- Email: [email protected]
Nonunions of the surgical neck of the humerus: Surgical treatment with an intramedullary bone peg, internal fixation, and cancellous bone grafting
ORIGINAL ARTICLES m
Nonunions of the surgical neck of the humerus: Surgical treatment with an intramedullary bone peg, internal fixation, and cancellous bone grafting G. Walch, MD, R. Badet, MD, L. Nov~-Josserand, MD, and C. Levigne, MD, Lyon, France
Twenty patients with pseudarthrosis of the upper humerus underwent surgery with the intramedullary bone peg technique. A 6 to 10 cm corticocancellous autogenous bone graft (11 lilac crest, 6 anterior fibial crest, 3 middle-third of the fibula) was pegged into the humerus and bridged the pseudarthrosis. Stability of the fracture site was obtained by plate osteosynthesis; an additional peripheral cancellous graft was performed. Our patient series included 15 women and five men with an average age at operation of 58 years; the dominant side was involved in 12 cases. Eleven had undergone 22 previous operations. The average delay between fracture and surgery was 12 months, (range 6 to 72 months). The patients were monitored an average of 42 months (range 12 to 120 months). Union was confirmed in 19 cases; the last case demonstrated no peripheral callus. No necrosis of the humeral head was seen. Active anterior elevation of the shoulder improved from an average of 60 ~ to an average of 131 o. According to Constant's scale adjusted according to age and sex, the results obtained averaged 81.2%. Subjectively, 65% of patients were very satisfied, 30% were satisfied, and 5% were disappointed. The rate of union (96%) is in contrast with the results reported in the literature, underlining the importance of an intramedullary bone graft in association with peripheral osteosynthesis in the treatment of pseudarthrosis of the surgical neck of the humerus. (J SHOULDERELBOWSURG 1996;5:161-8.) Pseudarthrosis of the upper humerus is rare and most disabling; it causes considerable pain and loss of active motion of the shoulder/' lo Direct surgical correction is difficult because of the absence of good-quality cancellous bone in the humeral shaft and because of softening, resorption, and osteopenia of the humeral head. 1~ 12 In treating upper humeral nonunions the surgeon must provide adequate immobilization of the frac-
From Clinique de Chirurgie Orthop~dique EmiJie de Vialar. Reprint requests: Gilles Wdch, MD, Clinique de Chirurgie Orthop6dique Emilie de Vidar, 116 Rue Antoine Charid, Lyon 69003, France. Copyright @ 1996 by Journal of Shoulder and Elbow Surgery Board of Trustees. 1058-2746/96/$5.00 + 0 32/1171631)
ture itself by both internal and external means while at the same time providing for early joint mobilization to prevent stiffness and restore muscle function.Z, ~r Open reduction and internal fixation with a plate gave a 70% failure rateT' 9. 12Fixation with a tension band and an intramedullary nail or a T plate led to union in more cases but required prolonged immobilization and additional surgery for arthrolysis, hardware removal, or both. 1~ The use of hemiarthroplasty of the shoulder carries the risk of poor functional results, and this problem has limited its application in posttraumatic conditions, especially if the glenohumeral joint is intact. 4' 12 We have reviewed our experience with a similar procedure used for the surgical correction of nonunions of the surgical neck of humerus. We aimed at bony realignment and stable fixation; but we 161
162
J. Shoulder Elbow Surg. May/June 1996
Walch et al.
Table I Data on the patients
Case
Age (yrs.)
Sex
Side
Fracture type
Initial treatment
Other operation
Delay (mo) between
injury-operation
Type of graft corticocancellous
1
44
F
R
surgical neck
conservative
6
tibia
2
73
F
R
3-part
9
8
tibia
3
65
F
L
surgical neck
conservative
6
tibia
4
68
F
R
3-part
conservative
12
tibia
5
60
F
L
surgical neck
conservative
6
tibia
6
45
M
L
3-part
ORIF
72
tibia
7 8 9 10 11 12 13 14 15 16
78 27 66 67 61 75 40 42 83 68
F F F F F M F M F M
L R R R L L L R L R
3-part 3-part surgical neck 3-part 3-part 3-part surgical neck surgical neck 3-part 3-part
conservative ORIF sepsis 9 ORIF conservative conservative conservative ORIF ORIF conservative
6 12 8 6 6 6 7 8 12 6
Fibula Fibula Fibula iliac-crest ihac-crest i liac-crest iliac-crest ilioc-crest iliac-crest i liac-crest
17
46
F
R
9
29
18
iliac-crest
18
54
F
R
surgical neck (deltoid palsy) surgical neck
10RiF 2 Ilizarov
18
i liac-crest
]9
39
M
R
surgicalneck
ORIF (deltoid palsy) ORIF
12
iliac-crest
R
{gunshot) 3-part
ORIF
8
iliac-crest
20
54
F
29
30RIF Hw removal 10RIF 20RIF
19
F, Female; M, male; R, right; L, left; ORIF, open reduction and internol fixation; Hw, metallic internal fixation. *Range of motion on last consultation. tlnterviewed over the phone, examined by his attending physician.
considered preservation of the patient's joint and functional rehabilitation to also be important. PATIENTS A N D METHOD Twenty-three patients with nonunion of the proximal humerus were seen in our unit between 1983 and 1993. Three patients refused surgery because they had no pain; 20 underwent surgery. The same technique was applied regardless of age, presence of osteoporosis, or necrosis of the humeral head; our main objective was to consolidate their pseudarthrosis. Only patients with an average period from injury of at least 6 months or those with failed surgery were included. The cases were limited to those in which the pseudarthrosis gap lay between the lesser tuberosity and the pectoralis major insertion. Fifteen women and five men with an average
age of 58 years (range 21 to 83 years) were studied. The series included 12 right and eight left shoulders. Twelve were on the dominant arm. Fourteen of the original fractures resulted from a fall, three from a car accident, and one each from gunshot, a helicopter accident, and a motorcycle accident. The initial fractures included nine displaced two-part fractures of the proximal humerus and 11 three-part fractures of the proximal humerus. 8 Union was seen for all the fractures of the greater tuberosity, with any persistent malunion offset less than 5 mm. Nineteen fractures were closed, and one fracture was due to a gunshot wound. Initial management had been by open reduction and internal fixation in 11 cases and by nonoperative treatment in nine. Eight patients had undergone one or several operations for pseudarthrosis, four of them in our unit; two patients had
J. Shoulder Elbow Surg. Volume 5, Number 3
Walch et al.
Follow-up (mo)
Final active elevation (~
yes
120
130"
60/60*
~lephone
None
no
40
110
60
Deceased
stress fracture of the tibia None
no
56
160
60
Deceased
yes
72
160
60
87.5
no
60
160
60
100 81
Type of graft cancellous
Type of synthesis
Complication
Su ~erior tibia epiphys~s Su ~erior tibia epiphysls Su ~erior tibia epiphysls Su ~erior tibia epiphysls Su ~erior tibia epiphysls Su ~erior tibia epiphysts iliac-crest iliac-crest
T plate
None
T plate T plate
lilac-crest
T plate
iliac--crest iliac-crest iliac-crest iliac-crest iliac-crest iliac-crest iliac-crest
Blade plate T plate T plate T plate T plate T plate T plate
Blade plate T plate
iliac-crest
Constant
score (%)
no
60
160
0
yes yes yes yes yes no no yes no no
24 42 35 64 53 39 44 27 23 31
60* 110 160 130 90 120 170 180 90 110
20/40* 20 60 45 20 50 60 60 20/80 20/70
yes
19
165
60/60
90
Blade plate
stress fracture of the tibia stress fracture of the tibia None None None None None None None None None fracture shaft humerus repair pectoralis major None
no
16
110
80
75
T plate
None
yes
20
110
40
70
Blade plate
None
yes
18
140
40
80
T plate T plate T plate
Blade plate iliac-crest
Hardware removal
Final external rotation (~
163
undergone three operations, two had had two operations, and four had had a single operation (Table I). The period belween fracture and our operative treatment averaged 12 months (range 6 to 72 months). At the time of treatment of the nonunion, the patients reported loss of function with inability to raise the arm more than 60 ~ actively; active external and internal rotation was not possible. The shoulder was not painful if the arm was not being used but was painful in the activities of daily living. The passive range of motion was preserved, but motion stemmed from both the pseudarthrosis and the joint. Four patients had complications after the accident or previous surgery. One man (case 19) was the victim of a shooting accident; he had received a gunshot to the upper humerus. Emergency surgery had been performed debriding wounds and inserting Rush pins. The anterior and middle por-
telephone 61 100 72 70 80 95 100t 95 80
tions of the deltoid were severely injured. One woman (case 17), who had undergone operation twice, had partial paralysis of the deltoid; this condition was of iatrogenic origin or occurred after trauma at the time of the accident. She also had a ruptured pectoralis major after her second operation; this had resulted from an extended approach. One woman (case 18), who had undergone four operations with insertion of a nail, a plate, and two external fixators according to the technique of Ilizarov, had total paralysis of the deltoid, probably resulting from passage of the wires. One woman (case 8), who had undergone emergency surgery, had an infection that had led to early additional surgery to remove the osteosynthetic material and for lavage and drainage. Infection was irradicated, leaving the patient with necrosis of the humeral head and pseudarthrosis of the surgical neck. At the time we performed our
164
J. Shoulder Elbow Surg. May/June 1996
Walch et al.
Figure | Techniquefor osteosynthesisassociated with intramedullary bone peg.
operation 12 months later, all pathologic and microbiologic test results were normal. Operative ter A deltopectoral approach was used, preserving the upper and lower deltoid insertions. The muscle was separated from the upper humerus, and the proximal and distal fragments of the pseudarthrosis were freshened, with excision of the interposed fibrous tissues. A 6 to 10 cm bone peg was taken 11 times from the iliac crest on the same side, six times from the upper third of the tibial crest, and three times from the middle-third of the fibula on the same side. Additional curettage of adjacent cancellous bone was carried out with a curet. The corticocancellous graft was first impacted into the distal humerus and was then plugged into the head like a peg-and-ball (Figure 1). Impaction was easily obtained by gently tapping the bended elbow. Our objective was to obtain as much surface contact as possible. Rotatory problems were avoided by using the long head of the biceps as a guide in patients who had had no previous surgery, but this was particularly difficult in patients who had undergone surgery. We endeavoured to keep 20 ~ of external rotation with the elbow near the side. An external T plate was then applied onto the outer humerus and fixed by 3.5 or 4.5 mm cortical screws and 4.0 or 6.5 mm cancellous screws at the level of the humeral
epiphysis. The screws generally bored into the bony corticocancellous peg graft. Complementary cancellous graft was placed around the fracture site. Surprisingly, glenohumeral motion was always relatively well preserved. We have never performed arthrolysis, and the rotator cuff has never been explored or repaired; our aim was to obtain bone union. The wound was drained for 48 hours. The arm was immobilized by a sling for 30 days. Pendulum motions, gentle mobilization with the patient lying on his back; and self-rehabilitation in water were started after 30 days. Active motion of the shoulder was not allowed before 3 months. Rehabilitation in water by rotational motions was carried on for as long as possible. Anterioposterior radiographs with internal, neutral, and external rotation were done every 45 days until solid formation of bony callus occurred. Final evaluation was done by personal examination and included clinical examination, radiographic evaluation (three rotations), range of active motion, and functional results. Points were scored according to the scale of Constant, 3 which assigns 15 points to pain, 20 points to daily activities, 40 points to range of motion, and 25 points to strength. The final result was scored on a scale of 100 points, adjusted according to age and sex. Results were excellent when the shoulder reached at least 90% of the normal score in relation to age and sex, good when the score was between 75% and 89%, fair when it was between 74% and 50%, and poor when it was below 50%. The subjective results were determined by the question, "Are you very satisfied, satisfied, disappointed, or dissatisfied?"
RESULTS Fifteen patients were reexamined. Three patients were unable to return, and they were interviewed over the phone and sent recent x-ray films; one had had his shoulder examined once by his attending physician. Two had died, but because they had undergone radiologic reexamination at 40 months and 56 months, we included their outcomes. The average follow-up was 42 months (range 12 to 120 months). Nineteen patients (96%) had radiologic union (Figures 2, 3, and 4); mean time to union was 4 months (range 3 to 10 months). An 84-year-old woman had good bone continuity because of the intramedullary plug, although the peripheral graft had failed, suggesting that bone union was not complete. No aseptic necrosis of the
J. Shoulder Elbow Surg. Volume 5, Number 3
Walch et al.
165
Figure 2 Case 5, 60year old woman who underwent surgery for pseudarthro sis 6 months after fracture (A). Union was successful(B). Five years later Constant score was 100%
Figure 3 Case 4, atrophic, unstable pseudarthrosiswith plate fixation after k,vo operations (A). Union was successful (B). Six years later Constant score was 875%
humeral head was observed, except for case 8, who had preoperative necrosis. Pain relief was achieved for all our patients, with total relief in 54% and persistent but less marked pain in 46%. None of our patients had aggravation of pain, and patients who had had no pain remained without pain. The mean subjective evalu-
ation of pain with Constant's scale was 11 out of 15. Range of motion of active anterior elevation of the shoulder had improved from an average of 60 ~ to an average of 131 ~ (60 ~ to 180~ The mean active external rotation with the elbow at the side was 45 ~ (0 ~ to 60~ With the scale of Con-
166
Walch et al.
J. ShoulderElbow Surg. May/June 1996
Figure 4 Case 20, atrophic, unstable pseudarthrosis after failed internal fixation (A). Union was successful (B). Eighteen months later Constant score was 80%.
stant the mean objective evaluation of motion was 30 out of 40. The overall score of Constant, adjusted according to age and sex, was 81.2% (61 to 100). Results obtained were excellent for six patients, good for six, and fair for four; no negative results occurred. The score of Constant was not recorded for the last four patients, because two had died and two had been interviewed over the phone with no clinical examination. Subjectively, 65% of our patients were very satisfied, 30% were satisfied, and 5% were disappointed. Three patients did not recover active anterior elevation motion of more than 90 ~ A 78-year-old woman (case 7) had a traumatic three-part fracture of the upper humerus. She used walking aids to counterbalance the injuries to her left leg resulting from the same accident. No rehabilitation was possible; active elevation was 60 ~. Moreover, medicolegal insurance problems complicated the situation. The subjective result was disappointing. The second case (15) concerned an 83-year-old woman who also had malunion of the upper epiphysis and associated greater tuberosity fracture. The intramedullary graft was secure, but peripheral callus failed to develop because of resorption of the peripheral cancellous graft. We did not wish to continue physiotherapy to avoid refracture. The patient is very satisfied, although her active elevation is only 90 ~ The third case (t 1) con-
cerned a 61-year-old woman who had nonunion after a traumatic three-part fracture. Rehabilitation was discontinued because of poor motivation. Three factors were responsible for the poor resuits obtained in restoring active motion. The first was the type of initial fracture. Restoration of the active elevation was 117 ~ after a three-part fracture of the upper humerus and 157 ~ for fractures of the surgical neck that preserved the tuberosities. The second factor was the age of the patient, which also had an influence on rehabilitation. Before the age of 65 years, active motion averaged 145~ after 65 years it fell to 121 o. The third factor was poor condition of the deltoid (cases 17, 18, and 19), although all three patients managed to recover an active motion of more than 90 ~. On the other hand, the presence and number of previous operations had no significant incidence on restoration of function, if the deltoid was not impaired. Complications. The bone peg was obtained six times from the anterior tibial crest. A pathologic fracture was observed three times at the lower end of the graft site at 2 months after operation. These fractures required a walking cast for 2 to 3 months, after which union was complete. Rupture of the pectoralis major (case 17) was not repaired at the same time as the treatment of the pseudarthrosis. Suture of the tendon into a
J. Shoulder Elbow Surg. Volume 5, Number 3
bony trough was performed 12 months later during removal of the internal fixation. Restoration of motion was full; the cosmetic result was good. A humeral fracture below the T-plate (case 16) was observed 30 days after surgery, after a fall occurred. The fracture showed little displacement, and union occurred after a further 45-day immobilization period. DISCUSSION It is established that successful bony union is easily obtained for fracture of the surgical neck, whether simple or complex, when suitable orthopaedic treatment is applied. Baux and Razemon 1 have reported only three cases of pseudarthrosis in a multicenter series of 1237 fractures of the upper humerus after conservative treatment. In 1964 Sorensen TM reported two cases of pseudarthrosis; he found only seven cases in his literature review. In 1978 Lerat and Trillat 6 reported three cases of pseudarthrosis in a series of 500 fractures of the surgical neck. In 1983 Neer 9 reported 42 cases of pseudarthrosis and found 29 cases described in the literature. In 1990 Healy et al. 4 reported 14 patients in a multicenter study. Our series included nine cases previously treated conservatively. It is not easy to establish whether the initial treatment invites pseudarthrosis, because many different approaches have been described: sling, spica, hanging cast, traction, or thoraco-brachial cast. Moreover, comparative series of fractures to which various treatments were applied revealed no fewer or no more pseudarthroses depending on the treatment method selected. Many factors are probably responsible for these pseudarthroses: absence of prolonged immobilization, age of the patient, poor quality of the bone, displacement of the inferior fragment produced by the pectoralis major, or interposition of the biceps or deltoid muscle. Similarly, pseudarthrosis occurring after immediate osteosynthesis is rare. Lerat and Trillat 6 report one case, Neer 9 eight cases, and Healy et al. 4 14 cases; we found 12 cases in our series. A rigorous analysis is not possible because of the multiple types of osteosyntheses used in emergencies (Kwires, screws, rods, or plates). Apart from the factors described previously, inadequate firmness of the fixation was one of the main causes of pseudarthrosis. Pseudarthrosis is unexpectedly difficult to manage; in our series 14 operations failed, four of which we had carried out personally. In 1983
Walch etal.
167
Neer 9 reported 35 failed attempts at surgical repair out of a series of 50 patients. Healy et al. 4 reported 12 unions in a multicenter study of 18 cases of pseudarthrosis treated by osteosynthesis plus bone graft. Original suggestions have been proposed to obviate such failures. In 1965 Leach and Premer5 carried out an osteosynthesis on three patients using an angulated blade plate and a nail placed successively through the head, the plate, and the medullary shaft. The technique they used led to union but did not significantly improve motion despite a short immobilization of 3 weeks. In 1982 Scheck ~3 offered five of his patients treatment of their pseudarthrosis by implanting the humeral shaft into the head. He placed a variety of osteosynthesis materials and an extramedullary, intracephalic bone plug to act as a stabilizer. This allowed 80% union, a constant lowering of pain, and an important increase in motion. However, this technique leads to loss of length, weakening the deltoid, rendering it too long, and causing inferior subluxation of the humeral head. In 1983 Neer 9 proposed a new technique with associated osteosynthesis by an intramedullary nail and tension band. Immobilization lasted 3 months, and the operation entailed additional surgery for arthrolysis and removal of hardware. He obtained 12 unions out of 13 cases, but he reported no precise results of function and motion. With a similar technique (figure-of-eight wires and rods), Bigliani et al. 2 have reported 10 cases (21 operations), with four persistent pseudarthrosis. According to Bigliani et al., "Open reduction and internal fixation for nonunions following primary fracture fixation attempts led to poor results in all cases." The results obtained in our series were more promising, because we obtained 96% union. The intramedullary bone peg provided solid bone that filled the cavity of the head, thus preventing motion between the shaft and the head. It also allowed solid osteosynthesis by transfixing the intramedullary support with screws and it increased the bone present by adding cancellous graft. The technique preserves the rotator cuff and does not aggravate joint stiffness, because no arthrotomy is required. Immobilization is limited to 30 days, avoiding further stiffening of the shoulder that often follows a long period of immobilization. Because of complications that occurred with bone fragments harvested from the anterior tibial crest, the site is not recommended. When relatively straight 6 to 8 cm corticocancellous bone grafts
168
Walch etal.
can be obtained from the iliac crest, this site is recommended. When the graft is too narrow, we advocate obtaining a bone graft from the middle third of the fibula. We have been offering this technique to our patients since 1983 whatever the state of their humerus, and even to patient 8, who had humeral head necrosis after sepsis. We consider the primary aim is to obtain bone union, and surprisingly we have observed no postoperative necrosis, which encourages us to propose the technique whatever the radiologic condition of the humeral head is. Motion is less improved in three-part fractures, but the results are better than those we obtained with humeral head prosthetic replacements in fracture malunions. Neer 1~ advocates total or humeral head replacement in cases of long-standing pseudarthrosis when there is extensive osteopenia or necrosis of the humeral head; his results were not reported. Bigliani et al. 2 reported 70% satisfactory results for 10 cases treated by humeral head replacement. In conclusion, although rare, nonunion after fracture of the surgical neck of the humerus remains a challenge for the orthopaedic surgeon. Surgical treatment with the intramedullary bone peg technique seems to afford good results with a 96% rate of bone union. REFERENCES 1. Baux S, Razemon jP. Les fractures et les fractures-luxafions de I'extr6mit6 sup6rieure de I'hum6rus. Rev Chir Orthop 1969; 55:387-496.
J. Shoulder Elbow Surg. May/June 1996 2. Bigliani LU, Nicholson GP, Pollock RG, Duralde XA, Self EB. Personal communication. Ninth Open Meeting of the American Shoulder and Elbow Surgeons. San Francisco, February 21, 1993. 3. Constant CR, Murley AHG. A clinical method of functional assessment of the shoulder. Clin Orthop 1987;214:160 4. 4. Healy WL, JupiterJB, Kristiansen TK, White RR. Nonunion of the proximal humerus. In: Post M, Morrey BF, Hawkins RJ, editors. Surgery of the shoulder. Chicago: Mosby-Year Book Inc., 1990:59 62. 5. Leach RE, Premer RF. Nonunion of the surgical neck of the humerus. Minn Med 1965;48:318-22. 6. Lerat JL, Trillat A. La pseudoarthrose du col hum@ral existe-telle? Lyon Chir 1978;74:105-9. 7. McCreath SW. Delayed union and nonunion in fractures of humeral shaft. J Bone Joint Surg 1955;57B:393. 8. Neer CS II. Displaced proximal humeral fractures. Part I: Classification and evaluation. J Bone Joint Surg Am 1970; 52A:1077 89. 9. Neer CS ik Nonunion of the surgical neck of lhe humerus. Orthop Trans 1983;7:389. 10. Neer CS II. Shoulder reconstruction. Philadelphia: WB Saunders Co. 1990:230-4. 11. Norris TR, TurnerJA. Surgical treatment of nonunions of the upper humerus shaft fracture in the elderly. Orthop Trans 1985;9:44. 12. Norris TR, Turner JA, Bovill D. Nonunion of the upper humerus: an analysis of the etiology and treatment in 28 cases. In: PostM, Morrey BF, Hawkins RJ, editors. Surgery of the shoulder. Chicago: Mosby Year Book Inc, 1990:63-7. 13. Scheck M Surgical treatment of nonunions of the surgical neck of the humerus. Chn Orthop 1982;167:255-9. 14. Sorensen KH. Pseudarthrosis of the surgical neck of the humerus. Acta Orthop Scand 1964;34:132-8.
Nonunions of the surgical neck of the humerus: Surgical treatment with an intramedullary bone peg, internal fixation, and cancellous bone grafting G. Walch, MD, R. Badet, MD, L. Nov~-Josserand, MD, and C. Levigne, MD, Lyon, France
Twenty patients with pseudarthrosis of the upper humerus underwent surgery with the intramedullary bone peg technique. A 6 to 10 cm corticocancellous autogenous bone graft (11 lilac crest, 6 anterior fibial crest, 3 middle-third of the fibula) was pegged into the humerus and bridged the pseudarthrosis. Stability of the fracture site was obtained by plate osteosynthesis; an additional peripheral cancellous graft was performed. Our patient series included 15 women and five men with an average age at operation of 58 years; the dominant side was involved in 12 cases. Eleven had undergone 22 previous operations. The average delay between fracture and surgery was 12 months, (range 6 to 72 months). The patients were monitored an average of 42 months (range 12 to 120 months). Union was confirmed in 19 cases; the last case demonstrated no peripheral callus. No necrosis of the humeral head was seen. Active anterior elevation of the shoulder improved from an average of 60 ~ to an average of 131 o. According to Constant's scale adjusted according to age and sex, the results obtained averaged 81.2%. Subjectively, 65% of patients were very satisfied, 30% were satisfied, and 5% were disappointed. The rate of union (96%) is in contrast with the results reported in the literature, underlining the importance of an intramedullary bone graft in association with peripheral osteosynthesis in the treatment of pseudarthrosis of the surgical neck of the humerus. (J SHOULDERELBOWSURG 1996;5:161-8.) Pseudarthrosis of the upper humerus is rare and most disabling; it causes considerable pain and loss of active motion of the shoulder/' lo Direct surgical correction is difficult because of the absence of good-quality cancellous bone in the humeral shaft and because of softening, resorption, and osteopenia of the humeral head. 1~ 12 In treating upper humeral nonunions the surgeon must provide adequate immobilization of the frac-
From Clinique de Chirurgie Orthop~dique EmiJie de Vialar. Reprint requests: Gilles Wdch, MD, Clinique de Chirurgie Orthop6dique Emilie de Vidar, 116 Rue Antoine Charid, Lyon 69003, France. Copyright @ 1996 by Journal of Shoulder and Elbow Surgery Board of Trustees. 1058-2746/96/$5.00 + 0 32/1171631)
ture itself by both internal and external means while at the same time providing for early joint mobilization to prevent stiffness and restore muscle function.Z, ~r Open reduction and internal fixation with a plate gave a 70% failure rateT' 9. 12Fixation with a tension band and an intramedullary nail or a T plate led to union in more cases but required prolonged immobilization and additional surgery for arthrolysis, hardware removal, or both. 1~ The use of hemiarthroplasty of the shoulder carries the risk of poor functional results, and this problem has limited its application in posttraumatic conditions, especially if the glenohumeral joint is intact. 4' 12 We have reviewed our experience with a similar procedure used for the surgical correction of nonunions of the surgical neck of humerus. We aimed at bony realignment and stable fixation; but we 161
162
J. Shoulder Elbow Surg. May/June 1996
Walch et al.
Table I Data on the patients
Case
Age (yrs.)
Sex
Side
Fracture type
Initial treatment
Other operation
Delay (mo) between
injury-operation
Type of graft corticocancellous
1
44
F
R
surgical neck
conservative
6
tibia
2
73
F
R
3-part
9
8
tibia
3
65
F
L
surgical neck
conservative
6
tibia
4
68
F
R
3-part
conservative
12
tibia
5
60
F
L
surgical neck
conservative
6
tibia
6
45
M
L
3-part
ORIF
72
tibia
7 8 9 10 11 12 13 14 15 16
78 27 66 67 61 75 40 42 83 68
F F F F F M F M F M
L R R R L L L R L R
3-part 3-part surgical neck 3-part 3-part 3-part surgical neck surgical neck 3-part 3-part
conservative ORIF sepsis 9 ORIF conservative conservative conservative ORIF ORIF conservative
6 12 8 6 6 6 7 8 12 6
Fibula Fibula Fibula iliac-crest ihac-crest i liac-crest iliac-crest ilioc-crest iliac-crest i liac-crest
17
46
F
R
9
29
18
iliac-crest
18
54
F
R
surgical neck (deltoid palsy) surgical neck
10RiF 2 Ilizarov
18
i liac-crest
]9
39
M
R
surgicalneck
ORIF (deltoid palsy) ORIF
12
iliac-crest
R
{gunshot) 3-part
ORIF
8
iliac-crest
20
54
F
29
30RIF Hw removal 10RIF 20RIF
19
F, Female; M, male; R, right; L, left; ORIF, open reduction and internol fixation; Hw, metallic internal fixation. *Range of motion on last consultation. tlnterviewed over the phone, examined by his attending physician.
considered preservation of the patient's joint and functional rehabilitation to also be important. PATIENTS A N D METHOD Twenty-three patients with nonunion of the proximal humerus were seen in our unit between 1983 and 1993. Three patients refused surgery because they had no pain; 20 underwent surgery. The same technique was applied regardless of age, presence of osteoporosis, or necrosis of the humeral head; our main objective was to consolidate their pseudarthrosis. Only patients with an average period from injury of at least 6 months or those with failed surgery were included. The cases were limited to those in which the pseudarthrosis gap lay between the lesser tuberosity and the pectoralis major insertion. Fifteen women and five men with an average
age of 58 years (range 21 to 83 years) were studied. The series included 12 right and eight left shoulders. Twelve were on the dominant arm. Fourteen of the original fractures resulted from a fall, three from a car accident, and one each from gunshot, a helicopter accident, and a motorcycle accident. The initial fractures included nine displaced two-part fractures of the proximal humerus and 11 three-part fractures of the proximal humerus. 8 Union was seen for all the fractures of the greater tuberosity, with any persistent malunion offset less than 5 mm. Nineteen fractures were closed, and one fracture was due to a gunshot wound. Initial management had been by open reduction and internal fixation in 11 cases and by nonoperative treatment in nine. Eight patients had undergone one or several operations for pseudarthrosis, four of them in our unit; two patients had
J. Shoulder Elbow Surg. Volume 5, Number 3
Walch et al.
Follow-up (mo)
Final active elevation (~
yes
120
130"
60/60*
~lephone
None
no
40
110
60
Deceased
stress fracture of the tibia None
no
56
160
60
Deceased
yes
72
160
60
87.5
no
60
160
60
100 81
Type of graft cancellous
Type of synthesis
Complication
Su ~erior tibia epiphys~s Su ~erior tibia epiphysls Su ~erior tibia epiphysls Su ~erior tibia epiphysls Su ~erior tibia epiphysls Su ~erior tibia epiphysts iliac-crest iliac-crest
T plate
None
T plate T plate
lilac-crest
T plate
iliac--crest iliac-crest iliac-crest iliac-crest iliac-crest iliac-crest iliac-crest
Blade plate T plate T plate T plate T plate T plate T plate
Blade plate T plate
iliac-crest
Constant
score (%)
no
60
160
0
yes yes yes yes yes no no yes no no
24 42 35 64 53 39 44 27 23 31
60* 110 160 130 90 120 170 180 90 110
20/40* 20 60 45 20 50 60 60 20/80 20/70
yes
19
165
60/60
90
Blade plate
stress fracture of the tibia stress fracture of the tibia None None None None None None None None None fracture shaft humerus repair pectoralis major None
no
16
110
80
75
T plate
None
yes
20
110
40
70
Blade plate
None
yes
18
140
40
80
T plate T plate T plate
Blade plate iliac-crest
Hardware removal
Final external rotation (~
163
undergone three operations, two had had two operations, and four had had a single operation (Table I). The period belween fracture and our operative treatment averaged 12 months (range 6 to 72 months). At the time of treatment of the nonunion, the patients reported loss of function with inability to raise the arm more than 60 ~ actively; active external and internal rotation was not possible. The shoulder was not painful if the arm was not being used but was painful in the activities of daily living. The passive range of motion was preserved, but motion stemmed from both the pseudarthrosis and the joint. Four patients had complications after the accident or previous surgery. One man (case 19) was the victim of a shooting accident; he had received a gunshot to the upper humerus. Emergency surgery had been performed debriding wounds and inserting Rush pins. The anterior and middle por-
telephone 61 100 72 70 80 95 100t 95 80
tions of the deltoid were severely injured. One woman (case 17), who had undergone operation twice, had partial paralysis of the deltoid; this condition was of iatrogenic origin or occurred after trauma at the time of the accident. She also had a ruptured pectoralis major after her second operation; this had resulted from an extended approach. One woman (case 18), who had undergone four operations with insertion of a nail, a plate, and two external fixators according to the technique of Ilizarov, had total paralysis of the deltoid, probably resulting from passage of the wires. One woman (case 8), who had undergone emergency surgery, had an infection that had led to early additional surgery to remove the osteosynthetic material and for lavage and drainage. Infection was irradicated, leaving the patient with necrosis of the humeral head and pseudarthrosis of the surgical neck. At the time we performed our
164
J. Shoulder Elbow Surg. May/June 1996
Walch et al.
Figure | Techniquefor osteosynthesisassociated with intramedullary bone peg.
operation 12 months later, all pathologic and microbiologic test results were normal. Operative ter A deltopectoral approach was used, preserving the upper and lower deltoid insertions. The muscle was separated from the upper humerus, and the proximal and distal fragments of the pseudarthrosis were freshened, with excision of the interposed fibrous tissues. A 6 to 10 cm bone peg was taken 11 times from the iliac crest on the same side, six times from the upper third of the tibial crest, and three times from the middle-third of the fibula on the same side. Additional curettage of adjacent cancellous bone was carried out with a curet. The corticocancellous graft was first impacted into the distal humerus and was then plugged into the head like a peg-and-ball (Figure 1). Impaction was easily obtained by gently tapping the bended elbow. Our objective was to obtain as much surface contact as possible. Rotatory problems were avoided by using the long head of the biceps as a guide in patients who had had no previous surgery, but this was particularly difficult in patients who had undergone surgery. We endeavoured to keep 20 ~ of external rotation with the elbow near the side. An external T plate was then applied onto the outer humerus and fixed by 3.5 or 4.5 mm cortical screws and 4.0 or 6.5 mm cancellous screws at the level of the humeral
epiphysis. The screws generally bored into the bony corticocancellous peg graft. Complementary cancellous graft was placed around the fracture site. Surprisingly, glenohumeral motion was always relatively well preserved. We have never performed arthrolysis, and the rotator cuff has never been explored or repaired; our aim was to obtain bone union. The wound was drained for 48 hours. The arm was immobilized by a sling for 30 days. Pendulum motions, gentle mobilization with the patient lying on his back; and self-rehabilitation in water were started after 30 days. Active motion of the shoulder was not allowed before 3 months. Rehabilitation in water by rotational motions was carried on for as long as possible. Anterioposterior radiographs with internal, neutral, and external rotation were done every 45 days until solid formation of bony callus occurred. Final evaluation was done by personal examination and included clinical examination, radiographic evaluation (three rotations), range of active motion, and functional results. Points were scored according to the scale of Constant, 3 which assigns 15 points to pain, 20 points to daily activities, 40 points to range of motion, and 25 points to strength. The final result was scored on a scale of 100 points, adjusted according to age and sex. Results were excellent when the shoulder reached at least 90% of the normal score in relation to age and sex, good when the score was between 75% and 89%, fair when it was between 74% and 50%, and poor when it was below 50%. The subjective results were determined by the question, "Are you very satisfied, satisfied, disappointed, or dissatisfied?"
RESULTS Fifteen patients were reexamined. Three patients were unable to return, and they were interviewed over the phone and sent recent x-ray films; one had had his shoulder examined once by his attending physician. Two had died, but because they had undergone radiologic reexamination at 40 months and 56 months, we included their outcomes. The average follow-up was 42 months (range 12 to 120 months). Nineteen patients (96%) had radiologic union (Figures 2, 3, and 4); mean time to union was 4 months (range 3 to 10 months). An 84-year-old woman had good bone continuity because of the intramedullary plug, although the peripheral graft had failed, suggesting that bone union was not complete. No aseptic necrosis of the
J. Shoulder Elbow Surg. Volume 5, Number 3
Walch et al.
165
Figure 2 Case 5, 60year old woman who underwent surgery for pseudarthro sis 6 months after fracture (A). Union was successful(B). Five years later Constant score was 100%
Figure 3 Case 4, atrophic, unstable pseudarthrosiswith plate fixation after k,vo operations (A). Union was successful (B). Six years later Constant score was 875%
humeral head was observed, except for case 8, who had preoperative necrosis. Pain relief was achieved for all our patients, with total relief in 54% and persistent but less marked pain in 46%. None of our patients had aggravation of pain, and patients who had had no pain remained without pain. The mean subjective evalu-
ation of pain with Constant's scale was 11 out of 15. Range of motion of active anterior elevation of the shoulder had improved from an average of 60 ~ to an average of 131 ~ (60 ~ to 180~ The mean active external rotation with the elbow at the side was 45 ~ (0 ~ to 60~ With the scale of Con-
166
Walch et al.
J. ShoulderElbow Surg. May/June 1996
Figure 4 Case 20, atrophic, unstable pseudarthrosis after failed internal fixation (A). Union was successful (B). Eighteen months later Constant score was 80%.
stant the mean objective evaluation of motion was 30 out of 40. The overall score of Constant, adjusted according to age and sex, was 81.2% (61 to 100). Results obtained were excellent for six patients, good for six, and fair for four; no negative results occurred. The score of Constant was not recorded for the last four patients, because two had died and two had been interviewed over the phone with no clinical examination. Subjectively, 65% of our patients were very satisfied, 30% were satisfied, and 5% were disappointed. Three patients did not recover active anterior elevation motion of more than 90 ~ A 78-year-old woman (case 7) had a traumatic three-part fracture of the upper humerus. She used walking aids to counterbalance the injuries to her left leg resulting from the same accident. No rehabilitation was possible; active elevation was 60 ~. Moreover, medicolegal insurance problems complicated the situation. The subjective result was disappointing. The second case (15) concerned an 83-year-old woman who also had malunion of the upper epiphysis and associated greater tuberosity fracture. The intramedullary graft was secure, but peripheral callus failed to develop because of resorption of the peripheral cancellous graft. We did not wish to continue physiotherapy to avoid refracture. The patient is very satisfied, although her active elevation is only 90 ~ The third case (t 1) con-
cerned a 61-year-old woman who had nonunion after a traumatic three-part fracture. Rehabilitation was discontinued because of poor motivation. Three factors were responsible for the poor resuits obtained in restoring active motion. The first was the type of initial fracture. Restoration of the active elevation was 117 ~ after a three-part fracture of the upper humerus and 157 ~ for fractures of the surgical neck that preserved the tuberosities. The second factor was the age of the patient, which also had an influence on rehabilitation. Before the age of 65 years, active motion averaged 145~ after 65 years it fell to 121 o. The third factor was poor condition of the deltoid (cases 17, 18, and 19), although all three patients managed to recover an active motion of more than 90 ~. On the other hand, the presence and number of previous operations had no significant incidence on restoration of function, if the deltoid was not impaired. Complications. The bone peg was obtained six times from the anterior tibial crest. A pathologic fracture was observed three times at the lower end of the graft site at 2 months after operation. These fractures required a walking cast for 2 to 3 months, after which union was complete. Rupture of the pectoralis major (case 17) was not repaired at the same time as the treatment of the pseudarthrosis. Suture of the tendon into a
J. Shoulder Elbow Surg. Volume 5, Number 3
bony trough was performed 12 months later during removal of the internal fixation. Restoration of motion was full; the cosmetic result was good. A humeral fracture below the T-plate (case 16) was observed 30 days after surgery, after a fall occurred. The fracture showed little displacement, and union occurred after a further 45-day immobilization period. DISCUSSION It is established that successful bony union is easily obtained for fracture of the surgical neck, whether simple or complex, when suitable orthopaedic treatment is applied. Baux and Razemon 1 have reported only three cases of pseudarthrosis in a multicenter series of 1237 fractures of the upper humerus after conservative treatment. In 1964 Sorensen TM reported two cases of pseudarthrosis; he found only seven cases in his literature review. In 1978 Lerat and Trillat 6 reported three cases of pseudarthrosis in a series of 500 fractures of the surgical neck. In 1983 Neer 9 reported 42 cases of pseudarthrosis and found 29 cases described in the literature. In 1990 Healy et al. 4 reported 14 patients in a multicenter study. Our series included nine cases previously treated conservatively. It is not easy to establish whether the initial treatment invites pseudarthrosis, because many different approaches have been described: sling, spica, hanging cast, traction, or thoraco-brachial cast. Moreover, comparative series of fractures to which various treatments were applied revealed no fewer or no more pseudarthroses depending on the treatment method selected. Many factors are probably responsible for these pseudarthroses: absence of prolonged immobilization, age of the patient, poor quality of the bone, displacement of the inferior fragment produced by the pectoralis major, or interposition of the biceps or deltoid muscle. Similarly, pseudarthrosis occurring after immediate osteosynthesis is rare. Lerat and Trillat 6 report one case, Neer 9 eight cases, and Healy et al. 4 14 cases; we found 12 cases in our series. A rigorous analysis is not possible because of the multiple types of osteosyntheses used in emergencies (Kwires, screws, rods, or plates). Apart from the factors described previously, inadequate firmness of the fixation was one of the main causes of pseudarthrosis. Pseudarthrosis is unexpectedly difficult to manage; in our series 14 operations failed, four of which we had carried out personally. In 1983
Walch etal.
167
Neer 9 reported 35 failed attempts at surgical repair out of a series of 50 patients. Healy et al. 4 reported 12 unions in a multicenter study of 18 cases of pseudarthrosis treated by osteosynthesis plus bone graft. Original suggestions have been proposed to obviate such failures. In 1965 Leach and Premer5 carried out an osteosynthesis on three patients using an angulated blade plate and a nail placed successively through the head, the plate, and the medullary shaft. The technique they used led to union but did not significantly improve motion despite a short immobilization of 3 weeks. In 1982 Scheck ~3 offered five of his patients treatment of their pseudarthrosis by implanting the humeral shaft into the head. He placed a variety of osteosynthesis materials and an extramedullary, intracephalic bone plug to act as a stabilizer. This allowed 80% union, a constant lowering of pain, and an important increase in motion. However, this technique leads to loss of length, weakening the deltoid, rendering it too long, and causing inferior subluxation of the humeral head. In 1983 Neer 9 proposed a new technique with associated osteosynthesis by an intramedullary nail and tension band. Immobilization lasted 3 months, and the operation entailed additional surgery for arthrolysis and removal of hardware. He obtained 12 unions out of 13 cases, but he reported no precise results of function and motion. With a similar technique (figure-of-eight wires and rods), Bigliani et al. 2 have reported 10 cases (21 operations), with four persistent pseudarthrosis. According to Bigliani et al., "Open reduction and internal fixation for nonunions following primary fracture fixation attempts led to poor results in all cases." The results obtained in our series were more promising, because we obtained 96% union. The intramedullary bone peg provided solid bone that filled the cavity of the head, thus preventing motion between the shaft and the head. It also allowed solid osteosynthesis by transfixing the intramedullary support with screws and it increased the bone present by adding cancellous graft. The technique preserves the rotator cuff and does not aggravate joint stiffness, because no arthrotomy is required. Immobilization is limited to 30 days, avoiding further stiffening of the shoulder that often follows a long period of immobilization. Because of complications that occurred with bone fragments harvested from the anterior tibial crest, the site is not recommended. When relatively straight 6 to 8 cm corticocancellous bone grafts
168
Walch etal.
can be obtained from the iliac crest, this site is recommended. When the graft is too narrow, we advocate obtaining a bone graft from the middle third of the fibula. We have been offering this technique to our patients since 1983 whatever the state of their humerus, and even to patient 8, who had humeral head necrosis after sepsis. We consider the primary aim is to obtain bone union, and surprisingly we have observed no postoperative necrosis, which encourages us to propose the technique whatever the radiologic condition of the humeral head is. Motion is less improved in three-part fractures, but the results are better than those we obtained with humeral head prosthetic replacements in fracture malunions. Neer 1~ advocates total or humeral head replacement in cases of long-standing pseudarthrosis when there is extensive osteopenia or necrosis of the humeral head; his results were not reported. Bigliani et al. 2 reported 70% satisfactory results for 10 cases treated by humeral head replacement. In conclusion, although rare, nonunion after fracture of the surgical neck of the humerus remains a challenge for the orthopaedic surgeon. Surgical treatment with the intramedullary bone peg technique seems to afford good results with a 96% rate of bone union. REFERENCES 1. Baux S, Razemon jP. Les fractures et les fractures-luxafions de I'extr6mit6 sup6rieure de I'hum6rus. Rev Chir Orthop 1969; 55:387-496.
J. Shoulder Elbow Surg. May/June 1996 2. Bigliani LU, Nicholson GP, Pollock RG, Duralde XA, Self EB. Personal communication. Ninth Open Meeting of the American Shoulder and Elbow Surgeons. San Francisco, February 21, 1993. 3. Constant CR, Murley AHG. A clinical method of functional assessment of the shoulder. Clin Orthop 1987;214:160 4. 4. Healy WL, JupiterJB, Kristiansen TK, White RR. Nonunion of the proximal humerus. In: Post M, Morrey BF, Hawkins RJ, editors. Surgery of the shoulder. Chicago: Mosby-Year Book Inc., 1990:59 62. 5. Leach RE, Premer RF. Nonunion of the surgical neck of the humerus. Minn Med 1965;48:318-22. 6. Lerat JL, Trillat A. La pseudoarthrose du col hum@ral existe-telle? Lyon Chir 1978;74:105-9. 7. McCreath SW. Delayed union and nonunion in fractures of humeral shaft. J Bone Joint Surg 1955;57B:393. 8. Neer CS II. Displaced proximal humeral fractures. Part I: Classification and evaluation. J Bone Joint Surg Am 1970; 52A:1077 89. 9. Neer CS ik Nonunion of the surgical neck of lhe humerus. Orthop Trans 1983;7:389. 10. Neer CS II. Shoulder reconstruction. Philadelphia: WB Saunders Co. 1990:230-4. 11. Norris TR, TurnerJA. Surgical treatment of nonunions of the upper humerus shaft fracture in the elderly. Orthop Trans 1985;9:44. 12. Norris TR, Turner JA, Bovill D. Nonunion of the upper humerus: an analysis of the etiology and treatment in 28 cases. In: PostM, Morrey BF, Hawkins RJ, editors. Surgery of the shoulder. Chicago: Mosby Year Book Inc, 1990:63-7. 13. Scheck M Surgical treatment of nonunions of the surgical neck of the humerus. Chn Orthop 1982;167:255-9. 14. Sorensen KH. Pseudarthrosis of the surgical neck of the humerus. Acta Orthop Scand 1964;34:132-8.