.I. Bzomechanics Vol. 25. No. 8, pp. 891-901, Printed in Great Britam

0021-9290192

1992.

c8 1992 Pergamon

$5 00 + .oO Press Ltd

THE INFLUENCE OF DYNAMIC FACTORS ON TRIAXIAL NET MUSCULAR MOMENTS AT THE L5/Sl JOINT DURING ASYMMETRICAL LIFTING AND LOWERING DENIS GAGNON*? and MICHELINE GAGNON~ tFacult6 d’kducation physique et sportive, Universitd de Sherbrooke, Sherbrooke, Qubbec, Canada, JI K 2Rl and $Dtpartement d’tducation physique, Universitk de Montrkal, Montrkal, Quebec. Canada,

H3C 357

Abstract-Asymmetrical lifting and lowering are predominant activities in the workplace. Mechanical causes are suggested for many back injuries and the dynamic conditions within which spine loading occurs are related to spine loading increase. More data on tridimensional biomechanical lumbar spine loading during asymmetrical lifting and lowering are needed. A tridimensionaf dynamic multisegment model was developed to compute spinal loading for asymmetrical box-handling situations. The tridimensional positions of the anatomical markers were generated by a direct linear transformation algorithm adapted for the processing of data from two real and two virtual views (mirrors). Two force platforms measured the external forces. Five male subjects performed three variations (slow, fast and accelerated) of asymmetric lifting and two variations (slow and fast) of asymmetric lowering. The torsional, extension/flexion and lateral bending net muscular moments at the LS/Sl joint were computed and peak values selected for statistical analysis. For the lifting task, the fast and accelerated conditions showed significant increases over the slou condition for torsion, extension/flexion and lateral-bending moments. The accelerated condition also showed significant increases over the fast condition for extension. A comparison between lifting and lowering tasks showed equivalent loadings for torsion and extension. The moments were compared to average maximal values measured on equivalent male subject populations by isokinetic dynamometry. This showed torsional and extension values of 30 and 83% of the maximal possible subject capacity, respectively. These results demonstrated that dynamic factors do influence the load on the spine and highlighted the influence of both lifting and lowering on the loading of the spine. This suggested that for a more complete analysis of asymmetrical handling, the maximal velocity and acceleration produced during lifting should be included.

more data on biomechanical lumbar spine loading during asymmetrical MMH are needed (Chaffin, 1988; Chaffin and Andersson, 1984). The dynamic conditions within which spine loading occurs are related to lumbar spine loading increase (Drury et al., 1989; Lavender et al., 1989). Theoretical considerations also suggest that the rate of loading of the spine may lead to an accelerated onset of degenerative changes of the lumbar structures (Sandover, 1983). These findings are consistent with the recommendations for the use of dynamic biomechanical modelling. Dynamic models should improve the understanding and accuracy of lumbar spine loading estimation during MMH activities (Marras et nl., 1987; McGill and Norman, 1985). Effects of the type of task (e.g. lifting vs lowering) are rarely assessed. Most biomechanical studies consider lifting and few studies, the lowering part of a task. Nevertheless, lowering tasks are frequent in the workplace (Drury et al., 1982; Lamonde, 1987). In fact, the loading on the spine is quantitatively similar for the lifting and lowering parts of sagittal-plane MMH tasks (Gagnon and Smyth, 1991; Leskinen et ul., 1983). No similar comparisons are found in the literature for asymmetrical MMH tasks. The purpose of the present study was to estimate the influence of the rate of task execution (dynamic face tors) for two types of asymmetrical MMH task (lifting

INTRODUCTION Manual materials handling (MMH) remains a frequent activity in modern industries. Most MMH tasks are physically demanding to the worker. Epidemiological data from the United States show that 23% of all the compensated injuries to the back are related to MMH (Snook, 1978; Snook et al., 1978). Mechanical causes are suggested as being responsible for many back injuries (Andersson, 1981). The tasks involving asymmetrical body motions are predominant in the workplace (Drury et al., 1982). Studies assessing mechanical variables for these types of effort are sparse. Most studies related to asymmetrical MMH focus on physiological or psychophysiological stresses (Drury et ul., 1989; Garg and Banaag, 1988; Kumar, 1980, 1984; Mital and Fard, 1986). Evaluations of compression and shear forces at the L5/Sl joint during symmetrical and asymmetrical box lifting (Kromodihardjo and Mital, 1987; Mital and Kromodihardjo, 1986) show significant effects of box size and handles on spine loading. More compression is observed for symmetrical lifting while more shearing is present for asymmetrical lifting. This suggests that Received in $nal form 26 November 1991. *Data acquisition and preliminary numerical conducted at Universitk de Montr&al.

analyses

891

D. GAGNONand M. GAGNON

892

and lowering) on the net muscular moment components at the L5/Sl joint. A dynamic tridimensional multisegment model was used to compute the torsional, extension/flexion and lateral-bending net muscular moments at the selected joint. It was hypothesized that a higher rate of task execution may increase the net muscular moment components at the L5jSl joint. It was also postulated that asymmetrical lowering may impose equally high moments on the L5jS.l joint as asymmetrical lifting under similar dynamic conditions.

EXPERIMENTAL

PROCEDURES

bending. Once this position was determined, the location of each foot was sketched on the respective force platform. These individual footprints standardized the feet positions of each subject for all the trials. The average lateral distance between the toes was 538 mm (S.D. 30 mm) and between the heels was 392 mm (31 mm). Each subject adopted a diagonal manual grip during the practice phase (Drury et al., 1982). This grip was maintained the same for all the tasks in order to standardize the hand position on the box. The box was not equipped with handles because they are not frequently present on the boxes handled in industry (Drury et al., 1982).

Subjects

Measurement techniques

Five male physical education students volunteered to participate in the study. Their average age was 29.8 yr (range 22-39 yr), their mass, 77.4 kg (68.1-87.1 kg), and their height, 1.79 m (1.71-1.88 m). None of the subjects had previous MMH experience or a medical history of low-back or spine injury.

Stereocinematography and force platforms were used to collect the tridimensional position and force data, respectively. The tridimensional positions of the anatomical markers were obtained from the coordinates digitized from processed tine films by a direct linear transformation (Marzan, 1975). The computational algorithm was adapted from the work of Walton (1981). Two Locam (Redlake Corporation, SantaClara, CA) 16 mm cameras linked to a synchronization unit were used to film the subjects. The cameras had a macro lens (focal length adjustable from 12 to 120 mm) and were powered by a direct-current unit. A filming speed of 100 frames s- 1 was selected. A preliminary study demonstrated that the Locam cameras were not synchronizing properly below this speed. Each camera was coupled with a large adjustable mirror (width 2.4 m and height 1.8 m). This produced a real and a virtual view of the subject on the same image. Calibration of the four available views was done by filming 24 control points symmetrically distributed in a 2.2 m3 volume (height 2.0 m, width 0.9 m and depth 1.2 m). The control points were previously measured to an accuracy of 1 mm and a precision of 0.5 mm. These targets were reconstructed with maximal root mean square (rms) errors of 3.4,4.7 and 3.5 mm for the X, Y and Z axes, respectively. The calibrated volume enclosed the subject during the execution of the asymmetrical box-handling tasks. Following the processing of tine films, images were digitized on a NAC motion analyzer (Instrumentation Marketing Corporation, Burbank, CA) linked to a PDPll/23 computer (Digital Equipment Corporation, Maynard, MA). For each view, the control points were digitized twice and the average value was retained in the data file. The anatomical markers were then digitized at every other frame (effective rate 50 frames s-i). Ten extra frames were included at the beginning and at the end of the trial. The external forces and their points of application were measured with two AMTI (Advanced Mechanical Technology, Inc., Newton, MA) force platforms (one under each foot). A PDPl l/73 computer collected the force signals at a frequency of 100 Hz via

Tasks

Each subject performed two repetitions of each of the three variations of an asymmetric box-lifting task and of each of the two variations of an asymmetric box-lowering task (Fig. 1). The initial and final locations of the box were adjusted to each subject while he was standing in the anatomical position. For the lifting task, the top of the box was located 1 to 5 cm below the hip (94+4 cm from the floor) in the initial position (Fig. 1A). The bottom of the box was located 1 to 5 cm above shoulder height (160 + 5 cm from the floor) in the final position (Fig. 1B). A rotation angle of 90” about the vertical axis was estimated between the initial and final box positions. For the lowering task, the initial and final box positions were inverted. The box was of standard size and mass for the industry (height 230 mm, width 350 mm, depth 243 mm and mass 11.6 kg). The classification of the tasks was based on the rate of execution. The lifting tasks included a slow, a fast and an accelerated condition. The lowering task consisted of a slow and a fast condition. For the slow conditions, the subjects controlled the box from the beginning to the end of the task. For the fast conditions, the subjects maintained the box velocity for the lifting tasks and guided the box for the lowering tasks. For the accelerated conditions, the subjects gave a strong initial acceleration to the box. The five experimental conditions were randomly administered for each subject. Sufficient time to practice the tasks was allowed to each subject before performing experimental trials. The same instructions were given to all subjects. Each subject selected one position for a comfortable execution of all the tasks. To permit some flexibility in this choice, no other instructions were given to the subjects regarding specific joint torsion, flexion or

Fig. 1. Illustration of the initial and final positions for the lifting and the lowering tasks. For the lifting task, the box was moved in a continuous diagonal movement starting on the left side of the subject (A), the top of the box being just above hip height, and finishing to a shelf in front of the subject slightly above shoulder height(B). The inverse trajectory was followed for the box-lowering task.

893

895

Dynamic factors in asymmetrical lifting interactive software and saved the data in binary format on a hard-disk drive for further processing. The same interactive software controlled the synchronization of the film and the force data during the acquisition session. At the initiation of the force recording, the program generated an electrical pulse. This pulse was used to trigger two light emitting diodes (LEDs) for each camera. One was inside the camera and the other outside but in its field of view. These signals were then used during the digitizing phase to synchronize the beginning of the force recording with the appropriate film sequence. Statistical analyses

Repeated measures analyses of variance were used the effect of the independent variables (rates of lifting and lowering, and the type of task) on the dependent variables (kinematic and kinetic variables). The equations of a standard ANOVA model (Daniel, 1974; Keppel, 1982) were macro-programmed into a spreadsheet (Lotus l-2-3, Cambridge, MA). This automated the analysis process and provided a summary table of the statistical results. A difference was considered significant with p < 0.05. All the experimental conditions were first evaluated to detect any significant difference between the two repetitions of each task. The average values for two repetitions of the same task were then used to locate the differences between the experimental treatments. For the lifting tasks, comparisons were made to assess the differences between the slow, fast and accelerated conditions. For the lowering tasks, the dynamic effects were assessed by comparing the slow and the fast conditions. Comparisons between the lifting and the lowering tasks were executed for the slow and for the fast conditions. First, comparisons were performed for five dependent kinematic variables. These were: the two maximal velocities (linear of the C7/Tl joint and angular of the trunk), the two maximal accelerations (linear of the C7/Tl joint and angular of the trunk) and the task duration. Second, comparisons were made for seven dependent kinetic variables. These were: the three maximal positive net muscular moment components (torsion, extension/flexion, lateral bending), the three maximal negative net muscular moment components (torsion, extension/flexion, lateral bending), and the maximal resultant net muscular moment at the L5/Sl joint. The concept of net muscular moment assumes that the net moment balances the moment generated by the muscle forces (Crowninshield and Brand, 1981). This simplifying assumption presumes the absence of muscular co-contractions. It also neglects the moments generated by the other anatomical structures (i.e. ligaments, capsule and the non-contractile portion of muscles) and by the intra-abdominal force (Cappozzo, 1983). For some cases, more than two means were compared. A post hoc analysis based on the Scheffe method

was utilized to locate the conditions that were significantly different. A pdO.05 significance level was accepted. THE DYNAMIC

TRIDIMENSIONAL

MULTISEGMENT

MODEL

The model included eight segments. These segments were: both feet (from the fifth metatarsal to the ankle), both shanks (from the ankle to the knee), both thighs (from the knee to the hip), the pelvis (from the hips to the L5/Sl joint) and the trunk (from LSjSl to C7/Tl). Each segment was considered as a rigid body. The inverse dynamic calculations of the net forces and moments at each joint proceeded segment by segment from each foot toward the L5/Sl joint.

to evaluate

Coordinate systems

Two types of coordinate systems (Fig. 2) were used, namely, the fixed and the anatomical coordinate systems (Paul, 1981). The fixed coordinate system was defined during the camera calibration process using the position data from the control points of the calibration structure (Walton, 1981). The spatial orientations of the anatomical axes were computed utilizing three markers on each segment. Two of these three

Fig. 2. Fixed {0, X, Y, Z} and anatomical {N, L, T. S} coordinate systems for the eight link segments of the model and sign conventions. The longitudinal &,) axis points from the proximal to the distal extremity of the segment; the transverse (TN)axis points from the left to the right side of the segment; the sagittal (S,) axis points from the back to the front of the segment.

896

D. GAGNON and M. GAGNON Table 1. Positive sign (+) net muscular moments conventions Anatomical axis Joint

Segment

L5/Sl

Trunk

Longitudinal Torsion of left shoulder toward right hip

markers were located at the proximal and distal extremities of the limb. A third marker was located on each segment, not collinearly with the first two, so as to provide the two remaining anatomical axes via vector products. Details on marker location and procedure are available in Gagnon (1990). In this study, the net muscular moments at the L5/Sl joint were expressed about the axes of the anatomical coordinate system of the trunk. Sign conventions for the net muscular moments about the L5/Sl joint are presented in Table 1.

Dynamic equilibrium

A tridimensional inverse dynamic analysis was performed on each segment starting with the feet and proceeding sequentially upwards through the shanks, thighs, pelvis and trunk (L5/Sl). The segmental model included the net reaction forces and moments at each joint, the external forces measured by one distinct

Extension

Sagittal Lateral bending to the right

force platform under each foot, the segment’s weight and moments of inertia. The inertial forces and moments were determined from the kinematic analysis. The mass, center of mass and moments of inertia of each segment of the model were estimated by the regression equations provided by Zatsiorsky and Seluyanov (1981, 1983), which are based on subject’s mass and height. Similar tridimensional analytic approach is used for gait analysis (Antonsson, 1982). The dynamic equilibrium equations are detailed in the Appendix.

Linear and angular kinematics

The linear displacements of the anatomical markers were obtained through cinematography. The first and second derivatives of the linear displacement data were computed by numerical differentiation using the central difference method (Winter, 1979; Wood, 1982). Linear displacement data were first filtered and then differentiated twice to obtain the linear velocities and accelerations. These procedures were repeated for each anatomical marker (n = 19) and for each axis (X, Y, Z) of the fixed coordinate system. A fourth-order zero-phase Butterworth filter was used (Winter, 1990) with an adjusted cutoff frequency for each marker and each axis. The spectral characteristics of each marker, computed with a fast Fourier transform using a Parzen data window (Press et al., 1988), were inspected. The cutoff frequencies were selected utilizing a residual analysis (Winter, 1990) for each marker and axis. The cutoff frequency values ranged from 0.6 Hz for the metatarsal joint Z axis to 3.1 Hz for the C7/Tl joint Y axis. The angular kinematics were calculated from the linear kinematics. The angular velocities were determined by the product of the derivative of the segment’s direction cosine matrix as a function of time and the transpose of the direction cosine matrix (Ramey and Yang, 1981). Numerical differentiation was employed to calculate the angular accelerations (Kromodihardjo and Mital, 1986,1987). The procedures for axes transformation are described in the Appendix.

Transverse

RESULTS Influence

of the dynamic factors

The kinematic parameters describing the lifting conditions (slow, fast and accelerated) and the lowering conditions (slow and fast) were computed (Table 2). Within the limitations of the present dynamic tridimensional multisegment model, the C7/Tl joint was selected, among the available anatomical points, as the more representative marker for the different lifting rates. Statistically significant linear and angular velocity and angular acceleration increases were present between most of the lifting and lowering rates. Longer durations were measured for the slow conditions and shorter durations were measured for the fast and for the accelerated conditions. Coefficients of determination (r’) for the significant effects ranged from 0.51 to 0.75 and from 0.20 to 0.46 for the lifting and lowering, respectively. Lifting tasks. All the agonist components of the L5/Sl moment were significantly higher for the fast and accelerated conditions than for the slow condition (Table 3). For the antagonist moment components, only the flexion component of the accelerated condition was significantly higher than the flexion components for the other two conditions (rZ =0.39-0.69). The presence of this antagonist moment component could be attributed to a counteraction produced by the abdominal muscles at the end of the lifting task. Lowering tasks. Similarities were observed between the execution conditions for the lowering task (Table 3). Only the torsional muscular moment components showed a significant increase for the fast over the slow lowering condition (r’ = 0.23-0.40). Influence of the type of task Comparisons between lifting and lowering were performed on the basis of each distinct dynamic factor

897

Dynamic factors in asymmetrical lifting Table 2. Descriptive statistics for selected kinematic variables (n = 5) Lifting Variable

Lowering

Slow

Fast

Accel.

Slow

Fast

Maximal linear velocity at the C7/Tl joint (m s- ’ )

w

0.69 (0.11)

0.95s (0.11)

1.20s.r (0.14)

0.76 (0.24)

1.03s (0.29)

Maximal linear acceleration at the C7/Tl joint (ms-*)

aw

4.32 (1.07)

6.62 (1.43)

8.19s (2.03)

5.70 (2.48)

6.47 (3.00)

Maximal angular velocity of the trunk (rads-‘)

w3

1.79 (0.14)

2.37 (0.38)

3.01S (0.47)

1.80 (0.25)

2.43s (0.44)

Maximal angular acceleration of the trunk (rad s-‘)

as

5.73 (1.01)

8.22s (2.22)

12.26S,F (2.31)

6.28 (1.95)

9.81s (1.91)

Execution time (s)

ax

1.33 (0.16)

o.88s (0.04)

0.75S’F (0.04)

1.50 (0.19)

o.94s (0.20)

SD.

S.D.

S.D.

S.D.

SD.

pGO.05; ‘different from slow; Fdifferent from fast; no significant difference was found between slow lifting and lowering and between fast lifting and lowering for the maximal angular and linear velocity and acceleration at the C7/Tl joint and for execution time. Table 3. Descriptive statistics and ANOVA results about the effects of dynamic factors and the type of task on the maximal net muscular moments (Nm) at the L5/Sl joint during lifting and lowering (n = 5)

Task

Net muscular moment

Antagonist Slow

Agonist

Fast

Accel.

-15 (3)

-12 (3) -9s.r (5) -4 (7) N.A.

Slow

Fast

Accel.

Lifting avg S.D. avg S.D. avg S.D. avg S.D.

-14 (5) -1T (1) N.A.

-2 (3) N.A.

ML.

avg SD.

M*

avg S.D.

28 (6) 133 (26)

avg S.D.

(::)

avg S.D.

138 (28)

39’ (8) 136 (18) 37’ (8) 143 (20)

M, Mr Ms Ma

123 (16) 43 (8) 127 (18)

37s (6) 139s (18) 54sT (8) 147s (19)

164sF (21) 65’ (15) 175sF (22)

Lowering

Ms Ms

N.A.

-14 (4)

N.A. N.A. N.A.

(0”) -14’ (8) N.A.

-18s (4) -1 (1) -13 (11) N.A.

N.A. N.A. N.A. N.A.

p GO.05; ‘different from slow; Fdifferent from fast; %fting different from lowering; M,: moment about the longitudinal axis (torsion); MT: moment about the transverse axis (flexion or extension); M,: moment about the sagittal axis (lateral bending); M,: resultant moment; N.A.: not applicable; agonist muscular moment acts in the same direction as the global trunk movement while antagonist muscular moment acts in opposition to the global trunk movement.

(slow or fast). Statistically equivalent lifting and lowering maximal linear (C7/Tl joint) and angular (trunk) resultant velocities and accelerations values were recorded (Table 2). Slow execution conditions. The results emphasized the similarities between the types of task (Table 3). The only component of the maximal net muscular moment that showed a significant difference between the slow

lifting and the slow lowering conditions was the negative lateral-bending component (r2 = 0.54). A similar, but not significant tendency, was noticed for the maximal positive lateral-bending moment component. Fast execution conditions. These results also showed the similarities between the types of task (Table 3). Only the positive lateral-bending moment

D. GAGNON and M. GAGNON

898

component presented a significant difference between the fast lifting and lowering tasks (r2 =0.52). The lack of significant differences for the resultant maximal net muscular moments between the lifting and the lowering tasks (slow and fast conditions) reinforced the tasks equivalence argument.

DISCUSSION

Limitations of the method It was necessary to assess the sensitivity to the formulation of the mechanical model and validity of the net muscular moment approach to ensure confidence in the results. The sensitivity of the net muscular moments at the L5/Sl joint was evaluated for a static box-handling task. Variations of 3-10 mm and of 3-22 N on each of the three axes of the fixed coordinate system were induced in the joint center position and external force values, respectively. Maximal sensitivity values of 3 Nm about the longitudinal axis (torsion), 12Nm about the transverse axis (extension/flexion) and 10 N m about the sagittal axis (lateral bending) were generated. These sensitivity values were comparable to other results from the literature obtained at the lumbar spine level (Cappozzo, 1983) and were smaller for all the anatomical axes than the statistically significant differences obtained in the present study. The validity of the results was indirectly estimated by comparing the maximal muscular moment components of this study with those found in the literature for asymmetrical lifting under comparable experimental conditions. A direct validation of the muscular moment results was not feasible. This validation limitation originated from the absence of the upper arm and arm segments in the tridimensional multisegment model. The only data available to our knowledge for such comparisons were published by Kromodihardjo and Mital(1987); see also Kromodihardjo (1986). For a lifting task performed with a 11.3 kg box without handles, their torsional, extension and lateral-bending moments were 20, 150, 10 N m, respectively. These values were comparable to the average values found in this study for the slow lifting task (Table 3). In the present study, negligible muscular co-contraction was assumed. Co-contractions occur during activities involving heavy physical exercise and/or muscular fatigue. It is impossible to get a quantitative evaluation of co-contractions without dynamic forcecalibrated electromyographic (EMG) measurements. In this study, appropriate rest was allowed to the subjects between each trial to minimize the development of muscular fatigue. The lifting and lowering tasks were not perceived as heavy physical activities by any one of the subjects according to their recorded comments after each experimental session. The question of the contribution of passive and active anatomical structures to the net intersegmental moment is also of interest. Recent results from Vrahas et al. (1990)

support the hypothesis that this contribution is minimal. In this study, this contribution is less than 5 N m for the net intersegmental moments at the hip joint during normal walking. Nevertheless, very few studies reported results on this matter. Injlluence of the dynamic factors

The results of the present study show the influence of the dynamic factors on the components of the maximal net muscular moment at the L5/Sl joint during lifting tasks. A study by Bush-Joseph et al. (1988) investigating different lifting rates for symmetrical tasks in the sagittal plane demonstrates the same trends. A general outcome of the results of the present study is the need for a reevaluation of some of the standard MMH recommended guidelines for symmetrical lifting (NIOSH, 1981). For example, the relation between the weight of the load handled and its distance from the body should also consider the velocity and the acceleration of the load. This should decrease the values for the safe load and of its distance from the body in proportion to the increase in velocity and acceleration of the load. This is supported by findings on the measurement of dynamic back strength, which indicate a strong relationship between the increase in velocity of lifting and the decrease in back strength (Kishino et al., 1985; Kumar et al., 1988). The loading rate of the lumbar spine may increase the risk of injury to the intervertebral joint structures by altering their mechanical properties (Smeathers and Joanes, 1988). Estimates of the level of risk for the more constraining lifting tasks performed in the actual study were calculated. The measured maximal muscular moment values about two (longitudinal and transverse) of the three anatomical axes were selected from two different studies using modified isokinetic dynamometric equipment at a constant velocity of 120” s-l. This velocity was comparable to the average maximal angular velocity of the trunk in torsion (120” s - ‘) and extension (126” s- 1) for the accelerated lifting conditions. The torsional moments were equivalent to 30% of the average value for the maximal muscular moments obtained from 62 male subjects by Smith et al. (1985). For the extension moments, this figure was 83% compared to the average value from 98 male subjects gathered by Davies and Gould (1982). These percentages give an estimate of the relative demand of the accelerated tasks. Despite their asymmetrical nature, extension dominates the muscle action. In fact, the extension moment components were always higher than the other components by a factor ranging from three to five. This could be attributed to a greater vertical than horizontal translation of the box, thus requiring a greater activity from the muscles acting to extend the trunk. Moreover, it was observed that the subjects were rotating their hips to partially face the load origin and destination (Fig. 1). This minimized the amount of trunk torsion and lateral bending and promoted the extensor muscles’ action.

Dynamic

factors

1

2037

Fast

h

‘:.

IGO-

E

I

n.

I

1.

y).

_.-L

I

1,.

Ld

250

ia -;a*

150

Accelerated

4..

-‘“%.

n.

1c.l

Movement Time (%) Fig. 3. Typical net muscular moment curves for one subject during slow, fast and accelerated lifting. Positive-sign moment components denote that (1) the rotator muscles pulled about the longitudinal axis to bring the left shoulder toward the right hip; (2) the extensor muscles pulled about the transverse axis to extend the trunk toward the vertical position; (3) the lateral-bend muscles pulled about the sagittal axis to bend the trunk to the right. The negative-sign moment components produce action in the opposite direction about each anatomical axis. The curves illustrate that the transition from slow to fast and to accelerated lifting primarily affected the magnitude of each moment component. The evidence of a trunk flexion moment for the accelerated lifting condition indicates the presence of an antagonist muscular action to control the initial acceleration transmitted to the trunk by a trunk extension moment applied during a short period. The net muscular moment curves (Fig. 3) were consistent with these observations and also with the results of studies reporting the measured EMG of trunk muscles during lifting activities (Andersson et al., 1980; Zetterberg et al., 1987). In a review, Andersson and Winters (1990) report that during lateral flexion and twisting the EMG activity of posterior back and abdominal muscles increases mainly on the side contralateral to the direction of the lateral bend. This type of pattern was observed on the net muscular moment curves (Fig. 3).

InJuence

in asymmetrical

of the type of task

A high frequency of box-lowering tasks is observed in industry (52% of the 2000 MMH tasks, Drury et al., 1982). High injury risks are associated with the handling of boxes with the aid of gravity (Lamonde, 1987).

899

lifting

Few studies compared directly the lifting and lowering tasks. Studies comparing sagittal-plane lifting and lowering (Gagnon and Smyth, 1991; Leskinen et cd.. 1983) demonstrate slight differences between the two types of tasks, the lifting task creating 10P15% more lumbar spine loading. These findings concur with the results of the present study. From another perspective, lowering tasks require mostly eccentric strength and are less demanding in muscular mechanical energy production (Gagnon and Smyth, 1991). This makes the lowering tasks apparently easier to execute. However, a higher risk of low-back disorders for lowering tasks is related to significant reductions in the eccentric strength capability of the trunk at low angular velocities (Marras and Mirka, 1989). This suggests that more caution may be required for lowering tasks. In conclusion, a higher rate of task execution significantly increased the net muscular moment at the L5/Sl joint for asymmetrical lifting but had negligible effect for asymmetrical lowering. The comparisons between asymmetrical lifting and lowering showed that spine loading was quantitatively comparable under similar dynamic conditions. These results suggest that the rate of lifting and the type of task must be considered in biomechanical analyses of MMH situations. Acknowledgements-Denis

Gagnon was supported by Ph.D. research grants from the Institut de recherche en Sante et securite du travail (IRSST) du Quebec and from the Faculte des etudes superieures de I’Universite de Montreal. The project was also partly supported by grants to Micheline Gagnon from the Natural Sciences and Engineering Research Council of Canada (NSERCC # OGPO020812) and from IRSST. REFERENCES

Andersson, G. B. J. (1981) Epidemiological aspects on lowback pain in industry. Spine 6, 53-60. Andersson, G. B. J., Ortengren. R. and Schultz, A. B. (1980) Analysis and measurement of the loads on the lumbar spine during work at a table. J. Biomechanics 13, 5 133520. Andersson, G. B. J. and Winters, J. M. (1990) Role of muscles in postural tasks: spinal loading and postural stability. In Multiple Muscle Systems: Biomechanics and Movement Organization (Edited by Winters, J. M. and Woo, S. L.-Y.), pp. 377-395. Springer-Veriag. New York. Antonsson, E. K. (1982) A three-dimensional kinematic acquisition and intersegmental dynamic analysis system for human motion. Unpublished doctoral dissertation, Massachusetts Institute of Technology. Beer, F. P. and Johnston, E. R. (1988) Vector Mechanicsjiw Engineers: Dynamics. McGraw-Hill, New York. Bush-Joseph, C., Schipplein, 0.. Andersson. G. B. J. and Andriacchi, T. P. (1988) Influence of dynamic factors on the lumbar spine moment in lifting. Ergonomics 31. 211-216. Cappozzo, A. (1983) The forces and couples in the human trunk during level walking. J. Biomechanics 16, 265-277. Chaffin, D. B. (1988) Biomechanical modelling of the low back during load lifting. Ergonomics 31, 685-697. Chaffin, D. B. and Andersson, G. B. J. (1984) Occupational Biomechanics. Wiley, New York. Crowninshield, R. D. and Brand, R. A. (1981) The prediction of forcesin joint structures: distribution of intersegmental

900

D.

GAGNON

and M.

resultants. In Exercise and Sport Sciences Reviews, Vol. 9 (Edited by Miller, D. I.), pp. 159-181. The Franklin Institute Press, Philadelphia. Daniel, W. W. (1974) Biostatistics: A Foundationfor Analysis in the Health Sciences. Wilev. New York. Davies, G. J. and Gould J. A.-(1982) Trunk testing using a prototype Cybex II isokinetic dynamometer stabilization system. J. Orthop. Sports Phys. Therapy 3, 164-170. Drury, C. G., Deeb, J. M., Hatrman, B., Wooley, S., Drury, C. E. and Gallagher, S. (1989) Symmetric and asymmetric manual materials handling-Part 1: physiology and psychophysics. Ergonomics 32,467-489. Drury, C. G., Law, C. H. and Pawenski, C. S. (1982) A survey of industrial box handling. Hum. Factors 24, 553-565. Gagnon, D. (1990) Analyse cintmatique et cin6tique tridimensionnelle de tlches asymttriques de manutention. Unpublished doctoral dissertation, Universitt de Mont&al. Gagnon, M. and Smyth, G. (1991) Muscular mechanical energy expenditure as a process for detecting potential risks in manual materials handling. J. Biomechanics 24, 191-203.

Garg, A. and Banaag, J. (1988) Maximum acceptable weights, heart rate and RPE’s for one hour’s repetitive asymmetric lifting. Ergonomics 31, 77-96. Keppel, G. (1982) Design and Analysis: A Researcher’s Handbook (2nd Edn). Prentice-Hall, Englewood Cliffs, NJ. Kishino, N. D., Mayer, T. G., Gatchel, R. J., McRate Parrish, M., Anderson, C., Gustin, L. and Mooney, V. (1985) Quantification of lumbar function. Part 4: isometric and isokinetic lifting simulation in normal subjects and lowback dysfunction patients. Spine 10, 921-927. Krag, M. H., Byrne, K. B., Gilbertson, L. G. and Haugh, L. D. (1986) Failure of intra-abdominal pressurization to reduce erector spinae loads during lifting tasks. In Proceedings of the North American Congress on Biomechanics, Vol. I (Edited by Allard, P. and Gagnon, M.), pp. 87-88. Published by the Organizing Committee, Montrkal. Kromodihardjo, S. (1986) A 3-D kinetic biomechanical model of symmetrical and asymmetrical lifting tasks. Unpublished doctoral dissertation, University of Cincinnati (University Microfilms International No. 8712716). Kromodihardjo, S. and Mital, A. (1986) Kinetic analysis of manual lifting activities: Part I-Development of a threedimensional computer model. Int. J. Ind. Ergonomics 1, 77-90. Kromodihardjo, S. and Mital, A. (1987) Biomechanical analysis of manual lifting tasks. J. biomech. Engng 109, i32-138. Kumar, S. (1980) Physiological responses to weight lifting in different planes. Eraonomics 23, 987-993. Kumar, S. il984) The physiological cost of three different methods oflifting in sagittal and lateral planes. Ergonomics 27,425-433.

Kumar, S., Chaffin, D. B. and Redfern, M. (1988) Isometric and isokinetic back and arm lifting strengths: device and measurement. J. Biomechanics 21, 35-44. Lamonde, F. (1987) Analyse des accidents de manutention dans le secteur transpqrt ginBra et entreposage. Unpublished master’s thesis. Ecole Polvtechniaue de MontrCal. Lavender, S. A., Mirka; G. A., Schbenmarklin, R. W., Sommerich, C. M., Sudhakar, L. R. and Marras, W. S. (1989) The effects of preview and task symmetry on trunk muscle response to sudden loading. Hum. Factors 31, 101-115. Leskinen, T. P. J., Stalhammar, H. R. and Kuorinka, I. A. A. (1983) A dynamic analysis of spinal compression with different lifting techniques. Ergonomics 26, 595-604. Marras, W. S. and Mirka, G. A. (1989) Trunk strength during asymmetric trunk motion. Hum. Factors 31, 667-677. Marras, W. S., Rangarajulu, S. L. and Lavender, S. A. (1987) Trunk loading and expectation. Ergonomics 30, 551-562. Marzan, G. T. (1975) Optimum configuration ofdata acquisition in close-range photogrammetry. In Symposium on Close-Range Photogrammetric Systems, pp. 558-573.

GAGNON

American Society of Photogrammetry, Falls Church, Illinois. McGill, S. M. and Norman, R. W. (1985) Dynamically and statically determined low back moments during lifting. J. Biomechanics 18, 877-885. Mital, A. and Fard, H. F. (1986) Psychophysical and physiological responses to lifting symmetrical and asymmetrical loads symmetrically and asymmetrically. Ergonomics 29, 1263-1272.

Mital, A. and Kromodihardjo, S. (1986) Kinetic analysis of manual lifting activities: Part II-biomechanical analysis of task variables. Int. J. Ind. Ergonomics 1, 91-101. National Institute for Occupational Safety and Health (NIOSH) (1981) Work practices guide for manual lifting. Report no. 81-122, US. Dept. of Health and Human Services, Cincinnati, OH. Paul, R. P. (1981) Robot Manipulators: Mathematics, Programming, and Control. The MIT Press, Cambridge, Massachusetts. Press, W. H., Flannery, B. P., Teukolsky, S. A. and Vetterling, W. T. (1988) Numerical Recipes in C. Cambridge University Press, New York. Ramey, M. R. and Yang, A. T. (1981) A simulation procedure for human motion studies. J. Biomechanics 14, 203-213. Sandover, J. (1983) Dynamic loading as a possible source of low-back disorders. Soine 8. 652-658. Smeathers, J. E. and Joanes,‘D. N. (1988) Dynamic compressive properties of human lumbar intervertebral joints: a comparison between fresh and thawed specimens. J. Biomechanics 21, 425-433.

Smith, S. S., Mayer, T. G., Gatchel, R. J. and Becker, T. J. (1985) Quantification of lumbar function. Part 1: isometric and multispeed isokinetic trunk strength measures in sagittal and axial planes in normal subjects. Spine 10, 157-764.

Snook, S. M. (1978) The design of manual handling tasks. Ergonomics 21,963-985.

Snook, S. H., Campanelli, R. A. and Hart, J. W. (1978) A study of three preventive approaches to low back injury. J. Occupational Med. 20,478-481.

Vrahas, M. S., Brand, R. A., Brown, T. D. and Andrews, J. G. (1990) Contribution of passive tissues to the intersegmental moments at the hip. J. Biomechanics 23, 357-362. Walton, J. S. (1981) Close-range tine-photogrammetry: a technique for quantifying gross human motion. Unpublished doctoral dissertation, Pennsylvania State University (University Microfilms International No. 8120471). _ Winter. D. A. (1979) Biomechanics of Human Mooement. Wiley, New k;ork. Winter, D. A. (1990) Biomechanics and Motor Control of Human Movement. Wiley-Interscience, New York. Wood, G. A. (1982) Data smoothing and differentiation procedures in biomechanics. In Exercise and Sport Sciences Reviews, Vol. 10 (Edited by Terjung, R. L.), pp. 308-362. The Franklin Institute Press, Philadelphia. Zatsiorsky, V. and Seluyanov, V. (1981) The mass and inertia characteristics of the main segments of the human body. In Biomechanics VIII-B (Edited by Matsui, H. and Kobayashi, K.), pp. 1152-l 159. Human Kinetics, Champaign. Zatsiorsky, V. and Seluyanov, V. (1983) Estimation of the mass and inertia characteristics of the human body by means of the best predictive regression equations. I&I Biomechanics IX-B (Edited bv Winter. D. A.), __ pu. 233-239. Human Kinetics, Champaign. Zetterberg, C., Andersson, G. B. J. and Schultz, A. B. (1987) The activity of individual trunk muscles during heavy physical loading. Spine 12, 1035-1040. APPENDIX

Axes transformation

The angular velocity and acceleration vectors were transformed from the fixed coordinate system (0, X, Y, Z} to the

Dynamic factors in asymmetrical lifting

901

Rxl. R,L RZL 1

anatomical coordinate system of the segment {N, L, T, S} using the following rotation matrix (see Fig. 2):

CT1O,N= Rx, RYT &T

1

641)

[ Rxs RYS &s

CTIO.Nis the rotation matrix from the fixed coordinate system 0 to the anatomical coordinate system N; the variables R,,, are the direction cosines of each axis of the fixed coordinate system (0, X, Y, Z} about each axis of the anatomical coordinate system {N, L, T, S}. The following equations were then used to compute the angular velocity and acceleration vectors about the anatomical coordinate system: ~,=CTl,,.x~m

642)

a~= CTIO.N x QO,

(A3)

where wn and w. are the column vectors of the angular velocity about the anatomical coordinate system N and fixed coordinate system 0, respectively, and an and a, are the column vectors of the angular acceleration about the anatomical coordinate system N and fixed coordinate system 0, respectively. Dynamic equilibrium

equations

The spatial motion of each segment was defined first by a general equation governing the center of mass translation of a segment: xF=mi.

(A4)

The sum ofall force vectors c F acting on a segment equals the product of the mass m of this segment by its linear acceleration vector ii. The expansion of equation (A4), according to the symbols used in Fig. Al, yields the following three scalar equations to solve the spatial translational equilibrium of the Nth segment: F,, + F,, = m&

(As)

F,,+F,,=mr.r,

(A@

F,,+F,,-mg=m~z.

(A7)

Then, the spatial motion of each segment was defined by a second general equation governing the segment rotation about its center of mass and known in mechanics as the Euler’s equation of motion (Beer and Johnston, 1988). For a coordinate system corresponding to the fixed coordinate system and having as origin the center of mass (C) of the segment, the spatial rotational equilibrium was defined by the following equation: ~M,=H,.

(A@

The sum of all the moment-of-force vectors CM, about the center of mass of the segment equals the rate of change of the angular momentum vector Hc of the segment. The expansion of this equation (A8), according to the symbols used on Fig. Al, yields the following three scalar equations to solve the spatial rotational equilibrium of segment N:

(A9)

(AlO)

Fig. Al. Free body and mass-acceleration diagrams for one segment of the model with the sign convention of the fixed coordinate system (0, X, Y, Z}. The net reaction force (F), the moment (M), the osition (r), the rate of change of the angular momentum (Ipc) and the inertial force (mi) vectors acting at the proximal(P), distal(D) or at the center of mass(C) of the segment are included. The weight of the segment (mg) was considered as acting negatively along the vertical (Z) axis.

=I,a,-(I,-l,)w,w,.

(All)

The left-hand side of each equation consists of the sum of the axial components of the proximal (P) and distal (D) net muscular moments on segment N about the fixed coordinate system (0, X, Y,Z} as well as the moments created by each of the net reaction force components about the center of mass (C) of the segment. The right-hand side of each equation includes the moment of inertia components I,, I, and I, about the center of mass of the segment, the angular velocity components ox, wy and wa and the angular acceleration components a,, a,. and aa. The numerical solution of each equation for a complete movement was obtained by solving sequentially for the unknown forces and moments for each axis and each time sample. Once solved, the unknown proximal net reaction forces and moments on segment N (F,, , F,, , F,, and M,, , M,,, M,,) became the respective distal components of the next (N + 1)th segment. The process was the same for every segment of the chain up to the proximal end of the trunk segment. The net force and moment vectors were transformed from the fixed coordinate system (0, X, Y, Z} to the anatomical coordinate system {N, L, T, S} using the appropriate transformation matrix [TIO.n as illustrated for the angular velocity and acceleration vectors [equations (Al) and (A2)]. The temporal and segmental indices were intentionally omitted from each variable of all the previous equations.

S1 joint during asymmetrical lifting and lowering.

Asymmetrical lifting and lowering are predominant activities in the workplace. Mechanical causes are suggested for many back injuries and the dynamic ...
2MB Sizes 0 Downloads 0 Views