,3 The aetiology of periodontal diseases is complex and includes dysbiosis of the oral microbiome towards pathogenic microbes, as well as the host's imbalanced immune response, which can harm supporting tissues.^2,4 The symptoms of periodontal disease are swollen and bleeding gums, gum retraction, and bone absorption. Dental caries is a polymicrobial and multifactorial oral disease that can be triggered by dysbiosis. It causes tooth decay in both the enamel and dentin layers. Without improved oral hygiene or the treatment of carious lesions, dental caries will progress, resulting in extensive tooth destruction, loss of tooth vitality, and potentially even complete tooth loss.⁵

Over the years, evidence linking oral diseases to systemic diseases, such as atherosclerotic cardiovascular disease, diabetes, and adverse pregnancy outcomes, has accumulated.⁶ In dermatology, poor oral status, and periodontitis are risk factors for psoriasis.⁷ However, data concerning oral health in AD is currently limited and conflicting. Children with AD were found to have an increased frequency of malocclusion (incorrect relation between the teeth)⁸ and an increased risk of developing tooth decay by the age of three.⁹ Studies using data from a population-based national survey conducted in Korea found an association between oral symptoms (eg, sensitive teeth, toothache, bleeding gums, and bad breath) and current AD in adolescents^10,11 and an increased rate of caries in adults.¹² By contrast, data from a 46-year follow-up study conducted in Northern Finland showed periodontal disease to be associated with seborrheic dermatitis and nummular eczema but not with AD.¹³ The oral microbiome was not assessed in these studies. Recently, Li et al¹⁴ studied the structure and function of microbial communities within the skin, oral cavity, and gut of AD patients and healthy controls. They found that AD patients’ skin and oral cavities showed a reduction in microbial diversity, which correlated with their AD severity.¹⁴ Nonetheless, to our knowledge, no previous work has examined the oral microbiome of AD patients concerning their concordant clinical dental status. Since oral microbiome dysbiosis is a crucial catalyst of several oral diseases such as periodontal disease, caries, and halitosis, further investigation of oral microbiome in clinically and dentally defined AD patients was needed.

In the present study, we sought to examine the clinical status and microbiome complex of the oral cavities of adults with AD.

Materials and methods

Design

This was an age- and sex-matched (1:1) case-control study approved by the ethics board for human subjects of the Sheba Medical Center (IRB#6713-19), Rabin Medical Center (IRB#050119), and Hadassah Medical Center (IRB#0649-17). The study was conducted following the Helsinki Declaration 1975, revised in 2013. All participants provided their written informed consent to participate in the study.

Study population

Participants with AD aged 18 to 75 years old were recruited from the outpatient dermatology clinics of two tertiary centres in Israel: the Sheba Medical Center and the Rabin Medical Center. The control group participants with no dermatological diagnoses were volunteers recruited from dental clinics at the Faculty of Dental Medicine of the Hebrew University of Jerusalem for dental examination. Participants were generally excluded for recent oral, topical, or systemic medication use that could interfere with the microbiome assessment. The detailed inclusion and exclusion criteria are presented in the Supplemental (Appendix S1) and applied for both test and control groups.

Medical and demographic data recording

A complete medical history, including demographics, history of tobacco, alcohol, and nicotine use, comorbidities, and medical treatments, was taken for all the participants. In addition, AD-related medical history, including a history of food allergy, asthma, allergic rhinitis, and a family history of atopic diseases, was taken. Lifestyle factors affecting the oral microbiome, such as pet ownership, country of birth, race/ethnicity, and special diet, were also documented.

Oral clinical examination

A certified dentist assessed the oral health condition of each participant. The gingival index (Loe and Silness)¹⁵ was used to score the participants’ gingival and periodontal status. The plaque index¹⁵ and the oral hygiene index (Greene)¹⁶ were used to assess the participants’ oral hygiene status, while the caries severity index¹⁵ was used to evaluate the severity of their carries. Higher scores represent higher disease severity for all four instruments. The participants’ prosthodontic status, including restorations, crowns, implants, and dentures, was also recorded. The details of the oral examination are available in the Supplemental Methods section. Each patient was categorized into one of three defined oral diseases according to their oral status: (1) gingivitis (healthy to bleeding gingiva without alveolar bone resorption), (2) periodontitis stage II to IV according to periodontal disease classification, including alveolar bone resorption and tooth mobility, and (3) caries (clinically observed tooth decay or cavity).

Scoring of AD severity

Disease severity on the day of the visit was assessed using three validated scores: the validated investigator's global assessment for AD (vIGA-AD) (ranging from 0 = clear to 4 = severe),¹⁷ the eczema area and severity index¹⁸ (ranging from 0 to 72), and the patient-oriented eczema measure questionnaire¹⁹ (ranging from 0 to 28). Higher scores represent higher disease severity for all these instruments.

Microbiome samples and sequencing

From each participant, on the day of oral and dermatologic examination, one whole oral microbiome sample was obtained using sterile swabs to scratch from the tongue and with a curette to collect supra and subgingival plaque from the incisors and first molars. Until DNA extraction, samples were stored in test tubes with phosphate-buffered saline at −20°C. DNA extraction was performed using a microbiome DNA isolation kit (Microbiome DNA isolation kit, Norgen Biotek Corp), and the DNA samples were then stored at –20°C. Library preparation and next-generation sequencing of the 16S rRNA were performed by the interdepartmental equipment unit of the Hebrew University of Jerusalem, and this was followed by a sequencing analysis to map the oral microbiota phylogenetic populations in each group.²⁰ Detailed descriptions of the methods used for the microbiome sequencing and analysis are available in the Supplemental (Appendix S1).

Statistical analyses and bioinformatics

The sample size was calculated with the following assumptions: type 1 error of 5%, power of 80%, 1:1 allocation, and moderate effect size for the difference between the groups (odds ratio = 2.0). All clinical data in this study were analysed using SPSS software version 25. The sample size was calculated based on an anticipated moderate effect size, following a 1:1 allocation between the case and control groups. However, because of practical considerations, such as participant availability and consent, the final numbers in each group varied slightly. This discrepancy was noted and was accounted for in our statistical analyses to ensure the validity of the results.

The effect size used for the sample size calculation was primarily based on differences in oral microbial diversity, which we considered a key indicator of overall oral health status and its potential link to systemic conditions. Oral microbial diversity was selected as the primary outcome because of its comprehensive nature, encompassing various factors, such as plaque, bleeding, and caries, all of which contribute to and are influenced by the microbial environment. Descriptive statistics were performed using means and standard deviations for continuous variables, while frequencies were used for discrete variables. In addition to adjusting for age and sex differences between the groups, the propensity score was used for matching purposes, which produced similar groups in terms of the main baseline characteristics. Univariate correlations were performed using the chi-square (χ^2) test for the discrete variables and the t test for the continuous variables. A P value lower than 5% was considered to indicate statistical significance. Detailed descriptions of the bioinformatics test used for the microbiome analysis are available in the Supplemental (Appendix S1).

Results

A total of 86 participants (45 AD patients and 41 controls) were included in this study. The participants in the two groups had no significant differences in demographic and clinical characteristics, except for the higher rate of atopic comorbidities in the AD group (P < .001). The mean age of the participants was 36 ± 15.7 years, and 62% were female. All were Caucasian (Table S1). Most participants had mild or moderate AD severity according to the investigator-based scores (vIGA-AD and eczema area and severity index) and moderate or severe disease according to the participant-based scores (patient-oriented eczema measure). Almost all participants in the AD group (91.1%) experienced disease flares during the previous year. Some 90.2% were being treated with emollients, and 95.1% were being treated with topical corticosteroids for their AD (except for the washout period for the study). Of note, none of the AD participants were previously treated with cyclosporine, a drug with known adverse effects on periodontal health.²¹ Forty-four per cent of the AD participants had facial/head and neck dermatitis, and 16% had cheilitis. Allergic rhinitis was the most common atopic comorbidity (48.9%) in the AD group (Table 1).

Table 1.

Clinical characteristics and severity scores of atopic dermatitis participants (N = 45).

		N	%
VIGA2 AD score	Clear (0)	0	0
	Almost clear (1)	6	13.3
	Mild (2)	17	37.8
	Moderate (3)	20	44.4
	Severe (4)	2	4.4
EASI3 score (mean = 8.43, SD = 7.58)	clear (0)	2	4.4
	Almost clear (0.1-1)	5	11.1
	Mild (1.1-7)	19	42.2
	Moderate (7.1-21)	17	37.7
	Severe (21.1-50)	2	4.4
	Very severe (50.1-72)	0	0
POEM4 score (mean = 15.29, SD = 8.47)	Almost clear (0-2)	4	9.8
	Mild (3-7)	3	7.3
	Moderate (8-16)	15	36.6
	Severe (17-24)	13	31.7
	Very severe (25-28)	6	14.6
	Missing	4	9.8
Atopic dermatitis flares Within the past year		41	91.1
Facial/head and neck dermatitis		11	44.0
Cheilitis		4	16.0
1st or 2nd degree relative with atopic dermatitis		20	48.8
1st or 2nd degree relative with allergies		20	52.6
Skin infection treated with pharmacological treatment in the last year		4	8.9
Atopic comorbidities	Asthma	11	24.4
	Allergic rhinitis	22	48.9
	Food allergy	5	11.1
	Allergic conjunctivitis	12	26.7
	Allergic contact dermatitis	6	13.3
	Nasal polyps	2	4.4
	Eosinophilic esophagitis	0	0
Atopic dermatitis treatments*	Currently treated	40	97.6
	Emollients	37	90.2
	Bleach/potassium permanganate bath	2	4.9
	Topical steroids	39	95.1
	Topical calcineurin Inhibitors	20	48.8
	Phototherapy	21	51.2
	Methotrexate	4	9.8
	Cyclosporine	0	0
	DUPILUMAB	10	22.2

EASI, Eczema Area and Severity Index; POEM, Patient-Oriented Eczema Measure; VIGA, Validated Investigator's Global Assessment for Atopic Dermatitis.

^⁎See Supplemental Methods for wash-out periods.

Increased severity of oral health status in AD patients

Oral examination revealed that the AD and control groups differed in terms of plaque index scores (P = .04), with higher rates of severe plaque found among the AD patients (24.4%) than among those in the control group (7.3%). In addition, oral hygiene was found to be inferior in the AD group, with a higher rate of poor oral hygiene recorded among the AD participants (20% vs 2.4%, P = .04). Numerically, more participants with AD had higher gingival index scores (moderate and severe), with a trend towards significance (44.4% vs 24.4%, P = .05). The two groups did not differ in terms of the caries severity index scores, the prevalence of oral disease (gingivitis, periodontitis, or caries, P = .41), or the prevalence of prosthodontics (restorations, crowns, or dentures, P = .36). Taken together, these findings indicate that AD patients had poorer oral hygiene and higher plaque index scores. Moreover, they exhibited a trend towards an increased severity of gingivitis (Table 2).

Table 2.

Comparison between atopic dermatitis and control groups’ oral clinical characteristics.

Variable		Control (N = 41) N (%)	Atopic dermatitis (N = 45) N (%)	P value
Plaque index	No plaque	6 (14.6%)	1 (2.2%)	.04
	Mild	22 (53.7%)	20 (44.4%)
	Moderate	10 (24.4%)	13 (28.9%)
	Severe	3 (7.3%)	11 (24.4%)
Plaque index by groups	No plaque – mild	28 (68.3%)	21 (46.7%)	.04
	Moderate – severe	13 (31.7%)	24 (53.3%)
Oral hygiene	Good	19 (46.3%)	16 (35.6%)	.04
	Fair	21 (51.2%)	20 (44.4%)
	Poor	1 (2.4%)	9 (20%)
Gingival index score	No inflammation	7 (17.1%)	4 (8.9%)	.23
	Mild	24 (58.5%)	21 (46.7%)
	Moderate	7 (17.1%)	15 (33.3%)
	Severe	3 (7.3%)	5 (11.1%)
Gingival index score by groups	No inflammation – mild	31 (75.6%)	25 (55.6%)	.05
	Moderate – severe	10 (24.4%)	20 (44.4%)
Caries severity score	No caries	25 (55.6%)	29 (70.7%)	.31
	Mild	13 (28.9%)	9 (22.0%)
	Moderate	2 (4.4%)	2 (4.9%)
	Severe	5 (11.1%)	1 (2.4%)
Caries severity score	No caries – mild	38 (92.7%)	38 (84.4%)	.23
	Moderate – severe	3 (7.3%)	7 (15.6%)
Oral diagnosis	Gingivitis	29 (70.7%)	25 (55.6%)	.21
	Periodontitis	2 (4.9%)	2 (4.4%)	.28
	Caries	12 (29.3%)	20 (44.4%)	.18
Prosthodontics	Clear	14 (34.1%)	10 (22.2%)	.36
	Restorations	16 (39%)	24 (53.3%)
	Crowns	0 (0%)	2 (4.4%)
	Restorations + crowns	9 (22%)	7 (15.6%)
	Dentures	2 (4.9%)	2 (4.4%)

Higher oral microbial diversity and altered microbial composition in AD

The results of the Faith phylogenetic diversity (FDR) test²² indicated that the oral microbial diversity in the AD group was significantly higher (α diversity, FDR-adjusted P = .0007) than that in the control group (Figure 1A). Interestingly, the analysis of the different age groups revealed that the participants with AD had higher microbial diversity than those in the control group in the 20- to 39-year-old age group (P value = .0008). In contrast, the participants in the older age groups did not show any differences (Figure S1).

Fig. 1.

Oral microbiome biodiversity and composition. The whole group atopic dermatitis (AD) versus the control group (n = 45; n = 41, respectively) is in green, and in purple, the AD group. (A) Microbial richness and evenness using faith PD (phylogenic diversity) test (adjusted P value **<.000). The Kruskal–Wallis test was used to determine significance. Boxplots represent the medians and interquartile ranges with whiskers, as determined by Tukey's method. (B) Comparison of microbial composition and communities between the samples (β-diversity) of the two groups, double principal coordinates analyses (DPCoA) based on weighted UniFrac distances is shown along the first two principal coordinate (PC) axes (adjusted P value .001). Each point represents a single patient.

The characterization of the similarities and differences in the composition of the microbial communities between the samples (β diversity) from the AD patients and the controls indicated that they clustered separately. The groups showed significant dissimilarity and distinct clustering of microbial abundance (Figure 1B, adjusted P = .001). High community diversity was previously linked to periodontitis and gingival disease.^23,24

Increased abundance of gingivitis- and periodontitis-associated bacteria in AD patients

To determine which taxa are enriched in AD, the differential abundance of species in the oral cavity of AD patients versus the controls was investigated. Forty taxa were found with significantly (adjusted P < .05) different levels of abundance between the AD and control groups using the DESeq test (Figure 2A). Some of the taxa found in higher abundance in the AD group correlated with periodontitis and gingivitis (Figure 2B) (eg, Tannerella, Treponema, Filifactor, Peptostreptococcus, Porphyromonas endodontalis, Selenomonas, Leptotrichia, Fusobacterium, TM7-3, Mogibacteriaceae, and Prevotella).24, 25, 26, 27 Moreover, a higher abundance of the taxa Capnocytophaga ochracea, which is associated with the early phase of dental plaque formation and periodontal disease,²⁸ was observed among AD patients. By contrast, the taxa Haemophilus and Gemellaceae, associated with periodontal health,^24,27,29 showed decreased abundance in AD patients (Figure 2B).

Fig. 2.

Differential abundance of taxa in atopic dermatitis (AD) compared to the control group. (A) Correlation Barplot of the 40 taxa that have significantly (adjusted P value <.05 and |log2foldchange|>=0.58) different abundance in AD compared to control with blocking the effect of age using DESeq2 test. The length of the bar represents the log2 fold change of the taxa in the AD group; the colour represents the adjusted P value. (B) Heatmap showing the differential abundance correlation of the 40 taxa between all the samples. In purple AD patients, in the green control group, the colour scale represents the abundance of taxa in each sample, from red, indicating very high abundance, to blue, indicating very low abundance.

The comparison of the relative abundance of the top-related bacteria between the AD and control groups revealed a significantly different relative abundance for 20 taxa. Similarly, the relative abundance of the gingivitis- and periodontitis-associated bacteria increased significantly (Figure 3A-J), while that of health-associated bacteria decreased (Figure 3K, L). Interestingly, the relative abundance of the caries-associated bacteria Streptococcus and Veillonella^30,31 was lower in the AD group (Figure 3M, N). Furthermore, the relative abundance of some commensal bacteria, including Cardiobacterium, Comamonadaceae, Paludibacter, Clostridium, and Blvi28, was found to be increased in the AD group. In contrast, the abundance of Prevotella (family Paraprevotellaceae) was decreased (Figure S2).

Fig. 3.

Relative abundance of taxa in atopic dermatitis (AD) patients compared to the control group. Relative abundance Boxplot of 14 taxa with significantly (adjusted P value <.05) different abundance in AD compared to control with blocking the effect of age using Wilcox test. Boxplots represent the medians and interquartile ranges with whiskers. (A-J) Gingivitis and periodontitis-associated bacteria (K and L), health-associated bacteria (M and N), and caries-associated bacteria. In purple AD patients, in green control group.

The oral microbiome diversity of the AD participants with asthma (N = 11) was not statistically different from that of the AD participants without asthma (N = 34). Of all the microbiota found to be significantly different between the AD and control groups, only Prevotella abundance was increased in AD patients with asthma compared with AD patients without asthma and the control group (Figure S3).

Discussion

This study aimed to determine whether AD is associated with oral health status and dysbiosis of the oral microbiota. While the adult participants with AD were not found to have an increased prevalence of oral disease, they displayed poorer oral hygiene status, evident from the significant differences in the plaque and oral hygiene index scores between the two groups and a tendency towards increased gingival index scores. The analysis of the oral microbiome revealed significantly higher microbial diversity in the AD group than in the control group (both α and β diversity). In addition, the AD patients exhibited an abundance of taxa correlated with periodontal disease (eg, Treponema and Tannerella) and a decreased abundance of taxa associated with periodontal health (eg, Haemophilus, Gemellaceae).

These clinical findings align with population-based surveys that found an association between oral symptoms (eg, sensitive teeth, toothache, bleeding gums, and bad breath) and AD in adolescents^10,11 and adults.¹² In contrast to studies that only assessed symptoms, in the present study, the oral and dermatological exams were performed directly by certified dentists and dermatologists, strengthening the link between AD and poor oral health status.

Our main finding was that AD participants exhibited poorer oral hygiene with a tendency towards increased gingival index scores, which might be primary or secondary to impaired oral hygiene. Thus, poor dental habits (eg, ineffective tooth brushing, inadequate flossing, and consumption of harmful food) possibly caused the clinical and microbial differences between the two groups. A possible factor contributing to poor oral hygiene habits in the AD group might be sleep disturbances associated with AD.³² Indeed, a sleep duration of ≤5 hours per night was previously shown to be related to poor oral health compared with a sleep duration of 6 to 8 hours per night.³³

The mechanism underlying the correlation between skin disorders and oral status remains to be elucidated. Moreover, as the effect of periodontal disease on systemic disease is well established,³⁴ as is the association between poor oral hygiene practices and chronic diseases,^35,36 investigating whether the effective treatment of oral disease can alleviate the severity of AD (or vice versa) is essential. AD is one of several skin disorders associated with allergic diathesis. Previous work on psoriasis patients reported several predictors of the disease, including oral pain or discomfort, within the last 12 months. In addition, poor gum health and speech difficulties caused by dental problems were positively correlated with severe psoriasis.⁷ These findings and our study's results suggest that a link between oral health status and dermatological inflammatory disorders may exist, warranting further research on the subject.

Each participant's oral microbiome was obtained for oral microbiome analysis on the day of the examination visit. Although the oral microbiome may be affected by several daily factors, several studies have shown that the oral microbial environment, in general, can be relatively stable for up to a year. While intraindividual variations might occur, they pale in significance compared to interindividual variability.37, 38, 39, 40, 41 Our finding of higher microbial diversity in the AD group correlates with poorer oral health status, as communities of higher diversity than those found in healthy status are associated with periodontitis and gingival disease.23, 24, 25 By contrast, Li et al¹⁴ found that the oral cavity of AD patients showed reduced microbial diversity, which was distinctly correlated with disease severity. This discrepancy in the findings may have several explanations. The first is the methodology used for microbiome sampling. In the study by Li et al,¹⁴ the oral samples were a mixture of microbiota from the buccal mucosa, keratinized gingiva, supragingival plaque, and tongue dorsum, including all the major sites of the oral cavity. In our study, the samples were taken from the tongue and the cervical part of the molars to focus on the periodontal microbiome. Second, the reduced diversity in the Li et al¹⁴ study was found in a heterogeneous subgroup aged 13 to 59 years, including teenagers and adults (AD, n = 35; control, n = 35). Periodontal characteristics differ between adolescents and adults, with a gradual increase in gingival activity occurring from early childhood to adulthood.⁴² Consequently, our study focused on adult AD patients. Third, the dissimilarities in the findings may stem from the different populations tested (ie, Chinese vs Caucasian).

There are several limitations to our study. The AD group predominantly included patients with mild and moderate severity; whether the results are generalizable to other disease severities is unknown. Furthermore, the AD group had a higher number of patients with atopic comorbidities, such as asthma or allergic rhinitis. These diseases and their related medications (ie, inhalers) were previously reported to affect oral health and microbiome positively or negatively.43, 44, 45 In addition, this study included relatively small groups, and the results require verification in extensive studies.

The findings of this study highlight the need for awareness of the increased severity of oral disease in AD patients. We encourage dermatologists treating these patients to discuss oral hygiene and consider referrals to dentists, particularly periodontal disease experts.

Funding

The study was funded by a research grant from Pfizer 53899769 (ASPIRE 2019). The funder has played no role in the research.

IRB approval status

This study was approved by the human subject's ethics board of Sheba Medical Center (IRB#6713-19) and Rabin Medical Center (IRB#050119) and Hadassa Medical Center (IRB#0649-17) and was conducted in accordance with the Helsinki Declaration of 1975, as revised in 2013. The privacy rights of patients have been observed and informed consent was obtained for experimentation.

Author contributions

Aaya Shahin: Contributed to conception, design, data acquisition interpretation, and analysis, drafted and critically revised the manuscript. Yael Houri-Haddad, Shoshana Greenberger, Yael A. Leshem, and Ronen Hazan: Contributed to conception, design, data acquisition interpretation, and analysis, critically revised the manuscript. Omry Koren: Contributed to data analysis, critically revised the manuscript. Efrat Sharon: Contributed to data analysis, drafted the manuscript. Sharon Baum and Aviv Barzilai: Contributed to conception, critically revised the manuscript. Yossi Taieb, Danielle Jedda, Rinat Tzach-Nahman: Participated in analyzing and interpreting the data and contributed to the research's overall findings, and Shunit Coppenhagen-Glazer: Contributed to data acquisition, drafted the manuscript. All authors have contributed substantially to the study's conception and design, critically revised the manuscript, gave their final approval, and agreed to be accountable for all aspects of the work.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Yael Houri-Haddad, Shoshana Greenberger reports financial support was provided by Pfizer. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this article.