A spine motion segment has 6° of freedom, and therefore, there are theoretically 21 coupling coefficients. However, owing to the geometrical and ligamentous symmetry about the sagittal plane the coupling coefficients for flexion and extension are zero, and only coupling between rotation, lateral-bending and translation is observed (Panjabi et al, 1974). The early observations of segmental motion were developed by Fryette who, in 1954, undertook experiments upon ‘a spine mounted in soft rubber’ and subsequently published three laws of coupling motion (Gibbons & Tehan, 1998). The first law stipulates that with the spine neutral, side-bending produces rotation to the opposite side. The second law refers to the condition of flexion or extension of the vertebrae where the rotation and side-bending are both directed to the same side. The third law indicates that motion of a vertebral joint in one plane automatically reduces its mobility to the other two planes (Gibbons & Tehan 1998, 2000).

Contrary to Fryette’s laws, several textbooks report that in the lumbar spine there is contralateral rotation side-bending in extension and ipsilateral rotation side-bending in flexion (Huijbregts, 2004). However, these textbooks lack support from primary references (Huijbregts, 2004). In 1978, White and Panjabi studied the coupling directions using two-dimensional imagery. They found that the direction of the coupling did not depend on the position of the spine, (flexion–extension) as Fryette’s laws would predict, but on which movement would be initially engaged. Therefore, when side-bending was initially engaged the coupling would be contralateral rotation, but in the case of rotation being engaged first the coupling would be ipsilateral. However, two-dimensional radiographic methods of examining coupling behaviour have been criticized as they can lead to inaccurate and misleading results (Coleman et al, 1999; Evans & Lissner, 1959; Harrison et al, 1998).

A number of studies have been published that used three-dimensional measurement of the coupling motions, both in vivo (Hindle et al, 1990; Pearcy & Tibrewal, 1984) and in vitro (Cholewicki et al, 1996; Oxland et al, 1992; Panjabi et al, 1989, 1994; Vincenzino & Twomey, 1993) which are thought to provide more accurate readings that reflect the true motion of the spine (Harrison et al, 1998; Rab & Chao, 1977). All the above studies used non-pathological, asymptomatic subjects.

Coupled motion with side-bend initiation with the spine in neutral

The study of Pearcy and Tibrewal (1984) showed no coupling rotation at the level of L1–L2 which is in agreement with the studies of Panjabi et al (1989, 1994) and Cholewicki et al (1996). However, the studies of Cholewicki et al (1996) indicated that there might be cases of rotation coupling in the opposite direction; in the study of Panjabi (1989) a similar tendency was indicating rotation occurring on the same side as side-bending. Although there is general agreement that the coupling rotation occurs in the opposite direction to the side-bend (Hindle et al, 1990, Oxland et al, 1992; Panjabi et al, 1989, 1984; Pearcy & Tibrewal, 1984), the study of Cholewicki et al (1996) reports an ipsilateral coupling of rotation to side-bending at the L4–L5 motion segment.

Coupled motion with rotation initiation with the spine in neutral

Under this condition there seems to be more agreement amongst the researchers. The studies of Pearcy and Tibrewal (1984), Panjabi et al (1989, 1994) and Cholewicki et al (1996) have shown the contralateral coupling of side-bend for the L1–L2, L2–L3 and L3–L4 motion segments. The same studies have shown ipsilateral side-bend coupling for the L4–L5 and L5–S1 which is in agreement with the study of Oxland et al (1992). It is worth mentioning though that the study of Pearcy and Tibrewal (1984) reported some subjects who exhibited coupling of side-bend to the same side as rotation at the L4–L5 level.

Coupled motion with side-bend initiation with the spine in extension

There are two studies that explored the coupling behaviour of the lumbar spine under this condition (Panjabi et al, 1989, Vincenzino & Twomey, 1993). Both studies report that at the L3–L4 motion segment, the direction of coupling rotation is contralateral to the side-bend. However, the studies reach opposite conclusions for the rest of the levels of the lumbar spine with the L5–S1 segment displaying extreme variety in the study of Vincenzino and Twomey (1993) where coupled rotation was identified in both directions or none at all.

Coupled motion with side-bend initiation with the spine in flexion

Under this condition the studies of Panjabi et al (1989) and Vincenzino and Twomey (1993) reported agreement for the L2–L3 and L4–L5 motion segments where the rotation was contralateral to the side-flexion. For the L1–L2 level, Panjabi reported occurrences where the rotation was ipsilateral or contralateral, while Vincenzino and Twomey reported occurrences with ipsilateral coupling or none at all. For the L3–L4 level, the coupling was found to be ipsilateral by Vincenzino and Twomey (1993) and ipsilateral or non-existent for L5–S1, whereas for Panjabi et al (1989) both the above levels displayed contralateral coupling.

Coupled motion with rotation initiation with the spine in extension and flexion

The only study that explored the coupling motion of side-flexion with rotation in extension of flexed lumbar spine was that of Panjabi et al (1989). In the extended spine the coupled side-flexion is ipsilateral for the L1–L2, L4–L5 and L5–S1 levels and contralateral for the L2–L3. The L3–L4 level exhibited both ipsilateral and contralateral coupling. With the spine in flexion the L1–L2, L2–L3 and L3–L4 exhibited contralateral side-bend coupling while the L4–L5 and L5–S1 segments ipsilateral (Panjabi et al, 1989). It is worth mentioning that lateral translations were found to be in the range of 1 to 2 mm in all directions (Pearcy & Tibrewal, 1984) and when associated with lateral-bending to the same side, and to be at an average of 1.1 mm at all levels (Panjabi et al, 1994) and therefore of doubtful clinical significance.

Flexion and extension were found to cause minimal to negligible coupling rotation or side-flexion (Cholewicki et al, 1996, Hindle et al, 1990). However, Cholewicki et al (1996) have found flexion motion associated with side-bending in all lumbar levels.

Pearcy and Tibrewal (1984) have reported extension of all levels from L1 to L4 during lateral-bending. The L4–L5 level showed variable behaviour extending in general but occasionally flexing. The L5–S1 generally flexed.

In contrast to the above studies on coupling movements that studied non-pathological asymptomatic subjects, there are few studies that looked at the coupling movement of subjects with low back pain. The study of Lund et al (2002) showed inconsistencies of coupled motions during both axial rotation and side-bending in subjects with low back pain (LBP). Axial rotation to either side produced side-flexion to the right in some of the majority of the subjects, with the rest displaying ipsilateral side-bending and some no coupled side-flexion. During lateral-bending, axial rotation coupling was observed in 50% of the patients studied. In 8 of the 22 cases there was no coupled axial rotation and contralateral rotation in 3 cases. The results of the coupling behaviour during side-bending are of interest as they are not in agreement with the consistent pattern of contralateral axial rotation in normal subjects in the majority of the studies (Panjabi et al, 1989, 1994; Pearcy & Tibrewal, 1979). This can be attributed to the altered role of the neuromuscular system during LBP as it may be an important contributor to the coupling behaviour (Cholewicki et al, 1996; Pearcy & Tibrewal, 1984). Furthermore, the study of Oxland et al (1992) on cadavers showed that acute injury of the disc and facet joints altered the magnitude of the coupling motions but not the direction of the coupling.