The British Journal of Psychiatry (2015)

207, 299–305. doi: 10.1192/bjp.bp.114.156323

Review article

Association between poor oral health and eating

disorders: systematic review and meta-analysis
Steve Kisely, Hooman Baghaie, Ratilal Lalloo and Newell W. Johnson

Background
There is a well-established link between oral pathology and Patients with an eating disorder had five times the odds of
eating disorders in the presence of self-induced vomiting. dental erosion compared with controls (95% CI 3.31–7.58);
There is less information concerning this relationship in the odds were highest in those with self-induced vomiting (odds
absence of self-induced vomiting, in spite of risk factors such ratio (OR) = 7.32). Patients also had significantly higher DMFS
as psychotropic-induced dry mouth, nutritional deficiency or scores (mean difference 3.07, 95% CI 0.66–5.48) and reduced
acidic diet. salivary flow (OR = 2.24, 95% CI 1.44–3.51).

Aims Conclusions
To determine the association between eating disorder and These findings highlight the importance of collaboration
poor oral health, including any difference between patients between dental and medical practitioners. Dentists may be
with and without self-induced vomiting. the first clinicians to suspect an eating disorder given
patients’ reluctance to present for psychiatric treatment,
Method whereas mental health clinicians should be aware of the oral
A systematic search was made of Medline, PsycINFO, consequences of inappropriate diet, psychotropic medication
EMBASE and article bibliographies. Outcomes were dental and self-induced vomiting.
erosion, salivary gland function and the mean number of
decayed, missing and filled teeth or surfaces (DMFT/S). Declaration of interest
None.
Results
Ten studies had sufficient data for a random effects meta- Copyright and usage
analysis (psychiatric patients n = 556, controls n = 556). B The Royal College of Psychiatrists 2015.

The complex interrelationships between physical and psychiatric The association between oral pathology and eating disorders
illnesses have been the focus of much research.1 Depression, for is most clearly established in cases with frequent self-induced
example, is not only a serious chronic illness but also a major vomiting, regardless of whether the diagnosis is anorexia or
risk factor for heart disease and cancer. At the same time there bulimia, and is characterised by dental erosion on palatal surfaces
is strong evidence that mental health is important for maintaining (the inner surfaces of teeth in the upper jaw).7,8 Dental caries and
good physical health and healthy lifestyle practices. However, dry mouth secondary to salivary gland dysfunction also occur.
the relationship between oral and mental health is a relatively Oral pathology is less clearly established in patients with an eating
neglected area. In the only meta-analysis of the association disorder but without self-induced vomiting, particularly in the
between the two, patients with psychiatric disorders such as case of dental caries.7 On one hand, people with anorexia without
schizophrenia, bipolar disorder and dementia were over three self-induced vomiting might be more at risk through additional
times as likely to have lost all their teeth.2 However, this study factors such as nutritional deficiency or the use of carbonated
did not include eating disorders. drinks as appetite suppressants;6 on the other hand, the frequent
Eating disorders are divided into three main diagnoses.3 co-occurrence of obsessional personality traits might mean such
Anorexia nervosa is characterised by low body weight and food people are more fastidious in their oral hygiene.6 We therefore
restriction. Bulimia nervosa is characterised by binge eating and undertook a meta-analysis to determine the association between
inappropriate compensatory behaviours such as self-induced eating disorders and poor oral health including any differences
vomiting, use of laxatives and excessive exercise. Eating disorder between patients with and without self-induced vomiting.
not otherwise specified includes a mixture of anorexia- and Although there have been several systematic reviews, none
bulimia-like atypical disorders.3 The impact of eating disorders included any meta-analysis.3,7,8
on oral health was initially reported by Hellstrom and Hurst et
al in the late 1970s.4,5 There are three main types of oral Method
pathology.3,6–9 Dental erosion or pathological wear on tooth
surfaces is defined as loss of dental tissue without the involvement The review was registered with PROSPERO, an international database
of bacteria;8 risk factors include the consumption of large of prospectively registered systematic reviews in health and social care
amounts of citrus fruit, soft drinks and sports drinks, as well as based in the UK.10 In addition, we followed recommendations for the
the presence of gastric reflux or frequent vomiting. In contrast, reporting of Meta-analyses of Observational Studies in Epidemiology
dental caries is the result of bacterial action;8 organic acids (MOOSE), including background, search strategy, methods, results,
produced by microorganisms in dental plaque cause decalcification discussion and conclusions.11
of the tooth enamel and subsequent destruction of enamel and
dentin. Finally, self-induced vomiting or starvation can lead to
hyposalivation and xerostomia (dry mouth). This may be accentuated Oral health outcomes
by psychotropic medication.8 Hyposalivation and xerostomia are The primary outcome of this study was dental erosion. This can be
risk factors for both dental caries and erosion. expressed as either a continuous or a dichotomous variable. In

299
Kisely et al

either situation, the area of the mouth that is worst affected Study quality
determines the overall score. The pattern of erosion is described We assessed the quality of included studies using the Newcastle–
by the direction in which the tooth surface faces (see online Fig. Ottawa Scale (NOS).15 This assesses the quality of non-randomised
DS1). In terms of the outer surfaces, buccal surfaces are adjacent studies in meta-analyses in the three following areas:
to the cheeks, whereas labial ones face the lips. Inner surfaces
adjacent to the tongue are called lingual;12 sometimes these (a) selection of the study groups in terms of case definition,
surfaces are called palatal when referring to the upper teeth representativeness (e.g. all eligible cases with the outcome of
adjacent to the palate. The top areas are divided into the occlusal interest over a defined period of time or from a defined
surfaces (the chewing surfaces of posterior teeth) and the incisal catchment area), source of controls (ideally the community)
edges (the biting edge of anterior teeth).12 and checks that the controls did not have an eating disorder;
Secondary outcomes were dental decay and salivary gland (b) comparability of the groups such as the use of matching or
function. Dental caries was assessed by the number of decayed, multivariate techniques;
missing and filled teeth or surfaces. Both scores are expressed as (c) ascertainment of outcome such as the use of standardised or
a continuous variable that accumulates over a person’s lifetime,
validated measures with masking to psychiatric status.
reflecting the individual’s overall experience of dental caries.13
This is because both dental decay and its treatment leave
permanent marks, either through the continued presence of Statistical analysis
carious lesions, the presence of fillings or the loss of affected teeth We used Review Manager version 5.0, a statistical software
by extraction. The total number of teeth (T) and surfaces (S) that package for analysing a Cochrane Collaboration systematic review,
are decayed (D), missing because of pathology (M) or filled (F) for our analysis. For each outcome we divided the eating disorder
are measures referred to as the DMFT and DMFS respectively. group into three subgroups regardless of whether the diagnosis
In both measures an increase in score means greater cumulative was anorexia or bulimia: one in which self-induced vomiting
dental decay; however, DMFS scores are higher than DMFT scores was frequent, a second in which it was absent and a third group
because the former measure counts damage to each surface of that had a mixture or where it was not explicitly stated. Where
every tooth rather than counting the tooth as a single unit. This we encountered a situation where data for the same outcome were
can be four or five surfaces depending on the tooth. The presented in some studies as dichotomous data and in others as
maximum possible DMFT score is therefore 32, whereas the continuous data, we combined them using statistical approaches
maximum DMFS score is 148. Salivary gland function was as recommended in the Cochrane Handbook.16 These techniques
assessed where possible by measurement of unstimulated salivary re-express odds ratios as standardised mean differences (and vice
flow, otherwise by report of dry mouth by the patient. Both versa), allowing dichotomous and continuous data to be pooled.
outcomes are usually reported as dichotomous variables; in the case Once standardised mean differences (or log odds ratios) and their
of saliva flow, cut-off values can be either less than 0.1 ml/min or standard errors were computed for all studies in the meta-analysis,
less than 0.2 ml/min.14 they were combined using the generic inverse variance method in
RevMan. For dental decay and salivary flow, most studies reported
dichotomous outcomes. We therefore converted continuous data
Inclusion and exclusion criteria to dichotomous variables and then calculated odds ratios given
We included studies of the oral health of people with eating that the studies were a cross-sectional design. Odds ratios also
disorders that included a control group of people without eating have the advantage of being easier to understand and more
disorders, ideally matched by age, gender, socioeconomic status clinically meaningful than standardised mean differences (SMD).
and education level. Psychiatric status could be determined by Dental decay was consistently reported using continuous data.
clinical diagnosis or diagnostic criteria. Studies of people with We calculated the mean differences (as opposed to SMD) for
severe mental illness, primary alcohol or substance use disorders, studies that used the same scale for each outcome (e.g. DMFT,
intellectual disability and other psychological disorders were DMFS).
excluded. As our focus was on dental erosion, decay and salivary We assessed heterogeneity by using the I 2 statistic. This
gland function, we excluded studies of other dental outcomes such provides an estimate of the percentage of variability due to hetero-
as poor oral hygiene. geneity rather than chance alone. An I 2 estimate of 50% or above
indicates possible heterogeneity, and scores of 75–100% indicate
considerable heterogeneity.16 The I 2 statistic is calculated using
Search strategy the chi-squared statistic (Q) and its degrees of freedom. It has
We searched Medline, PsycINFO and EMBASE from January 1951 several advantages over the Q statistic alone in that it does not
until June 2014 using the following text, MeSH or Emtree terms as depend on the number of studies in the meta-analysis and so
appropriate: bulimia, binge eating, eating disorder, overeating, has greater power to detect heterogeneity where the number of
appetite disorder, binge eating disorder, binge eating disorders, studies is relatively low.16 The I 2 statistic can also be interpreted
anorexia, anorexia nervosa, oral health, dental health survey, similarly irrespective of whether outcome data are dichotomous
dental care, dental health services, edentulous mouth, dental or continuous.
caries, dental erosion, toothloss and tooth wear. Other descriptive We used a random effects model throughout as there was
words associated with the above MeSH terms were also used as key significant heterogeneity in the majority of our analyses. This
terms. We searched for further publications by scrutinising the model assumes that variations in effect among different studies
reference lists of initial studies identified and other relevant review are due to differences in samples or paradigms and have a normal
papers. We made attempts to contact selected authors and experts. distribution, i.e. that heterogeneity exists. In addition, where
Two reviewers (H.B. and S.K.) independently assessed titles, possible, we investigated heterogeneity in sensitivity analyses
abstracts and papers, as well as extracting and checking the data omitting each study in turn. Other sensitivity analyses included
for accuracy. In cases of disagreement consensus was reached on investigating the effects of setting (e.g. in-patient or out-patient)
all occasions. Authors R.L. and N.W.J. provided content expertise, and of excluding studies where there were concerns about data
especially in relation to oral and dental health issues. quality. Where there were at least 10 studies we tested for

300
 Oral health and eating disorders

publication bias using both the fail-safe N statistic and funnel plot the presence of morbidity using the Eating Disorders Inventory
asymmetry. We used WinPepi version 11.34.17 The fail-safe N and Examination, as well as the Eating Attitudes Test.24 None of
statistic is the number of non-significant studies that would be the studies used community controls; three recruited dental pa-
necessary to reduce the odds ratio or effect size to a negligible tients,19,23,27 and the remainder used staff or university students.
value. In tests for a skewed funnel plot, low P-values suggest In one study the source of controls was not stated.26 One study
publication bias. excluded the presence of eating disorders in the control group
through the use of a standardised questionnaire,19 whereas
another asked about past psychiatric history.26
Results In terms of group comparability, all the studies either used
age- and gender-matched controls or checked that there was no
We found 1085 citations of interest in the initial electronic significant difference between the two groups at baseline. Two
searches, from which 112 abstracts were screened. Of these, 39 studies checked that participants in the eating disorder cases
full-text papers were potentially relevant and assessed for eligibility. and control groups were of similar socioeconomic status,22,23
Assessing the references in all papers yielded one more full-text and a third that they were similar in terms of ethnicity and
paper that was deemed relevant, giving a total of 40 full-text medical history.27
papers. Of these, 30 papers were excluded, most because they were Ascertainment of oral status in all the studies was by trained
not prevalence studies of oral health and eating disorder, or did dental examiners. In the case of erosion, this was a clinical
not include a relevant dental outcome (Fig. 1). This left 10 studies assessment sometimes guided by an established classification.
that could be included in the meta-analysis. Four were from Two studies used the Tooth Wear Index.28 Five studies supplemented
Nordic countries,18–21 two studies were from England,22,23 and the clinical examination with dental impressions, radiographs
there was one study each from Australia,24 Israel,25 Germany,26 and/or intraoral photographs. In the case of caries, all the studies
and the USA.27 The most common diagnosis was bulimia, used some or all of the Decayed, Missing and Filled classification.29
followed by anorexia and eating disorders not otherwise specified. In two studies, radiographs and/or intraoral photographs were
Ages ranged from 10 years to 50 years. The studies are summarised also taken. However, only one study made specific mention of
in online Table DS1. assessor calibration and the measurement of interrater reliability,
Study quality was not optimal, particularly in the areas of and another measured agreement on a random subsample of 10
selection and ascertainment of outcome. Only one study stated participants.19,21 The dental assessor was masked to psychiatric
that eating disorder cases were consecutive admissions,24 and no status in four studies.18,19,21,22 Five studies assessed salivary gland
study gave details about participation rates. Six studies defined function, and saliva flow was measured in all cases. (Full details
psychiatric caseness using diagnostic criteria such as the DSM or are given in Table DS1.)
ICD, but this was by clinical assessment not a standardised Data for meta-analyses were available for 556 patients with
psychiatric interview (Table DS1). One further study assessed eating disorder and 556 controls (total n = 1112). Gender data
were available for 868 participants, of whom 852 (98%) were fe-
Total papers yielded:
male. There were two studies where a mean and range were
abstracts searched reported.18,19 In this situation we used the range rule to estimate
electronically for the standard deviation through dividing the range by four.30
key terms
1085 papers However, given this is not universally accepted,16 we did a
sensitivity analysis of the effect of excluding these two studies.
Did not meet
7 inclusion criteria
973 papers
6 Dental erosion
Possible inclusion:
abstracts scrutinised
Participants with an eating disorder had five times the odds of
112 papers dental erosion (95% CI 3.31–7.58) compared with controls (Fig.
Did not meet 2). Patients with self-induced vomiting had the highest
7 inclusion criteria, likelihood (odds ratio (OR) = 7.32) whereas those without vomit-
e.g. not a prevalence study ing had the lowest (OR = 3.10), although this still remained signif-
72 papers
6 icantly greater than for controls (95% CI 1.67–5.77). Excluding
Full-text papers the two studies where the s.d. was estimated from ranges made
scrutinised in detail no difference to these results. The same applied when we included
40 papers
only studies that used diagnostic criteria to define the psychiatric
cases (OR = 4.95, 95% CI 3.13–7.84). In terms of sensitivity
Excluded: 30 papers
Not a prevalence study of oral health analyses of the effect of setting, it was not always clear whether
and eating disorder or did not include studies were of in-patients, day patients or out-patients. However,
a relevant dental outcome (25) limiting the analyses to studies that were clearly restricted to out-
Eating disorders were self-reported
7 or suspected but not diagnosed (1) patients made no difference to the results (OR = 3.75, 95% CI
Follow-up of previous study 2.11–6.70). We undertook two sensitivity analyses of study quality.
with original being included (1)
Insufficient data or no controls (1)
Including only studies that matched or checked for confounding
Cause of vomiting not specified (1) variables such as socioeconomic status, ethnicity and medical
Duplicate data (1) history,22,23,27 or those where the dental assessment was made
6 masked to psychiatric status,18,19,21,22 also did not alter the results.
Meta-analysis
10 papers
Dental caries
Fig. 1 Number of papers yielded by search strategy.
Four studies used some or all of the DMFS classification, and a
fifth used the DMFT (Fig. 3). Patients with an eating disorder

301
Kisely et al

Eating disorders +ve Controls OR OR

Study or subgroup log (OR) s.e. Total Total Weight IV, random, 95% CI IV, random, 95% CI
Eating disorder, no vomiting
Emodi-Perlman et al (2008)25 1.02 0.64 36 24 5.2% 2.77 (0.79, 9.72)
Liew et al (1991)24 0.51 0.73 15 15 4.5% 1.67 (0.40, 6.96)
Milosevic & Slade (1989)22 0.73 0.37 25 25 7.6% 2.08 (1.00, 4.29)
Philipp et al (1991)26 2.86 0.74 11 25 4.5% 17.46 (4.09, 74.47)
Robb et al (1995)23 – AN 1.09 0.36 54 54 7.7% 2.97 (1.47, 6.02)
Subtotal (95% CI) 141 143 29.6% 3.10 (1.67, 5.77)
Heterogeneity: t2 = 0.22; w2 = 7.29, d.f. = 4 (P = 0.12); I 2 = 45%
Test for overall effect Z = 3.57 (P = 0.0004)

Combined group
Altshuler et al (1990)27 2.13 0.46 40 40 6.8% 8.41 (3.42, 20.73)
Johansson et al (2012)19 2.68 0.4 54 54 7.3% 14.59 (6.66, 31.94)
Ohm et al (1999)20 0.66 0.23 81 52 8.9% 1.93 (1.23, 3.04)
Rytomaa et al (1998)21 1.12 0.2 35 105 9.1% 3.06 (2.07, 4.54)
Subtotal (95% CI) 210 251 32.1% 4.84 (2.16, 10.84)
Heterogeneity: t2 = 0.57; w2 = 23.41, d.f. = 3 (P50.0001); I 2 = 87%
Test for overall effect Z = 3.83 (P = 0.0001)

Self-induced vomiting
Dynesen et al (2008)18 2.43 0.64 20 20 5.2% 11.36 (3.24, 39.82)
Emodi-Perlman et al (2008)25 1.38 0.63 43 24 5.3% 3.97 (1.16, 13.66)
Milosevic & Slade (1989)22 0.93 0.36 33 25 7.7% 2.53 (1.25, 5.13)
Philipp et al (1991)26 2.37 0.4 41 25 7.3% 10.70 (4.88, 23.43)
Robb et al (1995)23 – BN 2.49 0.53 39 39 6.1% 12.06 (4.27, 34.08)
Robb et al (1995)23 – vomiting AN 2.5 0.48 29 29 6.8% 12.18 (4.95, 30.01)
Subtotal (95% CI) 205 162 38.4% 7.32 (3.92, 13.67)
Heterogeneity: t2 = 0.36; w2 = 12.84, d.f. = 5 (P = 0.02); I 2 = 61%
Test for overall effect Z = 6.24 (P50.00001)

Total (95% CI) 556 556 100.0% 5.00 (3.31, 7.58)

Heterogeneity: t2 = 0.45; w2 = 54.87, d.f. = 14 (P50.00001); I 2 = 74%
Test for overall effect Z = 7.61 (P = 0.00001) 0.02 0.1 1 10 50
Test for subgroup differences: w2 = 3.68; d.f. = 2 (P = 0.16); I 2 = 45.3% Controls Eating disorders

Fig. 2 Dental erosion. AN, anorexia nervosa; BN, bulimia nervosa.

had significantly more decayed, missing and filled surfaces than

Discussion
controls. The study that used the DMFT reported no difference
but this was confined to non-vomiting patients. There were
To our knowledge, this is the first meta-analysis of the association
insufficient studies to undertake any sensitivity analyses.
between eating disorders and poor oral health including any
difference between patients with and without self-induced
Salivary gland function vomiting. The most frequent finding was the presence of erosion
or pathological wear on tooth surfaces. The risk of dental erosion
Five studies assessed salivary gland function in terms of dry is increased both in individuals who consume large amounts of
mouth or reduced salivary flow and there was a significant citrus fruits, soft drinks or sports drinks, and in the presence of
association with eating disorders (Fig. 4). Again, there were gastric reflux or vomiting.8 This is consistent with our finding that
insufficient studies to undertake any sensitivity analyses. patients with self-induced vomiting have the greatest degree of
oral pathology, but even in patients in whom vomiting is not a
prominent symptom erosion still occurs. Dental caries and
Heterogeneity
reduced salivary gland function, although less marked, remained
All but two of the results had an I 2 estimate of 50% or more, significantly greater than in controls.
indicating possible heterogeneity. The two exceptions were erosion Early reports of the oral consequences of eating disorders
in the absence of self-induced vomiting and DMFS scores. There emphasised the role of self-induced vomiting as an aetiological
was no difference in I 2 values in sensitivity analyses of the effect factor, with wear occurring especially on palatal (inner) surfaces
of omitting each study in turn. of teeth in the upper (maxillary) arch.4–6,31 However, the
relationship is not simple because the frequency, duration and
total number of vomiting episodes were not linearly associated
Publication bias with erosion.6 In addition, later work has shown that erosion
We were only able to test for publication bias for dental erosion as occurs in patients with an eating disorder who do not vomit,
there were insufficient studies for the other two outcomes. The although the pattern is different with greater involvement of
fail-safe N of additional ‘null’ studies needed to reduce the overall buccal and or labial sites (i.e. the outer surfaces of the teeth facing
odds ratio to 1.1 was 201, suggesting that the findings for erosion the cheeks and lips).6 It may therefore be simplistic to ascribe all
were reasonably robust against publication bias. Tests for funnel dental effects as being secondary to vomiting when other risk
plot asymmetry gave a P-value of 0.35 (Fig. 5). factors such as acidic food and drink, and the pattern of their

302
 Oral health and eating disorders

Eating disorders Controls Mean difference Mean difference

Study or subgroup Mean s.d. Total Mean s.d. Total Weight IV, random, 95% CI IV, random, 95% CI
Decayed surfaces
Altshuler et al (1990)27 4.1 4 40 1.5 2 40 18.9% 2.60 (1.21, 3.99)
Ohm et al (1999)20 1.6 2.5 81 1 1.5 52 35.1% 0.60 (70.08, 1.28)
Rytomaa et al (1998)21 1.2 0.6 35 0.6 0.8 105 46.0% 0.60 (0.35, 0.85)
Subtotal (95% CI) 156 197 100.0% 0.98 (0.21, 1.75)
Heterogeneity: t2 = 0.32; w2 = 7.78, d.f. = 2 (P = 0.02); I 2 = 74%
Test for overall effect Z = 2.49 (P = 0.01)

Decayed filled surfaces

Ohm et al (1999)20 13.5 9.8 81 9.4 8.6 52 100.0% 4.10 (0.93, 7.27)
Subtotal (95% CI) 81 52 100.0% 4.10 (0.93, 7.27)
Heterogeneity: not applicable
Test for overall effect Z = 2.54 (P = 0.01)

DMFS
Altshuler et al (1990)27 16.7 13.4 40 17.7 16.5 40 13.4% 71.00 (77.59, 5.59)
Milosevic & Slade (1989)22 33.4 10 33 30.8 10 50 30.1% 2.60 (71.80, 7.00)
Ohm et al (1999)20 15.3 10.9 81 10.8 9.1 52 49.5% 4.50 (1.07, 7.93)
Rytomaa et al (1998)21 23.6 25.9 35 20.8 16 105 7.0% 2.80 (76.31, 11.91)
Subtotal (95% CI) 189 247 100.0% 3.07 (0.66, 5.48)
Heterogeneity: t2 = 0.00; w2 = 2.18, d.f. = 3 (P = 0.54); I 2 = 0%
Test for overall effect Z = 2.50 (P = 0.01)

DMFT no vomiting
Liew et al (1991)24 4.4 4.7 15 5.2 4.1 15 100.0% 70.80 (73.96, 2.36)
Subtotal (95% CI) 15 15 100.0% 70.80 (73.96, 2.36)
Heterogeneity: not applicable
Test for overall effect Z = 0.50 (P = 0.62)

710 75 0 5 10
Controls Eating disorders

Fig. 3 Dental caries (tooth decay). DMFS/T, decayed, missing and filled surfaces/teeth.

Eating disorders Controls OR OR

Study or subgroup log (OR) s.e. Total Total Weight IV, random, 95% CI IV, random, 95% CI
Altshuler et al (1990)27 2.17 0.64 40 40 8.2% 8.76 (2.50, 30.70)
Dynesen et al (2008)18 1.81 0.61 20 20 9.8% 6.11 (1.85, 0.20)
Johansson et al (2012)19 0.61 0.18 54 54 27.2% 1.84 (1.29, 2.62)
Ohm et al (1999)20 0.28 0.15 81 52 28.7% 1.32 (0.99, 1.78)
Rytomaa et al (1998)21 0.74 0.22 35 105 25.1% 2.10 (1.36, 3.23)

Total (95% CI) 230 271 100.0% 2.24 (1.44, 3.51)

Heterogeneity: t2 = 0.18; w2 = 14.78, d.f. = 4 (P = 0.005); I 2 = 73%
Test for overall effect Z = 3.54 (P = 0.0004) 0.01 0.1 1 10 100
Controls Eating disorders

Fig. 4 Dry mouth or reduced salivary flow.

consumption, have not been studied.6,7 One theory is that discomfort or pain, poor aesthetic quality and reduced quality of
intrinsic (gastric) acid results in palatal erosion, whereas extrinsic life.9 In turn, deteriorating facial appearance may further alter
(dietary) acids from fruit or carbonated drinks lead to labial body perception and/or self-esteem and hence contribute to a
erosion.6 dangerous vicious cycle.9
Dental decay and salivary gland dysfunction are closely
related. Dry mouth is a major risk factor for decay, possibly
compounded by an increased risk of opportunistic infections as Limitations
a result of nutritional deficiencies.9 Changes in salivary secretion There are a number of limitations to our study. Study quality
may be secondary to structural change within the gland, and was not optimal. For instance, none of the studies established
benign parotid enlargement has been frequently described in psychiatric caseness using the gold standard of a structured
patients with bulimia.6 Dry mouth may also be a side-effect of standardised interview. Although all the studies used age- and
commonly used psychotropic medications.6,8 gender-matched controls, only three matched or checked for other
Poor dental health can have major consequences for patients potential confounding variables such as socioeconomic status,
with eating disorder. These include oral function impairment, oral ethnicity and medical history.22,23,27 However, restricting the

303
Kisely et al

0–
Vomiting absent
5 Combined group
Vomiting present
0.2 – 5
5

0.4 –
s.e. (log odds ratio)

5
5

0.6 –

0.8 –

1–
0.02 0.1 1 10 50
Odds ratio

Fig. 5 Funnel plot of studies on dental erosion.

analyses to just these three studies did not change the results for They should brush their teeth gently with a small amount of
our primary outcome. Although dental status in all the studies desensitising or bicarbonate toothpaste; vigorous brushing after
was assessed by trained examiners, in only four was this done self-induced vomiting is inadvisable as the softened, demineralised
masked to psychiatric status.18,19,21,22 Again, a sensitivity analysis surface is more susceptible to toothbrush abrasion.22 Finally,
of the effects of only including masked (blinded) outcomes made medical practitioners should be aware that many psychotropic
no difference to the erosion results. Unfortunately, there were medications can exacerbate dry mouth with consequent adverse
insufficient studies for sensitivity analyses of the secondary effects on oral health.6 If this is unavoidable, they should prescribe
outcomes. neutral artificial saliva or sialogogue pastilles.6
There were other limitations in study quality that we could The increased focus on physical and psychiatric comorbidity
not attempt to address using sensitivity analyses, such as the should include consideration of oral health. Policy makers should
calibration or standardisation of dental assessments. In addition, consider providing free, accessible dental care for people with
many of our results showed heterogeneity. We explored this eating disorders. For example, Queensland’s strategy to improve
further through sensitivity analyses of the effect of omitting the physical health of people with psychiatric illness (Activate:
each study in turn, but this made no difference to the results. Mind and Body) includes both the promotion of oral hygiene
Accordingly, we used a random effects model throughout to and regular care from a dentist.32
incorporate heterogeneity into our analyses. However, although
we have tried to minimise the effects of heterogeneity, our results Steve Kisely, MD, PhD, School of Medicine, University of Queensland,
should still be treated with caution. Finally, we cannot exclude the Woolloongabba; Hooman Baghaie, BOH, School of Dentistry, University of
Queensland, Herston, Queensland; Ratilal Lalloo, MChD, PhD, Australian Centre for
possibility of publication bias even though the fail-safe N for the Population Oral Health, School of Dentistry, University of Adelaide, South Australia;
primary outcome was 201. This is because tests for funnel plot Newell W. Johnson, MDSc, PhD, Population and Social Health Research
Programme, Griffith Health Institute, Gold Coast, Queensland, Australia
asymmetry tend to be underpowered when the number of studies
is relatively low. There were also insufficient studies to test for Correspondence: Dr Steve Kisely, Level 4, Building 1, Princess Alexandra
Hospital, 199 Ipswich Road, Woolloongabba, QLD 4102, Australia. Email:
publication bias for the other two outcomes. s.kisely@uq.edu.au

First received 14 Aug 2014, accepted 14 Mar 2015

Implications
These findings highlight the importance of collaboration between
dental and other health workers such as dieticians, general
Funding
practitioners, psychiatrists and other mental health clinicians. This
applies to all patients with eating disorders, not just those who This project was supported by the University of Queensland Winter Research programme.
present with self-induced vomiting. Dentists may be the first
clinicians to suspect the diagnosis, given the reluctance of some
people with eating disorders to present for treatment.9 In References
established cases collaboration might help to minimise the
harmful effects of inappropriate diet and self-induced vomiting;6 1 Lawrence D, Hancock KJ, Kisely S. The gap in life expectancy from
preventable physical illness in psychiatric patients in Western Australia:
for instance, patients should be advised to reduce their intake of retrospective analysis of population based registers. BMJ 2013; 346: f2539.
acidic drinks and food such as citrus fruit, as well as alcohol. After
2 Kisely S, Quek LH, Pais J, Lalloo R, Johnson NW, Lawrence D. Advanced
episodes of self-induced vomiting they should chew gum and dental disease in people with severe mental illness: systematic review and
rinse their mouth with water, milk or an antacid preparation.6 meta-analysis. Br J Psychiatry 2011; 199: 187–93.

304
 Oral health and eating disorders

3 Romanos GE, Javed F, Romanos EB, Williams RC. Oro-facial manifestations in 18 Dynesen A, Bardow A, Petersson B, Nielsen LR, Nauntofte B. Salivary
patients with eating disorders. Appetite 2012; 59: 499–504. changes and dental erosion in bulimia nervosa. Oral Surg Oral Med Oral
4 Hellstrom I. Oral complications in anorexia nervosa. Scand J Dent Res 1977; Pathol Oral Radiol Endod 2008; 106: 696–707.
85: 71–86. 19 Johansson AK, Norring C, Unell L, Johansson A. Eating disorders and oral
5 Hurst PS, Lacey LH, Crisp AH. Teeth, vomiting and diet: a study of the dental health: a matched case-control study. Eur J Oral Sci 2012; 120: 61–8.
characteristics of seventeen anorexia nervosa patients. Postgrad Med J 1977; 20 Ohrn R, Enzell K, Angmar-Mansson B. Oral status of 81 subjects with eating
53: 298–305. disorders. Eur J Oral Sci 1999; 107: 157–63.
6 Milosevic A. Eating disorders and the dentist. Br Dental J 1999; 186: 109–13. 21 Rytomaa I, Jarvinen V, Kanerva R, Heinonen OP. Bulimia and tooth erosion.
7 Frydrych AM, Davies GR, McDermott BM. Eating disorders and oral health: Acta Odontol Scand 1998; 56: 36–40.
a review of the literature. Aust Dental J 2005; 50: 6–15. 22 Milosevic A, Slade PD. The orodental status of anorexics and bulimics.
8 Bretz W. Oral profiles of bulimic women: diagnosis and management. What is Br Dent J 1989; 167: 66–70.
the evidence? J Evid Based Dent Pract 2002; 2: 267–72. 23 Robb ND, Smith BG, Geidrys-Leeper E. The distribution of erosion in the
9 Lo Russo L, Campisi G, Di Fede O, Di Liberto C, Panzarella V, Lo Muzio L. dentitions of patients with eating disorders. Br Dent J 1995; 178: 171–5.
Oral manifestations of eating disorders: a critical review. Oral Dis 2008; 24 Liew VP, Frisken KW, Touyz SW, Beumont PJ, Williams H. Clinical and
14: 479–84. microbiological investigations of anorexia nervosa. Aust Dent J 1991; 36:
10 Booth A, Clarke M, Dooley G, Ghersi D, Moher D, Petticrew M, et al. The nuts 435–41.
and bolts of PROSPERO: an international prospective register of systematic 25 Emodi-Perlman A, Yoffe T, Rosenberg N, Eli I, Alter Z, Winocur E. Prevalence
reviews. Syst Rev 2012; 1: 2. of psychologic, dental, and temporomandibular signs and symptoms among
11 Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. chronic eating disorders patients: a comparative control study. J Orofac Pain
Meta-analysis of observational studies in epidemiology: a proposal for 2008; 22: 201–8.
reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) 26 Philipp E, Willershausen-Zonnchen B, Hamm G, Pirke KM. Oral and dental
group. JAMA 2000; 283: 2008–12. characteristics in bulimic and anorectic patients. Int J Eat Disord 1991; 10:
12 Metivier A, Bland K. Dental Anatomy: A Review. Dental Care.com, 2014 423–31.
(http://www.dentalcare.com/media/en-US/education/ce421/ce421.pdf). 27 Altshuler BD, Dechow PC, Waller D, Hardy BW. An investigation of the oral
13 Slade GD, Spencer AJ, Roberts-Thomson KF. Australia’s Dental Generations: pathologies occurring in bulimia nervosa. Int J Eat Disord 1990; 9: 191–9.
The National Survey of Adult Oral Health 2004–06 (AIHW cat. no. DEN165). 28 Smith BGN, Knight JK. An index for measuring the wear of teeth. Br Dent J
Australian Institute of Health and Welfare, 2007. 1984; 156: 435–8.
14 Dawes C. Physiological factors affecting salivary flow rate, oral sugar 29 World Health Organization. Oral Health Survey – Basic Methods (4th edn).
clearance, and the sensation of dry mouth in man. J Dent Res 1987; 66: WHO, 1997.
648–53. 30 Taylor C. Range Rule for Standard Deviation. How to Estimate the Standard
15 Wells G, Shea B, O’Connell D, Peterson J, Welch V, Losos M, et al. Deviation. Available at: http://statistics.about.com/od/Descriptive-Statistics/
The Newcastle-Ottawa Scale (NOS) for Assessing the Quality of a/Range-Rule-For-Standard-Deviation.htm.
Nonrandomised Studies in Meta-analyses. Available at: http://www.ohri.ca/ 31 Roberts MW, Li SH. Oral findings in anorexia nervosa and bulimia nervosa:
programs/clinical_epidemiology/oxford.asp. a study of 47 cases. J Am Dent Assoc 1987; 115: 407–10.
16 Higgins JPT, Green S (eds). Cochrane Handbook for Systematic Reviews 32 General Practice Queensland. Activate: Mind & Body. General Practice
of Interventions Version 5.0.2. Cochrane Collaboration, 2009. Queensland, 2009 (http://www.gpqld.com.au/page/Programs/Mental_Health/
17 Abramson JH. WINPEPI updated: computer programs for epidemiologists, Improving_the_Physical_Health_of_People_with_a_Severe_Mental_Illness_
and their teaching potential. Epidemiol Perspect Innov 2011; 8: 1. Project/).

305
Online supplement to Kisely et al. Association between poor oral health and eating
disorders: systematic review and meta-analysis. Br J Psychiatry doi:
10.1192/bjp.bp.114.156323

Fig. DS1 Dental surfaces

1
Table DS1 Studies included in the review
Study Year Country Setting n Gender Age, years Primary psychiatric Erosion Decay Salivary
Mean (range) diagnosis flow
27
Altshuler 1990 USA Eating 40 Psychiatric 40F 23.8 BN (n=40) (DSM III R) Clinical assessment DMFS Measured
disorders 40 Controls 40F
unit/out-
patients
18
Dynesen 2008 Denmark Out-patients 20 Psychiatric 20F 23.8 BN (n=20) (DSM IV) Clinical assessment DMFS Measured
20 Controls 20F Silicone impressions
Radiographs
Emodi- 2008 Israel In-patients 79 Psychiatric 79F 23.5 (18–35) BN (n=29) Clinical assessment using DMFS
25
Perlman 48 Controls 48F AN (n=24) Johansson ‘s classification

EDNOS (n=16)
Mixed (n=10)
19
Johansson 2012 Sweden Out-patients 54 Psychiatric 50F, 4M 21.5 (10–50) EDNOS (n=32) Clinical assessment DMFS
54 Controls 50F, 4M AN (n=14) Dental impressions Dental radiographs

BN (n=8)
24
Liew 1991 Australia Eating 15 Psychiatric 15F 20.1 AN (n=15) (EDE, EAT, EDI) Clinical assessment DMFT
disorders unit 15 Controls 15F
22
Milosevic 1989 UK Out-patients 58 Psychiatric 56F, 2M 24 BN (n=40) (DSM III) Tooth Wear Index DMFS Measured
50 Controls 50F AN (n=18)
20
Ohrn 1999 Sweden Out-patients 81 Psychiatric 79F, 2M 25 (17–47) BN (n=46) (DSM III-R) Clinical assessment with DMFS Measured
52 Controls 48F, 4M Mixed (n=7) Eccles system for Tooth Dental radiographs
AN (n=3) Wear Dental impressions
EDNOS (n=25) Radiographs Photos
2
 Dental impressions
Photographs
26
Philipp 1991 Germany In-patients 52 Psychiatric 52F BN: 25 (17–39) BN (n=41) (DSM III) Clinical assessment DMFS
and out- 50 Controls 50F AN: 22 (18–27) AN (n=11)
patients Control: 27 (17–37)
23
Robb 1995 UK Not stated 122 Psychiatric Not stated Not stated BN (n=39) Tooth Wear Index DMFS
122 Controls Vomiting AN (n=29)
Abstaining AN (n=54)
21
Rytomaa 1998 Finland Out-patients 35 Psychiatric 35F BN: 25.3 BN (n=35) (DSM-III-R) Clinical assessment DMFS Measured
105 Controls 105F Control: 25.7 X-rays
Dental impressions
Photographs
AN, anorexia nervosa; BN, bulimia nervosa; DFMT/S, decayed, filled and missing teeth/surfaces; DSM, Diagnostic and Statistical Manual; EDE, Eating Disorder Examination; EAT, Eating Attitude Test;
EDI, Eating Disorder Inventory; EDNOS, eating disorder not otherwise specified; F, female; M, male.

3
Association between poor oral health and eating disorders:
systematic review and meta-analysis
Steve Kisely, Hooman Baghaie, Ratilal Lalloo and Newell W. Johnson
BJP 2015, 207:299-305.
Access the most recent version at DOI: 10.1192/bjp.bp.114.156323

Supplementary Supplementary material can be found at:

Material http://bjp.rcpsych.org/content/suppl/2015/10/01/207.4.299.DC1

References This article cites 25 articles, 5 of which you can access for free at:
http://bjp.rcpsych.org/content/207/4/299#BIBL
Reprints/ To obtain reprints or permission to reproduce material from this paper, please
permissions write to permissions@rcpsych.ac.uk

You can respond /letters/submit/bjprcpsych;207/4/299

to this article at
Downloaded http://bjp.rcpsych.org/ on February 28, 2017
from Published by The Royal College of Psychiatrists

To subscribe to The British Journal of Psychiatry go to:

http://bjp.rcpsych.org/site/subscriptions/