o

ral Candidiasis in Patients With

Diabetes Mellitus: A Thorough Analysis
GARY A. BARTHOLOMEW, DMD, BRAD RODU, DPS, AND DAVID S. BELL, MD

It has generally been assumed that oral candidiasis occurs with increased frequency in patients with
diabetes mellitus. To evaluate this, we compared the frequency and severity of oral Candida colonization
in 60 patients with insulin-dependent diabetes mellitus (IDDM) admitted to a low-intensity-care diabetes
unit with those in 57 age- and sex-matched controls. Swabs taken from the tongue and buccal mucosa
were examined by cytology rather than culture because of the discrimination provided by the former.
Cytological smears were classified according to the presence and morphology of the Candida organisms.
Overall, a significant difference in Candida species colonization was found between patients with diabetes
(75.0%) and controls (35.1%) (P < .005). In the diabetic group, no relationship was found to recent
use of antibiotics, total or differential white blood cell count, serum glucose, presence of diabetic
retinopathy, or glycosylated hemoglobin values. We conclude that in IDDM there is a predisposition
to oral candidiasis and that this predisposition is independent of glucose control. Diabetes Care 10:60712, 1987

andidiasis is an infection caused by any of several

species of the fungus Candida. By culture, Candida
is considered normal oral flora at a frequency of
44-55% (1). This observation is supported by
well-controlled studies with exfoliative cytology (1,2). Although numerous candidal species can be isolated from the
oral cavity, the predominant species affecting hospitalized
patients is Candida albicans (3,4).
Clinical diagnosis of oral candidiasis relies on the recognition of granular, erosive, and pseudomembranous forms of
the infection, with the easily removed curdlike plaques of
the latter being the most common (5). However, substantial
colonization can exist in the absence of clinical lesions (2).
Although cultures have been favored for confirmation of
clinical infections, periodic acid-Schiff-stained cytologic
smears are also an excellent method; they hold a marked
time advantage over and are less costly than cultures (1). In
fact, cytology offers a further advantage in facilitating a distinction on morphological grounds between a carrier state
and active infection. Because the pseudohyphal phase is considered the invasive phase of the fungus, the diagnosis of
mucosal candidiasis relies on the demonstration of these
forms, as well as blastospores (6). The high carrier rate of
Candida in a normal population emphasizes the advantage of
cytology because it has been stated that positive cultures by

themselves are inadequate for the diagnosis of oral candidiasis (7).

Groups classically considered at increased risk for candidal
infections include cancer patients and those receiving antibiotics and supraphysiologic doses of corticosteroids or other
immunosuppressants (5). In addition, candidal infections are
believed to be more frequent in people with diabetes (5,810). It has been suggested that the highest rate of colonization
occurs in diabetic patients with poor serum glucose control
(11), although proof supporting this association is lacking
(8,9).
The purpose of this study was to define the prevalence of
candidal colonization in people with insulin-dependent diabetes mellitus (IDDM) versus controls and to identify factors
predisposing to colonization in the diabetic patient. The
variables evaluated include absolute white blood cell counts
and differentials, glycosylated hemoglobin (GHb) levels,
serum glucose, recent antibiotic or immunosuppressant therapy, duration of diabetes, and diabetic retinopathy.
MATERIALS AND METHODS

Oral cytology was performed on 60 patients with IDDM on

admission to a low-intensity-care diabetes unit and 57 ageand sex-matched controls (from the Oral Diagnosis Clinic,

DIABETES CARE, VOL. 10 NO. 5, SEPTEMBER-OCTOBER 1987

607

ORAL CANDIDIASIS IN IDDM/G. A. BARTHOLOMEW, B. RODU, AND D. S. BELL

B
608

DIABETES CARE, VOL. 10 NO. 5, SEPTEMBER-OCTOBER 1987

ORAL CANDIDIASIS IN 1DDM/G. A. BARTHOLOMEW, B. RODU, AND D. S. BELL

FIG. 1. Cytological preparations showing negative smear (A, x 100), scattered budding yeast forms (B, X 160), and proliferation
of pseudohyphae (C, X 125).

University of Alabama School of Dentistry). Mucosal scrapings were obtained from the buccal mucosa and posterior
dorsum of the tongue of each patient with a tongue blade
moistened with water. These scrapings were smeared over a
glass microscope slide and sprayed with a commercially prepared fixative (Surgipath, Medical Industries, Grays Lake,
IL). The slides were coded and submitted to the Oral Pathology Laboratory, where they were stained with periodic
acid-Schiff reagent.
All slides were examined microscopically for the presence
of Candida. Screening was performed at X 100 with verification of yeast and pseudomycelial forms completed at X 450.
All evaluations were performed in a blind manner by one of
us (B.R.).
Microscopic findings were grouped into the following categories in the same manner as previous investigations
(1,12,13; Fig. 1):
0: negative, adequate numbers of epithelial cells with
no evidence of fungi.
+ C: noninvasive colonization, scattered collections of
yeast forms in association with epithelial cells.
+ 1: invasive colonization, variable numbers of intertwining pseudomycelial forms.
On admission, the following information was obtained
from the study group: age, sex, type and duration of diabetes,
and reasons for admission. A detailed history was taken with

regard to current and recent (within 1 mo before admission)

medications, with particular attention to the use of antibiotics or immunosuppressant therapy. An oral soft tissue
examination was performed to rule out clinical candidiasis.
Each patient was assessed by an ophthalmologist for the presence of diabetic retinopathy. A white blood cell count and
differential, serum glucose (performed by glucose oxidase
method), and GHb value (from ion-exchange chromatography, Isolab Fast Hemoglobin method) were also obtained
on admission.
A complete history and oral examination were obtained
from the control group. Controls who were on immunosuppressant drugs or had a history of recent antibiotic use or
malignancy were excluded from the study.
The prevalence of colonization was compared by x 2 'dis'
tribution with the appropriate degrees of freedom. The G
TABLE 1
Comparison of negative, noninvasive, and invasive colonization
Diabetic (n = 60)
Negative (0)
Noninvasive ( + C)
Invasive ( + 1 )

15 (25.0)
16(26.7)
29 (48.3)

Control (n

::

57)

37 (64.9)
8 (14.0)
12 (21.1)

G = 19.5. x:.eos,2 10.6. Values in parentheses are percentages.

DIABETES CARE, VOL. 10 NO. 5, SEPTEMBER-OCTOBER 1987

609

ORAL CANDID1ASIS IN 1DDM/G. A. BARTHOLOMEW, B. RODU, AND D. S. BELL

TABLE 2
Factors associated with candidal colonization in diabetic subjects

Negative (0)
Noninvasive ( + C)
Invasive ( + 1)

Glycosylated
hemoglobin (%)

Blood glucose
(ing/100 ml)

Duration of diabetes (yr)

Total white
blood cells

Segmented leukocytes

Lymphocytes

10.14
10.55
10.96

297.2
238.9
232.1

9.9
10.2
14.8

7078
8219
7853

7062
5134
5658

2141
2467
2378

statistic for a 3 X 2 contingency table was calculated. The

variables were classified and compared according to the original categories (0, + C , +1) by calculating a linear discriminant function based on an assumption of normal occurrence
within each variable. Antibiotic history and retinopathy were
evaluated by calculating the G statistic for a 3 X 2 contingency table.

(G = 23.06). In the group with diabetes, retinopathy did

not show any correlation with colonization status (Table 3).
DISCUSSION

uch has been written about candidiasis in the

diabetic patient. Most statements made have
been in reference to vulvovaginal infections.
These observations are primarily supported by
RESULTS
empirical clinical observations, and there has been a paucity
The mean age of the study population was 31.8 yr (range 9 of properly controlled experimental studies. We thought that
76 yr), and the controls had a mean age of 30.1 yr (range the oral mucosa of the IDDM patient is another likely site
9-66 yr). In both the diabetic and control populations, the of candidal colonization and chose to investigate this further.
We have used oral cy to logical smears as a simple, repromale-to-female ratio was 3:2.
In the study population the primary reasons for admission ducible means of demonstrating a significant increase in
were poorly controlled diabetes mellitus (39 of 60), diabetic frequency of candidal colonization in diabetic patients
ketoacidosis (13 of 60), surgical treatment of ophthalmologic compared with age- and sex-matched healthy controls. With
complications (6 of 60), and acute infection (2 of 60). Other well-defined microscopic criteria, this technique can semicommon secondary problems on admission included hyper- quantitatively illustrate actual candidal colonization. Theretension, peripheral vascular disease, retinopathy, neuropa- fore, we believe the cytological approach is not only valuable
for diagnosis but may play an equally important role in dethy, and chronic infections.
Differences in the distribution of Candida colonization be- fining more fundamental pathogenetic aspects of the infectween diabetic patients and controls were statistically sig- tion.
nificant (P < .005; Table 1). Oral examinations revealed
We are aware of the potential pitfalls in comparing a hosno clinical infections in either the study or control groups, pitalized population with an outpatient group. However, we
indicating that the colonization was subclinical. In the group attempted to minimize these disadvantages by I) conducting
with diabetes, all variables in Table 2, when subjected to a the study in a low-intensity-care unit, 2) cytologically screenlinear discriminant function analysis, resulted in an ex- ing patients on admission, and 3) analyzing numerous potremely high miscalculation rate, indicating that the predic- tentially complicating factors. Several parameters were
tive quality of the collective body of variables or any subset evaluated to identify overt factors responsible for this inthereof was poor.
Evaluation of variables related to hospital admission, such TABLE 3
as antibiotic therapy, retinopathy, and ketoacidosis, was perAntibiotic treatment, retinopathy, and ketoacidosis in diabetic group
formed (Table 3). Many patients with diabetes were on mulAntibiotic
tiple drugs treating either acute or long-term cardiovascular
history*
Ketoacidosist
Retinopathyt
complications of diabetes. In general, most of these medications also had no significant impact on the study. ComPositive Negative Positive Negative Positive Negative
parison of antibiotic history among microscopic categories
showed no significance. The two patients with acute infec- Negative
4
14
6
12
6
9
(0)
tions had negative and noninvasive cytology results. No paNoninvasive
tient in the control group was under therapy.
3
11
2
11
2
14
(+Q
The 13 patients in the study group with ketoacidosis were
Invasive
distributed fairly evenly among the microscopic categories
9
8
20
20
5
24
(+D
(negative, 6; noninvasive colonization, 2; and invasive colonization, 5; Table 3). Elimination of this subgroup from the *G = 0.36 (NS).
study population in Table 1 resulted in an increased colo- tG = 1.53 (NS).
nization rate in patients with diabetes compared with controls tG = 3.81 (NS).

610

DIABETES CARE, VOL. 10 NO. 5, SEPTEMBER-OCTOBER 1987

ORAL CAND1DIASIS IN 1DDM/G. A. BARTHOLOMEW, B. RODU, AND D. S. BELL

creased colonization in diabetic patients. Although examination of white blood cell counts revealed no correlation to
colonization status, the complex nature of the defense mechanisms against the Candida organism suggests that covert
functional defects of polymorphonuclear leukocytes (14-19)
or cell-mediated immunity (20,21) may play an important
role in this population.
The predisposition of the diabetic patient to infection by
pathogenic fungal species has been explained in terms of
enhancement of yeast growth by elevated tissue fluid glucose
levels (22). In addition, a good correlation between salivary
glucose and Candida growth has been demonstrated in diabetic patients (23). Thus, a correlation between diabetic
control and extent of oral mucosal yeast colonization would
be expected. We evaluated serum glucose levels on admission
because subsequently these levels were rapidly adjusted for
optimal control. GHb levels were evaluated as an index of
control over a more extended period of 2-3 mo (24). However, we were unable to correlate diabetic control with frequency of colonization, as other studies have shown (25).
In addition, diabetic ketoacidosis was not important in determining colonization patterns (Table 3).
The availability of salivary glucose may influence Candida
growth during antibiotic administration due to a selective
reduction in oral microflora and subsequent decrease in competition for the nutrient (25). Diabetic patients receiving
antibiotics might be expected to demonstrate increased candidal colonization. We were unable to establish any correlation, although only 25% of our diabetic patients were in
this group (Table 3).
Vascular compromise has been related to the frequency
and severity of certain infections in patients with diabetes
(10). This may be due to an exaggeration of immunologic
deficits by proliferative changes in the capillary endothelial
basement membrane causing impedance of leukocyte movement and diffusion of necessary nutrients into tissues (18).
In addition, it has been shown that diabetic retinopathy is
associated with duration and age at onset of diabetes while
also indicating the degree of vascular compromise (5). However, we failed to correlate candidal colonization with either
the duration of diabetes or the presence of diabetic retinopathy.
Despite solid evidence that candidal colonization is more
prevalent in patients with IDDM than in age- and sexmatched controls, the factors responsible are largely unknown. Although fragmentary and circumstantial evidence
of immunologic and metabolic defects has been found, a
broader perspective on this infection remains elusive.
ACKNOWLEDGMENTS:

We thank Carl Russell, DMD, for the

statistical analysis.
From the University Hospital Dental Clinic (G.A.B), the Department of Pathology (B.R.), and the Division of Endocrinology
Metabolism (D.S.B.), University of Alabama Schools of Medicine
and Dentistry, Birmingham, Alabama.

Address correspondence and reprint requests to Brad Rodu,

DDS, Box 440, University Station, U.A.B., Birmingham, AL
35294.

REFERENCES

1. Arendorf TM, Walker DM: The prevalence and intraoral distribution of Candida albicans in man. Arch Oral Biol 25:1-10,
1980
2. Rodu B, Griffin IL, Gockerman JP: An assessment of oral
candidiasis in cancer patients. South Med ] 77:31214,
1984
3. Shipman B: Clinical evaluation of oral Candida in cancer chemotherapy patients. ] Prosthet Dent 41:63-67, 1979
4- Odds FC, Evans E: Distribution of pathogenic yeasts and humoral antibodies to Candida among hospital patients. J Clin
Pathol 33:750-56, 1980
5. Bennett JE: Candidiasis. In Harrison's Principles of Internal Medicine. 9th ed. Isselbacher KJ, Adams RD, Braumwald E, PetersdorfRG, Wilson JD, Eds. New York, McGraw-Hill, 1980,
p. 741-42
6. Hoffman H: Mycology. In Oral Microbiohgy. 2nd ed. Nolte
WL, Ed. St. Louis, MO, Mosby, 1973, p. 206
7. Kobayashi GS: Fungi. In Microbiology. 3rd ed. Davis BD, Dulbecco R, Eisen HN, Ginsberg HS, Eds. Philadelphia, PA,
Harper & Row, 1980, p. 818-50.
8. Wheat JL: Infection and diabetes mellitus. Diabetes Care 3:18797, 1980
9. Bagdade JD: Infection in diabetes. Postgrad Med 59:160-64,
1976
10. Edwards JE, Tillman DB, Miller ME, Pitchon HE: Infection
in diabetes mellitus. West J Med 130:515-21, 1979
11. Rayfield EJ, Auk MJ, Kensch GT, Brothers MT, Nechemias
C, Smith A: Infection and diabetes: the case for glucose control. Am J Med 72:439-50, 1982
12. Budtz-Jorgensen E, Stenderup A, Grabowski M: An epidemiologic study of yeasts in elderly denture wearers. Community Dent
Oral Epidemiol 3:115-19, 1975
13. Davenport JC: The oral distribution of Candida in denture
stomatitis. Br Dent J 129:151-56, 1970
14- Jackson RA, Bryan CS, Weeks BA: Phagocytosis of Candida
albicans by polymorphonuclear leukocytes from normal and diabetic subjects. Adv Exp Med Biol 121:33-37, 1979
15. Bybee MD, Rogers DE: The phagocytic activity of polymorphonuclear leukocytes obtained from patients with diabetes
mellitus. J Lab Clin Med 64:1-13, 1964
16. Miller ME, Baker L: Leukocyte function in juvenile diabetes
mellitus: hormonal and cellular aspects. ] Pediatr 81:979-82,
1972
17. Bagdade JD, Mielson KL, Bulger RJ: Reversible abnormalities
in phagocyte function in poorly controlled diabetic patients.
AmJ Med Sci 263:452-56, 1972
18. Mowat AG, Baum J: Chemotaxis of polymorphonuclear leukocytes from patients with diabetes mellitus. N Engl J Med
284:621-27, 1971
19. Hill HR, Sauls HS, Dettloff JL, Quie PG: Impaired leukotactic
responsiveness in patients with juvenile diabetes mellitus. Clin
Immunol Immunopathol 2:395-403, 1974
20. Brody JI, Merlie K: Metabolic and hiosynthetic features of
lymphocytes from patients with diabetes mellitus: similarities
to lymphocytes in chronic lymphocytic leukemia. BrJ llaemotol
19:193-201, 1970

DIABETES CARE, VOL. 10 NO. 5, SEPTEMBER-OCTOBER 1987

611

ORAL CANDIDIASIS IN IDDM/G. A. BARTHOLOMEW, B. RODU, AND D. S. BELL

21. Delespesse G, Duchateau J, Bastenic PA: Cell mediated immunity in diabetes mellitus. Clin Exp Immunol 18:461-67,
1974
22. Knight L, Fletcher J: Growth of Candida aibicans in saliva:
stimulation of glucose associated with antibiotics, corticosteroids and diabetes mellitus. ] Infect Dis 123:371-77, 1971
23. Odds FC, Evans EGV, Taylor MAR, Wales JK: Prevalence of

612

pathogenic yeasts and humoral antibodies to Candida in diabetic

patients. J Clin Pathol 31:840-44, 1978
24. Gabbay KH: Glycosylated hemoglobin and diabetes mellitus.
Med Clin N Am 66:1309-15, 1982
25. Tapper-Jones LM, Aldred MJ, Walker DM, Hayes TM: Candidal infections and populations of Candida aibicans in mouths
of diabetics. ] Clin Pathol 34:706-11, 1981

DIABETES CARE, VOL. 10 NO. 5, SEPTEMBER-OCTOBER 1987