Management of the Failed Arthroplasty for Proximal Humerus Fracture

Djuro Petkovic, MD; David Kovacevic, MD; William N. Levine, MD; Charles M. Jobin, MD


J Am Acad Orthop Surg. 2019;27(2):39-49. 

In This Article

Authors' Preferred Techniques for Revision of Failed Arthroplasty for Fracture

When performing a revision of a failed arthroplasty for fracture, we typically perform a comprehensive diagnostic workup to correctly identify contributing failure modes. We first rule out PSI on all cases and routinely obtain an ESR, CRP, and peripheral CBC with differential. If either the ESR or the CRP is increased, or if concerning features such as a loose humeral stem even with normal labs are present, then we perform a fluoroscopic-guided arthrocentesis and send the fluid for gram stain and culture. Typically, there is not enough fluid aspirated for cell count. We routinely perform arthroscopic or open biopsy to obtain culture results before revision if any part of the workup is worrisome for deep infection. During revision surgery, we use frozen section and consider >5 PMNs per HPF in any sample to be suggestive of infection and subsequently implant an antibiotic-impregnated cement spacer.

We typically use a deltopectoral approach. We avoid an anterosuperior deltoid-splitting approach for revision surgery because this approach is nonextensile and the axillary nerve may be difficult to identify in cases with notable scar, putting it at risk for injury. When a hemiarthroplasty is revised to an anatomic shoulder arthroplasty, we typically perform a lesser tuberosity osteotomy if sufficient bone stock exists. In the setting of poor bone quality, we perform a subscapularis tenotomy. Lesser tuberosity osteotomy is preferable because it improves the glenoid exposure when the soft tissues are less compliant during revision. When revision to a RTSA is done, we perform a subscapularis peel technique.

If the humeral stem requires removal for infection, revision to reverse, or malposition, we prefer to attempt stem removal without humeral osteotomy. However, if the stem is well cemented and cannot be removed with proximal osteotomy preparation using osteotomes, stem extractors, and slap hammers, then a vertical humeral osteotomy is preferred.[19] Once the controlled osteotomy has been made and an osteotome is used to dislodge the intramedullary cement, we cerclage wire the proximal humerus to prevent humeral fracture or further propagation of the controlled osteotomy before attempted stem extraction.

If an infection is diagnosed or presumed, we prefer a prefabricated antibiotic spacer because the cement head is smooth and may limit glenoid erosion. If the spacer stem is not long enough, we will either fabricate our own cement spacer with gentamicin-impregnated cement and a large bore threaded Steinmann pin (Figure 4, B) or augment the prefabricated spacer with an extended stem using a chest tube and a threaded Steinmann pin to create a shape and length suitable for the proximal humerus. A second batch of cement is commonly required to fill the metaphyseal void and provide rotational and axial fixation to the cement spacer. We typically place the spacer in 30° to 40° retroversion and repair the anterior tissues and/or subscapularis to prevent anterior dislocation of the spacer. In cases of infection, this repair should be done with monofilament suture. If patients have satisfactory function and symptomatic relief and the patient wants to avoid surgery, it is appropriate to leave these spacers in permanently.

When revision to a RTSA is performed, we consider soft-tissue releases of the capsule to ensure adequate soft-tissue tensioning with RTSA and to prevent postoperative dislocation. To safely release the soft-tissue contracture that is common to failed hemiarthroplasty, we routinely identify the axillary nerve anterior to the inferior subscapularis. We attempt to maintain the teres minor and infraspinatus attachment to the greater tuberosity to help with postoperative external rotation. It is reasonable to resect the supraspinatus tendon to allow for proper prosthetic tensioning.

Baseplate implantation requires careful attention to glenoid bone stock and type of glenoid erosion. Superior glenoid erosion may require structural grafting or a superiorly eccentric baseplate to correct the superior tilt to the glenoid. If notable medialization of the glenoid joint line is present, a lateralized glenosphere may help restore offset, improve the deltoid wrap over the proximal humerus, and provide compression of the articulation. However, the soft tissues are often poorly compliant, and increasing lateral offset or lengthening may be difficult as a result of the excessive soft-tissue tightness. In cases with contained glenoid deficiency with central erosion, we prefer to use a long-post baseplate that has purchase in and may even penetrate the medial vault cortical bone. Cancellous impaction bone grafting can be performed within the contained deficiency. In situations with uncontained glenoid bone loss, we will place structural bone graft with autograft from the iliac crest, femoral head allograft, or consider metallic augmented baseplate designs.

Humeral stem revision to RTSA typically requires cementation because of proximal humeral calcar and metaphyseal bone loss. If notable proximal humeral bone loss exists, then we prefer long-stem fixation with structural autografting or metallic replacement and try to avoid alloprosthetic reconstruction because of concern for infection risk. We typically do not repair the subscapularis because it is frequently deficient or noncompliant. In the setting of intraoperative anterior instability, we assess for component mechanical impingement posteroinferiorly, because unreleased posteroinferior capsule/soft tissue prevents proper rotation and lengthening. We also assess for soft-tissue tension including conjoint tendon tightness, axillary nerve tension, axial shuck, and lateral dislocation force and may consider a retentive polyethylene liner despite the risk for increased notching. Other options can include a thicker or lateralized polyethylene, as well as a glenosphere with greater diameter.

Postoperatively, we keep RTSA patients immobilized in an abduction sling for 4 to 6 weeks followed by active assisted motion as tolerated.