- Email: [email protected]
A metallurgical examination of fractured stainless-steel ASIF tibial plates
Injury, 8, 13-l
13
9
A metallurgical stainless-steel Marc
H. Richman,
Joel
examination of fractured ASI F tibia1 plates K. Weltman
and Alan
Cole
Divisions of Engineering and Biomedical Sciences, Brown University, The Miriam Hospital, Providence, Rhode Island
Jan Thunold* Department
of Surgery, University of Bergen, Haukeland
Summary
Between 1970 and 1973 99 tibia1 fractures were treated by rigid internal fixation with ASIF plates. The fractures were all regarded as sufficiently stable for exercise without weight bearing, thus needing no additional external-support during the healing period. Four of the plates.broke late in the healing period, after the onset of weight bearing. These fractures had some degree of delayed union withslight resorption of the bone ends, resulting in cyclical bending of the plate. Examination of 2 of the fractured plates by scanning electron microscopy, electron microprobe analysis and optical metallography revealed that the primary cause of plate fracture was fatigue. There was no evidence that corrosion fatigue or inclusion content were factors leading to plate fracture. INTERNAL fixation with ASIF stainless-steel plates using the compression technique is widely used for treatment of tibia1 fractures (Miiller et al., 1970). The compression technique facilitates an early return of function to the injured limb with rapid restoration of weight bearing capacity. As with any method of fracture treatment, this method also has its complications, one of which is mechanical failure due to fracture of the implanted plate. Breakage of stainless-steel or Vitallium plates has previously been ascribed to corrosion fatigue or to structural defects in the metal (Cohen, 1966; Greene and Jones, 1966; Cahoon and Paxton, 1968; Colangelo and Greene, 1969; Brettle and Hughes, 1970; Cohen and Wulff, 1972; Rose et al., 1972). In contrast *Present appointment: Department Hospitalet Betanien, Bergen.
of
Surgery,
Hospital, Bergen
to these findings, we present data here which show that mechanical failure of 316L stainlesssteel tibia1 plates is due to pure fatigue when associated with non-union or delayed union of the fracture, rather than poor design or poor material. MATERIALS
AND
METHODS
Between 1970 and 1973 a total of 99 tibia1 fractures were operated on with rigid internal fixation according to the ASIF method (Miiller et al., 1970), using tibia1 plates in 82 instances. Four of these plates fractured in situ late in the healing period in patients with signs of delayed union, and a second operation was necessary to achieve union in all 4 patients. The cause of plate fracture was assumed to be a bending force on the plate due to bone resorption at the fracture site patient a (Thunold et al., 1975). In another broken screw was detected on removing the implant some 18 months after the operation, after a normal clinical healing course. The first 2 broken plates were rejected before the present study, but the next 2 plates and the broken screw (from Cases 3, 4 and 5) were available for further The removed specimens were examination. examined by scanning electron microscopy and electron microprobe analysis with an ARLEMX electron microbeam probe. Optical metallography was also performed on polished transverse sections of the plates and the screw to determine inclusion content. A control study was undertaken on an unused dhole ASIF tibia1 plate. It was subjected to
14
Injury: the British Journal
3-point cyclical bending in an lnstron machine, using a repeated load/unload mode with periodically increasing stress until the plate partially cracked, and the final fracture was then effected manually. The stress for the initiation of failure was I77 942 psi. Scanning electron microscopy and electron microprobe analysis were performed on the control plate.
of Accident
Surgery
Vol. ~/NO. 1
and the patient was able to move about with crutches and leave the hospital some 6 weeks after admission. He was, however, unable to follow the instructions
about avoiding weight bearing, and less than IO weeks after the injury he returned to hospital with a refracture (Fig. 3e and f).
RESULTS Fig. 2 shows
CASE
the fractured 8-hole plate removed from Case 3. Arrow A shows the flat surface area adjacent to the countersunk hole where crevice corrosion was visible to the naked eye. Arrow B denotes the transverse fracture surface. Fig. 4 shows a scanning electron micrograph (SEM) of the region indicated by arrow A in Fig. 2. The SEM shows a broken and cracked surface with the formation of large grains of corrosion
REPORTS
Case 3
A 40-year-old road worker was struck by a lorry, sustaining open fractures of his right tibia and fibula (Fig. la and b). Shortly after admission he was operated on, using an S-hole ASIF tibia1 plate. The reduction was not anatomical, with deficient contact between the main fragment and a slight posterolateral
b
a Fig. 1. X-rays
4 months postoperative,
d
e
a and b; immediately postoperative,
f c and d:
e and f.
angulation of the tibia (Fig. Ic and d). After 4 months, union of the tibia was seriously delayed (Fig. 1e and f). Reoperation was necessary, and when the plate was removed we found it broken through the fourth hole at the site of the fracture (Fig. 2). Case 4 A 74-year-old
c
of Case 3 showing delayed union. Preoperative,
pedestrian was struck by an automobile and admitted with a serious head injury and an open, comminuted fracture of his right tibia and fibula (Fig. 3a and b). He was unconscious for 2 weeks, but we operated on his tibia 8 days after admission using a 9-hole ASIF tibia1 plate and 2 compression screws. The reduction was only fair (Fig. 3c and d), the postoperative course uneventful,
product. This large grained surface was not seen everywhere within the region of crevice corrosion. As the interface between the plate and the corrosion product was approached, smaller grain size with more particulate matter was found on the surface. As shown in Fig. 4, there are grain boundaries and cracks in the product. Such cracks may have formed during corrosion because of an unfavourable Pilling-Bedworth ratio (Uhlig, 1948) or by contraction due to dehydration. Electron microprobe analysis of the corroded regions of countersunk screw holes is given in Table I. The elements Fe, Ni, Cr, Mn, MO and Si are characteristic of stainless steel. The
Richman et al.
: Metallurgical
Examination
of Plates
15
a fatigue failure. The fatigue striations in Fig. 5 are not all parallel. It has been shown that fatigue striae need not all run in the same direction over the failure surface. Electron microprobe analysis of the plate surface (Table I) revealed only the presence of elements consistent with the basic constitution of the stainless-steel plate. Even with the broadest diameter beam, there was no trace of Ca, K, Cl, P, Na or 0. Data indistinguishable from those given in Figs. 2, 4 and 5 and Table I were obtained from the tibia1 plate removed from Case 4. Scanning electron micrographs of the fracture surface of the screw fragment removed from Case 5 are shown in Fig. 6, which reveals some striations, conchoidal fracture, void coalescence and, possibly, some inclusions. Fig. 6b shows as
Fig. 2. ASIF tibia1 plate which fractured in situ. Arrow A and arrow B point to area of crevice corrosion and fracture surface respectively.
a
b
c
d
e
Fig. 3. X-ravs of Case 4 showing fracture of ulate in situ. Preoperative, d IlO weeks postoperative, e anh f. elements Ca, K, Cl, P, Na and 0 are components of blood, interstitial fluid and bone, and are to be expected in products of steel corroded in viva. The electron microprobe analysis of the corrosion around the countersunk hole of the plate is indicative of crevice corrosion caused by a differential aeration cell at the screw-countersink interface. Fig. 5 shows a SEM of the fracture surface indicated by arrow B in Fig. 2. The micrograph shows that some of the fracture surface was plastically deformed by wear between the fragments after the plate had broken in the patient. However, there is ample evidence of striations on the original fracture surface to characterize this
f
a and b; immediately postoperative,
Table 1. Electron
microprobe
c and
analysis
Crevice-corroded screw-plate interface of plate
Fe, Ni, Cr, Mn, Cl, P, Na, 0
MO, Si, Ca, K,
Plate fracture surface*
Fe, Ni, Cr, Mn,
MO, Si
Screw fracture surface?
Fe, Ni, Cr, Mn,
MO, Si
*Data shown for plate removed from Case 3. Identical results were obtained with plate from Case 4. iScrew
fractured
in situ in Case 5.
Injury : the British Journal
Fig. 4. SEM of crevice-corroded ( x 2000).
region
of plate
a
of Accident
Surgery
Vol. ~/NO. 1
Fig. 5. SEM of fracture surface of plate (X 1000).
b
Fig. 6. SEM of fracture surface of screw ( x 1000). a and b show different regions of the surface. fatigue striations on the screw fracture surface which are similar to those in the broken surfaces of the plates removed from Cases 3 and 4 (e.g. Fig. 5). Microprobe analysis of the screw fracture surface revealed only the constituents of stainless steel (Table I). Optical metallography was performed on transverse sections from the tibia1 plate broken in situ in Case 3 (Fig. 7a), the broken screw of Case 5 (Fig. 7b), a new unused plate (Fig. 7c) and an unbroken, unused screw (Fig. 7d). The metallography was done to determine the nature, morphology and distribution of the inclusion content of these various specimens since SEM of some of the in vivo fracture surface revealed inclusions and/or possibly inclusion-related microvoid coalescence. The typical inclusions shown in Fig. 7 are similar in amount and distribution in each of the samples studied. Results indistinguishable from those shown in Fig. 7 were
also obtained by optical metallography of an unimplanted plate broken in the Instron machine, a plate routinely removed intact from a patient, and the fractured plate removed from Case 4. The inclusions detected by optical metallography were found by microprobe analysis to be metal carbides, silicides, and A1203, resulting from the ‘killing’ process in steel manufacture. An unused 6-hole tibia1 plate was subjected to 3-point bending in a repeated load/unload mode with periodically increasing stress until failure occurred. The initial failure occurred in the plate at a stress level of 177 942 psi. This value agrees closely with that reported by Ludwigson (1956) for cold-worked 3 16 stainless steel. The plate was only partially cracked in the repeated bending apparatus, however, and the final fracture was effected manually. The fracture surfaces were examined by scanning electron microscopy in those regions of initial failure produced by the
Richman et al. : Metallurgical Examination of Plates
b
d Fig. 7. Optical metallography
of a, broken plate, b, screw, c, an unused plate and d, unused screw (X 100).
repeated bending. The SEM of the fracture surface shows fatigue striations as well as small pits which may be associated with inclusions and/ or microvoid coalescence (Fig. 8). The striations in Fig. 8 show the same non-parallel nature as those in the tibia1 plates which failed in viva (Fig. 5).
DISCUSSION Compression plates are currently in widespread use for the internal fixation of fractures (Miiller et al., 1970). The main materials used for these plates are type 316L stainless steel and Vitallium (Brettle, 1970; Brettle et al., 1971). Both materials possess a high yield and tensile strength, a high Young’s modulus and good fatigue resistance. Vitallium is more resistant to corrosion than stainless steel. Stainless steel, however, possesses certain properties which might be advantageous in bone compression plates. Most notable among
Fig. 8. SEM of 6-hole ASIF tibia1 plate fractured by repeated 3-point bending ( x 500). these properties is the ability of stainless steel to be cold worked. Thus, stainless-steel plates not only may be made by metal working rather than by
18
but they may be bent in the operating theatre to conform to the bone under repair and thereby make possible a more efficient reduction of the fracture. Cold working of ASIF tibia1 plates in the lnstron machine increased the fatigue threshold from 138 000 psi to 180 000 psi in unworked samples. In contrast, neither cast nor wrought Vitallium is easily worked. Vitallium plates are also much more expensive than those made of stainless steel. Because of its great strength and its ability to be bent in the operating theatre to conform to bone, stainless steel would appear to be the material of choice for a compression plate for a weight-bearing bone such as the tibia. However, stainless-steel implants are thought to be susceptible to corrosion and corrosion fatigue. The present study was undertaken to determine whether corrosion or corrosion fatigue played a significant role in the failure of stainless-steel tibia1 compression plates in patients whose tibia1 fractures were treated by the ASIF method of internal fixation at the Haukeland Hospital. Failure of the plate was considered, in this study, to be fracture of the plate in situ. Fracture of a tibia1 compression plate is rare and occurred in only 4 of our 99 fractures operated on between 1970 and 1973. Two of the broken plates were available for the present metallurgical study. All 4 plate failures, however, were associated with delayed union of the fracture (Thunold et al., 1975). All these patients persisted in weight bearing in spite of warnings from the surgeons based on radiological and clinical evidence of non-union. The importance of an adequately united fracture that can allow weight bearing by the tibia in preventing fracture of the tibia1 plates is shown in the following paragraph. Assuming that a plate is rigidly fixed to both fragments of a tibia, and that the fracture has either not been properly reduced or some resorption has taken place, the effect of weight bearing on the plate may be calculated. A person weighing 150 lb (685 kg) when walking or bearing all his weight on one leg will produce a bending moment and a stress of 180 000 psi in the tibia1 plate at a section through each of the screw holes adjacent to the fracture. This stress of 180 000 psi exceeds the ultimate tensile strength of 142 000 psi of the 316L stainless steel of which the plate is made, and failure of the plate will occur through one of the screw holes. If only partial weight is exerted on the leg with the fracture, the applied stress will be less than the tensile strength but greater than the yield strength, and the plate will be subjected to cyclical loading castmg,
Injury: the British Journal ot Accident Surgery Vol. ~/NO. 1
on walking. This will lead to failure of the tibia1 plate by fatigue, as in the plates in this study. The SEM of the interface between the screw and countersink hole (Fig. 4) demonstrates crevice corrosion. Electron microprobe analysis of the corroded area revealed Ca, K, Cl, P, Na and 0 in addition to the elements normally found in AIST 316L stainless steel (Table I). Crevice corrosion occurs normally in 42 per cent of stainless-steel implants and does not usually contribute to fracture failure (Colangelo and Greene, 1969). In none of our cases did crevice corrosion appear to be linked to plate failure. Fig. 5 shows a typical SEM of the fracture surface of a broken plate. Features of the SEMs of both the plates that broke in the patient were essentially the same as those shown in Fig. 5. In both cases fatigue striations were found. Scanning electron microscopy and microprobe analysis revealed no corrosion products on the fracture surface (Table I). Electron microprobe analysis of the fracture surfaces indicated that the surfaces contained the elements characteristic of AISI 316L stainless steel but not the corrosion products of the type found at the screw-plate interface. The fracture surface of the screw which broke in situ without serious inconvenience to the patient is indistinguishable from the surfaces of the broken plates (Fig. 6). The fracture surface appearances made a pure fatigue rather than a corrosion fatigue mechanism a more likely cause of plate and screw failure in the cases studied. The SEMs of a plate fatigue-fractured by cyclical 3-point loading showed fatigue striations, microvoids, pits and small inclusions (Fig. 8). Regions of ductile failure induced by manual separation of the plate fragments are not shown. The plate began to break at 178 000 psi after 39 000 cycles. The presence of inclusions might have caused the failure, but the metallography chiefly revealed fatigue failure when the tensile strength of the stainless steel had been raised by cold working and prior deformation over the normal value of 138 000 psi. While attention has been called to the role of inclusions in fracture of stainless-steel plates (Cahoon and Paxton, 1968), the plates which failed in situ in our series were found to have the same inclusion content as both those removed without failure and new plates (Fig. 7). The inclusion content cannot therefore
be responsible
for the fracture of these plates. Though inclusions may potentially increase the rate of corrosion, the data presented here show that corrosion did not contribute to the failure of the plates, but that in an ASIF screw and in 2 plates which fractured in
Richman et al.
: Metallurgical
Examination
19
of Plates
situ, the cause of failure was pure fatigue. The plate failures were preceded by delayed bone union (Thunold et al., 1975). The primary problem associated with failure of ASIF plates appears to be, therefore, delayed or non-union, perhaps induced by some bio-incompatibility of the steel or by imperfect surgical technique. A broken plate is certainly a complication of importance in a patient operated on for a fracture in the lower limb. It was, however, a direct consequence of delayed healing in this series and was not connected with defects in the implant itself. Plate fracture should not be regarded as an argument against the ASIF technique for reduction and fixation of tibia1 fractures. The clinical course after a second operation, using the same ASIF method, was uneventful and all fractures were united within II months of the primary injury; thus, bioincompatibility of the tibia1 plates was unlikely to have caused the original non-union. Acknowledgement The authors thank Dr John J. Sipe for providing ASIF tibia1 plates for the cyclical bending study. REFERENCES J. (1970) A survey of the literature on metallic surgical implants. Injury 2,26.
Brettle
~cquests for reprints 02906, USA.
should be addressed to: Dr J. K. Weltman,
Brettle J. and Hughes A. N. (1970) A metallurgical examination of surgical implants which have failed in service. Injury 2, 143. Brettle J., Hughes A. N. and Jordan B. A. (1971) Metallurgical aspects of surgical implant materials. Injury 2, 225. Cahoon J. R. and Paxton H. W. (1968) Metallurgical analysis of failed orthopaedic implants. J. Biomed. Mater. Res. 2, I. Cohen J. (1966) The performance and failure in performance of surgical implants in orthopaedic surgery. J. Mater. 1,354. Cohen J. and Wulff J. (1972) Clinical failure caused by corrosion of a vitallium plate. J. Bone Joinr Surg. 54A, 617. Colangelo V. J. and Greene N. D. (1969) Corrosion and fracture of type 316 SMO orthopaedic implants. J. Biomed. Mater. Res. 3, 247. Greene N. D. and Jones D. A. (1966) Corrosion of surgical implants. J. Mater. 1,345. Ludwigson D. C. (1956) Requirements for metallic surgical implants and devices. Metals Engng Q. 5, 1. Mtiller M. E.. Allgower M. and Willenegaer H. (1970) Man&l of internal Fixation (AO-tec&ique). Berlin, Springer-Verlag. Rose R. M., Schiller A. L. and Radin E. L. (1972) Corrosion accelerated mechanical failure of a vitallium nailplate. J. Bone Joint Surg. 54A, 854. Thunold J.. Varhaug J. E. and Bierkeset T. (1975) Tibia1 shaft fractures treated by rigid internal fixation. Injury 7, 125. Uhlig H. H. (ed.) (1948) Corrosion Handbook. New York, Wiley, p. 11.
The Miriam
Hospital,
164 Summit
Avenue,
Providence,
Rhode
Island
13
9
A metallurgical stainless-steel Marc
H. Richman,
Joel
examination of fractured ASI F tibia1 plates K. Weltman
and Alan
Cole
Divisions of Engineering and Biomedical Sciences, Brown University, The Miriam Hospital, Providence, Rhode Island
Jan Thunold* Department
of Surgery, University of Bergen, Haukeland
Summary
Between 1970 and 1973 99 tibia1 fractures were treated by rigid internal fixation with ASIF plates. The fractures were all regarded as sufficiently stable for exercise without weight bearing, thus needing no additional external-support during the healing period. Four of the plates.broke late in the healing period, after the onset of weight bearing. These fractures had some degree of delayed union withslight resorption of the bone ends, resulting in cyclical bending of the plate. Examination of 2 of the fractured plates by scanning electron microscopy, electron microprobe analysis and optical metallography revealed that the primary cause of plate fracture was fatigue. There was no evidence that corrosion fatigue or inclusion content were factors leading to plate fracture. INTERNAL fixation with ASIF stainless-steel plates using the compression technique is widely used for treatment of tibia1 fractures (Miiller et al., 1970). The compression technique facilitates an early return of function to the injured limb with rapid restoration of weight bearing capacity. As with any method of fracture treatment, this method also has its complications, one of which is mechanical failure due to fracture of the implanted plate. Breakage of stainless-steel or Vitallium plates has previously been ascribed to corrosion fatigue or to structural defects in the metal (Cohen, 1966; Greene and Jones, 1966; Cahoon and Paxton, 1968; Colangelo and Greene, 1969; Brettle and Hughes, 1970; Cohen and Wulff, 1972; Rose et al., 1972). In contrast *Present appointment: Department Hospitalet Betanien, Bergen.
of
Surgery,
Hospital, Bergen
to these findings, we present data here which show that mechanical failure of 316L stainlesssteel tibia1 plates is due to pure fatigue when associated with non-union or delayed union of the fracture, rather than poor design or poor material. MATERIALS
AND
METHODS
Between 1970 and 1973 a total of 99 tibia1 fractures were operated on with rigid internal fixation according to the ASIF method (Miiller et al., 1970), using tibia1 plates in 82 instances. Four of these plates fractured in situ late in the healing period in patients with signs of delayed union, and a second operation was necessary to achieve union in all 4 patients. The cause of plate fracture was assumed to be a bending force on the plate due to bone resorption at the fracture site patient a (Thunold et al., 1975). In another broken screw was detected on removing the implant some 18 months after the operation, after a normal clinical healing course. The first 2 broken plates were rejected before the present study, but the next 2 plates and the broken screw (from Cases 3, 4 and 5) were available for further The removed specimens were examination. examined by scanning electron microscopy and electron microprobe analysis with an ARLEMX electron microbeam probe. Optical metallography was also performed on polished transverse sections of the plates and the screw to determine inclusion content. A control study was undertaken on an unused dhole ASIF tibia1 plate. It was subjected to
14
Injury: the British Journal
3-point cyclical bending in an lnstron machine, using a repeated load/unload mode with periodically increasing stress until the plate partially cracked, and the final fracture was then effected manually. The stress for the initiation of failure was I77 942 psi. Scanning electron microscopy and electron microprobe analysis were performed on the control plate.
of Accident
Surgery
Vol. ~/NO. 1
and the patient was able to move about with crutches and leave the hospital some 6 weeks after admission. He was, however, unable to follow the instructions
about avoiding weight bearing, and less than IO weeks after the injury he returned to hospital with a refracture (Fig. 3e and f).
RESULTS Fig. 2 shows
CASE
the fractured 8-hole plate removed from Case 3. Arrow A shows the flat surface area adjacent to the countersunk hole where crevice corrosion was visible to the naked eye. Arrow B denotes the transverse fracture surface. Fig. 4 shows a scanning electron micrograph (SEM) of the region indicated by arrow A in Fig. 2. The SEM shows a broken and cracked surface with the formation of large grains of corrosion
REPORTS
Case 3
A 40-year-old road worker was struck by a lorry, sustaining open fractures of his right tibia and fibula (Fig. la and b). Shortly after admission he was operated on, using an S-hole ASIF tibia1 plate. The reduction was not anatomical, with deficient contact between the main fragment and a slight posterolateral
b
a Fig. 1. X-rays
4 months postoperative,
d
e
a and b; immediately postoperative,
f c and d:
e and f.
angulation of the tibia (Fig. Ic and d). After 4 months, union of the tibia was seriously delayed (Fig. 1e and f). Reoperation was necessary, and when the plate was removed we found it broken through the fourth hole at the site of the fracture (Fig. 2). Case 4 A 74-year-old
c
of Case 3 showing delayed union. Preoperative,
pedestrian was struck by an automobile and admitted with a serious head injury and an open, comminuted fracture of his right tibia and fibula (Fig. 3a and b). He was unconscious for 2 weeks, but we operated on his tibia 8 days after admission using a 9-hole ASIF tibia1 plate and 2 compression screws. The reduction was only fair (Fig. 3c and d), the postoperative course uneventful,
product. This large grained surface was not seen everywhere within the region of crevice corrosion. As the interface between the plate and the corrosion product was approached, smaller grain size with more particulate matter was found on the surface. As shown in Fig. 4, there are grain boundaries and cracks in the product. Such cracks may have formed during corrosion because of an unfavourable Pilling-Bedworth ratio (Uhlig, 1948) or by contraction due to dehydration. Electron microprobe analysis of the corroded regions of countersunk screw holes is given in Table I. The elements Fe, Ni, Cr, Mn, MO and Si are characteristic of stainless steel. The
Richman et al.
: Metallurgical
Examination
of Plates
15
a fatigue failure. The fatigue striations in Fig. 5 are not all parallel. It has been shown that fatigue striae need not all run in the same direction over the failure surface. Electron microprobe analysis of the plate surface (Table I) revealed only the presence of elements consistent with the basic constitution of the stainless-steel plate. Even with the broadest diameter beam, there was no trace of Ca, K, Cl, P, Na or 0. Data indistinguishable from those given in Figs. 2, 4 and 5 and Table I were obtained from the tibia1 plate removed from Case 4. Scanning electron micrographs of the fracture surface of the screw fragment removed from Case 5 are shown in Fig. 6, which reveals some striations, conchoidal fracture, void coalescence and, possibly, some inclusions. Fig. 6b shows as
Fig. 2. ASIF tibia1 plate which fractured in situ. Arrow A and arrow B point to area of crevice corrosion and fracture surface respectively.
a
b
c
d
e
Fig. 3. X-ravs of Case 4 showing fracture of ulate in situ. Preoperative, d IlO weeks postoperative, e anh f. elements Ca, K, Cl, P, Na and 0 are components of blood, interstitial fluid and bone, and are to be expected in products of steel corroded in viva. The electron microprobe analysis of the corrosion around the countersunk hole of the plate is indicative of crevice corrosion caused by a differential aeration cell at the screw-countersink interface. Fig. 5 shows a SEM of the fracture surface indicated by arrow B in Fig. 2. The micrograph shows that some of the fracture surface was plastically deformed by wear between the fragments after the plate had broken in the patient. However, there is ample evidence of striations on the original fracture surface to characterize this
f
a and b; immediately postoperative,
Table 1. Electron
microprobe
c and
analysis
Crevice-corroded screw-plate interface of plate
Fe, Ni, Cr, Mn, Cl, P, Na, 0
MO, Si, Ca, K,
Plate fracture surface*
Fe, Ni, Cr, Mn,
MO, Si
Screw fracture surface?
Fe, Ni, Cr, Mn,
MO, Si
*Data shown for plate removed from Case 3. Identical results were obtained with plate from Case 4. iScrew
fractured
in situ in Case 5.
Injury : the British Journal
Fig. 4. SEM of crevice-corroded ( x 2000).
region
of plate
a
of Accident
Surgery
Vol. ~/NO. 1
Fig. 5. SEM of fracture surface of plate (X 1000).
b
Fig. 6. SEM of fracture surface of screw ( x 1000). a and b show different regions of the surface. fatigue striations on the screw fracture surface which are similar to those in the broken surfaces of the plates removed from Cases 3 and 4 (e.g. Fig. 5). Microprobe analysis of the screw fracture surface revealed only the constituents of stainless steel (Table I). Optical metallography was performed on transverse sections from the tibia1 plate broken in situ in Case 3 (Fig. 7a), the broken screw of Case 5 (Fig. 7b), a new unused plate (Fig. 7c) and an unbroken, unused screw (Fig. 7d). The metallography was done to determine the nature, morphology and distribution of the inclusion content of these various specimens since SEM of some of the in vivo fracture surface revealed inclusions and/or possibly inclusion-related microvoid coalescence. The typical inclusions shown in Fig. 7 are similar in amount and distribution in each of the samples studied. Results indistinguishable from those shown in Fig. 7 were
also obtained by optical metallography of an unimplanted plate broken in the Instron machine, a plate routinely removed intact from a patient, and the fractured plate removed from Case 4. The inclusions detected by optical metallography were found by microprobe analysis to be metal carbides, silicides, and A1203, resulting from the ‘killing’ process in steel manufacture. An unused 6-hole tibia1 plate was subjected to 3-point bending in a repeated load/unload mode with periodically increasing stress until failure occurred. The initial failure occurred in the plate at a stress level of 177 942 psi. This value agrees closely with that reported by Ludwigson (1956) for cold-worked 3 16 stainless steel. The plate was only partially cracked in the repeated bending apparatus, however, and the final fracture was effected manually. The fracture surfaces were examined by scanning electron microscopy in those regions of initial failure produced by the
Richman et al. : Metallurgical Examination of Plates
b
d Fig. 7. Optical metallography
of a, broken plate, b, screw, c, an unused plate and d, unused screw (X 100).
repeated bending. The SEM of the fracture surface shows fatigue striations as well as small pits which may be associated with inclusions and/ or microvoid coalescence (Fig. 8). The striations in Fig. 8 show the same non-parallel nature as those in the tibia1 plates which failed in viva (Fig. 5).
DISCUSSION Compression plates are currently in widespread use for the internal fixation of fractures (Miiller et al., 1970). The main materials used for these plates are type 316L stainless steel and Vitallium (Brettle, 1970; Brettle et al., 1971). Both materials possess a high yield and tensile strength, a high Young’s modulus and good fatigue resistance. Vitallium is more resistant to corrosion than stainless steel. Stainless steel, however, possesses certain properties which might be advantageous in bone compression plates. Most notable among
Fig. 8. SEM of 6-hole ASIF tibia1 plate fractured by repeated 3-point bending ( x 500). these properties is the ability of stainless steel to be cold worked. Thus, stainless-steel plates not only may be made by metal working rather than by
18
but they may be bent in the operating theatre to conform to the bone under repair and thereby make possible a more efficient reduction of the fracture. Cold working of ASIF tibia1 plates in the lnstron machine increased the fatigue threshold from 138 000 psi to 180 000 psi in unworked samples. In contrast, neither cast nor wrought Vitallium is easily worked. Vitallium plates are also much more expensive than those made of stainless steel. Because of its great strength and its ability to be bent in the operating theatre to conform to bone, stainless steel would appear to be the material of choice for a compression plate for a weight-bearing bone such as the tibia. However, stainless-steel implants are thought to be susceptible to corrosion and corrosion fatigue. The present study was undertaken to determine whether corrosion or corrosion fatigue played a significant role in the failure of stainless-steel tibia1 compression plates in patients whose tibia1 fractures were treated by the ASIF method of internal fixation at the Haukeland Hospital. Failure of the plate was considered, in this study, to be fracture of the plate in situ. Fracture of a tibia1 compression plate is rare and occurred in only 4 of our 99 fractures operated on between 1970 and 1973. Two of the broken plates were available for the present metallurgical study. All 4 plate failures, however, were associated with delayed union of the fracture (Thunold et al., 1975). All these patients persisted in weight bearing in spite of warnings from the surgeons based on radiological and clinical evidence of non-union. The importance of an adequately united fracture that can allow weight bearing by the tibia in preventing fracture of the tibia1 plates is shown in the following paragraph. Assuming that a plate is rigidly fixed to both fragments of a tibia, and that the fracture has either not been properly reduced or some resorption has taken place, the effect of weight bearing on the plate may be calculated. A person weighing 150 lb (685 kg) when walking or bearing all his weight on one leg will produce a bending moment and a stress of 180 000 psi in the tibia1 plate at a section through each of the screw holes adjacent to the fracture. This stress of 180 000 psi exceeds the ultimate tensile strength of 142 000 psi of the 316L stainless steel of which the plate is made, and failure of the plate will occur through one of the screw holes. If only partial weight is exerted on the leg with the fracture, the applied stress will be less than the tensile strength but greater than the yield strength, and the plate will be subjected to cyclical loading castmg,
Injury: the British Journal ot Accident Surgery Vol. ~/NO. 1
on walking. This will lead to failure of the tibia1 plate by fatigue, as in the plates in this study. The SEM of the interface between the screw and countersink hole (Fig. 4) demonstrates crevice corrosion. Electron microprobe analysis of the corroded area revealed Ca, K, Cl, P, Na and 0 in addition to the elements normally found in AIST 316L stainless steel (Table I). Crevice corrosion occurs normally in 42 per cent of stainless-steel implants and does not usually contribute to fracture failure (Colangelo and Greene, 1969). In none of our cases did crevice corrosion appear to be linked to plate failure. Fig. 5 shows a typical SEM of the fracture surface of a broken plate. Features of the SEMs of both the plates that broke in the patient were essentially the same as those shown in Fig. 5. In both cases fatigue striations were found. Scanning electron microscopy and microprobe analysis revealed no corrosion products on the fracture surface (Table I). Electron microprobe analysis of the fracture surfaces indicated that the surfaces contained the elements characteristic of AISI 316L stainless steel but not the corrosion products of the type found at the screw-plate interface. The fracture surface of the screw which broke in situ without serious inconvenience to the patient is indistinguishable from the surfaces of the broken plates (Fig. 6). The fracture surface appearances made a pure fatigue rather than a corrosion fatigue mechanism a more likely cause of plate and screw failure in the cases studied. The SEMs of a plate fatigue-fractured by cyclical 3-point loading showed fatigue striations, microvoids, pits and small inclusions (Fig. 8). Regions of ductile failure induced by manual separation of the plate fragments are not shown. The plate began to break at 178 000 psi after 39 000 cycles. The presence of inclusions might have caused the failure, but the metallography chiefly revealed fatigue failure when the tensile strength of the stainless steel had been raised by cold working and prior deformation over the normal value of 138 000 psi. While attention has been called to the role of inclusions in fracture of stainless-steel plates (Cahoon and Paxton, 1968), the plates which failed in situ in our series were found to have the same inclusion content as both those removed without failure and new plates (Fig. 7). The inclusion content cannot therefore
be responsible
for the fracture of these plates. Though inclusions may potentially increase the rate of corrosion, the data presented here show that corrosion did not contribute to the failure of the plates, but that in an ASIF screw and in 2 plates which fractured in
Richman et al.
: Metallurgical
Examination
19
of Plates
situ, the cause of failure was pure fatigue. The plate failures were preceded by delayed bone union (Thunold et al., 1975). The primary problem associated with failure of ASIF plates appears to be, therefore, delayed or non-union, perhaps induced by some bio-incompatibility of the steel or by imperfect surgical technique. A broken plate is certainly a complication of importance in a patient operated on for a fracture in the lower limb. It was, however, a direct consequence of delayed healing in this series and was not connected with defects in the implant itself. Plate fracture should not be regarded as an argument against the ASIF technique for reduction and fixation of tibia1 fractures. The clinical course after a second operation, using the same ASIF method, was uneventful and all fractures were united within II months of the primary injury; thus, bioincompatibility of the tibia1 plates was unlikely to have caused the original non-union. Acknowledgement The authors thank Dr John J. Sipe for providing ASIF tibia1 plates for the cyclical bending study. REFERENCES J. (1970) A survey of the literature on metallic surgical implants. Injury 2,26.
Brettle
~cquests for reprints 02906, USA.
should be addressed to: Dr J. K. Weltman,
Brettle J. and Hughes A. N. (1970) A metallurgical examination of surgical implants which have failed in service. Injury 2, 143. Brettle J., Hughes A. N. and Jordan B. A. (1971) Metallurgical aspects of surgical implant materials. Injury 2, 225. Cahoon J. R. and Paxton H. W. (1968) Metallurgical analysis of failed orthopaedic implants. J. Biomed. Mater. Res. 2, I. Cohen J. (1966) The performance and failure in performance of surgical implants in orthopaedic surgery. J. Mater. 1,354. Cohen J. and Wulff J. (1972) Clinical failure caused by corrosion of a vitallium plate. J. Bone Joinr Surg. 54A, 617. Colangelo V. J. and Greene N. D. (1969) Corrosion and fracture of type 316 SMO orthopaedic implants. J. Biomed. Mater. Res. 3, 247. Greene N. D. and Jones D. A. (1966) Corrosion of surgical implants. J. Mater. 1,345. Ludwigson D. C. (1956) Requirements for metallic surgical implants and devices. Metals Engng Q. 5, 1. Mtiller M. E.. Allgower M. and Willenegaer H. (1970) Man&l of internal Fixation (AO-tec&ique). Berlin, Springer-Verlag. Rose R. M., Schiller A. L. and Radin E. L. (1972) Corrosion accelerated mechanical failure of a vitallium nailplate. J. Bone Joint Surg. 54A, 854. Thunold J.. Varhaug J. E. and Bierkeset T. (1975) Tibia1 shaft fractures treated by rigid internal fixation. Injury 7, 125. Uhlig H. H. (ed.) (1948) Corrosion Handbook. New York, Wiley, p. 11.
The Miriam
Hospital,
164 Summit
Avenue,
Providence,
Rhode
Island