Pressures Generated During Spinal Manipulation and Their Association With Hand Anatomy

Methods 

Methods 

After their provision of written informed consent, each subject donned a protective lead vest and then was instructed to place the pad of the left index finger on the target. The pisiform of the right hand was then located and placed on the left index fingernail, and the left index finger removed from under the pisiform allowing the right pisiform to fall onto the target. This technique has been used successfully in the past to improve accuracy in placing the pisiform on the required target.26 The subject was then asked to use an arched hand configuration while performing an SMT (Fig 1). The arched hand configuration is much like the support hand used when shooting billiards. Pressures resulting from this manipulation were recorded directly from the pressure array to a computer at a sampling rate of 50 Hz. Immediately after the arched hand SMT, a dorsal-palmar radiograph of the hand/sensor array was obtained using standard radiographic techniques. In this manner, the location of the skeletal landmarks could be determined with respect to the radiopaque sensor array (Fig 2). This entire procedure was then repeated with each subject using the flat hand configuration (Fig 1).

View Large Image Download to PowerPoint

Fig 1. Hand configurations (postures) used for manipulation trials demonstrating the pressure sensor pad: arched configuration (left) and flat configuration (right).

View Large Image Download to PowerPoint

Fig 2. Plain film radiographs demonstrating superimposition of osseous anatomy and radiopaque sensor elements imbedded in the pressure sensor pad. The arched configuration is shown on the left and the flat configuration is shown on the right.

The resultant radiographic films were assessed by 2 radiologists who identified 2 osseous landmarks: one carpal bone (pisiform) and a prominence of another (hook of the hamate). The radiologists located the specific sensor where each landmark resided by identifying the sensor that contained more than 50% of the visualized landmark. To assess agreement between radiologists, the κ statistic was used.

For the purposes of this study, data pertaining to pressure and contact area were taken from the point during SMT application where the greatest pressure was generated. These data (eg, peak pressure magnitudes) were analyzed in pairs (arched and flat hand configuration) for each subject using a standard t test (α = .05).

Results 

Sixteen practicing chiropractors agreed to participate in the study. The 16 participants were split equally between sexes. The mean age of the subjects was 34.18 ± 1.38 years whereas the mean years in practice was 6.44 ± 1.35.

Peak Pressure 

During any manipulation, the highest pressure magnitudes observed for any one sensor within the array (peak pressure) were generated with the arched hand posture (44.81 ± 14.54 N/cm2) compared to those performed with the flat hand posture (32.13 ± 12.86 N/cm2) (Table 1, Table 2). Not only were the peak pressures in the arched configuration significantly larger (P < .000) (Table 3), they also occurred over a smaller spatial area than the flat hand manipulations. The arched configuration activated, on average, 0.25 additional sensors whose magnitudes were within 25% of the peak. In comparison, the flat posture activated almost 2 additional sensors within 25% of the peak pressure. This relation between peak pressure and the number of additional active sensors is displayed in Fig 3. In addition to this result, a significant difference was observed (P = .008) between arched and flat manipulations in that the arched configuration had a 37.67 ± 17.54 N/cm2 difference between the sensor displaying peak pressure and the sensor with the next highest pressure magnitude. In comparison, the flat configuration displayed a difference between the peak sensor and the next highest sensor of 14.37 ± 19.98 N/cm2.

Table 1. Data for manipulation trials using an arched hand posture

	Subject	Total pressure in array when peak pressure is generated (N/cm2)	No. of sensors activated when peak pressure is generated	Peak pressure (N/cm2)	Sensor with next highest pressure (N/cm2)	% Difference	No. of active sensors within 25% of peak pressure
	1	270	17	43	35	18.605	2
	2	268	15	45	27	40.000	0
	3	294	20	62	40	35.484	0
	4	101	11	31	17	45.161	0
	5	113	12	26	16	38.462	0
	6	155	14	37	30	18.919	1
	7	209	15	52	29	44.231	0
	8	133	12	46	25	45.652	0
	9	237	22	56	20	64.286	0
	10	203	16	63	21	66.667	0
	11	283	24	62	34	45.161	0
	12	207	17	64	28	56.250	0
	13	177	12	51	50	1.961	0
	14	148	16	27	17	37.037	0
	15	176	15	30	22	26.667	0
	16	88	12	22	18	18.182	1
	Mean value	191.38	15.63	44.81	26.81	37.67	0.25
	SE	16.56	0.94	3.64	2.36	4.39	0.14

Table 2. Data for manipulation trials using a flat hand posture

	Subject	Total pressure in array when peak pressure is generated (N/cm2)	No. of sensors activated when peak pressure is generated	Peak pressure (N/cm2)	Sensor with next highest pressure (N/cm2)	% Difference	No. of active sensors within 25% of peak pressure
	1	402	49	23	23	0.000	3
	2	384	27	41	39	4.878	3
	3	267	26	25	25	0.000	3
	4	108	13	20	20	0.000	0
	5	172	25	19	18	5.263	3
	6	390	44	25	23	8.000	0
	7	379	32	60	27	55.000	0
	8	207	24	32	28	12.500	1
	9	253	26	25	24	4.000	3
	10	361	31	43	33	23.256	2
	11	329	24	45	44	2.222	2
	12	252	25	26	23	11.538	3
	13	206	25	56	17	69.643	0
	14	219	27	26	24	7.692	2
	15	256	31	26	24	7.692	2
	16	264	37	22	18	18.182	2
	Mean value	278.06	29.13	32.13	25.63	14.37	1.81
	SE	21.870	2.127	3.214	1.860	4.995	0.306

Table 3. P values for tests of significance (paired t test) between data in Table 1, Table 2

	Subject	Total pressure in array when peak pressure is generated (N/cm2)	No. of sensors activated when peak pressure is generated	Peak pressure (N/cm2)	Sensor with next highest pressure (N/cm2)	% Difference	No. of active sensors within 25% of peak pressure
	P (α = .05)	<.000	<.000	.002	.670	.008	<.000

View Large Image Download to PowerPoint

Fig 3. Peak and total pressures with SE bars for manipulations performed with different hand configurations. The size of the circle (with associated value) represents the mean number of active sensors falling within 25% of the peak pressure. Differences between hand configurations were statistically significant for peak pressure, total pressure, and the number of active sensors within 25% of peak pressure.

Total Pressure 

With respect to total pressures, the flat hand posture showed a significantly greater pressure magnitude (278.06 ± 87.48 N/cm2) over a larger number of active squares (29.13 ± 8.51 SE) vs the arched contact which had a smaller total pressure (191.38 ± 66.23 N/cm2) distributed over a fewer number of active sensors (15.63 ± 3.76 SE). This relation is displayed in Fig 3, Fig 4. The differences in total pressure and total active sensors between the arched and the flat configurations were statistically significant (P < 2.000).

View Large Image Download to PowerPoint

Fig 4. Two-dimensional and 3-dimensional isobar representations of pressures recorded from a single subject during arched hand configuration (A) and flat hand configuration (B). The smaller peaks in the image relate to areas where the fingers made contact with the pressure sensor array.

Discussion 

The results of this study suggest that hand configuration (posture) influences the magnitude, location, and distribution of pressure generated by the hand during simulated SMT. Furthermore, specific hand configurations generate peak pressures associated spatially with specific anatomic structures. For the arched hand configuration, the peak pressure developed during SMT was localized under the pisiform in most subjects. When the hand was placed in a flattened configuration, the location of peak pressure was more varied although most often associated spatially with the hamate. Our results also describe significant alterations in the pressure magnitudes and pressure distributions between the arched and flat hand configurations. This is the first study that we are aware of that relates the location of the peak pressure generated during SMT to different anatomic structures within the hand.

These results suggest that pressures generated during SMT may not develop at the location intended by the clinician, but instead, may vary as a result of the prominence of different osseous landmarks, a function of hand configuration. The ability of rigid bodies (ie, bone) to facilitate force transmission compared to the attenuating effect of the surrounding soft tissues is a likely explanation for these results.

This study, performed on a flat rigid surface, showed that migration of peak pressure during an arched hand configuration was minimal. We speculate that this observation is the result of (1) a hand configuration that isolates structures other than the pisiform from involvement in the contact interface and (2) a relative absence of soft tissue overlaying the pisiform itself. In contrast, peak pressure migration was significantly greater in the flat hand configuration. We speculate that this result is a function of a greater number of osseous structures in the flat hand being available at the contact interface as potential pressure transducers. In addition, as the soft tissues of the hand become increasingly more compressed and widened as SMT progresses temporally, the rigid constituents of the flat hand involved at the contact interface may change, thereby resulting in peak pressure migration. Indeed, an increased variability in peak pressure location was seen in flat hand SMT in addition to a greater area of pressure distribution. These observations allude to the possibility that if human subjects replaced the rigid surface used in this study, even greater peak pressure migration may occur as a result of the contribution of nonrigid, heterogeneous soft tissues of the subject to the contact interface. Indirect support for this possibility is provided by the results of others from the University of Calgary who described up to 9.8 mm of peak force migration during SMT delivery in human subjects.25 Unfortunately, the authors of this previous study did not attempt to determine the cause of this migration although they discussed the possibility that their results may be artifacts due to bending of the pressure array caused by subject deformation during SMT. Such an artifact is not possible in this study because of our use of a flat nondeforming surface beneath the pressure array.

As with other forms of physical treatment involving injections or incision, accurate site specificity in SMT is thought to improve clinical outcome and is therefore desirable.1, 27, 28 However, our study reveals the mechanism of a potential source of variation that reduces the site specificity of SMT: peak pressure migration associated with specific hand configurations. Unfortunately, this potential source of SMT variation is but another identified recently. A number of investigators have now shown that clinicians are generally inaccurate when identifying the location of internal spinal targets when using only surface-based anatomy to guide them.29, 30, 31 Others have shown similar findings when clinicians were asked to locate vertebrae targeted for SMT.32, 33, 34, 35 In total, these sources of variation conspire against the ability of clinicians to apply SMT accurately. With this in mind, it is not difficult to imagine that these sources of variation may offer a partial explanation as to why SMT outcomes are often inconsistent.36, 37

Given the results, this study suggest that use of the arched hand contact by clinicians would decrease pressure variations generated during SMT delivery. Unfortunately, there is a suggestion that the arched configuration, with its smaller area of pressure distribution, results in greater patient discomfort. Kirstukas et al38 found that subjects manipulated with smaller pressure distributions reported that the SMT was less comfortable than SMT performed with a more distributed contact area (written communication from Jerrilyn A. Cambron, DC, MPH, July 16, 2003). The arched contact is also a less stable position for the provider's wrist and potentially exposes the wrist to harmful forces.20 At present, there is no literature of which we are aware that describes if these 2 factors affect the clinical frequency in which the arched or flat hand configuration is applied.

The results of this and other studies suggest that the accuracy of providing SMT to a specific location is marginal. Although it may be argued that this inaccuracy has consequences related to SMT outcomes and patient safety, there is a contrary evidence that suggests that attempting to create a highly accurate application of SMT may not be important clinically. Haas et al39 have found that the clinical effectiveness of cervical manipulation was independent of whether the manipulation was based upon examination findings or randomized matched motion segments in the cervical spine. In addition, Reggars and Pollard40 found that the phenomenon of joint cavitation associated with SMT occurred in joints other than those that were targeted. Although it has been acknowledged that the effect of cavitation produced by SMT is unknown, it is one indication used by practitioners to determine if the application of SMT was successful.1 Similarly, Ross et al41 found that during both thoracic and lumbar spine manipulations, cavitation often occurred remote to the targeted articulation and that the chances of the target site cavitating were greatly improved by producing manipulations having multiple cavitations. Therefore, the impact of SMT location accuracy or lack thereof is unknown at this time. However, given that others have assumed that a physiologically relevant SMT contact area is 25 mm,2, 25 it is possible that inaccurate SMT will have a significant effect on clinical outcome and/or patient safety. Future studies are necessary to further characterize sources of SMT variation and assess its significance in terms of treatment efficacy and patient safety.

Conclusions 

This study shows that hand configuration influences the magnitude, location, and distribution of pressure generated by the hand during manipulation. Specifically, this is the first study that we are aware of that relates the location of the peak pressure generated during SMT to different anatomic structures within the hand. In this study, flat hand configurations tended to transmit pressure through the hook of the hamate whereas arched hand configurations transmitted pressure almost exclusively through the pisiform. These findings may be of significance for a practitioner's ability to deliver a specific force magnitude to a specific target and therefore possibly influence treatment efficacy and patient safety.