Biomechanical consequences of lumbar belt design for the therapeutic treatment of low back pain

1

Keywords:

• Finite element analysis
• Laplace law
• low back pain
• lumbar belt
• medical device

1. Introduction

Low back pain induces high cost to the society in developed countries. It concerns more than 50% of the French population each year (Rossignol et al. 2009). One of the possible treatments for low back pain is the use of lumbar belts; the objective is to reduce mobility and to decrease pain (Calmels et al. 2009). Even if many clinical studies show the efficiency of lumbar belt treatment, many questions are still opened, in particular in relation with the belt design. Recently, Bonnaire proposed a numerical 3 D model of the human trunk to study the biomechanical effects of lumbar belts (Bonnaire 2015). This trunk is a simplification of a subject’s MRI so as to make it possible to carry out a parametric study by varying morphological parameters of the patient, characteristic parameters of the lumbar belt and mechanical parameters of body. The choice of belt appeared to be very important according to the patient's morphology, but in this first study, only a few different design elements were present. This study is carried out with different lumbar belts on four typical morphologies of patients with hyper-lordosis so as to identify the key parameters of lumbar belt for their design purpose.

2. Materials and methods

Four typical patient morphologies are chosen: tall or small, fat or slim. The same material properties of the body structures (soft tissues, vertebrae and intervertebral discs) and boundary conditions are taken for all models so as to focus only on the morphological effects of lumbar belts. In particular, all virtual patients suffer hyperlordosis with a lordosis angle set to 53°.

Eight different belts, from 4 manufacturers present in the French market, were analysed. Lumbar belts can be splited in two groups depending on the belt construction (Figure 1). In the 1^st group, the belts consist in a wide width fabric cut and sewn, while in the 2^nd group they consist in narrow fabrics assembled together along the length. They all have the same height (26 cm). Material properties of lumbar belts are obtained by mechanical testing.

Figure 1. Fabric-based belt (upper) and assembled narrow fabrics belts (lower).

The FE model is the same as that of (Bonnaire 2015) with two major changes on the belt description and on the model outputs: belts are described in the FE model as pressure boundary conditions; the pressure is calculated by Laplace’s law, taking into account the belt shape, the body shape and the belt stiffness. All belts were closed with a 20% mean circumferential strain, according to manufacturer’s recommendation.

Comparing various belts implies the definition of their efficiency. This can be seen in many ways: the mechanical effect on the lumbar disk (global decompression, or decompression on the back) or the lordosis angle variation could characterize the active effect on the patient’s pain; the mean and maximum pressure applied on the trunk should give an indication of the comfort and compliance to the treatment – even if pain, comfort or compliance is an undoubtedly complex phenomenon difficult to quantify.

Pearson correlation test (PCC) is chosen to qualify the correlation level between possible efficiency parameters.

3. Results and discussion

The FE results after processing and presented in Figure 2. Depending on the patient and the belt, belts generate a mean pressure varying from 12.7 to 48.8 mmHg, and a maximum (last decile) from 26.1 to 316.6 mmHg. The variation of lordosis angle ranges from −0.5% to −8.6%.

Figure 2. Synthetic representation of a lumbar belt effect on intervertebral disks.

The bending moment has its highest delordosing effect between T12/L1 or L1/L2. This must be considered as a drawback of all the belts in the test, low back pain being associated usually to lower disk (L3/L4, L4/L5). Considering the overall mechanical configuration, it seems reasonable to imagine that the spine is bent on two supports with a force distributed between them. In this case, the shear moment, and therefore the angle of rotation between two sections, has its maximum between the supports. A change in belt height and/or position could possibly solve this problem.

Mechanical parameters inside the disk are strongly correlated with the lordosis angle, and only this parameter will be considered hereafter. The mean pressure inside the discs is less correlated than other parameters (PCC = 0.53). This could indicate that global decompression can occur without major change in lordosis angle. Correlation between lordosis angle and the mean pressure applied on the body is highly correlated (PCC = -0.81), establishing clearly the link between external action to its internal effects.

The belt design is characterised here by its global stiffness and by the choice of a narrow or wide width fabric. Even if the belt stiffness is highly correlated to the mean pressure applied to the trunk, its effect on the lordosis angle is low (PCC = -0.1).

The peak pressure ratio, that characterizes the intensity of a local pressure around the trunk, is significantly lower if the belt is made with narrow fabrics for every body shape, except the small-fat morphology, which has its maximum pressure on the abdomen (Table 1). The peak pressure ratio has no significant effect on the lordosis angle.

Table 1. Mean and max (last decile) pressure applied on the body (mmHg).

		Tall-Fat		Tall-Slim		Small-Fat		Small-Slim
		Mean	Max	Mean	Max	Mean	Max	Mean	Max
Narrow fabric	Belt A	29,5	53,4	35,3	142,1	27,0	178,4	35	116,5
	Belt B	13,3	26,1	16,1	61,6	12,7	78,7	16,1	53,9
	Belt C	18,3	30,8	21,6	78,2	17,2	98,7	21,6	75,5
	Belt D	27,3	46	32,3	116,6	25,8	147,5	32,2	110,6
Fabric-based	Belt E	23,8	52,5	29,0	114,8	22,8	93,6	28,5	117,4
	Belt F	28,0	56,7	34,4	192	26,1	98,3	36,7	316,7
	Belt G	38,9	82,4	48,8	231	33,6	123,6	47,4	192,2
	Belt F	31,4	78,2	38,5	167	28,5	109,1	36,8	127,9

These results show that it is possible to increase the belt efficacy without decreasing its comfort, and ensure a better treatment, but they are limited with such study. In particular, it is difficult to define a threshold level for angular or disk pressure variation. More generally, the model is simple on purpose: linear elastic, without thermal effects, idealised geometries, and without any friction between the body and the belt.

4. Conclusions

This study illustrates in a mechanical way the effect of design choice on the belt action. The best solution seems to be a narrow fabric-based belt. The numerical model can serve as a basis for more in-depth studies on the analysis of lumbar belts efficiency in low back pain. However, FE simulations on real patient’s trunks are necessary to propose patient specific solutions. The model will be soon improved by using more realistic body shapes and by modelling the contact conditions between the belt and the body.

Acknowledgements

This study is carried out with Thuasne Co.