Objective evaluation of clubbing on shadow images of index fingers: a study of patients with pulmonary disease and of normal individuals

Abstracts

BACKGROUND: Normal diagnosis of clubbing is clinical; however use of objective criteria may improve the accuracy of findings OBJECTIVE: To present a simple method of obtaining finger images for the purpose of studying clubbing. METHOD: Shadow images of the index fingers obtained by projection through a transparent glass plate virtually without distortion and displayed on a common sheet of paper yielded the profile (PA) and hyponychial (HA) angles; as well as the ratio between distal phalangean and interphalangean depths (DPD/IPD). Upon physical examination of 306 adult bearers of pulmonary disease, 116 disclosed presence of clubbing (YES); 126 absence (NO); and 64 were doubtful cases (DBT). Also studied were 452 normal adult individuals. Among these 71.0% of the bearers and 33.4% of the controls were smokers. RESULTS: Values found in normal individuals and in patients bearers of clubbing (YES) were, respectively, 172.8±5.9° vs. 183.4±5.9° for PA, 181.5±5.0° vs. 201.4±6.5° for HA, and 0.904±0.029 vs. 1.014±0.062 for DPD/IPD (significant differences, p<0.05). CONCLUSIONS: This is a simple method of obtaining clear index finger images. The hyponychial angle determined on the images was the most useful measurement to discriminate digits clinically with and without clubbing (sensitivity of 76.7%, specificity of 83.2%, predictive positive value of 95.5% and predictive negative value of 96.9%).

Clubbing; profile angle; hyponychial angle; relation DPD/IPD

INTRODUÇÃO: O diagnóstico do hipocratismo digital é clínico, mas pode tornar-se mais acurado pelo uso de critérios objetivos de determinação. OBJETIVO: Mostrar um método simples de obtenção de imagens de dedos para estudo do hipocratismo digital. MÉTODO: Em imagens de dedos indicadores em perfil, obtidas praticamente sem distorção em folha de papel comum, projetadas através de lâmina de vidro, determinaram-se os ângulos do perfil e hiponiquial, e a relação entre as espessuras falangeana distal e interfalangena. Estudaram-se 306 pneumopatas adultos, clinicamente com hipocratismo presente, ausente ou duvidoso, e 452 indivíduos adultos normais. Eram fumantes 71,0% dos pacientes e 33,4% dos controles. RESULTADOS: Os valores encontrados nos indivíduos normais e nos pacientes nos quais havia a presença clínica de hipocratismo foram, respectivamente, 172,8 ± 5,3º e 183,4 ± 5,0º para ângulo de perfil, 181,5 ± 4,8º e 201,4 ± 6,5º para ângulo hiponiquial, e 0,904 ± 0,029 e 1,014 ± 0,062 para a relação entre as espessuras falangeana distal e interfalangeana. As diferenças foram significativas. Os casos duvidosos também tiveram valores maiores que os verificados nos controles. No grupo controle, observou-se que os fumantes masculinos apresentaram os três valores significativamente superiores aos exibidos pelos não fumantes, enquanto que nas mulheres fumantes desse grupo apenas a relação entre as espessuras falangeana distal e interfalangeana se mostrou superior. CONCLUSÃO: Imagens nítidas de dedos indicadores foram facilmente obtidas pelo método empregado. O ângulo hiponiquial, determinado a partir das imagens, foi a medida que se mostrou com maior capacidade para discriminar casos com e sem hipocratismo (sensibilidade de 76,7%, especificidade de 83,2%, valores preditivos positivo de 95,8% e negativo de 96,9%).

Hipocratismo digital; baqueteamento; ângulo do perfil; ângulo hiponiquial; relação entre espessuras do dedo

ORIGINAL ARTICLE

Objective evaluation of clubbing on shadow images of index fingers. A study of patients with pulmonary disease and of normal individuals^*

José da Silva Moreira; Nelson da Silva Porto; Ana Luiza Schneider Moreira

^{Correspondence}

ABSTRACT

BACKGROUND: Normal diagnosis of clubbing is clinical; however use of objective criteria may improve the accuracy of findings

OBJECTIVE: To present a simple method of obtaining finger images for the purpose of studying clubbing.

METHOD: Shadow images of the index fingers obtained by projection through a transparent glass plate virtually without distortion and displayed on a common sheet of paper yielded the profile (PA) and hyponychial (HA) angles; as well as the ratio between distal phalangean and interphalangean depths (DPD/IPD). Upon physical examination of 306 adult bearers of pulmonary disease, 116 disclosed presence of clubbing (YES); 126 absence (NO); and 64 were doubtful cases (DBT). Also studied were 452 normal adult individuals. Among these 71.0% of the bearers and 33.4% of the controls were smokers.

RESULTS: Values found in normal individuals and in patients bearers of clubbing (YES) were, respectively, 172.8±5.9° vs. 183.4±5.9° for PA, 181.5±5.0° vs. 201.4±6.5° for HA, and 0.904±0.029 vs. 1.014±0.062 for DPD/IPD (significant differences, p<0.001). In the doubtful) cases (DBT) the three values were also higher than in normal controls (p<0.001). Furthermore, it was shown that among controls PA, HA and the DPD/IPD ratios were significantly larger in male smokers (p<0.005) while only the DPD/IPD ratio was larger in female smokers (p<0.05).

CONCLUSIONS: This is a simple method of obtaining clear index finger images. The hyponychial angle determined on the images was the most useful measurement to discriminate digits clinically with and without clubbing (sensitivity of 76.7%, specificity of 83.2%, predictive positive value of 95.5% and predictive negative value of 96.9%).

Key words: Clubbing, profile angle, hyponychial angle, relation DPD/IPD.

Abbreviations used in this paper:

COPD – Chronic obstructive pulmonary disease

DPD – Distal phalangeal depth

HA – Hyponychial angle of index finger

HOA – Hypertrophic osteoarthropathy

IPD – Interphalangeal depth

PA – Profile angle of index finger

Introduction

Digital clubbing is a valid clinical indication of intrathoracic, pulmonary or cardiac disease.^(1,2) It may sometimes appear due to chronic liver diseases,⁽³⁾ intestinal diseases,⁽⁴⁾ thyroid problems,⁽⁵⁾ or heredity.⁽⁶⁾ Clubbing has been recognized since the time of Hippocrates⁽⁷⁾ and has long been related to tuberculosis. In 1832, however, Pigeaux established its relationship with something broader, as a “corruption of the hematosis”.⁽⁸⁾ Clubbing may be idiopathic or may appear as a result of hypertrophic osteoarthropathy (HOA). To date, no clear definitions of the nature of clubbing or HOA has been offered. Although some theories have been proposed, none have yet been proven. Recent studies suggest that cytokines acting as growth factors, with or without megakaryocytes, may be implicated in the pathogenesis of these abnormalities.^(11,12)

Although various clinical criteria⁽¹³⁾ and names are used,⁽¹⁴⁾ clubbing is easily diagnosed when it is clinically obvious, appearing as rough alterations of the fingertips. However, the diagnosis may become difficult when the alterations are subtle.^(15,16) In such cases, objective criteria are quite useful for clinical diagnosis and facilitate the compilation of data for further studies.

The profile angles of index fingers⁽¹⁷⁾ or thumbs,⁽¹⁸⁾ the hyponychial angles of index fingers,⁽¹⁹⁾ and the ratio between the distal phalangeal depth (DPD) and the interphalangeal depth (IPD) of index fingers^(20,21) are the most dependable criteria for the evaluation of digital clubbing. However, profile images or rigid casts of the fingers are necessary for their determination.⁽²³⁾ Other criteria, such as nail curvature measurement,^(24,25) radiographic imaging⁽²⁷⁾ and thermographic assessment⁽²⁸⁾ have also been explored as means of evaluating clubbing and hypertrophic osteoarthropathy. All methods attempt to determine the amount of soft-tissue growth in the affected areas, especially in subungual areas.^(29-31)

In the literature, normal index finger values for profile angles have been reported as 168.3° ± 3.6°⁽³²⁾ and 171.4° ± 5.5°⁽³³⁾ and for hyponychial angles as 186.0° ± 2.0°,⁽¹⁹⁾ 180.1° ± 4.2°⁽³²⁾ and 180.7° ± 5.2°.⁽³³⁾ Values that have been considered normal for the ratio between DPD and IPD are 0.895 ± 0.041 for children,⁽¹⁶⁾ 0.889 ± 0.037 for adults of African-descent and 0.911 ± 0.049 for adult Caucasians.⁽³⁴⁾ These values are significantly higher for individuals presenting clubbing, a fact which helps discriminate between normal and abnormal fingers.

The purpose of this study was to present a quick and simple method for the production of profile images of index fingers in order to arrive at a diagnosis of clubbing through the use of the objective criteria mentioned above. The study group was composed of 306 patients with pulmonary disease and 452 normal individuals.

Methods

Image production

The shadow of the index finger in profile is projected through a flat glass pane onto a white sheet of paper placed atop the glass. The index finger is positioned in correct profile directly under (and pressed against) the glass, and the light source is positioned at 70 cm directly below the finger. The light source can be an ordinary incandescent light bulb or a simple 2-battery flashlight.

In practice, the glass pane is placed on a table and the lamp on the floor. Figure 1 shows how to set up the system in order to produce the image. The back of the hand must be in contact with the underside of table and the thumb pressed against the glass in order to prevent movement of the index finger.

The sheet of paper is marked where the shadow appears upon in order to determine the hyponychial angle (HA), profile angle (PA) and DPD/IPD ratio. These are the study parameters: distal dorsal cutaneous fold; eponychium; highest point of the dorsal surface of the nail hyponychium; ventral point, diametrically opposite to the eponychium; and ventral point diametrically opposed to the distal dorsal interphalangeal cutaneous fold (Figure 2). Figure 3 shows the group of marks on the paper (without the projected shadow), as well as the angles and distances between those marks. The right index finger was used in almost all the determinations. The left index finger was used only when the right one had a lesion or was missing (usually due to trauma). Demarcations were drawn using a sharp pencil, a ruler calibrated to 0.5 mm and a 360° protractor calibrated to 0.5°. The same researcher took all readings, in patients and in normal individuals alike.

The reproducibility of the method was initially tested through determination of PA, HA, and DPD/IPD ratio in 12 normal individuals. These parameters were measured 6 times for each individual, and index finger values did not vary.

Patients and controls

The method was applied in 306 patients (241 males and 65 females), ranging in age from 15 and 88 (mean, 51.3), who suffered from various pulmonary diseases, and in 452 healthy adults (233 males and 219 females), with ages ranging from 15 and 80 (mean, 40.2). The patients were all selected from a university hospital specialized in lung diseases.

Of the 306 patients, 217 (71%) were smokers; 116 (38%) were diagnosed with clubbing, 126 (41%) had no sign of clubbing, and the results were inconclusive in 64 (21%). The most common diseases found among the patients (either admitted or examined as outpatients) were bronchial carcinoma, chronic obstructive pulmonary disease (COPD), bronchiectasis, and pulmonary fibrosis.

The 452 individuals in the control group were all asymptomatic. None were clinically diagnosed with digital clubbing, and 301 (66.6%) had never smoked – a significant difference (p < 0.001) in comparison to the study group (patients). Of the 151 smokers in the control group, 82 were male, 69 were female, and all had smoked for more than 5 years. The males smoked 15.7 ± 7.2 cigarettes a day, whereas the females smoked 13.0 ± 7.2 cigarettes a day – a significant difference (p < 0.05). Of the 452, 354 (78.3%) were Caucasians and 98 (21.7%) were of African descent. Individuals were also evaluated as to manual labor (especially involving the fingertips), which was classified as either light (physicians, nurses, and students), moderate (workers in general who do not specifically use their fingertips), and heavy (surgeons, pianists, organists, and typists). Manual labor was considered light in 164 individuals (36.3%), moderate in 220 (48.7%), and heavy in 68 (15%). Among the controls, 385 (85.2%) reported having had a conventional chest radiograph with normal results, 341 of which were taken within the last 5 years.

Experienced physicians, trained in pulmonary diseases, clinically diagnosed clubbing according to the following classifications: positive, negative and inconclusive. In the case of inconclusive results, clinicians were obliged to decide whether the clinical evidence showed a tendency toward the positive or toward the negative. The objective data obtained from the measurements were contrasted against the subjective clinical evidence, a method similar to that described by Regan et al.⁽¹⁹⁾

Written informed consent was obtained from all study participants.

Mean tests (Student’s t-tests) were used in the comparison of numerical values, and chi-square tests were used for proportional values.

A significance level of 5.0% was adopted. Sensitivity, specificity, and positive predicted value were determined for the angles and the ratio between depths. Clinical diagnoses were used as a pattern for comparison. Therefore, the 64 inconclusive cases were distributed into the subgroups positive (n = 31) and negative (n = 33), according to the tendency demonstrated by the clinical evidence.

Results

The reproducibility of the method using the coefficient of variation showed variability lower than 2.5% for the repeated determinations of PA, HA, and DPD/IPD ratio of the same finger. The lowest degree of variability was found in the DPD/IPD ratio (Table 1).

The analysis of PA, HA, and DPD/IPD values obtained from the 452 healthy individuals in the control group was made using the previously described method, and the results are presented in Table 2.

The control group results show values that are independent of gender, age, race, or level of manual labor of the participants. However, the PA, HA and DPD/IPD values of the 82 male smokers in this group were significantly higher (p < 0.005) than those of the non-smokers (Table 3).

Comparing the 69 female smokers to the 150 female non-smokers, the only difference found was in the DPD/IPD ratios, which were slightly higher in the smokers (0.911 ± 0.029) than in the non-smokers (0.902 ± 0.030) – a significant difference (p = 0.037)

Table 4 shows the PA, HA and DPD/IPD results, divided among the three groups of patients (positive diagnosis of clubbing, negative diagnosis and inconclusive) and the controls. Patients diagnosed with clubbing or whose results were inconclusive presented values that were significantly different from those of patients not diagnosed with clubbing (p < 0.001). Study group patients who were not clinically diagnosed with clubbing also presented HA and DPD/IPD ratio results that were higher than those for individuals in the control group (p < 0.05).

We assumed the means of the angles and of the depth ratios as cutoff points and reclassified the inconclusive cases as either positive (when clinical evidence showed a higher positive tendency; n = 31) or negative (when clinical evidence showed a higher negative tendency; n = 33). We were then able to determine sensibility, sensitivity, positive predictive value and negative predictive value for the PA, HA, and DPD/IPD values of all 306 patients (Table 5).

The HA results proved to be the best parameter for the differentiation in the diagnosis of clubbing, with a cutoff point of 191.5° (Figure 4).

Discussion

Shadow images of fingers have been previously used for the study of clubbing by Bentley and Cline in 1970 and by Bentley et al. in 1976.^(22,32) They used a device which projected the amplified shadow on a screen, forming a “shadowgram” by tracing the profile of the extremity of the finger. In 1979, Siniah and Omar created shadow images of fingers in profile for the study of clubbing using an overhead projector and a screen.⁽³³⁾ However, our method obtains – or rather collects – the shadow images of fingers in a manner that differs from those employed by the above mentioned researchers.

The method presented in this study can be easily repeated. No special device is needed, and a great number of individuals can be evaluated within a short period of time. It can even be performed in homes and workplaces. The shadow images are clear and are little affected by projecting distortions since the finger is positioned near the target and far from the light source. The reference marks used in the determination of the angles and depths are easily traced on the image and can be reproduced. The demarcations traced on sheets of paper can be easily stored and are especially useful for comparative studies – with objective measured data – to determine clinical evolution, particularly for the documentation of the remission of clubbing after treatment of the baseline disease.

The PA, HA and DPD/IPD ratio values for the healthy individuals (controls) are in agreement with those from other studies (using other methods) in the literature.^{(16,21,33,34)} These controls were carefully selected, and most were non-smokers presenting normal chest radiographs. The considerable number of healthy individuals in the study assures the consistency of the results. The pertinent values were significantly lower in these control individuals than in the study group patients, especially when compared to those of patients whose clinical signs and symptoms led to a classification of positive or inconclusive for clubbing.

The HA presented the highest sensitivity, specificity, positive predictive value, and negative predictive value in the differentiation between patients with and without digital clubbing. This can be attributed to the fact that both the increase in subungual soft-tissue depth and nail curvature are taken into consideration in the determination of this angle.⁽¹⁹⁾ The range of the clinically inconclusive results in the study group was between 189.0° and 195.0°. However, most of the cases with values greater than 191,5° (cutoff point) are very likely to present digital clubbing, especially if we consider the control group values, which were significantly lower. The sensibility, specificity, and predictive values for the HA obtained using this methodology indicate that the degree of this angle, taken together with the clinical diagnosis, is the most appropriate objective criterion for the diagnosis of digital clubbing.

Among the control subjects, we also observed that male smokers presented higher values for the 3 parameters under study than did male non-smokers, whereas female smokers presented higher values than female non-smokers only for the DPD/IPD ratio. This suggests that pulmonary changes caused by smoking, which was more intense in males in the control group, may cause changes in the fingertips, even if only to a subclinical degree. It is possible that the changes in the small airways of smokers⁽³⁵⁾ also play a role. A recent study analyzing lung biopsy samples, described adult males, diagnosed with respiratory bronchiolitis-associated interstitial lung disease, who presented severe dyspnea, exercise-induced hypoxemia and digital clubbing.⁽³⁶⁾ The authors found that symptoms and results of functional tests significantly improved a few months after patients quit smoking. In our study, the minor differences detected between the measured values of control group smokers and those of control group non-smokers may be due to the large number of participants.

In the present study, the values for HA and the DPD/IPD ratio among patients not clinically diagnosed with clubbing were significantly higher than those seen among the controls. However, all of the patients presented some lung disease, and most were smokers. In 1972, Sly et al.⁽¹⁶⁾ reported that the DPD/IPD ratio, objectively measured in rigid casts of index fingers, was higher in asthmatic children than in normal children. The authors named it “minimum clubbing”. Goyal et al.⁽¹⁵⁾ stated that clubbing could be “precociously detected” by objective determinations, for example, the morphometric analysis of fingernails and toenails.

References

Submitted: 11 June 2003.

Accepted, after revision: 19 December 2003

• 1. Dickinson CJ. Lung diseases associated with digital clubbing. Clin Exp Rheumatol 1992;10(Suppl 7):23-5.
• 2. Moreira JS, Rubin AS, Silva LCC, Silva FAA, Hetzel JL. Clubbing: frequency in several pulmonary diseases. Eur Respir J 2000; 16(Suppl 31):422.
• 3. Stoller JK. As the liver does, so goes the lung. Chest 1990;97:1028-30.
• 4. Fielding JF, Cooke WT. Finger clubbing and regional enteritis. Gut 1971;123:442-4.
• 5. Gotte D. Thyroid acropachy. Arch Dermatol 1980;116:201-8.
• 6. Fisher DS, Singer DH, Feldman SM. Clubbing, a review with emphasis on hereditary acropachy. Medicine 1964;43:459-79.
• 7. Hippocrates: the book of prognostics. In: Major RH. Classic descriptions of disease. 3rd ed. Springfield: Charles C Thomas; 1945. p.4-5.
• 8. Pigeaux J. Recherches nouvelles sur l'étiologie, la symptomatologie et le mécanisme du developpement fusiforme de l'extremité des doigts. Arch Gen Med 1832;29:174-84.
• 9. Martinez-Lavin M. Digital clubbing and hyperthrophic osteoarthropathy: a unifying hypothesis. J Rheumatol 1987;14:6-8.
• 10. Shneerson JM. Digital clubbing and hypertrophic osteoartropathy: The underlying mechanisms. Br J Dis Chest 1981;75:113-29.
• 11. Silveri F, De Angelis R, Argenmtati F, Muti S, Cervini C. Hypertrophic osteoathropathy: endothelium and platelet function. Clin Rheumatol 1996;15:435-9.
• 12. Yosipovitch G, Weinberger A. Cytokines. A unifying concept in pathogenesis of clubbing. Med Hypotheses 1991;36:122-5.
• 13. Myers KA, Farquhar DRE. Does this patient have clubbing? JAMA 2001;282:341-7.
• 14. Mendlowitz M. Clubbing and hypertrophic osteoarthropathy. Medicine 1942;21:269-306.
• 15. Goyal S, Griffiths AD, Omarouayache S, Mohammedi R. An inproved method of studying fingernail morphometry: application to the early detection of fingernail clubbing. J Am Acad Dermatol 1998;39:640-2.
• 16. Sly RM, Fuqua G, Matta G, Waring WW. Objective assessment of minimal digital clubbing in asthmatic children. Ann Allergy 1972;30:575-8.
• 17. Trousseau A. Forme hipocratique des doigts des tuberculeux. J Conaiss Med Cir 1834;1:351-2.
• 18. Lovibond JL. Diagnosis of clubbed fingers. Lancet 1938;1:363-4.
• 19. Regan BM, Tagg B, Thomson ML. Subjective assessment and objective measurement of finger clubbing. Lancet 1967;1:530-2.
• 20. Rice RE, Rowland PW. A quantitative method for the estimation of clubbing. Scientific Session of the Senior Class, Tulane University Medical School 1961;11:299.
• 21. Waring WW, Wilkinson W, Wiebe RA, Faul BC, Hilman BC. Quantitation of digital clubbing in children. Am Rev Respir Dis 1971;104:166-74.
• 22. Bentley D, Cline J. Estimation of clubbing by analysis of shadowgraph. Br Med J 1970;2:43.
• 23. Mellins RB, Fishman AP. Digital casts for study of clubbing of the fingers. Circulation 1966,33:143-5.
• 24. Neufeld O, Wallbank WL. Clubbed fingers. Ohio Med J 1952;48:834-7.
• 25. Staven P. Instrument for estimation of clubbing. Lancet 1959;2:7-8.
• 26. Cudkowicz L, Wraith DG. An evaluation of the clinical significance of clubbing in common lung disorders. Br J Tuberc 1957;51:14-31.
• 27. Pineda C. Diagnostic imaging in hypertrophic ostearthropathy. Clin Exp Rheumatol 1992;10(Suppl 7):27-33.
• 28. Rush PJ, Giorshev C, Shore A, Levinson H. The use of thermography in clubbing. Respir Med 1992;86:257-9.
• 29. Bigler C. The morphology of clubbing. Am J Pathol 1958; 34:237-61.
• 30. Lovell RRH. Observations on the structure of clubbed fingers. Clin Sci 1950;9:299-321.
• 31. Ward RW, Chin R Jr, Keyes JW Jr, Haponik EF. Digital clubbing demonstration with positron emission tomography. Chest 1995;107:1172-3.
• 32. Bentley D, Moore A, Swachman H. Finger clubbing. A quantitative survey by analysis of the shadowgraph. Lancet 1976;1:164-7.
• 33. Sinniah D, Omar A. Quantitation of digital clubbing by shadowgram technique. Arch Dis Child 1979;54:145-6.
• 34. Sly RM, Ghazanshai S, Buranakul B, Puapan P, Gupta S, Warren R, et al. Objective assessment for digital clubbing in caucasian, negroe, and oriental subjects. Chest 1973;64:687-9.
• 35. Niewoehner DE, Kleinerman J, Rice DB. Pathologic changes in the peripheral airways of young cigarette smokers. N Engl J Med 1974;291:755-8.
• 36. Sadikot RT, Johnson J, Loyd JE, Christman JW. Respiratory bronchiolitis associated with severe dyspnea, exertional hypoxemia, and clubbing. Chest 2000;117:282-5.

*

Study carried out at the Pavilhão Pereira Filho (Serviço de Doenças Pulmonares da Santa Casa de Misericórdia de Porto Alegre).

Publication Dates

• Publication in this collection
  08 June 2004
• Date of issue
  Apr 2004

History

• Accepted
  19 Dec 2003
• Received
  11 June 2003