Nonunion and Reoperation Following Proximal Interphalangeal Joint Arthrodesis and Associated Patient Factors

Abstract

Background

Proximal interphalangeal joint (PIPJ) arthrodesis can provide reliable pain relief and restore hand function in patients with PIPJ arthritis. However, there is a paucity of literature on patient-specific preoperative risk factors that are associated with adverse outcomes after PIPJ arthrodeses. Therefore, the primary purpose of this study was to assess preoperative predictors of nonunion and reoperation after PIPJ arthrodesis.

Methods

This study identified all patients who underwent PIPJ arthrodesis at a single community practice between 1987 and 2013. The final analysis included 415 PIPJs treated with arthrodesis. The mean follow-up was 1.3 years. Data on preoperative diagnosis, demographics, comorbidities, and operative techniques were recorded, as well as the occurrence of nonunions and reoperations. Logistic regression models were used to identify independent risk factors of nonunion and reoperation.

Results

There were 40 nonunions (10%) and 62 reoperations (15%). Of the reoperations, there were 39 incidences of isolated hardware removal, 9 irrigation and debridement, 8 amputations, 5 revision arthrodeses, and 1 corrective osteotomy. The highest number of nonunions occurred in the traumatic diagnosis group (37%), followed by the rheumatoid group (23%) and the scleroderma group (15%). The highest number of reoperations occurred within the traumatic joint disorder group (40%), followed by the rheumatoid group (24%) and the scleroderma group (11%). Multivariate analysis revealed that male sex (P < .01) and hepatic disease (P = .03) were significant risk factors of nonunion. Male sex was also significantly associated with increased reoperation risk (P < .01).

Conclusion

Risks of nonunions and reoperations after PIPJ arthrodeses are low; however, these findings may guide clinicians and patients in the preoperative decision-making process and help with targeted postoperative surveillance to mitigate these risks.

Introduction

Disorders of the small joints of the hand and wrist have been associated with significant decline in patient’s quality of life. ¹ Arthritis of the proximal interphalangeal joint (PIPJ) in particular limits functionality of the hand and is second only to basilar thumb arthritis when it comes to incidence of hand pain. ² Treatment of PIPJ arthritis ranges from nonoperative, symptomatic management to arthroplasty or arthrodesis.^3,4 Although limiting in terms of range of motion, PIPJ arthrodesis can provide reliable pain relief and restore hand function. ⁵ However, small joint arthrodesis is not without complications and can be numerous, including: dorsal skin necrosis, infection, painful hardware, parenthesis, cold intolerance, and nonunion. ⁶ Previous studies have investigated rates of reoperation by comparing different surgical techniques. The rates of reoperation have only been reported in a small series of patients.^7-11 To our knowledge, no study has investigated possible patient-specific preoperative risk factors that could predict adverse outcomes with nonunion and reoperation. Therefore, the primary purpose of this study was to assess preoperative predictors of nonunion and reoperation after PIPJ arthrodesis.

Materials and Methods

Study Design

Following full institutional review board approval, all patients who underwent index, middle, ring, and small finger PIPJ arthrodesis at a single community hand surgery practice between 1987 and 2013 were identified for review. All surgeries were performed by 8 fellowship-trained hand surgeons. A total of 376 patients undergoing 506 PIPJ arthrodeses were identified. Patients with a follow-up of less than 6 weeks were excluded, resulting in a final investigation of 288 patients with 415 PIPJs treated, and the mean follow-up was 1.3 years.

Digits treated were grouped into 11 main principal preoperative diagnoses. The most common diagnosis was rheumatoid arthritis (RA) (34%), followed by traumatic causes (22%) and soft tissue contractures (20%). Digits from men accounted for 82% (75 of 92 digits) of PIPJs treated from the traumatic arthrodesis group. Men from the traumatic group accounted for 57% (75 of 132) of all male PIPJs treated. Regarding surgical technique, 382 PIPJs (92%) were operated using the Kirschner wire (K-wire) tension band technique. Other methods of fixation included screw (7), intramedullary rod (4), isolated wire fixation only (8), and isolated pin fixation (4).

Preoperative Variables

The primary preoperative diagnosis assigned to each digit was identified using the documented International Classification of Diseases (ICD) codes. Codes with fracture, dislocation, amputation, and soft tissue trauma were consolidated into 1 category, “traumatic.” Codes for contractures due to nerve or brain injury, Dupuytren disease, and ischemia were consolidated into 1 category, “soft tissue contracture.” All other diagnosis groups included ICD codes specific for their diagnosis. Additional preoperative information was recorded, including demographic information and documented baseline medical comorbidities. Operative technique was collected from operative reports.

Postoperative Complications

Postoperative complications including nonunion of fusions and reoperations were collected. Nonunion was defined by the clinical and/or radiographic determination of each treating surgeon at the time of evaluation. Reoperation was defined as any surgical procedure performed on the digit after the initial surgery. Operations were categorized into 2 groups, either isolated removal of hardware (ROH), or major operation (amputation, revision arthrodesis, irrigation and debridement [I&D], and osteotomy).

Data Analysis

Following data collection, bivariate analysis with Pearson coefficient and Spearman ρ was used to identify associations between principal preoperative diagnoses and incidences of nonunion or reoperation. Multivariate analysis was performed with separate logistic regression models using reoperation and nonunion as binary, dependent variables. Independent variables in the models included: sex, age, smoking, endocrine disease, cardiovascular disease, renal disease, liver disease, and the principal operative diagnosis. The threshold for statistical significance was a value of P < .05.

Results

Of the 415 digits treated, 68% of the digits belonged to female patients. The average patient age was 53 years. The most common joint fused was the small finger PIPJ (33%), followed by the ring finger PIPJ (25%), middle finger PIPJ (25%), and index finger PIPJ (17%) (Table 1). The most common preoperative diagnosis among women was RA and among men was trauma (Table 2). The most common class of medical comorbidity was endocrine (17%). Approximately 26% of all digits originated from smokers. The average follow-up of all patients was 1.3 years (range, 6 weeks to 16 years).

Table 1.

Patient Demographics of All Digits Included for Analysis.

Variable	Mean	No. of digits (%)
Sex
Female		283 (68)
Male		132 (32)
Age, y	53 (range, 17-87)
Follow-up	1.3 y (range, 6 wk to 16 y)
Smoking history
Smokers		106 (26)
Nonsmokers		309 (74)
Medical comorbidity
Cardiac		52 (13)
Renal		13 (3)
Endocrine		72 (17)
Hepatic		8 (2)
Finger
Index		71 (18)
Middle		105 (25)
Ring		105 (25)
Small		134 (32)

Table 2.

Preoperative Diagnoses by Sex.

Principal preoperative diagnosis	Women, %	Men, %
Rheumatoid	36	7
Traumatic	9	50
Soft tissue contracture	16	15
Osteoarthritis	28	9
Scleroderma	4	1
Psoriatic arthritis	2	1
PIPJ arthroplasty failure	2	0
Gout	0	1

Note. PIPJ = proximal interphalangeal joint.

Overall, there were 40 nonunions (10%) and 62 reoperations (15%). There were 39 incidences of isolated hardware removal, 9 I&D, 8 amputations, 5 revision arthrodeses, and 1 corrective osteotomy (Table 3). Of the 62 total reoperations, 21 (32%) came from nonunion digits. These 21 reoperations comprised 14 major reoperations and 7 isolated ROH. Three scleroderma nonunions, 2 traumatic nonunions, and 1 gouty digit nonunion underwent I&D. Revisions were performed in 2 scleroderma nonunions, 1 psoriatic nonunion (needed 2 revision surgeries), and 1 traumatic nonunion. Two amputations and 1 corrective osteotomy were completed in 3 traumatic nonunions. Twenty-one percent of reoperations were performed on middle fingers, 23% on index fingers, 26% on small fingers, and 31% on ring fingers.

Table 3.

Nonunions and Reoperations in Total Study Population.

Operative technique	Total, No. (%) (N = 415)
Nonunions	40 (10)
Reoperations for nonunions	21 (5)
All reoperations	62 (15)
Isolated hardware removal	39
Irrigation and debridement	9
Amputations	8
Revision arthrodesis	5
Corrective osteotomy	1

Preoperative Diagnoses

The highest number of nonunions occurred in the traumatic diagnosis group (37%), followed by the rheumatoid group (23%) and the scleroderma group (15%). The rate of nonunions per diagnosis was highest in the scleroderma group (33% or 6 of 18 digits), the failed PIPJ arthroplasty group (33% or 2 of 6 digits), and the gouty digit group (33% or 1 of 3 digits). The lowest nonunion rate per diagnosis was seen in osteoarthritis (OA) (7% or 4 of 58 digits), RA (6% or 9 of 143), and soft tissue contractures (1% or 1 of 79 digits) (Table 4).

Table 4.

Nonunion Rate by Principal Preoperative Diagnosis.

Principal preoperative diagnosis	Total nonunions, No. (%) (N = 40)	OR (95% CI)	P value
Rheumatoid	9 (23)	0.1 (0.01-1.3)	.07
Traumatic	15 (37)	0.3 (0.03-3.7)	.4
Soft tissue contracture	1 (2)	0.04 (0.002-0.66)	.02
Osteoarthritis	4 (10)	0.1 (0.01-1.7)	.1
Scleroderma	6 (15)	0.8 (0.1-9.9)	.8
Psoriatic arthritis	2 (5)	0.9 (0.5-13)	.9
PIPJ arthroplasty failure	2 (5)	0.4 (0.02-10)	.6
Gout	1 (2)	0.8 (0.8-9)	.5

Note. OR = odd ratio; CI = confidence interval; PIPJ = proximal interphalangeal joint. Statistically significant P values (<.05) are in bold.

The highest number of reoperations occurred within the traumatic joint disorder group (40%), followed by the rheumatoid group (24%) and the scleroderma group (11%). The highest rate of reoperation within a diagnosis group was 39% in the scleroderma group (7 reoperations per 18 digits), 38% in the psoriatic group (3 of 5), and 33% in the gouty digit group (1 of 3). Excluding those who did not have repeat surgery, the lowest rate of reoperation was in the OA (9% or 5 of 58 digits) and soft tissue contracture groups (9% or 7 of 78 digits) (Table 5).

Table 5.

Reoperation Rate by Principal Preoperative Diagnosis.

Principal preoperative diagnosis	Total reoperations, No. (%) (N = 62)	OR (95% CI)	P value
Rheumatoid	15 (24)	0.36 (0.03-4.2)	.4
Traumatic	25 (40)	0.66 (0.05-7.6)	.7
Soft tissue contracture	7 (11)	0.32 (0.03-3.9)	.4
Osteoarthritis	5 (8)	0.18 (0.01-2.4)	.2
Scleroderma	7 (11)	2.3 (0.2-32.6)	.5
Psoriatic arthritis	3 (5)	1.0 (0.05-18)	.9
PIPJ arthroplasty failure	0 (0)	—	—
Gout	1 (2)	0.8 (0.8-9)	.8

Note. OR = odd ratio; CI = confidence interval; PIPJ = proximal interphalangeal joint.

Patient Factors

Multivariate analysis revealed that male sex (P < .01) and hepatic disease (P = .03) were significantly associated with nonunion. Men were 6 times more likely to develop nonunion (odds ratio [OR], 6.0; 95% confidence interval [CI], 2.3-15; P < .01). Patients with hepatic disease were 7 times more likely to develop nonunion (OR, 7.2; 95% CI, 1.2-41; P = .03) (Table 6).

Table 6.

Incidence and Multivariate Analysis for Nonunion.

Variable	Union, No. (%) a	Nonunion, No. (%) b	OR (95% CI)	P value
Sex
Male	111 (27)	21 (51)	6.0 (2.3-15)	<.01
Female	263 (73)	19 (49)	—	—
Smoker	91 (24)	15 (37)	0.95 (0.4-2.6)	.30
Medical comorbidity
Cardiac	48 (13)	4 (10)	0.5 (0.1-2.2)	.40
Renal	9 (2)	3 (7)	0.86 (0.3-3.2)	.30
Endocrine	66 (18)	6 (15)	1.6 (0.4-5.6)	.50
Hepatic	6 (1.6)	2 (5)	7.2 (1.2-41)	.03

Note. OR = odds ratio; CI = confidence interval. Statistically significant P values (<.05) are in bold.

^aPercentage is based on the number of digits without nonunion per risk factor over the total number of no reoperation (356 digits).

^bPercentage is based on the number of digits with nonunion per risk factor over the total number of no reoperation (41 digits).

For reoperations, male sex was also associated with a significantly increased risk. Specifically, men were 3 times more likely to require reoperation when compared with women (OR, 3.2; 95% CI, 1.5-6.9; P < .01). Although smoking status or comorbidities were not significantly associated with reoperation risk, there was a trend observed toward greater risk with active smoking status and the presence of medical comorbidities (Table 7).

Table 7.

Incidence and Multivariate Analysis for Reoperation.

Variable	No reoperation, No. (%) a	Reoperation, No. (%) b	OR (95% CI)	P value
Sex
Male	99 (28)	33 (52)	3.2 (1.5-6.9)	<.01
Female	253 (72)	30 (48)
Smoker	84 (23)	22 (34)	1.4 (0.7-2.9)	.30
Medical comorbidity
Cardiac	43 (10)	9 (14)	1.2 (0.4-3.5)	.70
Renal	11 (3)	2 (3)	6.6 (0.9-47)	.06
Endocrine	59 (17)	6 (10)	1.3 (0.3-3.5)	.80
Hepatic	6 (1)	2 (3)	2.2 (0.3-14)	.40

Note. OR = odds ratio; CI = confidence interval. Statistically significant P values (<.05) are in bold.

^aPercentage is based on the number of digits without reoperation per risk factor over the total number of no reoperation (353 digits).

^bPercentage is based on the number of digits with reoperation per risk factor over the total number of no reoperation (62 digits).

Discussion

Previous studies have described complications following finger PIPJ arthrodesis.^7-11 Of these, 3 have reviewed PIPJ arthrodesis in more than 200 digits.^6,12,13 Carroll and Hill ¹³ treated 230 digits with PIPJ arthrodesis with the “cup and cone” technique and reported a 4% nonunion rate. The highest nonunion rate was related to patients with cerebral palsy. No nonunions were related to traumatic or arthritic PIPJ disorders; however, reoperations were not assessed. Stern et al reviewed 290 digits with small joint arthrodesis using the tension band technique. In all, 213 joints treated were finger PIPJs. The nonunion rate was 3% with an overall 9% rate of ROH. Only 1 PIPJ arthrodesis was found to be malrotated. The most common indication for arthrodesis was in RA in that study (33%). However, nonunion and reoperation were not classified by the principal operative diagnosis. ⁶ Leibovic et al retrospectively reviewed 224 PIPJs with multiple modes of arthrodesis fixation. A total of 45 radiographic and/or clinical nonunions were identified with an overall nonunion rate of 20%. However, 8 of 14 digits with psoriatic arthritis went onto nonunion, with none of these patients requiring reoperation. The reoperation rate was found to be 11%. These reoperations were only reported for patients with nonunion. ¹⁴ Nonunion rates of 0% to 50% have been described in smaller series. ¹⁵

In comparison, our study included more than 400 PIP joints. The nonunion rate of 10% reported here is consistent with other large investigations including those discussed above (3%-20%). Moreover, male sex was found to be a predictor of increased risk of nonunion and reoperation. This is likely not a true predictor of risk as 57% of all male patients treated belonged to the traumatic diagnosis group. Trauma accounted for the highest number of nonunions and reoperations, suggesting that the higher incidence in men is likely related to the overall higher incidence of men having traumatic accidents leading to failed arthrodesis of the PIPJ.

Bivariate analysis did not reveal traumatic PIPJ disorders to be a significant risk factor for either nonunion or reoperation. Furthermore, patients with soft tissue contractures had a significant protective risk for nonunion with an OR of <0.1. Although not statistically significant, RA and OA correlated toward protective risk for nonunion, with an OR of 0.1. These patients generally have good bone stock and stable soft tissues, which allows them to have a higher likelihood of fusion.

Although no diagnosis group was found to be a statistically significant risk factor for nonunion or reoperation, scleroderma and psoriatic arthritis correlated toward higher risk of reoperation with ORs of >1 and up to 30% reoperation rates in each group. In particular, systemic sclerosis is a chronic disease of connective tissue and blood vessels affecting multiple body systems, including the skin and skeletal system. ¹⁵ The extremities, particularly the hands and feet, are frequently affected by the disease process. Digital ulceration is also an unfortunate consequence of the disease process, occurring in up to 30% of scleroderma patients per year, with a lifetime incidence of 50%. ¹⁶ Digital ulcers occur in the fingertips and over the extensor apparatus covering the small joints of the hand. ¹⁷ The decreased perfusion and resultant soft tissue compromise with digital ulceration may correlate to a higher reoperation risk, particularly with I&D and ROH. Our reoperation rate is higher than that reported in other studies. Melone et al ¹⁸ reviewed 72 scleroderma patients with 211 PIPJ arthrodesis and reported uncomplicated wound healing in the immediate postoperative period. However, Gilbart et al ¹⁹ reported exposed hardware in 3 of 7 patients treated with K-wire figure tension band PIPJ arthrodesis. Hardware prominence increased over 6 to 24 months due to gradual skin contraction over the implant. They recommend using a cross K-wire construct that is easily removed in the office.^19,20

This study has some limitations, primarily related to the retrospective nature of our review. Roughly 20% of patients had incomplete follow-up and were not included for analysis. These patients may have been treated elsewhere following their operation, and their outcomes are unknown. Recovery from PIPJ arthrodesis is not thought to be intensive, which may account for patients who have a satisfactory result to forgo follow-up. Furthermore, assessments of nonunion likely varied as patients were evaluated and treated by 8 different fellowship-trained hand surgeons. There may also be a diagnostic confirmation bias for nonunion toward patients with conditions like psoriatic arthritis and scleroderma as these conditions are known to negatively affect the soft tissue of the fingers and the hands. Evaluators may have a lower threshold to diagnose nonunion in these patients. Although this is a large series on this topic, this study may be underpowered to find statistical significance. Moreover, the majority (>90%) of PIPJ arthrodesis were performed with a K-wire tension band construct; therefore, these findings may not be generalizable to patients undergoing other means of fixation.

Despite these limitations, this is the largest study to date and demonstrates that patients with PIPJ arthrodesis have low rates of nonunion and reoperation. This reaffirms that PIPJ arthrodesis is a reliable procedure in patients with OA, RA, soft tissue contractures, and traumatic injuries to the PIPJ. The rate of nonunion and reoperation with these conditions, predominantly after K-wire tension band arthrodesis, is low and consistent with the existing literature. However, there may be a higher risk of reoperation in patients with conditions that compromise soft tissues such as scleroderma and psoriatic arthritis. Although we could not demonstrate a significant association, there was a trend suggesting increased risk of reoperation in these patients. Perhaps surgeons could consider easily removable fixation to avoid hardware prominence, especially in patients predisposed to compromised soft tissue. At the very least, patients with scleroderma and psoriatic arthritis should be educated about potential complications prior to proceeding with surgery.

In conclusion, risks of nonunion and reoperation rates after PIPJ arthrodeses are low; however, these findings may guide clinicians and patients in the preoperative decision-making process and help with targeted postoperative surveillance to mitigate these risks.