Contents - Previous - Next

The mechanical factors which influence bone growth

1. Introduction
2. Historical perspective
3. Biomechanics
4. Clinical examples
5. Lengthening
6. Discussion

J.S.R. Golding

Department of Surgery, University of the West Indies, Kingston, Jamaica

1. Introduction

My task is to examine the effect that mechanical influences have on bone growth and form, with particular reference to the long bones of the legs since, I am told, in the so-called nutritional stunting that is the subject of this workshop, it is here that the deficit in length occurs. As an orthopaedic surgeon I shall describe various clinical conditions associated with both shortening and lengthening of the long bones of the legs, in the hope that they might perhaps provide clues to the processes that are operating in nutritional retardation of linear growth.

My own interest in the subject started in undergraduate days when we had the opportunity of visiting the Strangeways Research Institute, where Murray (1985) had been working on tissue culture while writing his definitive book on the subject of 'Bones'. He showed the importance of the intrinsic form of each bone and how it later became modified by the stresses and strains to which it was subjected externally as well as by the demands of the endosteal elements within.

2. Historical perspective

The idea that mechanical factors can influence growth seems to have been accepted from the earliest times which accounts for the ancient Chinese custom of binding the feet of aristocratic girls. This produced a marked reduction in both longitudinal and lateral growth as an aid to beauty and denoting social status.

All civilizations have been fascinated by abnormalities of growth, height and size. Abnormally short and deformed people have usually been treated in a special way by their contemporaries. Often they were kept as pets, jesters or acrobats. In Ancient Greece and Egypt they were used as dancers. During the Middle Ages they were often employed as confidential advisors or keepers of the crown jewels because it was easy to find them if they tried to abscond (Bailey, 1973). During the Renaissance it became so fashionable to have a midget in the homes of the wealthy that poor families would deliberately maim their children so that they might be sold. Catherine de Medici and Louis XIV both tried to inbreed a family of midgets but failed because so many of their offspring were born of normal size.

However, although the study of midgets is fascinating and important because of its association with nutritional and metabolic abnormality, dwarfism has received more attention clinically because the general and symmetrical smallness of the midget can only rarely be treated, whilst the disproportionate development of the dwarf frequently can be improved or prevented.

Our basic knowledge of the effects that mechanical factors have on bone growth are largely the result of the work of Julius Wolff who was born in West Prussia in 1836 (see Keith, 1919). He studied medicine in Berlin. After taking his degree in 1860, he came under the influence of Langenbeck, who advised him to study the regeneration of bone. His thesis was entitled "The law of bone transformation". It was largely based on the work of Henri-Louis Duhamel du Monceau (1700-1782), John Hunter and Jean Marie Pierre Flourens.

Duhamel had studied the uptake of dye 'madder' by bone and found it was only deposited where osteoblast activity was present. Duhamel, a bachelor, was a natural research worker delighting in his self-described role as 'Nature's detective'. He found that only parts of the bone became stained. The younger the animal the more bone would be stained because the madder was only deposited in newly formed active bone. By alternating a madder treated with a normal diet, he could produce successive layers of dyed bone, proving that bone grew by interstitial formation. By drilling holes in bone a measured distance apart, he proved that growth took place from the ends of long bones.

John Hunter was a boy while Duhamel's numerous papers on bone growth were being published. They had been met with great opposition which made them all the more interesting to Hunter who was able to show the wonderful way that a young bone could remodel. This implied that bone absorption was always present and an essential part of the remodelling process. Wolff found that periosteum was only weakly osteogenic. He confirmed the findings of John Hunter that the shaft of the bone actually expanded, so that the bone cells grew apart as interstitial bone formed. Wolff believed he was the first to understand that bone was a plastic, ever changing tissue, but, in fact, Flourens had come to the same conclusion. Hunter had studied the shape of the neck of the femur and Wolff continued his work. At the time, the architecture of this area of the femur was being studied independently by an anatomist, von Meyer, and Culmann, who was an engineer and mathematician. Culmann studied von Meyer's drawings and concluded that the architecture of the trabeculae was a perfect mathematical design for the transfer of weight 1. Thus the osteoblasts are the architects and do their work according to exact mechanical laws. Wolff then showed that in a deformed bone the internal structure was radically altered as a response to the static forces working on it. A normal bone will alter to meet a change in its function. If such change in mechanical environment is rectified, the bone will resume its former shape and structure. Wolff finally presented his Law as "Every change in the form and the function of a bone or of its function alone is followed by certain definitive changes in its internal architecture, and equally definite secondary alteration in its external conformation". This complicated formulation was reduced by John B. Murphy of Chicago to: "The amount of growth in a bone depends upon the need for it".

1 A drawing of the lines of stress in the neck of the femur after Culmann & Wolff, is reproduced in Thompson DAW, "On Growth and Form", abridged version, Cambridge University Press, 1969, pp. 233.

However, this is really an inadequate definition and should be expanded to include "the form of a bone being given, the bone elements place or displace themselves in the direction of functional pressure".

3. Biomechanics

Early in embryonic life the mesenchyme, which will later become cartilage and bone, begins to condense into recognizable shape. In the axial and long bones, mesenchyme first develops into cartilage, while in the cranium and clavicles it changes directly into bone. Mechanical factors will play a part even in bones that do not bear weight. Thus, the assumption of the right posture will be associated with a change of cranial length as the centre of gravity of the cranium is required to rest over the feet.

The hydroxy-apatite crystals are the pressure absorbing component of bone. The collagen fibres give the bone its tensile strength and elasticity. Compression and elongation give rise to electrical charges at the boundaries of the crystals and seem to produce appositional growth and reabsorption by triggering off the chemical processes leading to the formation of the metabolites which control these processes. Bone thus responds in its structure to the different forces such as compression, tension and torsion. The torsional strength of bone is about a third of its compressive strength, so it is not surprising to find that the tibia and humerus have a spiral arrangement of collagen fibres to protect them from the torsional forces to which they are particularly subjected. The bones of infants have a much lower modulus of elasticity than those of children, as though walking only becomes possible once the bone has stiffened.

The actual strength of a bone is much greater than might be expected. Thus, weight for weight, it is stronger than fibreglass. It seems that the individual apatite crystals rather than the long glass rods enable the cracks, which always develop before a material fails under stress, to remain isolated rather than spreading rapidly as they do in fibreglass. Thus any lack or excess of muscle pull or body weight will have a significant effect on bone growth. Most of the observed alterations in bone result from the effects of abnormal stress. Bones, like people, need stress for strength-but that stress has to be not only within reasonable limits but also in the proper direction if growth is to proceed in a normal and orderly way.

4. Clinical examples

In order to illustrate aspects of the factors which affect bone development clinically, I have selected a few conditions which show different ways that bone length may be influenced. Since my personal interest happens to be the development of the lower extremities, I will commence by showing how Wolff's Law works out in practice.

The effects of compression on bone growth vary according to whether the force is applied intermittently as in walking or continuously as in recumbency, paralysis or in any situation, including a plaster cast, maintained for a considerable time during a period of rapid growth. Sustained pressure on the growth plates of the distal femur and proximal tibia may compress the anterior portion and distract the posterior portion of the upper tibial growth plate (Stevenson, 1952). As the muscles waste, the knee falls back into hyper-extension. The blood flow in the vessels to the compressed anterior portion of the upper tibia slows. Eventually the growth plate may permanently close and growth will cease. This produces the well-known Recurvatum deformity which used to be commonly seen where prolonged recumbency was needed, as in the treatment of tuberculosis.

4.1. Genu valgum & obesity versus height

Morley (1957) examined 1100 children for abnormal growth in the lower extremities from the point of view of angular deformity. The only correlation discovered between the incidence of genu valgum and the other parameters he was looking for was with weight and height. It is now well recognized that obese children are above the average in height for their age group (Tupman, 1962). It will be interesting to study if the increase in height which follows dieting an obese child will also stimulate the correction of an angular deformity.

Any normal child may be expected to show a period of angular deformity during growth. About a third will show an appreciable degree of genu valgum, a third, genu varum and the remainder will stay more or less straight during the first five years of growth. Moreover a significant proportion will swing from valgus to varus and back before settling down to the usual straight alignment. Only a fraction of 1% of children finish up with a deformity great enough to require treatment.

The mechanisms for this spontaneous correction is probably a simple piezo-electric effect on the apatite bone crystals which respond to unusual pressure just as a crystal gramophone needle does. The effect is to stimulate growth on the side compressed by body weight until spontaneous correction occurs.

The question of remodelling is important because, although it may not have any effect on longitudinal growth, it certainly has on limb shape. For instance, remodelling is much more evident as far as alignment is concerned in young children. Thus, a birth fracture of the shaft of the femur will straighten itself rapidly without any treatment even when the angulation is extreme.

On the other hand, an angulated or rotated fracture of the lower end of the humerus, a non-weight-bearing bone, will not improve significantly, however long one waits. Remodelling depends on the direction of forces across the healing fracture site. Since rotary and angular deformities cause no great difference to the muscle length and therefore their effective power, and weight loading is not a factor, such deformities persist in the upper extremity and may become a serious cosmetic problem.

4.2. Growth plate damage (Golding & McNeil-Smith, 1963; Murray, 1985)

All leg discrepancies are due either to malunion of a fracture or to some abnormality associated with the growth plate. This may be part of a congenital failure in normal fetal development, which will cause the most severe type of growth abnormality, or to congenital lesions altering the blood supply of bone. It may also be due to trauma damaging the growth plate itself. However by far the most numerous and interesting results of mechanical factors appear to be associated with either excessive compressive or shear forces. This is well illustrated in the infantile form of tibia vara (Golding, 1990; Golding & McNeil-Smith, 1963). This condition is rather uncommon in Europe (apparently except in Finland), but remarkably common in the West Indies and peculiarly rare in West Africa, from whence the majority of our people emanate. The cause appears to be mechanical.

The normal child is born with the head of the femur directly over the shaft because otherwise the greater trochanters, being the widest part of the body, would obviously obstruct labour. In the next year the neck of the femur develops so that the angle between the neck and shaft changes from 180 to about 130. Thus, at birth the upper ends of the femoral shafts are closer together than at any other time in life. The shafts run downwards and slightly outwards because the knees, being covered by the thick layer of fat, so common in the lower extremities of infants, keep the bones apart. The normal adult alignment with the femora sloping downwards and inwards does not develop until 12-15 months.

The average English child is walking independently about that time, so that his leg alignment is either straight or valgus by the time he is really weight bearing. The West Indian child is unencumbered by bulky warm clothes and footwear. He is usually walking independently three months earlier, when the leg alignment is still varus and the ligaments very elastic. At this age the tibiae are commonly still showing the internal tibial torsion which is the normal result of the lower legs being moulded by pressure against the inner surface of the uterine wall. This gives the appearance of the so-called 'physiological' bow leg. When standing on such legs, the body weight is transmitted through the medial half of the upper tibial growth plate and the outer half is bearing little or no weight-it may even be distracted if the varus is severe.

The medial femoral condyle tends to force the upper tibial epiphysis laterally so that the cells which bud off from the epiphysis, on which all longitudinal growth depends, will not only be excessively compressed (Golding, 1990; Harris et al., 1971) but also subject to lateral sheer. The result may be a cessation of growth in the area where the abnormal stress is greatest. As the lateral half of the growth plate continues to grow, the result is a steadily increasing deformity and all the changes characteristic of tibia vara develop. The medial portion of the growth plate will become grossly abnormal and the histology shows not the normal beautifully ordered vertical columns of cells but total disorganization. I personally believe that this has a purpose, because the tongue of disordered cartilage and bone which is readily palpable, acts as a claw and, rather like a climber's hands, holds the epiphysis in place and prevents any further lateral shift.

It is so often the exception to a problem which proves the rule. It was known that the condition was uncommon in West Africa. In fact, this was the reason I first visited that area in 1959. I never saw a single case while working for a year at the Ibadan University Hospital. This was explained when an economics lecturer from Ghana brought his daughter for treatment shortly after he had transferred from Australia to Jamaica. She had classical early tibia vara. He told me that this was his third child. The first two had been born in Ghana, but this child had been born when he was working in Australia. The older children had been carried in the traditional way on their mother's back, while this child had not. The back-carry produces a constant valgus, external torsion pressure on the lower leg and feet against the mother's pelvis. Maybe the frequency of internal tibial torsion in West Indian children is built in to prevent the marked tendency to knock-knees which are so common in West Africa (Spray, personal communication).

Excessive compression, therefore, can be seen to result in thickening of the weight-bearing bone and eventually slowing of growth across the growth plate. Confirmation of the mechanical nature of the condition is the remarkable permanent correction which will usually follow adequately performed surgery designed to restore the normal mechanical forces across the growth plate. This usually results in a resumption of normal growth.

4.3. The necessity of movement

There is another mechanism which can cause the most marked shortening of the extremities in quite a different way. Some years ago, an assistant nurse was admitted to our intensive care unit. She had become pregnant while still under training and attempted to secure an illegal abortion. She became infected with tetanus when the fetus was about three months old. She finished up on a respirator to save her life from the spasms. At that time, such patients were kept fully curarized. Fortunately she made a complete recovery. However, it was not realized that the curare had paralysed the fetus as well as his mother. As soon as it was judged that the child was viable, a Caesarean section was performed and the child found to be grossly deformed. Examination that the muscles of the extremities had not developed properly and were almost impalpable. There were gross contractures of most of the joints of the extremities which proved resistant to passive stretching. The leg lengths were different being shorter on the more affected side.

This appearance was very similar to that seen in infants with arthrogryposis congenita multiplex. This condition usually affects the lower extremities alone, although the arms or all four extremities may be involved. The affected limbs are fixed in the position they had in utero. Frequently the child grows up unable to walk, with severe flexion deformities in the knees. They become disproportionate dwarfs with the lower extremities about half the length they should be, based on their expected height derived from measurement of their arm span.

Normal movement in utero is essential for normal limb development and future growth.

4.4. Pain leading to disuse and immobilization (Geisner & Trueta, 1958; Rang, 1969; Ring, 1961; Staheli, Duncan & Shaeffer, 1968)

As already stated, prolonged recumbency may result in damage to growth plates. Then various deformities may develop such as chest deformity, and even changes in the shape of the skull. The changes which are seen in limbs following severe prolonged pain are believed to result from the normal reflex immobilization of the part that pain brings. There is no definite evidence that pain alters growth in any other way.

Ring (1961) showed experimentally that the sensory denervation of a limb did not alter bone growth, unless the motor nerves were also divided, and then bone growth slowed as function was lost. Clinically a similar lack of longitudinal growth is seen to accompany any condition where normal movement is prevented whatever the cause may be.

4.5. Poliomyelitis (Gill, 1944; Ratlif 1959; Stinchfield, Reidy & Barr, 1931)

Barr (1949) found that 35% of patients with a limb weakened by poliomyelitis develops as much as 1 1/2 inches of shortening. Green (1949) found that 8% developed > 2" of shortening. Ring (1961) surmised that it was the age of onset which determined the outcome-the earlier the onset, the greater would be the amount of shortening. Fig. 1, reproduced from Stinchfield et al. (1949), represents the sort of correlation between degree of paresis and shortening which was found in adults who had contracted the disease in childhood. It was found impossible to predict how much inequality of leg length was to be expected in any individual case. Blount, who was an outstandingly accurate observer, felt that there was only a fair degree of correlation between the severity of the paralysis and the shortening, but the results of Ratlif (1959) suggest a certain relationship, although not a very close one. Ratlif found, amongst 225 untreated cases of anterior poliomyelitis, that 215 showed shortening, and that most of the shortening would be found in the tibia rather than the femur. He also noted that, although a mild degree of paresis produced a small amount of shortening, severe paresis might produce either a little or severe shortening. He observed that a limb with only one or two paralysed muscle groups was as likely to show marked shortening as a severely affected limb. In 7 patients the affected leg was longer. Ratlif did not find that the age of onset was particularly significant, nor did those with pronounced vascular skin changes, which commonly developed in these patients, show more shortening. The decisive factor appeared to be the particular standing and walking pattern which each individual found most comfortable. Where a brace was used, the more affected leg might occasionally bear weight for a longer time and then be actually longer than the less affected side.

Fig. 1. Muscle power and growth in length in the weaker extremity in 64 adults who had acute poliomyelitis before the age of 11 and were left with one normal and one involved leg. (From Stinchfield et al., 1949).

4.6. Folic acid deficiency

Whilst working in Nigeria I became involved in a study on the stunting which so commonly occurred there in patients presenting with symptoms and signs of sickle cell anaemia. In Jamaica the SS patients tend to be thin and spindly; they look tall because of their shape. In fact they are of a little less than average height. However, in Nigeria sickle cell anaemia is only one of the pathological conditions the patient is likely to have. Malaria and parasites are ubiquitous and cause great stress on the blood forming elements within the bone marrow. The clinical appearance of sickle cell anaemia was quite different from that seen in Jamaica. The patients were small and stunted. Their parents disliked their small stature, and we were frequently asked if anything could be done to correct it.

We investigated the cause of this dwarfism. The iliac crest marrow biopsies, which were part of the protocol, suggested that the cause was probably folic acid deficiency. Rapid growth ensued once this deficiency had been corrected.

4.7. Chronic osteomyelitis

If an attack of acute osteomyelitis is inadequately treated, there will be an increased inflammatory reaction which implies an increased blood supply. There will be overgrowth of bone length, which may be excessive, if the child is very young, and may require leg equalization at a later date. Only rarely is a growth plate permanently damaged by acute osteomyelitis, but it does occur, particularly in the rather acute haematogenous osteomyelitis of infancy, due not to the ubiquitous staphylococcus but to either haemophilus influenzae or the pneumococcus. Gross shortening or a severe angular deformity may result.

4.8. Long bone fracture in children (Meals, 1979; Rang, 1969)

It is commonly found that a growing child will show appreciable overgrowth after a lower limb fracture, particularly when the lower end of the femur is involved (Geisner & Trueta, 1958).

This is so constant a finding that the clinician allows the bone ends in such patients to unite with about 1-2 cm of overlap, knowing that the leg lengths will then equalize in the following year. This overgrowth is consequent on the inflammatory increase in vascularity which is part of fracture healing and affects the contiguous growth plate. It seems that growth stimulation is proportionate to the length of the affected bone. A fracture in very early childhood results in less overgrowth than might be expected.

5. Lengthening

Of more importance to the practical study of the effects on the mechanical behaviour of bone is the apparent growth stimulation which can occur in normal adolescents, particularly females developing idiopathic scoliosis. The average height of these patients has been found to be significantly greater than that of a similar matched population (Ashcroft & Lovell, 1966). This association was first observed in Sweden (Willner, 1975).

Amongst the many parameters we have studied in the course of a general investigation centered around the aetiology of idiopathic scoliosis, the height, arm span, upper and lower body segments were routinely measured. The results were compared with those obtained on a similar group of normal Jamaican school children of the same age. There is no doubt that excessive growth occurs. There seems little else to which to attribute this other than a changing diet.

The scoliosis study began in 1956 and includes cases presenting earlier. We seem to have had an epidemic of the condition, which started around 1975, reached its peak between 1978 and 1984, and has begun to decline. This decline is particularly impressive when one takes into account the fact that the number of adolescent children at risk has been steadily increasing at a rate of about 1.7% p.a. since the study started. Although the deformity, in the idiopathic variety, does not usually manifest itself until puberty, the background of stimulation must start earlier in childhood. We have tried to investigate this and came up against great difficulties, for the obvious reason that the suppliers of animal feeds, which are known to contain anabolic agents, are very loath to give the details of the amount and type of additive used. The fear of litigation prevents them from cooperating and has prevented us obtaining conclusive evidence. Large amounts of oestrogens were added to the feed of broiler hens in the 1960s and early 1970s before it was prohibited. Anabolic agents are still added to animal feeds, although it seems that the quantity has been decreased and the actual types used have been altered.

If one allows ten years between the onset of our 'epidemic' and the time that these agents apparently began to be added to the animal feed in Jamaica, there seems to be a degree of correlation, if only a loose one (Golding, 1991). There may be some confirmation for the hypothesis that the idiopathic form of scoliosis has an association with diet in three unrelated facts:

Firstly, we found that estimates of social class and income groups in these patients suggested strongly that it was the better off town patients who were particularly liable to develop the condition. The country patients who eat natural, locally produced food were less likely to develop idiopathic scoliosis.

Secondly, measurements of leg length have shown that a leg length discrepancy of more than a quarter inch occurs in almost 25% of patients with idiopathic scoliosis compared with 4% of normal adolescent children.

Thirdly, the rare form of idiopathic scoliosis which develops in infancy rather than at puberty is commoner in Holland than in most of the rest of Europe, and is peculiarly rare in North America, suggesting that an environmental rather than a genetic factor is at work.

6. Discussion

The shape, strength, growth and form of a bone is determined not only by heredity but by the work it has to perform. Although this will alter throughout life, the main effects will be seen during the period of growth, when remodelling is most evident. It is in this period that the actual length of a limb bone is being determined.

The actual amount of growth in a bone depends upon the need for it.

The final shape of an individual weight-bearing long bone will be modified by the general conditions existing in the body. If they are abnormal, the normal process of growth by absorption and remodelling may be expected to respond to these alterations.

The mechanical factors which most influence longitudinal growth and which are essential for normal growth are movement and weight bearing. Growth stimulation appears occasionally to be a possible cause of deformity. Any condition which is associated with an increased blood supply, the presence of abnormal nerve tissue or abnormal compression or alignment will alter the final form of a long bone.

It is probable that remodelling depends on the direction of forces in an actively growing bone. Since rotary and angular deformities cause no great difference to the muscle power, these deformities are likely to persist in the upper extremity where there is no weight-bearing to supplement the effects of muscle activity.

The general growth of the skeleton may be modified by prolonged recumbency and paralysis such as spinal injury in childhood. It will also alter as a response to metabolic factors, endocrine abnormality and hereditary factors. Only rarely will a chromosomal abnormality be found to be important.


Ashcroft MT & Lovell HG (1966): Heights and weights of Jamaican primary schoolchildren. J. Trop. Pediatr. 2, 37-43.

Bailey JA (1973): Disproportionate short stature. Philadelphia, PA: W.B. Saunders.

Barr JS (1949): The management of poliomyelitis: the late stage. 1st International Poliomyelitis Conference, p. 201. Philadelphia, PA: J. B. Lippincott.

Geisner M & Trueta J (1958): Muscle action, bone rarefaction and bone formation. J. Bone Joint Surg. 40B, 282-311.

Gill GG (1944): The cause of discrepancy in length of the limbs following tuberculosis of the hip in children. J. Bone Joint Surg. 26, 272-281.

Golding JSR (1990): A new look at tibia vara. In The epiphyseal growth plate, eds H Uthoff & R Whylie, pp. 295-298. New York: Raven Press.

Golding JSR (1991): Observations on idiopathic scoliosis. Aetiology and natural history in Jamaica. Cajanus 24, 31-38.

Golding JSR & McNeil-Smith JDG (1963): Observations on the aetiology of tibia vara. J Bone Joint Surg. 45B, 320-325.

Harris WH, Morrisy RT, Weinberg EMA & Mack PB (1971): The effect of increased and decreased stress on skeletal renewal. J. Bone Joint Surg. 53A, 797-798 (abstract).

Keith Sir Arthur (1919): Members of the maimed: the anatomical and physiological principles underlying the treatment of injuries to muscles, nerves, bones and joints. London: Hodder & Stoughton.

Meals RA (1979): Overgrowth of the femur following fractures in children. Influence of handedness. J. Bone Joint Surg. 61A, 381-384.

Morley AJM (1957): Knock-knee in children. Br. Med. J. 2, 976-979.

Murray PDF (1985): Bones: a study of the development and structure of the vertebrate skeleton. Cambridge Science Classics.

Rang M (1969): The growth plate and its disorders. Edinburgh/London: E & S Livingstone.

Ratlif AHC (1959): The short leg in poliomyelitis. J. Bone Joint Surg. 41B, 56-69.

Ring PA (1961): The influence of the nervous system on the growth of bones. J. Bone Joint Surg. 43B, 121-140.

Staheli LT, Duncan WR & Shaeffer E (1968): Growth alteration in the hemiplegic child. A study of 50 hemiplegic children. J. Bone Joint Surg. 50A, 1271 (abstract).

Stevenson HF (1952): The osteoporosis of immobilization in recumbency. J. Bone Joint Surg. 34B, 256-265.

Stinchfield AJ, Reidy JA & Barr JS (1931): Prediction of unequal growth of the lower extremities in anterior poliomyelitis. J. Bone Joint Surg. 31A, 42-67.

Tupman GS (1962): A study of bone growth in normal children and its relationship to skeletal maturation. J. Bone Joint Surg. 44B, 42-67.

Willner S (1975): A study of height, weight and menarche in girls with idiopathic structural scoliosis. Acta Orthop. Scand. 46, 71-83.

Contents - Previous - Next