The structure of the bony pelvis is maintained by several identifiable ligaments. When any of these ligaments is disrupted some amount of pelvic stability is lost. However, since the ligaments differ in size and location, some ligamentous injuries lead to major instabilities, while others leave the pelvis relatively stable. Tile1 outlined the relative contributions of the major pelvic ligaments to pelvic stability and, although his analysis has proven very useful clinically, there has been little experimental work to confirm his impressions. The purpose of Lhis project was to examine the relative contributions of the major ligamentous structures to pelvic stability.
MATERIALS AND METHODS
We tested 13 fresh frozen cadaver specimens comprised of pelvis and L4 and L5 vertebra. All specimens had intact pubic symphysis, anterior sacroiliac ligaments, and posterior sacroiliac ligaments (including interosseous ligaments), and a majority of specimens had intact sacrotuberous and sacrospinous ligaments. For testing, the potted L4 vertebra was loaded with the ram of a hydraulic MTS materials testing machine (Model). Positioning of the pelvis simulated double leg stance with the femora! prostheses transmitting load through the acetabula (Fig 1). Mounting the femomi prostheses on roller plates allowed free motion in the coronal and frontal planes, and partially constrained motion in the anterior-posterior plane.
For testing, each specimen was cyclic loaded to 600 N using a haver sine function at 1 Hz, and load displacements curves were recorded at stable hysteresis. A load cell mounted on the MTS ram measured loads, and the ram's linear excursion gave displacements. Each specimen was then loaded to 600 N under displacement control, one of the ligaments was cut, and the resultant drop in load was recorded. Following this the pelvis was again cyclic loaded and load displacement curves were recorded. This sequence was repeated until all the ligaments had been cut. The sequence of ligament sectioning was varied in each specimen. Because the sacrotuberous and sacrospinous ligaments were small, and not present in many specimens, these ligaments were always cut together.
Fig 1 : Loading configuration for pelvic specimens. Proximally the pelvis was loaded through a freely moving ball and socket joint, and distally the femoral prostheses were mounted on roller plates.
Fig 2: Load vs time curve recorded with the pelvis statically loaded under displacement control. The sharp drop in load occurred when the pubic symphysis was cut. The load time curves were difficult to interpret for other ligaments that could not be cut quickly.
The slopes of the load displacement curves were taken as indicative of pelvic stability, as they reflected the amount of load necessary to cause a given displacement. When certain combinations of ligaments had been sacrificed the specimens could not support 600 N. Therefore, to allow consistent comparisons among all specimens, slopes were calculated at loads between 100 N and 150 N. This was beyond the toe region of the curve, and the curves were approximately linear in this region.
All ligaments were cut with a simple pass of the scalpel except for the ligaments of the posterior sacroiliac complex (PSf). Because it was difficult to cut the posterior interosseous ligaments quickly using a scalpel, these ligaments were cut using an oscillating saw.
Cutting any of the ligaments with the pelvis loaded statically to 600 N resulted in a noticeable drop in the load supported by the pelvis (Fig 2). However, the actual data from these tests were difficult to interpret. The time required to transect ligaments varied not only from test to test but from ligament to ligament. It requires approximately 5 seconds to cut the sacrotuberous (ST) and sacrospinous (SS) ligaments. 15 seconds to cut the pubic symphysis (PS). 30 seconds to cut the anterior sacroiliac ligament (ASI), and 45 seconds to cut the PSI. Due to viscoelastic changes occurring during ligament sectioning, it was difficult to know how much load decrease to attribute to the loss of the ligament. For this reason we choose to make stability comparisons based on slopes of the load displacement curves obtained while cyclically loading the pelves. Representative load displacement curves are shown in Figure 3. With several ligaments sectioned, the pelves often could not support a 600 N load, and the peak loads were decreased. To make consistent comparisons we chose to calculate slopes from the load displacement curves at loads between 100 N and 150 N. The mean stiffness for the 13 test specimens with all ligaments intact was 129 ± 54 N/mrn (range: 34 N/mm to 259 N/mm).
Cutting the SS and ST had little effect on pelvic stiffness. Eight specimens had either the ST, SS, or both ligaments intact for the tests, and these ligaments were considered as one for the purposes of analysis. A mean 3% drop in stiffness resulted from cutting these ligaments with a range from 0% to 8%. Alone, these ligaments were never able to support the load on the pelves.
The composite stability was greatest with the pelvic ring intact (Fig 4). Cutting the ASI left the PS and PSI intact to support the pelvic ring and caused only a 9% (±12%) drop in pelvic stability. Similarly, cutting the PSI and leaving the PS and ASI intact to support the ring dropped the stability only 15% ( ±32%). On the other hand, cutting the PS and leaving the ASI and PSI intact disrupted the pelvic ring and dropped the stability 21% (±14%). Similarly, cutting the PSI and ASI but leaving the PS dropped the stability 35% (±13%). Although none of these groups were significantly different by one-way analysis of variance (ANOVA) at the .05 level, there was a trend suggesting mat ligament cuts decreased stability more if they disrupted the ring than if the ring remained intact.
A ligament contribution to pelvic stability varied depending on when in the sequence it was cut (Fig 5).
It is important to delineate how our measurements of pelvic stability relate to pelvic stability in a general and clinical sense. Clinically stability is usually defined as the pelvis' ability to allow only physiologic displacements under functional loads. Pelvic instabilities can result from displacements in any direction; however, clinically we generally concentrate on those displacements that are most significant to prognosis. For instance, for purposes of clinical classification Tile2 considers primarily rotational displacements of the iliac wings and vertical displacements at the SI joint.3 Our measurements of total displacement of the pelvis under load provide a measure of total pelvic stability. However, our measurements do not delineate where the displacements are occurring.
Cutting the PS decreased the stability of the pelvis by allowing the iliac wings to rotate in the horizontal plane, yet the clinically important Sl joint did not displace in the vertical plane. Therefore, although our measurements of stability are valid, they cannot be directly extrapolated to clinical situations. In additíon, we did not load the pelves to failure, so even though cutting the PS decreased stability substantially the pelvis may have been able to support functional loads without displacements at the clinically important SI joint. Finally, one must consider that we looked at ligament behavior only for one mode of pelvic loading (double leg stance). Nevertheless, our results do provide insight into how the pelvic ligaments interact to maintain overall pelvic stability.
Our results suggest that pelvic stability is maintained most effectively when the pelvicring remains intact. As long as the pubic symphysis and at least one of the two posterior ligaments (ASl and PSI) remain intact, the pelvic ring remains intact. If the pubic symphysis or both the posterior ligaments are cut, the pelvic ring is disrupted. With the pelvic ring intact, the pelves maintained an average of 89% of the intact pelvic stability; whereas with the ring disrupted, the pelves maintained only 73% of intact pelvic stability. It is difficult to conceive of an injury with isolated disruption of the ASI or PSI ligaments. However, these data do emphasize the importance of maintaining the pelvic ring for stability. In addition they suggest that the stability of the pelvis can be restored by simply restoring the integrity of the pelvic ring. Injuries in which the PS and ASI ligament are disrupted while the PSI ligament remains intact are not uncommon. Our data suggest that repairing the PS alone and thereby restoring the pelvic ring is adequate to repair pelvic stability.
Fig 3: Representative load displacement curves recorded before and after the pubic symphysis had been cut.
Fig 4: Figure showing the effect of disrupting the pelvic ring.
Fig 5: Histogram showing the effect of cutting sequence on a ligament contribution to pelvic stability. For example, cutting the PSI first in pelvis 1061 caused a negligible change in pelvic stability. On the other hand, cutting the PS! after the ASI had already been cut (pelvis 1 1 31) decreased pelvic stability substantially.
When the SS and ST ligaments were the only ligaments remaining intact, the pelves could not support the load. In addition, cutting these ligaments changed pelvic stiffness only slightly. It therefore seems unlikely that these ligaments by themselves play a major role in maintaining pelvic stability. This finding is somewhat inconsistent with current clinical thinking, as signs that these ligaments are disrupted (avulsion of their bony attachments) are usually indicative of major pelvic instabilities. It may be more reasonable to consider these ligaments not as isolated structures, but rather as integral parts of the pelvic floor. Indeed, when we harvested the pelves for testing it was difficult to distinguish these ligaments from the pelvic floor, which is why they were accidentally sacrificed in so many specimens. An intact pelvic floor may add greatly to the integrity of the pelvic ring, and thereby help maintain stability. Signs that these ligaments have been disrupted may simply suggest that the pelvic floor has been completely disrupted, implying complete disruption of the pelvic ring.
An intact PS provided more stability to the pelvic ring than might have been expected. Several biochemical studies comparing methods for stabilizing pelvic injuries with anterior and posterior disruptions have suggested that stabilizing the anterior pelvis alone does little to return pelvic stability.2,4"6 In our specimens when the PS alone remained intact, the pelves maintained 65% ( ± 12%) of intact pelvic stability. These results suggest that repairing the anterior pelvis in an injury with an anterior and posterior disruption would substantially improve pelvic stability. When analyzing these findings it is important to remember that we loaded our pelves in simulated double leg stance. Other studies looking at pelvic fixation have considered single leg stance modes of loading.2 s,6 When the pelvis is loaded in double leg stance, there are tensile forces at the PS and compressive forces at the posterior pelvis.7 When the PS is intact (either spared in the injury or repaired after the injury) the compression posteriorly would help to maintain pelvic stability.7-9 On the other hand, when the pelvis is loaded in a sitting position or in single leg stance, there are compressive forces at the PS and tensile forces posteriorly.6 With no compression posteriorly, the pelvis would be much less stable. When a patient has satisfactory anterior stability but questionable posterior stability, it would seem most reasonable to ambulate them in a swing-through gait.
A ligament's effect on pelvic stability depended not only on the ligament cut, but also on the other ligaments remaining intact. The behavior of specific ligaments provided further evidence of this phenomena. Cutting the PS first caused a 9% (±9%) reduction in pelvic stability. From this one might assume that the PS contributes 9% to pelvic stability. However, when the PS was the last remaining ligament it maintained 65% ( ± 1 2%) of intact pelvic stability. This suggests that the PS was able to increase its contribution to pelvic stability as other ligaments were sacrificed. Similar observations were made for the PSI and ASI. Clinically, this emphasizes the need to consider all structures intact and disrupted when trying to determine pelvic stability. Moreover, pelvic stability is the result of ligamentous interactions rather than simply the strength of individual ligaments. Thus, when one is trying to determine clinical stability it is important to consider possible ligamentous interactions and not just the ligaments disrupted.
Overall, this work must be considered preliminary. More definitive answers will require testing more specimens under different loading conditions. Our work suggests that the saeroiuberous and sacrospinous ligaments contribute little to stability directly. However, they may be important as part of the pelvic floor, and this possibility should be considered. Our work also suggests that the ligaments interact to create a composite stability. A ligament contribution to stability depends not only on the size of the ligament, but also on its position, the other ligaments intact, and the mode of loading. This implies that clinically pelvic stabilities cannot be neatly classified, and emphasizes Tile's point that pelvic stability must not be considered as black or white but as various shades of gray.
1. The M. Pelvic ring fractures: should they be fixed? J Bane Joint Surg. 1988: 7J)B: I- 1 2.
2. Tile M. Fractures of the Pelvis and Acetabulum. Baltimore. Md: Williams & Wilkins; 1984.
3. Pennal GF. Sutherland GO. Fractures of the Pelvis (motion picture |. Chicago, III: American Academy of Orthopaedic Surgeons Film Library: 1961.
4. Bell A. Smith AR. Brown TD. Nepola JV. Comparative study of the Orthofix and Pittsburgh frames for external fixation of unstable pelvic ring fractures. J Orthop Trauma. 1988;2:130-138.
5. Dahncrs LE. Jacobs RR. Jayaraman G. Cepulo AJ. A study of external skeletal fixation systems for unstable fractures. J Trauma. 1984; 24:876-881.
6. Mears DC. Rubasti HE. Pelvic and Acetabular Fractures. Thorolarc, NJ: SLACK Ine; 1986.
7. Pauwels F. Biomechanics of the Locomotor Apparatus. Berlin: Springer- Verlag: 1980.
8. Rolhkorier HJ. Berner W. Failure load and displacement of the human sacroiliac joint under in vitro loading. Arch Orthop Trauma Surf,. 1988: 107:283-287.
9. Slatis P. Karuharju EO. External fixation of unstable pelvic fractures: experiences in 22 treated with a trapezoid compression frame. Clin Orthop 1980; 151:73-80.