In 1912, Maurice Klippel and Andre Feil independently provided the first descriptions of Klippel-Feil syndrome in patients who manifested the following:
Short, webbed neck
Decreased range of motion (ROM) in the cervical spine
Feil subsequently classified the syndrome into the following three types:
Type I - Massive fusion of the cervical spine
Type II - Fusion of one or two vertebrae
Type III - Presence of thoracic and lumbar spine anomalies in association with type I or type II Klippel-Feil syndrome
Since the original description, other classification systems have been advocated to describe the anomalies, predict the potential problems, and guide treatment decisions.
Type I - Single-level fusion
Type II - Multiple, noncontiguous fused segments
Type III - Multiple, contiguous fused segments
Gray et al  described 462 patients with Klippel-Feil syndrome and found that the level of fusion did not greatly affect the incidence of neurologic symptoms. The most frequent level they identified was a defect of the occiput to C1, C2, and C3. These produced the most symptoms; lesions below C3 and 4 were slightly less likely to cause symptoms. Twenty-seven percent of symptoms occurred in the first decade.
Nagib et al  described three types and related the incidence of neurologic symptoms to each type as follows:
Type I - Two sets of block vertebrae with open intervening spaces that can sublux gradually or with acute trauma
Type II - Craniocervical anomalies with occipitalization of the axis and basilar invagination; this causes increased mobility at the craniocervical level and can lead to foramen magnum encroachment; it can be associated with Arnold-Chiari malformation and syringomyelia
Type III - Fusion of one or more levels with associated spinal stenosis 
Patients with Klippel-Feil syndrome usually present with the disease during childhood, but they sometimes present later in life. The challenge to the clinician is to recognize the associated anomalies that can occur with Klippel-Feil syndrome and to perform the appropriate workup for diagnosis.
Auerbach et al studied spinal cord dimensions in children with Klippel-Feil syndrome.  They reviewed magnetic resonance imaging (MRI) studies and clinical records of Klippel-Feil patients and age-matched controls. Torg ratios were measured, and the Torg-Pavlov ratios were found to be identical in the two groups.
The cross-sectional area of the spinal cord was smaller in Klippel-Feil syndrome patients at each level from C2-C7.  These differences were statistically significant, with no differences in the cerebrospinal fluid (CSF) column, suggesting that the cord size is smaller in children with Klippel-Feil syndrome than in control subjects. Four of the 12 children with Klippel-Feil syndrome presented with neurologic symptoms that improved after posterior cervical stabilization.
Samartzis et al studied the extent of fusion in the congenital K-F segment to evaluate the presence and extent of specific fusion patterns across the involved cervical segments.  In older patients, complete fusion was more prevalent in regard to C2-C7. In the absence of complete fusion, fusion of the posterior elements was noted more often than fusion of the anterior elements.
In another paper, Samartzis et al reviewed the role of the congenitally fused segments in 29 Klippel-Feil syndrome patients in relation to the space available to the cord (SAC) and associated cervical spine-related symptoms (CSS).  They suggested that an arrest of normal vertebral development may affect appositional bone development. The effect on vertebral body width may delay neurologic compromise resulting from the congenital fusion process and subsequent degenerative manifestations.
The etiology of Klippel-Feil syndrome and its associated conditions is unknown. The syndrome can present with a variety of other clinical syndromes, including fetal alcohol syndrome, Goldenhar syndrome, and anomalies of the extremities. [9, 10, 11]
Gunderson suggested that Klippel-Feil syndrome is a genetic condition, whereas Gray found a low incidence of inheritance. [12, 13] Two small studies suggested that mutations of the MEOX1 gene, which codes for mesenchyme homeobox 1, may cause a recessive subtype of the syndrome. [14, 15] A report by Karaca et al suggested that a homozygous frameshift mutation in RIPPLY2, a gene shown to play a crucial role in somitogenesis and participate in the Notch signaling pathway, could give rise to a type of autosomal recessive Klippel-Feil syndrome. 
Other investigators have considered Klippel-Feil syndrome to be some type of global fetal insult, which could explain the other associated conditions. Some have considered it to be a consequence of vascular disruption. [17, 18]
The true incidence of Klippel-Feil syndrome is unknown: To date, no one has studied a cross-section of healthy people to determine its actual frequency.
Two studies investigated the incidence of Klippel-Feil syndrome, using two different means. Gjorup and Gjorup reviewed all of the radiographic cervical spine films from a single hospital in Copenhagen.  From these films, they determined an incidence of 0.2 cases per 1000 people. Brown et al reviewed 1400 skeletons from the Terry collection, which at that time was located at the Washington University School of Medicine.  They found an incidence of 0.71%.
Nouri et el, using AOSpine data from the United States and other countries, reviewed MRI studies in a global cohort of 458 patients with degenerative cervical myelopathy and found that Klippel-Feil syndrome was present in 2.0%. 
The prognosis for Klippel-Feil syndrome depends on the specific anomalies present. Careful evaluation, consistent follow-up, and coordination with other providers are required to avoid pitfalls and to ensure that no diagnoses are missed. The classification system created by Samartzis et al (see Background) is useful for predicting which patients may develop symptoms.