Loads on an internal spinal fixation device during physical therapy. (Technical Report).

Spines that are unstable (eg, due to fracture or degenerative disorder) are often stabilized by an internal spinal fixation device. The 2 longitudinal rods of the paired implanted fixation devices are fixed to pedicle screws inserted in the vertebrae adjacent to the unstable segments. During anterior interbody fusion, 1 or 2 intervertebral disks are removed and replaced (eg, by autologous bone grafts from the iliac crest). Shortly after surgery, physical therapy should start. The exercises should improve muscle strength (force) but not endanger the implants or the clinical outcome. Pedicle screw breakage occurs in 6% to 7% of cases (1,2) and is a major complication associated with internal spinal fixation devices. The implant screws normally fail through fatigue fractures after a great number of loading cycles. (3)

Spinal loads and loads acting on the fixation devices during the different exercises are widely unknown. However, after spinal surgery, patients want to know when they are allowed to sit, to walk without a crutch, or to carry a weight. Intradiskal pressure (pressure in the nucleus pulposus) is a measure for the spinal load and was measured for several body positions and activities. (4-9) These measurements were performed in volunteers with healthy disks. Rohlmann and colleagues (10-18) have measured the loads acting on the fixation devices for many activities. These measurements could show:

* that the force components perpendicular to the longitudinal axis of the implant and the torsional moment are small for most exercises. (10,11)

* that the axial force (force in axial direction of the rod, such as tension and compression) component in the fixation device depends mainly on the spinal level to which the implant is fixed. Axial forces are high when the fixation devices are mounted on the concave side of the spine, and they are low when the devices are fixed to the convex side. (12-14)

* that the bending moment (caused by 2 parallel forces that are equal in amount but act in opposite directions) in the fixation device depends mainly on the surgical procedure. Bending moments are high when the bridged region is distracted (when the distance of the upper and lower pedicle screws is increased), and they are low when it is compressed (when that distance is decreased). (12-14)

* that the implant loads are altered but not necessarily reduced due to insertion of a bone graft. (15)

* that in most cases the implant loads stay nearly constant in the postoperative temporal course. (15) This explains the breakage of pedicle screws that can occur more than half a year after surgery and indicates that screw breakage does not prove that fusion has not occurred. Even after fracture healing, the implants may be highly loaded. Therefore, pedicle screw breakage has nothing to do with bony fusion or pseudarthrosis.

* that bending moments in the fixation devices are small when the patients are lying. The measured median values, related to that for standing, were 26% for the supine position, 32% for the prone position, and 34% for the side-lying position. (12,16)

* that bending moments in the fixation devices are on average for sitting relaxed only 87% of the corresponding value for standing. (12,16)

* that sitting erect and actively straightening the back, as taught in some back schools, causes bending moments about as high as in standing (101%). (16)

* that the type of seats (stool, physical therapy ball, knee stool), as well as a padded wedge, has a negligible effect on the fixation device loads. (16)

* that walking causes the highest bending moments in the fixation devices from all activities performed regularly. On average, the bending moment during walking is about 128% of that for standing. (16) The walking speed has only a minor influence on the fixation device loads. (11)

* that fixation device loads are similar to those in a lying position for exercises where the spine does not have to carry the weight above but below a certain level, as during hanging with the hands on wall bars, balancing the body on parallel bars with the legs in a vertical direction, or hanging on the feet with the head upside down. (16,17) The pull of the partial body weight below the fixation devices is, in these cases, obviously compensated for by muscle forces.

* that carrying a load has only a slight effect on the bending moments in the fixation devices. (13)

* that there is good agreement between intradiskal pressure and bending moments in the fixation devices for most activities when the results are related to the corresponding values for standing. (16) There is no biomechanical reason that pressure should be higher for sitting than for standing. Stadiometric studies showed that the pressure must be lower for sitting than for standing.

* that a brace or harness does not reduce implant loads.

The aim of this article is to describe the loads acting on an internal spinal fixation device during selected movements in different body positions, including lying, sitting, standing, and kneeling on hands and knees.

Material and Methods

A bisegmental internal spinal fixation device (19) was modified (Fig. 1). Six load sensors, a telemetric unit, and a coil for the inductive power supply were integrated in the longitudinal rod of the device. The instrumented implant allowed the measurement of 3 force and 3 moment components acting in the implant. For the measurements, a flat coil and a small wire antenna were placed on the patients' backs. During each measurement session, the patients were videotaped and the load-dependent signals of the 2 telemetries were stored on the same videotape. The signals could be read online or from the videotape by a computer where the forces and moments were calculated and shown on a monitor. Details of the telemeterized implant, the measuring equipment, and the accuracy of the measuring implant are given elsewhere. (20,21) Calibration constants were checked in the laboratory after implant removal. They had not changed while the fixation devices were in place.

[FIGURE 1 OMITTED]

Modified fixation devices were implanted in 10 patients (4 male and 6 female). The Table provides data on the patients and the surgical procedures. The average age was 51 years (range = 34-72). Three patients evidenced a degenerative instability with spinal stenosis at 2 adjacent levels, and 7 patients had a fractured vertebral body. The level of the bridged vertebra varied between T11 and L4. Anterior interbody fusion using iliac crest bone grafts was performed in a second session, normally 2 to 4 weeks after insertion of the devices. Instrumented fixation devices were removed an average of 12 months (range = 3-21) after implantation.

The Ethics Committee of the Free University of Berlin approved clinical implantation of the modified fixation devices in up to 10 patients. Prior to surgery, the procedure was explained to the patients, and they gave their written consent to implantation of instrumented internal fixation devices and subsequent load measurements.

The implant loads were measured 1 to 3 times a week during hospitalization and about once a month thereafter. The number of measuring sessions per patient varied between 13 and 25. When possible, fixation device loads for some common body positions and activities were measured during all measurement sessions. These included lying in different positions (ie, supine, prone, and side-lying), sitting, standing, walking, lifting an extended leg and lifting the pelvis in a supine position, abduction of a leg, lifting of only the knees and of only the feet while lying in a side-lying position, bending of the upper body in different directions, and rotation of the upper body while sitting and standing. However, for some exercises (eg, flexing the back so that it was concave on the ventral side and extending the back so that it was concave on the dorsal side while kneeling on hands and knees), fixation device loads were measured in only a few patients. We did not specify how an exercise had to be performed because we were interested in the interindividual variation of fixation device loads for an exercise.

The average resultant peak bending moment in each fixation device for both sides was determined for the different activities and related to the corresponding value for standing. For the different activities, the median and 25th and 75th percentiles were calculated from the averages of all patients. The bending moments were related to the corresponding value for standing in order to compare the loads for different exercises measured in several patients and to estimate the risk of a certain activity for screw breakage. Patients are normally allowed to stand shortly (ie, 3-4 days) after surgery. The intraindividual variation of the bending moment in the implant for standing was small, but the interindividual variation was large (range = 0.7-6.9 N*m). In none of the patients studied were the bending moments for standing higher than the strength of the implant guaranteed by the manufacturer (7.5 N*m in a dynamic test over 5 million loading cycles). The design strength is therefore 110% of the maximum bending load observed in patients for standing. Activities that cause loads below this level are therefore not likely to increase the risk of pedicle screw breakage.

The bending moment in the implant for standing is, in most patients, much lower than 6.9 N*m. In these patients, the implant will never break. It was important to show that bending moments in the implants are in the region of the fatigue strength only during a few movements, such as walking and ventral flexion and extension of the upper body. The implants are most frequently loaded during walking. Therefore, walking is the exercise that causes the highest risk for pedicle screw breakage. The kind of surgery performed has an influence on maximum implant load. When the bridged region is distracted, patients should avoid ventral flexion and extension of the upper body. When the implant stress caused by the load during walking is higher than the fatigue strength, it is unlikely that screw breakage can be prevented.

Results

The average bending moment in a fixation device while standing was set at 100% for comparing other exercises with standing for each patient. The mean of the average bending moment of all patients for standing was 3.6 N*m (SD = 1.75, range = 0.7-6.9) (25th percentile = 2.2 N*m, 75th percentile = 4.8 N*m). When not indicated otherwise, measurements were performed on all 10 patients.

Implant Loads for Lying Body Positions

The bending moments in the fixation devices were small when the patients were in lying positions. The patients were asked to contract their muscles to achieve tension of their body. Muscle contraction for tension of the body led to a slight increase of the bending moments in the implant. Lifting an extended leg in a supine position (Fig. 2) caused an increase of the bending moments in the fixation devices from 26% to 66% of the value for standing. When lifting both extended legs (9 patients), the related peak bending moment in the devices was higher (101%) than when lifting one leg. Lifting of the pelvis in a supine position led to peak bending moments in the devices of up to 89%. When the pelvis was only slightly lifted, the load increase was smaller. Lifting head and shoulders in a supine position (6 patients) caused bending moments in the devices of about 88%. Movements of a leg (such as those produced during bicycling) in a supine position led to a peak value of the bending moment in the implants of 63%.

[FIGURE 2 OMITTED]

Lifting a cranially extended arm while lying in prone position caused a peak bending moment in the fixation devices of 91% (Fig. 2). Lifting an extended leg in this position led to bending moments of about 75%. When the right arm and the left leg were lifted simultaneously, the peak bending moment was 91%.

Abduction of an extended leg while lying in a lateral position increased the peak bending moment to 82% (Fig. 2). Lifting both feet but no other body parts while lying in a side-lying position led to a peak bending moment in the fixation devices of 85%, whereas lifting both knees but no other body parts increased the bending moment to 89%.

Bending Moments in the Implant During Sitting

The bending moments in the fixation devices were, on average, 13% lower for sitting relaxed than for standing (Fig. 3). However, sitting erect and actively straightening the back caused bending moments about as high as those observed in a standing position (100%). Ventral flexion of the trunk (bending the trunk forward) while sitting increased the peak bending moment to 105%. The corresponding value for extension of the trunk was 107%. Lateral bending of the trunk increased the bending moment in the implant mounted on the concave side to 108% and decreased it for the implant mounted on the convex side of the bent spine. Axial rotation of the upper body in the transverse plane while sitting increased the bending moments in the devices to 108%. The changes of torsional moment measured in the devices for axial rotation were about 0.55 N*m, on average, which is low in comparison with the bending moment.

[FIGURE 3 OMITTED]

Bending Moments for Different Activities While Standing

Standing up and sitting down led to average peak bending moments on the fixation devices of 112% and 119%, respectively (Fig. 4). Getting up on tiptoes (9 patients) increased the fixation device loads to 110%. Ventral flexion (127%), extension (124%), lateral bending (118%), and axial rotation (115%) of the upper part of the body during standing led to higher peak bending moments in the devices than the corresponding activities while sitting. Elevation of an extended arm while standing increased the bending moment in the devices to 110%.

[FIGURE 4 OMITTED]

Bending Moments While Kneeling on Hands and Knees Kneeling on hands and knees (9 patients) caused a bending moment of 69% (Fig. 5). Flexing the spine in this position (3 patients) increased the peak value to 97%, whereas extending the spine (4 patients) caused a peak bending moment in the fixation devices of 80%. Cranially outstretching the right arm (extending the arm cranially until it was horizontal) while kneeling on hands and knees or outstretching the right leg (9 patients) increased the bending moment to 100%. Outstretching the right arm and the left leg while kneeling on hands and knees (9 patients) led to a peak bending moment of 106%.

[FIGURE 5 OMITTED]

Other Activities

Turning from a supine to a side-lying position caused bending moments of 110%, and turning from a lateral to a prone position led to bending moments of 103%. There were great differences between individuals because no advice was given directly before performing the motion.

Discussion and Conclusions

Patients exhibited a large range of bending moments on their internal spinal fixation device. Indications for surgery, bridged vertebral level, and surgical procedure varied in these patients. These factors have an effect on the loads taken by the implant. Another reason for the large range of bending moments found for an exercise is that we did not specify how the exercise had to be performed. Methods of exercise performed by the patients varied. We did not measure, for example, the flexion angle of the upper body or the abduction angle of the leg with the patients in a side-lying position. These values differed from patient to patient, and this is one reason for the variation in implant loads. However, for several exercises, the load changes in the fixation devices occurred mainly during the first half of the exercise, and they were small afterward. Most of the activities studied were performed by all 10 patients, and many of the activities were performed during each measurement session. Normally, the patients repeated an exercise several times during a session. For some exercises, the results of several hundred trials were used to determine median loads. During some measurement sessions, patients had wound pain or had different muscle strength (compared with shortly after surgery when the patients felt weak and could not exercise a lot), which led to lower bending moments in the devices and expanded the range of fixation device loads for an exercise.

For the different exercises, average relative bending moments in the fixation devices measured in the patients were used to calculate the median value and 25th and 75th percentiles. The total range of fixation device load for an exercise may be much higher because intraindividual variations were not taken into account. Additionally, the bending moment for standing varied greatly from patient to patient. A relative load increase of 20% may be negligible when the absolute value for standing is low. In another patient with a high absolute value for standing, it may increase the risk of a poor surgical outcome.

Fractures of pedicle screws are nearly always fatigue fractures. The highest loads on the fixation devices were measured for walking. This exercise is normally performed very often and therefore leads to a great number of loading cycles. Thus, we believe walking is the loading that plays the major role concerning pedicle screw breakage. Implant loads of 110% or less of that value for standing are below the fatigue strength of the implant. Therefore, exercises that cause bending moments below this level are not likely to increase the risk of pedicle screw breakage. Bending moments between 111% and 120% of the value for standing may slightly increase that risk, especially when the exercises are performed very often. The exercises standing up, sitting down, and lateral flexion and axial rotation of the upper body while standing led to bending moments in this range. We believe that exercises that cause bending moments in the devices higher than 120% may increase the risk of pedicle screw breakage considerably if they are performed very often and if the bridged region is distracted during surgery. Walking and anterior flexion and extension of the upper body while standing were the exercises during physical therapy that caused bending moments in this range.

In an upright body position, the spine has to carry the upper body. This is not the case in a lying position. Therefore, small bending moments were measured when the patients were lying relaxed. Lifting both stretched legs while lying in a supine position caused the highest fixation device loads found for an exercise in a lying position. It was the only exercise where the bending moments were, on average, higher than for standing.

Sitting caused lower implant loads than standing. This finding is in agreement with the intradiskal pressure measurements performed by Wilke et al (9) and with stadiometric measurements done by Althoff et al. (22) However, these results contradict those reported by Nachemson (5) and Nachemson and Morris. (6) Using a stiff needle instead of a flexible transducer, as was done in the more recent study, (9) they found 40% higher intradiskal pressure values for sitting than for standing. From the biomechanical point of view, we believe that patients with back problems should be allowed to sit as soon as they can get up. Sitting erect and consciously straightening the spine led to higher bending moments in the fixation devices than sitting relaxed. The same was found for the intradiskal pressure. (16) Actively straightening the spine requires muscle forces that, in turn, lead to higher spinal loads. This increase, however, is relatively small, and the maximum value is lower than the peak magnitude for walking.

Flexion of the upper body in the different directions caused higher bending moments when the activity was performed while standing rather than while sitting. The load changes during these exercises varied little between standing or sitting because the bending moment for sitting relaxed was about 13% lower for sitting than for standing. Flexion of the upper body increases the muscle forces required for balancing this moment and thus the spinal load. Because the fixation device loads rose only slightly during flexion of the upper body in the different directions, other structures had to take over the additional load. Therefore, high spinal loads can be expected for these exercises. During ventral flexion of the upper body while standing, the intradiskal pressure increased, for example, to 216% of the value for standing. (16) The stiffness of the anterior column increases strongly (nonlinear behavior) during the exercise, whereas the stiffness of the implant does not change. Therefore, a slight compression of the anterior column strongly increases intradiskal pressure but only slightly increases the bending moment in the implant.

Kneeling on hands and knees led to lower bending moments than standing erect. Flexing the spine in this position caused higher implant loads than extending the spine. Lifting an arm and a leg in this position led to bending moments that were only 6% higher than for standing erect.

Fixation device loads were measured in 10 patients for different movements and positions. None of the movements studied caused bending moments in the implants higher than those observed during walking. Of those pedicle screws that break, nearly all break due to fatigue. This requires very frequent loads, which causes implant stresses that are higher than the fatigue strength of the implant. Few movements, such as ventral flexion and extension of the upper body while standing, increase the load on pedicle screws above 110% of that caused by standing. Walking is the only exercise, to our knowledge, that may meet these requirements.

Table.

Data on Patients and Surgical Procedures

                             Patient 1        Patient 2

Sex                          Female           Male
Age (y)                            59               34
Weight (kg)                        75               90
Height (cm)                       170              185
Indication                   Degenerative     Compression
                               instability      fracture
                                                (old)
Bridged vertebra             L3               L4
Implant levels               L2-L4            L3-L5
Bone grafts                  L2/3 and         L3-L5
                               L3/4
Time between implantation          10               21
  and removal (mo)
No. of measuring sessions          19               25
Distraction                  Yes
Compression                                   Yes
Lordosis increased

                             Patient 3        Patient 4

Sex                          Female           Male
Age (y)                            54               72
Weight (kg)                        66               80
Height (cm)                       162              171
Indication                   Compression      Degenerative
                               fracture         instability
Bridged vertebra             L3               L4
Implant levels               L2-L4            L3-L5
Bone grafts                  L2/L3            L3/4 and
                                                L4/5
Time between implantation           8                3
  and removal (mo)
No. of measuring sessions          14               18
Distraction                  Yes              Yes
Compression
Lordosis increased           Yes

                             Patient 5        Patient 6

Sex                          Male             Male
Age (y)                            36               42
Weight (kg)                        75               81
Height (cm)                       173              164
Indication                   Compression      Degenerative
                               fracture         instability
Bridged vertebra             T11              L4
Implant levels               T10-T12          L3-L5
Bone grafts                  T10/T11          L3/4 and
                                                L4/5
Time between implantation           7               20
  and removal (mo)
No. of measuring sessions          14               24
Distraction                  Yes              Yes
Compression
Lordosis increased           Yes

                             Patient 7        Patient 8

Sex                          Female           Female
Age (y)                            46               62
Weight (kg)                        53               85
Height (cm)                       165              157
Indication                   Compression      Compression
                               fracture         fracture
                               (old)            (old)
Bridged vertebra             T12              L1
Implant levels               T11-L1           T12-L2
Bone grafts                  T11/T12          T12-L2

Time between implantation          12               15
  and removal (mo)
No. of measuring sessions          13               17
Distraction                                   Yes (a)
Compression                  Yes
Lordosis increased           Yes

                             Patient 9        Patient 10

Sex                          Female           Female
Age (y)                            47               54
Weight (kg)                        48               68
Height (cm)                       160              171
Indication                   Compression      Compression
                               fracture         fracture
Bridged vertebra             L1               T12
Implant levels               T12-L2           T11-L1
Bone grafts                  T12/L1           T11/T12
Time between implantation           8               14
  and removal (mo)
No. of measuring sessions          16               17
Distraction
Compression
Lordosis increased           Yes              Yes

(a) Plus anterior Zielke fixation device.

References

(1) Yahiro MA. Comprehensive literature review: pedicle screw fixation devices. Spine. 1994; 19 (suppl 20):S2297-S2299.

(2) Yuan HA, Garfin SR, Dickman CA, Mardjetko SM. A historical cohort study of pedicle screw fixation in thoracic, lumbar, and sacral spinal fusions. Spine. 1994; 19 (suppl 20):S2279-S2296.

(3) Rohlmann A, Bergmann G, Graichen F, Mayer H-M. Placing a bone graft more posteriorly may reduce the risk of pedicle screw breakage: analysis of an unexpected case of pedicle screw breakage. J Biomech. 1998;31:763-767.

(4) Nachemson AL. The load on lumbar disks in different positions of the body. Clin Orthop. 1966;45:107-122.

(5) Nachemson AL. Disc pressure measurements. Spine. 1981;6:93-97.

(6) Nachemson AL, Morris JM. In vivo measurements of intradiscal pressure. J Bone Joint Surg Am. 1964;46:1077-1092.

(7) Andersson GBJ, Ortengren R. Lumbar disc pressure and myoelectric back muscle activity during sitting, 3: studies on a wheelchair. Scand J Rehabil Med. 1974;6:122-127.

(8) Andersson GBJ, Ortengren R, Nachemson AL. Intradiskal presure, intra-abdominal pressure and myoelectric back muscle activity related to posture and loading. Clin Orthop. 1977;129:156-164.

(9) Wilke H-J, Neef P, Caimi M, et al. New in vivo measurements of pressures in the intervetebral disc in daily life. Spine. 1999;24:755-762.

(10) Rohlmann A, Bergmann G, Graichen F, Mayer H-M. Telemeterized load measurement using instrumented spinal internal fixators in a patient with degenerative instability. Spine. 1995;20:2683-2689.

(11) Rohlmann A, Bergmann G, Graichen F. Loads on an internal spinal fixation device during walking. J Biomech. 1997;30:41-47.

(12) Rohlmann A, Bergmann G, Graichen F. Loads on internal spinal fixators measured in different body positions. Eur Spine J. 1999;8: 354-359.

(13) Rohlmann A, Graichen F, Bergmann G. Influence of load carrying on loads in internal spinal fixators. J Biomech. 2000;33:1099-1104.

(14) Rohlmann A, Graichen F, Weber U, Bergmann G. 2000 Volvo Award winner in biomechanical studies: monitoring in vivo implant loads with a telemeterized internal spinal fixation device. Spine. 2000; 25:2981-2986.

(15) Rohlmann A, Bergmann G, Graichen F, Weber U. Changes in the loads on an internal spinal fixator after iliac-crest autograft. J Bone Joint Surg Br. 2000;82:445-449.

(16) Rohlmann A, Claes LE, Bergmann G, et al. Comparison of intradiscal pressures and spinal fixator loads for different body positions and exercises. Ergonomics. 2001;44:781-784.

(17) Rohlmann A, Bergmann G, Graichen F, Mayer H-M. Influence of muscle forces on loads in internal spinal fixation devices. Spine. 1998;23:537-542.

(18) Rohlmann A, Bergmann G, Graichen F, Neff G. Braces do not reduce loads on internal spinal fixation devices. Clin Biomech. 1999;14: 97-102.

(19) Dick W. Internal Fixation of Thoracic and Lumbar Spine Fractures. Bern, Switzerland: H Huber Publisher; 1989.

(20) Rohlmann A, Bergmann G, Graichen F. A spinal fixation device for in vivo load measurement. J Biomech. 1994;27:961-967.

(21) Graichen F, Bergmann G, Rohlmann A. Patient monitoring system for load measurement with spinal fixation devices. Med Eng Phys. 1996;18:167-174,

(22) Althoff I, Brinckmann P, Frobin W, et al. An improved method of stature measurement for quantitative determination of spinal loading: application to sitting postures and whole body vibration. Spine. 1992; 17:682-693.

A Rohlmann, Dr-Ing, is Research Fellow, Orthopaedic Biomechanics Laboratory, Free University of Berlin, UKBF, Hindenburgdamm 30, 12200 Berlin, Germany. Address all correspondence to Dr Rohlmann.

F Graichen, Dr-Ing, is Research Fellow, Orthopaedic Biomechanics Laboratory, Free University of Berlin.

G Bergmann, Dr-Ing, is Professor of Biomechanics, Orthopaedic Biomechanics Laboratory, Free University of Berlin.

All authors provided concept/project design, writing, and data collection and analysis. Dr Rohlmann and Dr Bergmann provided project management and fund procurement. Dr Graichen provided facilities/equipment. The authors acknowledge the friendly cooperation of their patients.

The study was approved by the Ethics Committee of Free University of Berlin.

This work was supported by a grant from the Deutsche Forschungsgemeinschaft, Bonn, Germany (Ro 581/7-2).

This article was submitted December 8, 2000, and was accepted April 6, 2001.