Since Hippocrates first described digital clubbing in patients with empyema, digital clubbing has been associated with various underlying pulmonary, cardiovascular, neoplastic, infectious, hepatobiliary, mediastinal, endocrine, and gastrointestinal diseases. Finger clubbing also may occur, without evident underlying disease, as an idiopathic form or as a Mendelian dominant trait. Clubbing is a clinically descriptive term, referring to the bulbous uniform swelling of the soft tissue of the terminal phalanx of a digit with subsequent loss of the normal angle between the nail and the nail bed.
Digital clubbing is classified into primary (ie, idiopathic, hereditary) and secondary forms. Digital clubbing may be symmetric bilaterally, or it may be unilateral or involve a single digit. Anatomic considerations, such as the classic measurement of the Lovibond angle or the more recently derived index of nail curvature by Goyal et al  , usually can be identified on simple physical examination and can be used to identify digital clubbing and to monitor this dynamic process objectively. Various imaging modalities have been used not only to evaluate clubbing but also to help identify possible clues to its development.
Clubbing is a feature of pachydermoperiostosis (PDP), a rare genodermatosis characterized by pachydermia, digital clubbing, periostosis, and an excess of affected males.  Although usually an autosomal dominant model with incomplete penetrance and variable expression, both autosomal recessive and X-linked inheritance have been suggested in some PDP families.
Primary hypertrophic osteoarthropathy (PHO), a rare hereditary disorder with digital clubbing, subperiosteal new bone formation, and arthropathy, has been linked mutations in the 15-hydroxy-prostaglandin dehydrogenase (15-PGDH) encoding gene HPGD, which causes PHO.  A mutation was identified in a large Pakistani family with isolated congenital nail clubbing.  In another study of homozygous mutations, an intragenic deletion that results in frameshift and a missense mutation was associated with a severe PHO phenotype. A heterozygous carrier of a stop mutation had isolated digital clubbing.
The specific pathophysiologic mechanism of digital clubbing remains unknown. Many theories have been proposed, yet none have received widespread acceptance as a comprehensive explanation for the phenomenon of digital clubbing. As stated best by Samuel West in 1897, "Clubbing is one of those phenomena with which we are all so familiar that we appear to know more about it than we really do."
Alterations in size and configuration of the clubbed digit result from changes in the nail bed, beginning with increased interstitial edema early in the process. As clubbing progresses, the volume of the terminal portion of the digit may increase because of an increase in the vascular connective tissue and change in quality of the vascular connective tissue, although some cases have been associated with spurs of bone on the terminal phalanx.
Although clubbing is a common physical finding in many underlying pathological processes, surprisingly, the mechanism of clubbing remains unclear. Different pathological processes may follow different pathways to a common end. Many studies have shown increased blood flow in the clubbed portion of the finger.
High-resolution magnetic resonance imaging uisng a contrast agent in 4 patients with finger clubbing and 4 healthy volunteers documented nail bed hypervascularization as linked with clubbed nails.  Most researchers agree that this results from an increase in distal digital vasodilation, the cause of which is unknown. Also unknown is the exact mechanism by which increased blood flow results in changes in the vascular connective tissue under the nail bed.
Many researchers agree that the common factor in most types of clubbing is distal digital vasodilation, which results in increased blood flow to the distal portion of the digits. Whether the vasodilation results from a circulating or local vasodilator, neural mechanism, response to hypoxemia, genetic predisposition, or a combination of these or other mediators is not agreed on currently.
Evidence that favors the presence of a circulating vasodilator derives from the association of clubbing with cyanotic congenital heart disease. Many potential vasodilators, which usually are inactivated as blood passes through the lungs, bypass the inactivation process in patients with right-to-left shunts. Patients with tetralogy of Fallot with substantial shunting have a high incidence of clubbing. After surgical correction diminishes the shunt, the clubbing improves. Also previously observed is clubbing confined to the feet in patients with late untreated patent ductus arteriosus in whom blood from the pulmonary artery bypasses the lungs and is shunted into the descending aorta. In the absence of a shunt, the circulating vasodilator may be produced by the lung tissue, or, possibly, it passes through the pulmonary circulation without becoming inactivated. Proposed vasodilatory factors include ferritin, prostaglandins, bradykinin, adenine nucleotides, and 5-hydroxytryptamine.
A neural mechanism has been proposed with particular consideration of the vagal system. An increased incidence of digital clubbing has been associated with the pathology and disease of vagally innervated organs. Furthermore, regression of clubbing after vagotomy has been reported. Although some factor related to the vagal system is a possible contributor to the development of clubbing, especially clubbing occurring with hypertrophic osteoarthropathy, the hypothesis of a neural mechanism has decreased in popularity because of the lack of evidence of clubbing in neurologic disorders and the presence of clubbing in diseases of organs not innervated by the vagal system.
Hypoxia has been proposed as an alternative explanation for clubbing in cyanotic heart disease and pulmonary diseases. An increase in hypoxia may activate local vasodilators, consequently increasing blood flow to the distal portion of the digits; however, in most cases, hypoxia is absent in the presence of clubbing, and many diseases with noted hypoxia are not associated with clubbing.
Genetic inheritance and predisposition also may play a role in digital clubbing. Hereditary clubbing is observed in 2 forms, including idiopathic hereditary clubbing and clubbing associated with pachydermoperiostosis. The 2 forms are believed to be separate entities. Both demonstrate autosomal dominant inheritance with incomplete penetrance.
More recently, platelet-derived growth factor released from fragments of platelet clumps or megakaryocytes has been proposed as the mechanism by which digital clubbing occurs.  The fragments are large enough to lodge in the vascular beds of the fingertips, and, subsequently, they release platelet-derived growth factor. This factor has been shown to have general growth-promoting activity and causes increased capillary permeability and connective tissue hypertrophy.
Idiopathic or primary clubbing is rare, while the occurrence of secondary clubbing depends on the underlying disease.
Primary digital clubbing has been reported to occur in 89% of patients diagnosed with pachydermoperiostosis. This syndrome most often occurs in young males.
Of patients with idiopathic pulmonary fibrosis, 65% have clinical digital clubbing. In these patients, an increased occurrence has been shown in patients with higher grades of smooth muscle proliferation in the lungs.
Clubbing has been reported in 29% of patients with lung cancer and is observed more commonly in patients with non–small cell lung carcinoma (35%) than in patients with small cell lung carcinoma (4%).
Digital clubbing was reported in 38% of patients with Crohn disease, 15% of patients with ulcerative colitis, and 8% of patients with proctitis. Clubbing was observed in up to one third of Ugandan patients with pulmonary tuberculosis.  It was not associated with stage of HIV infection, extensive disease, or hypoalbuminemia.
Racial predilection of secondary clubbing depends on the racial prevalence of the underlying disease.
Sex predilection of secondary clubbing of the fingers depends on the prevalence of the underlying disease.
Since clubbing of the fingers is a clinical finding, it is not an independent cause of mortality; however, mortality associated with underlying pathology in patients with secondary clubbing widely varies. 
Morbidity is minimal and typically is associated with the cosmetic appearance of clubbed digits.
Treatment of the underlying pathological condition may decrease the clubbing or, potentially, reverse it if performed early enough. Once substantial chronic tissue changes, including increased collagen deposition, have occurred, reversal is unlikely. Prognosis of the underlying disease should be determined on an individual basis.