Tendon continuity cannot be restored by means of medical therapy; in fact, the occurrence of tendon rupture may indicate that the current regimen of medical management is inadequate and that additional remittive therapy should be considered. Adequate medical control of rheumatoid disease is imperative to minimize synovial proliferation, which may otherwise jeopardize the results of surgical reconstruction, place other anatomic areas at risk, or both.
Pharmacologic preparation for surgery
Most authors recommend a single intravenous (IV) dose of an antibiotic (usually a cephalosporin) immediately before surgery, and some follow this with 24 hours of postoperative therapy as a prophylactic measure. The postoperative antibiotic can be administered orally on an outpatient basis. This author is not aware of a controlled study that demonstrates the advisability of any particular perioperative antibiotic regimen.
Patients treated with corticosteroids require increased doses in the perioperative period to protect against the possibility of addisonian crisis brought on by the stress of surgery. 
Adequate pain control must be ensured; occasionally, this necessitates postoperative inpatient hospital admission.
Prevention of tendon rupture
Prophylactic procedures intended to prevent tendon rupture are generally quite effective and provide function that is superior to the function provided by any method of tendon repair or reconstruction. Patients with risk factors such as dorsal prominence of the distal ulna, radiographic erosion involving the distal radioulnar joint (DRUJ), or persistent dorsal tenosynovitis that is unresponsive to medical management over a 6-month period should be considered candidates for such surgery. 
Weakness or extensor lag of the small-finger metacarpophalangeal (MCP) joint may indicate that one of the tendons that motor this joint has failed; this is usually caused by rupture of the extensor digiti minimi (EDM), with residual function resulting from the intact extensor digitorum communis (EDC) tendon. This warning sign or pain over the dorsum of the wrist associated with motion of the digital extensor tendons should also prompt early exploration. Tendon rupture after prophylactic surgery, which includes tenosynovectomy and DRUJ reconstruction, is rare (0-4%). 
Hsueh et al, in a retrospective review of 17 episodes of spontaneous tendon rupture (15 RA patients) and 14 episodes of tenosynovitis (12 RA patients), suggested that prophylactic extended tenosynovectomy to prevent more severe extensor tendon damage should be recommended in those with the following risk factors  :
Rheumatoid disease of more than 8 years' duration
Persistent tenosynovitis of more than 1 year's duration
Larsen grade higher than 4
Tendon continuity restoration
Surgical efforts to restore tendon continuity must always include thorough dorsal tenosynovectomy with retinaculum transposition and resection or reconstruction of the DRUJ sufficient to remove the bony prominences that produced the tendon rupture.  As noted previously, the choice of DRUJ reconstruction may vary, depending on the status of the wrist joint and other individual patient considerations.  After the loss of small-finger extension, surgical exploration should be undertaken without delay to avoid sequential rupture of the adjacent tendons, which would further complicate reconstruction. (See the image below.)
Direct repair of ruptured tendons is rarely possible, both because of the attritional nature of this process and because the zone of tendon injury is usually quite long. 
The use of bridge grafts to restore tendon continuity has been described with generally acceptable results. Bora et al reported 23 patients who recovered an average of 65° of MCP joint motion after free tendon grafting (average follow-up, 43 months).  However, this technique has not gained widespread acceptance, because tendon graft harvest may require additional operating time and surgical exposure, sometimes at a remote site.
Many authors have also expressed concern that tendon grafting results in two tenorrhaphy sites in what is often a less than optimal tendon bed, possibly increasing the risks of scar formation and motion-limiting adhesions.  In a study by Nakamura and Katsuki, grafting of multiple extensor tendon ruptures in 14 patients with rheumatoid arthritis (RA) resulted in good correction of extensor lag, but patients were dissatisfied with the accompanying loss of digital flexion. These authors postulated that this limitation of flexion is caused by contracture of the involved muscle and, thus, did not recommend tendon grafting. 
Chung et al reported that MCP joint extension lag correlated with patient satisfaction scores, whereas pulp-to-palm distance in flexion did not. Their study of 51 wrists with extensor tendon ruptures found no significant differences in outcome between tendon grafting and tendon transfers. 
Tendon transfer remains the most common method of restoring tendon continuity in the rheumatoid hand after extensor tendon rupture. However, it is associated with its own particular set of challenges, including the possibilities that the joints to be motored may be stiff or unstable, that the tendon beds may be less than ideal, that the motor tendons may have disease involvement or may be unavailable, and that tenodesis may be limited by wrist stiffness.
Despite these potential shortcomings, tendon transfer is often the best surgical alternative. The method of tendon transfer is determined primarily by the number of digits involved. 
Small-finger MCP joint extension may be lost after isolated rupture of the EDM in a patient without significant EDC tendon contribution to this digit, or it may follow rupture of both the EDM and a slip of the EDC. The distal tendon stump that produces strong MCP joint extension of the small finger (usually the EDM) may be transferred to the adjacent EDC of the ring finger. Occasionally, the distal stump is so short that such side-to-side suture tends to produce unacceptable abduction of the small finger. In these instances, the extensor indicis proprius (EIP) may be transferred to the EDM. 
Loss of ring- and small-finger MCP joint extension is usually best treated by using the EIP to motor both the EDM and the ring-finger EDC. Alternatively, the EIP may be transferred to the EDM, and the ring-finger EDC may be sutured in a side-to-side fashion to the intact long-finger EDC. 
Loss of long-, ring-, and small-finger MCP joint extension is usually treated by transfer of the EIP to both the EDM and the EDC of the ring finger and side-to-side suture of the long-finger EDC to the intact index-finger EDC. If the EIP is the only remaining extensor to the index finger, a flexor digitorum sublimis (FDS) tendon may be used to motor ring- and small-finger MCP joint extension in place of the EIP. Although one of the radial wrist extensors (preferably the extensor carpi radialis longus) can be considered as a motor tendon in this circumstance, it often is not long enough to reach the distal stumps of the ruptured tendons and lacks sufficient excursion to restore normal MCP joint motion. 
Loss of MCP joint extension of the index, long, ring, and small fingers constitutes the final stage of this process and should only be seen in neglected cases. Although one hesitates to remove power and joint stability from the volar surface of the rheumatoid hand, transfer of the FDS tendons to provide MCP joint extension, as originally described by Boyes in the treatment of radial nerve palsy,  is often the best alternative. Many patients feel that flexor function is improved as a result of decompressing the digital flexor sheaths by tendon harvest.
Most authors use the ring-finger FDS to motor the ring and small fingers and the long-finger FDS to motor the index and long fingers. The classic Boyes transfer routed these tendons through a window in the interosseous membrane. However, adhesions may be limited in the rheumatoid hand by using the alternative of a subcutaneous tunnel around the forearm.  The radial route is preferred because it tends to counteract the tendency for ulnar subluxation of the digital extensors on the dorsum of the MCP joints.  Other motor tendons may, of course, be used, depending on the specific clinical circumstance.
Inpatient hospital care is seldom required after surgical treatment of ruptured extensor tendons unless these procedures are combined with other major joint reconstruction procedures.
Surgery on a single upper extremity may significantly affect the functional capabilities of patients with RA. These patients frequently have limited function of the contralateral hand, as well as a need for ambulatory aids or other assistive devices that they may be incapable of using in the postoperative period. The need for increased support and home health assistance should be anticipated.
Appropriate hand therapy and splinting are crucial to the success of these procedures.  In the absence of supervised therapy, limitation of motion from scarring and adhesion formation may exceed the limitation that was present preoperatively. Joints adjacent to those that are being surgically treated are seldom entirely normal in the rheumatoid hand, and early motion is necessary to prevent stiffness and functional deterioration.
Although many authors advocate immobilization of the MCP joints in a position near full extension for 3-4 weeks after extensor tendon reconstruction, this author favors a program of graduated dynamic extension splinting in an attempt to maximize motion, particularly to avoid the loss of digital flexion. Therapy is usually initiated on postoperative day 3 or 4, after edema has declined and the wounds have settled. The overall pattern of patient disease and ongoing medical treatment may dictate accommodations in the usual course of postoperative therapy. 
Wound healing problems and infection are encountered in fewer than 5% of cases. 
Extension lag at the metacarpophalangeal (MCP) joint is a frequently mentioned occurrence,  though its incidence and magnitude are difficult to quantify. Most authors would agree, however, that extension contracture is very uncommon, and it generally is better to err on the side of increased tension when setting tendon transfers in the rheumatoid hand. 
Rerupture is distinctly uncommon, occurring in fewer than 5% of cases. This complication is usually related to inadequate reconstruction of the distal ulna, which allows recurrence of dorsal displacement and attritional tendon wear.