Family Practice Vol. 16, No.

6
© Oxford University Press 1999 Printed in Great Britain

Diagnosing dermatomycosis in general practice

D Lousbergh, F Buntinx and G Piérarda

Lousbergh D, Buntinx F and Piérard G. Diagnosing dermatomycosis in general practice. Family

Practice 1999; 16: 611–615.
Background. Diagnosing dermatomycosis from a clinical image is not always easy. Micro-
scopy of a potassium hydroxide preparation (KOH-test) and culturing are seldomly used in
general practice. Cyanoacrylate surface skin scraping (CSSS) is a new diagnostic tool that may
be useful and simple.
Objectives. We aimed to investigate the diagnostic value of signs and symptoms, the KOH-
test and the CSSS, in patients with erythematosquamous skin lesions, using the culture as the
gold standard. Our goal is to formulate an optimal algorithm for the diagnosis of mycosis, based
on one or more of these tests and including both optimal accuracy and costs.
Methods. Scales from 148 consecutive general practice patients were tested using a KOH-test,
CSSS and culture. Clinical data were collected using a questionnaire.
Results. Twenty-six (18%) positive fungal cultures were identified. The sensitivity of the clinical
diagnosis was 81% and its specificity 45%; for the KOH-test, these figures were 12 and 93%
respectively; and for the CSSS, 62 and 88%, respectively. The positive predictive value of the
clinical diagnosis was 24% and the negative predictive value 92%; for the KOH-test these figures
were 25 and 83%, respectively, and for the CSSS, 52 and 92%, respectively. Determining CSSS
in all patients proved to be the most accurate policy (accuracy = 83%). The likelihood ratio of
CSSS in all patients was 5.17 for a positive test result and 0.43 for a negative test result. An
approach in which CSSS is obtained in only those patients whom the physician considers by
clinical examination to have dermatomycosis, with no testing in other patients, results in
positive and negative likelihood ratios of 4.69 and 0.56, respectively. Such a policy would result
in an overall sensitivity of 50%, a specificity of 89%, a positive predictive value of 50% and a
negative predictive value of 89%.
Discussion. The clinical picture of dermatomycosis is not very reliable. The combination of a
clinical judgement if this is negative and an additional CSSS in the case of a positive clinical
judgement provides us with the best cost–benefit ratio, if both diagnostic accuracy and logistic
considerations are taken into consideration.
Keywords. Cyanoacrylate surface skin scraping, dermatomycosis, potassium hydroxide test.

Introduction The cyanoacrylate surface skin scraping (CSSS) is a strip

In spite of its high prevalence (5.4/1000),1,2 dermatomycosis
is frequently misdiagnosed in general practice.3,4 The
diagnostic value of signs and symptoms is doubtful; using
the potassium hydroxide test (KOH) could confirm the
diagnosis, but is often passed over.5–12
Classical culture is the generally accepted gold
standard, but has no impact on bedside diagnosis.6,13–15
Received 22 February 1999; Revised 21 May 1999; Accepted
22 June 1999.
Department of General Practice, University of Leuven and
aDepartment of Dermatopathology, CHU Sart-Tilman Univer-
biopsy where a layer of horn is removed together with
the adhesive.16–28
A correct diagnosis of erythematosquamous skin
lesions is important, as cheap and safe treatment is
available. Inappropiate treatment with combinations
of corticosteroids and antimycotics should be avoided.
The possible side-effects of oral treatment have to be
considered, together with its cost.29–32
Therefore we started a study to examine the diagnostic
value of the signs and symptoms, using the KOH-test
and the CSSS method, and using the culture as a refer-
ence test, in patients with an erythematosquamous
skin lesion. We intended to formulate an optimal

sity of Liège, Belgium. Correspondence to Dr D Lousbergh, algorithm for the diagnosis of mycosis in these patients,
Academisch Centrum voor Huisartsgeneeskunde, Kapucijn- including both diagnostic accuracy and cost in the
envoer 33, Blok J, B-3000 Leuven, Belgium. calculations.

611
612 Family Practice—an international journal

Methods treatment. Patients diagnosed as without mycosis are

The study was carried out in 27 general practices in
Flanders. Consecutive patients with an erythemato-
squamous lesion of the glabrous skin were eligible, with
exclusion of lesions already treated with topical or
systemic antifungal therapy.
Patient characteristics, local signs and symptoms were
identified. The GP registered his diagnosis as either a
mycosis or another disorder, as well as his treatment.
Scales were collected from the border of the lesion, for
microscopic testing with KOH and culture.
A CSSS was made from the same site. A drop of cyano-
acrylate adhesive (Loctite glue®) was deposited on a sheet
of terephthalate polyethylene (Melinex O®, ICI) and
pressed firmly onto the degreased lesion. After about 30
seconds, a sheet of uniform thickness of stratum corneum
was removed. Within 24 hours it was sent to the labora-
tory of dermatopathology of the University of Liège.
Skin scrapings were inoculated onto Mycoline® slides.
Examination followed after incubation at 28–30°C for
4 weeks. Microscopy was performed with KOH 10%,
using pen ink for staining. The CSSS was coloured with
polychrome multiple stain. All slides were read by the
same experienced technologist.
Results of the clinical examination, the KOH pre-
paration and the CSSS were compared against culture
results. A decision tree was produced, comparing all pos-
sible combinations signs and symptoms with or without
KOH-test and/or CSSS. The accuracy of each resulting
2 × 2 table was used to decide which was the best
algorithm. For the combinations with the highest
accuracy, likelihood ratios for a positive (LR+) and for a
negative test result (LR–) were calculated as a con-
firmation. Financial costs of the diagnosis and initial
treatment results from one of the algorithms were based
on £17.60 for a CSSS, and £3.70 for local antifungal
supposed to be treated with local corticoids at £3.40.
Prices related to one package of a generally used product
in Belgium. To make comparisons possible, all patients
on antifungal treatment are considered to be treated
with local drugs. Financial costs are balanced against the
sum of false positive and false negative diagnoses.
Results
Thirty-four GPs participated in the study. A total of 170
specimens from the same number of different patients
were sent to the laboratory; 148 were admitted for
analysis.
The most common diagnoses, apart from dermato-
mycosis were Pytiriasis rosea (10%) and Nummular
dermatitis (9%). There were 40 positive cultures. After
eliminating 14 cultures, considered as a contamination
or as non-pathogenic, 26 positive fungal cultures were
identified. The most frequently isolated dermatophyte
was Trichophyton rubrum (n = 9).
In 21 cases with a positive culture, the clinical diag-
nosis of the GP was correct (sensitivity = 81%); the KOH
correctly identified three cases (12%) and the CSSS 16
cases (62%).
Out of the 122 negative specimens, 67 (55%) were
falsely diagnosed as a mycosis based on the clinical
examination, 9 (7%) were read as false KOH positive
and 15 (12%) as false CSSS positive.
A lesion diagnosed by the GP as a dermatomycosis
resulted in a positive culture in 24%. The negative
predictive value of the clinical diagnosis was 92%. A
specimen read as KOH positive resulted in a positive
culture in 25%. Its negative predictive value was 83%.
The CSSS showed a positive predictive value of 52%
and a negative predictive value of 92% (Table 1).

TABLE 1 Diagnostic value of clinical examination, KOH-test and CSSS using culture as the gold standard (n = 148)

Test Culture

+ –

test+ test– test+ test– sens spec PPV NPV LR+ LR–
(95% CI) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI)

Clincial 21 5 67 55 0.81 0.45 0.24 0.92 1.47 0.42

examination
(0.60–0.93) (0.36–0.54) (0.16–0.34) (0.81–0.97) (1.15–1.88) (0.19–0.96)
KOH-test 3 23 9 113 0.12 0.93 0.25 0.83 1.71 0.95
(0.03–0.31) (0.86–0.96) (0.07–0.57) (0.76–0.89) (0.45–5.38) (0.82–1.11)
CSSS 16 10 15 107 0.62 0.88 0.52 0.92 5.17 0.43
(0.41–0.79) (0.80–0.93) (0.33–0.69) (0.85–0.96) (2.85–8.79) (0.27–0.72)

KOH = potassium hydrochloride test; CSSS = cyanocrylate surface skin scraping; 95% CI = 95% confidence interval; PPV = positive predictive
value; NPV = negative predictive value; LR = likelihood ratio.
 Diagnosing dermatomycosis in general practice 613
The sensitivity of individual signs and symptoms in all patients, irrespective of the clinical picture, proved
ranged between 4% (yellow colour) and 77% (minimal to be the most effective policy (accuracy = 83%). In
scaling), and the specificity between 20% (minimal patients with a positive clinical diagnosis, the accuracy of
scaling) and 96% (brown colour) (Table 2). the CSSS was considerably higher compared with the
After calculating the accuracy of each possible next best (KOH-test with or without CSSS). In patients
combination of test results (Table 3), determining CSSS with a negative clinical diagnosis, however, the difference

TABLE 2 Diagnostic value of clinical signs, using culture as the gold standard (n = 148, of which 26 dermatomycosis)

No. of positive test results Sensitivity Specificity PPV NPV LR+ LR–

Red colour 102 0.69 0.31 0.18 0.83 1.00 1.00

95% CI 0.48–0.85 0.23–0.40 0.11–0.27 0.68–0.92
Brown colour 7 0.08 0.96 0.29 0.83 2.00 0.96
95% CI 0.01–0.27 0.90–0.99 0.05–0.70 0.76–0.89
Yellow colour 7 0.04 0.95 0.14 0.82 0.80 1.01
95% CI 0.002–0.21 0.89–0.98 0.008–0.58 0.75–0.88
Mixed colour 32 0.19 0.78 0.16 0.82 0.86 1.04
95% CI 0.07–0.40 0.69–0.85 0.06–0.34 0.73–0.88
Desquamation absent 13 0.04 0.90 0.07 0.82 0.40 1.07
95% CI 0.02–0.22 0.83–0.95 0.04–0.38 0.74–0.87
Minimal scaling 118 0.77 0.20 0.17 0.80 0.96 1.15
95% CI 0.56–0.90 0.13–0.28 0.11–0.25 0.61–0.92
Thick plaque 17 0.19 0.90 0.29 0.84 1.90 0.90
95% CI 0.07–0.40 0.83–0.95 0.11–0.56 0.76–0.90
Waxy appearance 10 0.08 0.93 0.20 0.83 1.14 0.99
95% CI 0.01–0.27 0.87–0.97 0.04–0.56 0.75–0.88
Concentric rings 31 0.27 0.80 0.23 0.84 1.35 0.91
95% CI 0.12–0.48 0.72–0.87 0.10–0.42 0.76–0.90
Central healing 54 0.42 0.65 0.20 0.84 1.20 0.89
95% CI 0.24–0.63 0.56–0.73 0.11–0.34 0.75–0.91
Infiltration 41 0.31 0.73 0.20 0.83 1.15 0.95
95% CI 0.15–0.52 0.64–0.80 0.09–0.35 0.74–0.90
Exudation 26 0.19 0.83 0.19 0.83 1.12 0.98
95% CI 0.07–0.40 0.75–0.89 0.07–0.40 0.75–0.89
Incrustation 39 0.35 0.75 0.23 0.84 1.40 0.87
95% CI 0.18–0.56 0.67–0.83 0.12–0.40 0.76–0.90
Vesiculation 20 0.04 0.84 0.05 0.81 0.25 1.14
95% CI 0.002–0.22 0.77–0.90 0.003–0.27 0.72–0.87
Pustules 7 0.07 0.96 0.29 0.83 1.75 0.97
95% CI 0.01–0.27 0.90–0.99 0.05–0.70 0.76–0.89
Satellite lesions 64 0.39 0.56 0.16 0.81 0.89 1.09
95% CI 0.21–0.59 0.47–0.65 0.08–0.27 0.71–0.88
Fissures 24 0.23 0.85 0.25 0.84 1.53 0.91
95% CI 0.09–0.44 0.77–0.91 0.11–0.47 0.76–0.90
Maceration 27 0.27 0.84 0.26 0.84 1.69 0.87
95% CI 0.12–0.48 0.76–0.90 0.12–0.47 0.76–0.90
Pruritis or itching 87 0.54 0.40 0.16 0.80 0.90 1.15
95% CI 0.34–0.73 0.32–0.49 0.09–0.26 0.68–0.89
614 Family Practice—an international journal

TABLE 3 Results of each branch of the decision tree when diagnosing dermatomycosis starting from the clinical examination result in patients with
an erythematosquamous skin lesion (n = 148)

T+G+ T+G– T–G+ T–G–

Clinical picture only 21 67 5 55

(n = 148)
KOH-test only 3 9 23 113
(n = 148)
CSSS only 16 15 10 107
(n = 148)

Clinical diagnosis no additional tests 21 67 0 0

positive (n = 88)
KOH-test 2 7 19 60
CSSS 13 13 8 54
CSSS 2 7 0 0
KOH pos. (n = 9)
no additional test 2 7 0 0
CSSS 11 6 8 54
KOH neg. (n = 79)
no additional test 0 0 19 60

Clinical diagnosis no additional test 0 0 5 55

negative (n = 60)
KOH-test 1 2 4 53
CSSS 3 2 2 53
CSSS 1 2 0 0
KOH pos. (n = 3)
no additional test 1 2 0 0
CSSS 2 0 2 53
KOH neg. (n = 57)
no additional test 0 0 4 53

between all possible policies was minimal. In these
patients not implementing any additional test resulted in
only one additional false result. For the whole group, the
combination of accepting a negative clinical judgement,
and performing a CSSS in patients with a clinical picture
of dermatomycosis only, resulted in an accuracy that was
almost identical to performing a CSSS on all patients
(82%).
The likelihood ratio of CSSS on all patients was 5.17
(95% CI = 2.85–8.75) for a positive test result and 0.43
(95% CI = 0.27–0.74) for a negative test result. When
applying CSSS to patients with a positive clinical picture,
with no additional test in patients with a negative clinical
picture, these values were 4.69 (95% CI = 2.47–8.91) and
0.56 (95% CI = 0.38–0.83), respectively. Such policy
would result in an overall sensitivity of 50%, a specificity
of 89%, a positive predictive value of 50% and a negative
predictive value of 89%.
The total financial cost per patient for the actual
treatment, for the prescription of a CSSS test in all
patients and for the prescription of a CSSS test in
patients with a clinical diagnosis of mycosis only,
followed by the consequent local treatment, are
estimated at £3.60, £21.00 and £14.00, respectively.
Discussion
The prevalence of dermatomycosis in general practice is
high1,2 and commonly misdiagnosed by GPs.3,4 Elements
of the clinical picture are not very reliable. Although
its sensitivity (81%) is highly acceptable, more than
three out of four clinical diagnoses could not be con-
firmed by culture. According to our results, 71% of the
oral treatment and 76% of the topical treatment were
inappropiate.
The KOH test results in significant additional
diagnostic value if the GP suspects a positive diagnosis
based on the clinical picture. The optimal diagnostic
value was reached with a CSSS test performed on all
patients (LR+ = 5.17, LR– = 0.43). However, compared
with a clinical diagnosis only, the only benefit of this
test is for patients with a clinical picture suggesting
dermatomycosis.
 Diagnosing dermatomycosis in general practice
Considering the low extra diagnostic benefit, the ad-
ditional costs and the logistic disadvantages if all patients
were tested with CSSS, our results support a policy to
treat patients with a negative clinical diagnosis as a
patient without dermatophytosis. On the other hand,
patients with a positive clinical picture have to be tested
using CSSS. This algorithm results in a LR+ of 4.96 and a
LR– of 0.56.
Acknowledgements
We thank the 34 GPs for participating in this study,
Mr Pierre Stefan of the Laboratory of dermatopathology
of the University of Liège for his technical support,
Dr Janssens Monique and Mr Gillé Dirk of the Janssen
Research Foundation for their technical and scientific
support, Dr Van Cutsem J of Janssen Pharmaceutica for
his scientific advice, and Mr Blanckaert of ICI Belgium
and Mr De Filette of Loctite Belgium for their material
support.
References
1 Registratienet van huisartspraktijken. Gezondheidsproblemen
en diagnosen in huisartspraktijken 1993; 5: 1 (maart 1993).
Maastricht: Rijksuniversiteit Limburg, 1993.
2 Van den Hoogen HJM, Huygen FJA, Schellekens JWG et al.
Morbidity figures from general practice. Data from four general
pactices 1977–1982. Nijmegen: Nijmegen University Depart-
ment of General Practice, 1985.
3 Pariser DM. Superficial fungal infections. A practical guide for
primary care physicians. Postgrad Med 1990; 87: 205–214.
4 De Doncker P. Randomized, double-blind trial of the efficacy and
tolerability of itraconazole 220 mg compared with terbinafine
250 mg in the treatment of superficial dermatomycoses of the
glabrous skin. Janssen Research Foundation Belgium, Clinical
Research Report ITR-INT-47, 1995 Oct. (N 112826).
5 de Kock CA, Sampers GH, Knottnerus JA. Diagnosis and manage-
ment of cases of suspected dermatomycosis in the Netherlands:
influence of general practice based potassium hydroxide
testing. Br J Gen Pract 1995; 45(396): 349–351.
6 Masri-Fridling GD. Dermatophytosis of the feet. Dermatol Clin
1996; 14: 33–40.
7 Miller MA. Sensitivity and specificity of potassium hydroxide
smears in skin scrapings for the diagnosis of tinea pedis (letter).
Arch Dermatol 1993; 129: 510–511.
8 Salomon RJ. The sensivity and specificity of the potassium hydroxyde
smear (letter; comment). Arch Dermatol 1993; 129: 1342–1343.
9 Teunissen AKK, Van der Voort PHJ, Nieboer C, De Haan P, van
Dijk E. Diagnostiek van dermatomycosen door middel van de
sellotape-KOH-methode. Ned Tijdschr Geneeskd 1989; 133:
615–617.
615
10
Van de Poel GT, Lamberts H. Schimmelinfecties in de huisarts-
praktijk. Huisarts Wet 1981; 24(S5): 32–36.
11 Feld R, Jacobs A, Middag-Broekman J. Dermatomycosen van kop
tot teen. Beerse Janssen Pharmaceutica B.V., 1991.
12 Brodell RT, Helms SE, Snelson ME. Office dermatologic testing:
the KOH preparation. Am Fam Physician 1991; 43: 2061–2065.
13 Nielsen PG. A comparison between direct microscopy and
culture in dermatological mycotic material. Mykosen 1981; 24:
555–560.
14 Strauss JS, Kligman AM. An experimental study of tinea pedis and
onychomycosis of the foot. Arch Dermatol 1957; 76: 70–79.
15 van Dijk E. Mycologisch onderzoek van huidziekten: de kweek.
Ned Tijdschr Geneeskd 1984; 128: 513–515.
16 Gip LJ, Nilsson J. Skin sampling for dermatophytes with adhesive
tape. Castellania 1976; 4: 25–26.
17 Nielsen PG. Two different methods for direct microscopy of mycotic
material. Mykosen 1981; 25: 270–273.
18 Goldschmidt H, Kligman AM. Exfoliative Cytology of Human
Horny Layer; methods of cell removal and microscopic
techniques. Arch Dermatol 1967; 96: 572–576.
19
Marks R, Dawber RPR. Skin surface biopsy: an improved technique
for the examination of the horny layer. Br J Dermatol 1971, 84:
117–123.
20 Agache P, Mairey J, Boyer JP. Le stripping du stratum corneum au
cyanoacrylate; intérêt en physiologie et en pathologie cutanées.
J Médecine Lyon 1972; 53: 1017–1022.
21 Pierard-Franchimont C, Pierard GE. Skin surface stripping in diag-
nosing and monitoring inflammatory, xerotic and neoplastic
diseases. Pediatr Dermatol 1985; 2: 180–184.
22 Rurangirwa A, Pierard-Franchimont C, Pierard GE. Growth of
Candida albicans on the stratum corneum of diabetic and non-
diabetic patients. Mycoses 1989; 33: 253–255.
23 Rurangirwa A, Pierard-Franchimont C, Pierard GE. Culture of
fungi on cyanoacrylate skin surface strippings—a quantitative
bioassay for evaluating antifungal drugs. Clin Exp Dermatol
1989; 14: 425–428.
24 Pierard GE, Pierard-Franchimont C, Dowlati A. La biopsie de
surface en dermatologie clinique et expérimentale. Rev Eur
Dermatol MST 1992; 4: 455–466.
25 Nikkels AF, Pierard-Franchimont C, Pierard GE. Oppervlakte-
biopsie. Ned Tijdschrift voor Dermatologie & Venereologie
1993; 3: 362–364.
26 Lachapelle JM, Gouverneur JC, Boulet M, Tennstedt D. A modified
technique (using polyester tape) of skin surface biopsy.
Br J Dermatol 1977; 97: 49–52.
27 Whiting DA, Bisset EA. The investigation of superficial fungal
infections by skin surface biopsy. Br J Dermatol 1974; 91: 57.
28 Knudsen EA. Growth on Sabouraud’s agar of dermatophytes
obtained by conventional scraping and by stripping with vinyl
tape. Br J Dermatol 1974; 90: 163.
29 Bergus GR, Johnson JS. Superficial Tinea infections. Am Fam
Physician 1993; 48: 259–268.
30 Knight TE, Shikuma CY, Knight J. Ketoconazole-induced hepatitis
necessitating liver transplantation. J Am Acad Dermatol 1991;
25: 398–400.
31 Janssen PA, Symoens JE. Hepatic reactions during ketoconazole
treatment. Am J Med 1983; 74: 80–85.
32 Lewis JH, Zimmerman HJ, Benson GD, Ishak KG. Hepatic injury
associated with ketoconazole therapy. Analysis of 33 cases.
Gastroenterology 1984; 86: 503–513.