Separating Fact From Fiction: A Nationwide Longitudinal Examination of Complex Regional Pain Syndrome Following Treatment of Dupuytren Contracture

Abstract

Background:

One of the most feared complications following treatment of Dupuytren contracture is complex regional pain syndrome (CRPS). This study aims to provide a national perspective on the incidence of CRPS following treatment of Dupuytren contracture and identify patient factors to target for risk reduction.

Methods:

Using the Truven MarketScan databases from 2007 to 2016, individuals aged ≥18 years who developed CRPS within 1 year of treatment of Dupuytren contracture were identified using the International Classification of Disease diagnosis code for CRPS. Predictor variables included: age, sex, employment status, region, type of procedure, and concurrent carpal tunnel surgery. Multivariable logistic regression was used to analyze outcomes.

Results:

In all, 48 327 patients received treatment for Dupuytren contracture, including collagenase injection (13.6%); percutaneous palmar fasciotomy (10.3%); open palmar fasciotomy (3.9%); palmar fasciectomy with 0 (10.8%), 1 (29.2%), or multiple (19.6%) digit releases; or a combination of these procedures (12.8%). One hundred forty-five patients (0.31%) were diagnosed with CRPS at a mean of 3.4 months (standard deviation, 2.3) following treatment. Significant predictors of CRPS included female sex (odds ratio [OR], 2.02; P < .001), Southern region (OR, 1.80; P = .022), long-term disability status (OR, 4.73; P = .035), palmar fasciectomy with release of 1 (OR, 5.91; P = .003) or >1 digit (OR, 13.32; P < .001), or multiple concurrent procedures for Dupuytren contracture (OR, 8.23; P = .001).

Conclusions:

Based on national commercial claims data, there is a lower incidence of CRPS following treatment of Dupuytren contracture than previously reported. Risk factors identified should help with preoperative counseling and assist clinicians in targeting risk reduction measures.

Introduction

Complex regional pain syndrome (CRPS) is a rare and poorly understood constellation of symptoms that can occur following trauma and surgery. ¹ Type 1 CRPS occurs in the absence of direct injury to a peripheral nerve; in type 2 CRPS, there is an identifiable nerve injury. ² A leading hypothesis of the cause of CRPS is a dysregulation between the peripheral and central nervous systems related to trauma and inflammation at the time of injury. Clinically, CRPS can be profoundly debilitating, and there is still limited evidence to guide effective treatment.^3,4

The majority of literature describes CRPS after distal radius fracture,^5,6 yet CRPS is known to occur after routine hand surgery, including palmar fasciectomy for Dupuytren disease. ⁷ Dupuytren disease is common, with a prevalence ranging from 0.6% to 31.6% in different Western countries. ⁸ Treatment for Dupuytren contracture varies and includes collagenase injection, percutaneous fasciotomy, and open surgical procedures. A challenge in treating more complex Dupuytren disease is that the digital nerves are adjacent to the diseased palmar fascia, presenting a risk for irritation or direct injury during surgical excision. The proximity of the nerves to trauma can increase the risk of CRPS. Indeed, single-institution data have demonstrated CRPS occurring after 2.4% to 18.5% of surgical treatment of Dupuytren contracture.^9,10 A literature review of 28 papers reporting surgical complications associated with fasciectomy for Dupuytren disease yielded a postoperative CRPS rate of 5.5%. ¹¹

We aimed to expand the literature on CRPS and Dupuytren disease by investigating the incidence of CRPS following treatment of Dupuytren disease leveraging longitudinal US commercial insurance data. We also sought to determine patient- and treatment-level variables that contribute to CRPS in postprocedural Dupuytren disease, with the ultimate objective of better helping clinicians inform their patients and identify those at risk of CRPS. Finally, another controversy in the treatment of Dupuytren disease is the risk of CRPS or a pain “flare” when performing palmar fasciectomy concurrently with carpal tunnel release (CTR). We planned to assess the risk of CRPS after combined management of Dupuytren contracture and carpal tunnel syndrome.

Materials and Methods

Database

The Truven MarketScan Outpatient Services and Medicare Supplemental Insurance Outpatient Services databases were used to perform a retrospective cohort study of patients who underwent treatment for Dupuytren contracture from 2007 to 2016. Both databases include claims records from commercial insurance providers; the Outpatient Services database contains 8.7 billion individual claim records, whereas the Medicare Supplemental Insurance Outpatient Services database contains an additional 1.7 billion claims from supplemental commercial plans for Medicare recipients. ¹² Institutional review board approval was not required for use of these data sets.

Variables

Both databases were queried for individuals who underwent treatment for Dupuytren contracture based on Current Procedural Terminology (CPT) code (see Supplemental Tables S1 and S2) for procedures to treat Dupuytren contracture: closed (percutaneous) palmar fasciotomy; open palmar fasciotomy; palmar fasciectomy with 0, 1, or multiple digit releases; collagenase injection; or capsulectomy or capsulotomy of metacarpophalangeal or interphalangeal joints for contracture. These data sets were then merged to yield a comprehensive, combined data set of all unique procedural encounters. Using International Classification of Disease, Ninth Revision (ICD-9) and Tenth Revision (ICD-10) diagnosis codes, individuals with a diagnosis of CRPS type 1 or type 2 of the upper limb were then identified. The study population was limited to individuals aged ≥18 years with a diagnosis of CRPS within 1 year after treatment of Dupuytren contracture. Exclusion criteria were individuals who had a diagnosis of CRPS prior to or at the time of treatment for Dupuytren contracture, or had less than 1-year follow-up.

Predictor variables extracted from the database included: age, sex, employment status, region, type of procedure, and concurrent open or endoscopic CTR based on CPT code (64721, 29848). Time to CRPS development was calculated based on the dates of treatment for Dupuytren contracture and diagnosis of CRPS. In instances of multiple encounters with a postoperative CRPS diagnosis, the first encounter was considered to be the date of onset. In cases of multiple treatment dates for Dupuytren contracture, the last encounter was used in the analysis.

Patients who underwent a postoperative stellate ganglion block were also identified via CPT code. Treatment of CRPS is one of the relatively few indications for stellate ganglion block. ¹³ Thus, these patients represented potentially uncoded cases of CRPS. As such, these individuals were added to the CRPS cohort for a second analysis.

Statistical Analysis

Diagnosis of CRPS, based on the diagnosis of CRPS alone and CRPS and/or stellate ganglion block, following treatment of Dupuytren contracture was modeled using multivariable logistic regression. χ² and analysis of variance tests were used to assess the relationship between CRPS diagnosis and specific combinations of procedures for Dupuytren contracture (eg, palmar only vs digit fasciectomy, capsulotomy/capsulectomy vs fascial procedure only). Values of P < .05 were considered statistically significant. All analyses were conducted using Stata/SE Version 14.2 (StataCorp LLC, College Station, Texas).

Results

A total of 48 327 individuals underwent treatment for Dupuytren contracture from 2007 to 2015 (Table 1). The average age was 61.9 years (standard deviation [SD], 10.5; range, 18-97 years), with most patients aged ≥55 years (79.0%). Most patients were men (72.1%) with unknown employment status (46.7%), followed by active full-time employment (22.4%). Geographic breakdown revealed a difference of 14.0% between the region with the largest (South, 33.1%) and smallest (Northeast, 19.1%) proportion of procedures performed for Dupuytren contracture.

Table 1.

Summary Statistics of Dupuytren Contracture Cohort, N = 48 327.

Characteristic	Mean (SD) or No. (%)	Range
Age, y	61.90 (10.54)	18-97
18-34	571 (1.18%)	—
35-44	1693 (3.50%)	—
45-54	7897 (16.34%)	—
55-64	21 225 (43.92%)	—
65+	16 941 (35.05%)	—
Sex
Male	34 833 (72.08%)	—
Female	13 494 (27.92%)	—
Region
Northeast	9242 (19.12%)	—
North Central	11 517 (23.83%)	—
South	16 004 (33.12%)	—
West	10 461 (21.65%)	—
Unknown	1103 (2.28%)	—
Employment status
Active full time	10 844 (22.44%)	—
Active part-time or seasonal	220 (0.46%)	—
Early retiree	3544 (7.33%)	—
Medicare eligible retiree	8447 (17.48%)	—
Retiree (status unknown)	1914 (3.96%)	—
COBRA continuee	233 (0.48%)	—
Long-term disability	120 (0.25%)	—
Surviving spouse/Dependent	450 (0.93%)	—
Other/Unknown	22 555 (46.67%)	—
Procedure
Collagenase injection	6575 (13.61%)	—
Palmar fasciotomy, closed	4969 (10.28%)	—
Palmar fasciotomy, open	1859 (3.85%)	—
Palmar fasciectomy	5205 (10.77%)	—
Palmar fasciectomy, 1 digit	14 088 (29.15%)	—
Palmar fasciectomy, >1 digit	9471 (19.60%)	—
Combination	6160 (12.75%)	—
Concurrent carpal tunnel release	2431 (5.03%)	—
CRPS after Dupuytren procedure	149 (0.31%)	—
CRPS type 1	134 (89.93)%	—
CRPS type 2	15 (10.07%)	—
Time to CRPS, mo	3.43 (2.33)	0.23-11.08
Stellate ganglion block	97 (0.20%)	—
Time from Dupuytren procedure, mo	9.54 (17.11)	0.72-106.85
CRPS diagnosis	72 (74.23%)	—
Time from CRPS diagnosis, mo	2.99 (11.68)	0-98.73

Note. CRPS = complex regional pain syndrome; COBRA = Consolidated Omnibus Budget Reconciliation Act.

Procedures performed for Dupuytren contracture are described in Table 1. Procedures included: collagenase injection (13.6%), percutaneous palmar fasciotomy (10.3%), open palmar fasciotomy (3.9%), palmar fasciectomy (10.8%), palmar fasciectomy with 1 (29.2%) or multiple (19.6%) digit releases, or a combination of the aforementioned procedures (12.8%). Approximately 2.5% (n = 1213) of patients underwent concurrent capsulectomy or capsulotomy of metacarpophalangeal or interphalangeal joints. One hundred forty-nine patients (0.31%) were diagnosed with CRPS (type 1, 89.9%; type 2, 10.1%) at a mean of 3.4 months (SD, 2.3; range, 0.2-11.1 months; Figure 1) following treatment of Dupuytren contracture.

Figure 1.

Months until CRPS diagnosis after treatment of Dupuytren contracture. CRPS = complex regional pain syndrome.

Ninety-seven patients with a diagnosis of CRPS after treatment for Dupuytren contracture underwent postoperative stellate ganglion block at a mean of 9.5 months from Dupuytren treatment (SD, 17.1; range, 0.72-106.85 months) and 3.0 months from CRPS diagnosis (SD, 11.7; range, 0-98.73 months). Twenty-five of these patients did not have a diagnosis of CRPS. When added to the CRPS cohort, the postprocedural rate of CRPS and/or stellate ganglion block increased to 0.36%.

The rate of CRPS diagnosis following minimally invasive procedures (eg, collagenase injection or fasciotomy; CPT 20527, 26040, and 26045) was 0.05% compared with 0.41% for procedures involving fasciectomy (P < .001). The rate of CRPS diagnosis following procedures only involving the palmar fascia (CPT 26040, 26045, and 26121) was 0.12% compared with 0.44% for procedures involving digital fasciectomy (CPT 26123 and 26125; P < .001). The rate of CRPS after fascial-only procedures (CPT 26040, 26045, 26121, 16123, and 26125) was 0.22%, in contrast to 0.82% after those involving capsulectomy and/or capsulotomy (P < .001).

Outcomes of multivariable logistic regression to predict CRPS diagnosis are shown in Table 2. The c-statistic for the model was 0.75. Significant predictors included female sex (odds ratio [OR], 2.02; 95% confidence interval [CI], 1.44-2.82; P < .001), southern region (OR, 1.80; 95% CI, 1.09-2.98; P = .022), long-term disability status (OR, 4.73; 95% CI, 1.11-20.09; P = .035), palmar fasciectomy with release of 1 (OR, 5.91; 95% CI, 1.83-19.11; P = .003) or more than 1 digit (OR, 13.32; 95% CI, 4.17-42.53; P < .001), and a combination of procedures for treatment of Dupuytren contracture (OR, 8.23; 95% CI, 2.48-27.26; P = .001). Simultaneous open or endoscopic CTR was not a significant predictor of CRPS. Outcomes of a second regression to predict CRPS diagnosis and/or stellate ganglion block are shown in Table 3. The c-statistic for this model was 0.73.

Table 2.

Multivariable Predictors of CRPS After Treatment of Dupuytren Contracture.

Characteristic	Odds ratio	95% confidence interval	P value
Age	0.99	0.97-1.01	.210
Sex
Male	Ref.	—	—
Female	2.02	1.44-2.82	<.001
Region
Northeast	Ref.	—	—
North Central	1.28	0.73-2.25	.384
South	1.80	1.09-2.98	.022
West	1.04	0.57-1.92	.890
Unknown	2.39	0.95-6.02	.065
Employment status
Active full time	Ref.	—	—
Active part-time or seasonal	3.49	0.83-14.73	.089
Early retiree	0.95	0.47-1.90	.881
Medicare eligible retiree	0.67	0.34-1.33	.252
Retiree (status unknown)	1.02	0.38-2.76	.961
COBRA continuee	1.40	0.19-10.30	.742
Long-term disability	4.73	1.11-20.09	.035
Surviving spouse/Dependent	1.16	0.26-5.14	.842
Other/Unknown	1.05	0.69-1.59	.833
Procedure
Collagenase injection	Ref.	—	—
Palmar fasciotomy, closed	1.30	0.26-6.47	.746
Palmar fasciotomy, open	0.87	0.09-8.40	.901
Palmar fasciectomy	3.32	0.91-12.04	.068
Palmar fasciectomy, 1 digit	5.91	1.83-19.11	.003
Palmar fasciectomy, >1 digit	13.32	4.17-42.53	<.001
Combination	8.23	2.48-27.26	.001
Concurrent carpal tunnel release	1.14	0.58-2.27	.702

Note. CRPS = complex regional pain syndrome; Ref. = Reference (ie, baseline category that serves as standard of comparison for other categories); COBRA = Consolidated Omnibus Budget Reconciliation Act.

Table 3.

Multivariable Predictors of CRPS and/or Stellate Ganglion Block After Treatment of Dupuytren Contracture.

Characteristic	Odds ratio	95% confidence interval	P value
Age	0.98	0.97-1.00	.031
Sex
Male	Ref.	—	—
Female	2.00	1.47-2.73	<.001
Region
Northeast	Ref.	—	—
North Central	1.17	0.69-1.97	.560
South	1.78	1.12-2.82	.014
West	1.01	0.58-1.77	.967
Unknown	2.35	1.00-5.53	.049
Employment status
Active full time	Ref.	—	—
Active part-time or seasonal	2.84	0.68-11.87	.154
Early retiree	0.89	0.46-1.71	.721
Medicare eligible retiree	0.74	0.39-1.38	.344
Retiree (status unknown)	1.09	0.44-2.68	.857
COBRA continuee	1.14	0.16-8.34	.898
Long-term disability	3.91	0.93-16.50	.063
Surviving spouse/Dependent	1.08	0.25-4.70	.917
Other/Unknown	1.01	0.69-1.47	.967
Procedure
Collagenase injection	Ref.	—	—
Palmar fasciotomy, closed	1.04	0.28-3.89	.949
Palmar fasciotomy, open	1.95	0.52-7.38	.323
Palmar fasciectomy	2.41	0.86-9.78	.096
Palmar fasciectomy, 1 digit	3.97	1.58-10.00	.003
Palmar fasciectomy, >1 digit	8.82	3.55-21.92	<.001
Combination (except injection)	5.40	2.08-14.00	.001
Concurrent carpal tunnel release	1.39	0.78-2.47	.269

Note. CRPS = complex regional pain syndrome; Ref. = Reference (ie, baseline category that serves as standard of comparison for other categories); COBRA = Consolidated Omnibus Budget Reconciliation Act.

Discussion

Compared with prior studies reporting CRPS rates following surgical treatment of Dupuytren contracture ranging from 2.4% to 18.5%,^9,10 this claims database review found a lower postprocedural rate of CRPS of 0.31% for all treatment types. Although demographic information in both of the prior studies is limited, Sennwald found that 40% of female patients (6 of 15) developed postoperative CRPS compared with 14.8% of male patients (13 of 88), ¹⁰ consistent with the significant female gender predictor seen in this study. Long-term disability was also a significant predictor of postoperative CRPS in our analysis, in line with prior studies suggesting a significant relationship between employment status and outcomes in hand surgery.^14,15 Although geographic region does not have a well-established association with the pathophysiology or epidemiology of CRPS, the strong southern link observed in our study cannot be ignored and suggests an avenue for further inquiry.

Our analysis additionally indicates that more extensive operative procedures for treatment of Dupuytren contracture are associated with an increased risk of CRPS. Prior studies were limited to patients undergoing open surgical fasciectomy,^9,10 and thus could not appreciate the variations in risk along the spectrum of both nonoperative and operative treatment. Open fasciectomy is a more extensive and invasive procedure than collagenase injection or percutaneous palmar fasciectomy; theoretically, this may lead to a higher risk of postoperative CRPS. Despite the inclusion of nonoperative and operative procedures with less morbidity, our study yields an overall rate of postprocedural CRPS that is lower than that previously reported in studies of only fasciectomy. However, our findings support an increasing risk of CRPS with more extensive surgery, as open fasciectomy of 1 or more than 1 digit, as well as a combination of procedures to treat Dupuytren contracture, was a significant predictor of postoperative CRPS. Similarly, the addition of capsulectomy and/or capsulotomy to the procedure was found to be associated with a significantly higher rate of CRPS compared with fascial-only procedures.

Our findings additionally support the mounting evidence that concurrent treatment of Dupuytren contracture and carpal tunnel syndrome do not yield an increased risk of CRPS. Although popular opinion in hand surgery once shunned simultaneous surgical treatment of Dupuytren contracture and carpal tunnel syndrome due to concern of higher long-term complication rates, ¹⁶ including CRPS, recent studies have demonstrated that concurrent Dupuytren contracture release and CTR does not have higher complication rates.^17-19 Our study similarly suggests that combining treatment of Dupuytren contracture with open or endoscopic CTR does not increase the risk of CRPS diagnosis.

The severity of disease may contribute to rates of CRPS development. Sennwald reported that 68% (13 of 19) of CRPS cases occurred in patients with stage 3 or 4 disease. ¹⁰ Of note, Sennwald staged disease based on location of the affected fascia, rather than by Tubiana’s classic degree of flexion deformity. ²⁰ In the study by Sennwald, stage 3 disease was defined as extending beyond the proximal interphalangeal (PIP) joint, and stage 4 disease was defined as extending beyond the distal interphalangeal joint. Bulstrode et al ⁹ classified the severity of disease based on both the location of disease and the degree of flexion deformity; these authors similarly found that 83.3% (5 of 6) of CRPS cases occurred in patients with more severe disease, defined as disease involving ≥60° contracture at the PIP joint, extensor involvement, or ischemia with digit extension after fasciectomy. It is interesting to note that treatment of the patients with PIP disease would require a more extensive neurolysis of the digital nerves than disease isolated to the palm. However, due to the database nature of this study, we could not assess preoperative disease severity.

Treatment of CRPS can be challenging and requires a multidisciplinary approach. Although many pharmacologic agents have been studied for use in the treatment of CRPS, few have been shown to provide reproducible relief in quality studies. Gabapentin in doses ranging from 600 to 1800 mg per day may reduce pain symptoms in the acute phase of CRPS.^3,21 Subtherapeutic intravenous doses of ketamine were shown to reduce pain in a small cohort of patients with CRPS for an average duration of 9.4 months. ²² Free radical scavengers including dimethylsulphoxide and N-acetylcysteine were also likely to reduce CRPS pain.^3,23 A task force aimed at establishing evidence-based guidelines for treatment of CRPS found insufficient evidence that acetaminophen, nonsteroidal anti-inflammatory drugs, opioids, and local anesthetics are effective in reducing pain in CRPS. ³

Although most of the literature regarding medical therapies for CRPS is focused on treatment, there have been some efforts to reduce perioperative risk of CRPS. Although its use specifically in Dupuytren surgery has not been studied, vitamin C has been demonstrated as a potential means of reducing CRPS risk. A meta-analysis of 4 studies looking at the use of vitamin C following upper or lower extremity trauma or surgery showed a relative risk of 0.22 with daily administration of at least 500 mg of vitamin C starting immediately after the trauma or surgery and continued for at least 45 days. ²⁴ In addition, in patients with a history of CRPS who require additional surgery, the use of regional anesthetic blocks, ³ perioperative stellate ganglion blocks, ²⁵ and multimodal postoperative pain regimens can reduce the risk of relapsing or recurrent CRPS.

There are 2 main ways that our findings may promote risk reduction for CRPS development following treatment of Dupuytren contracture: first, by guiding selection of patients and surgical technique, and, second, by identifying patients who may benefit from risk-reducing interventions. We acknowledge that the significant predictors that we identify in our study are not modifiable on a patient level; however, considering these data, surgeons should think twice about performing fasciectomies of multiple digits in 1 procedure if the patient has additional demographic risk factors (eg, a woman on long-term disability). Instead, a staged approach may be more appropriate in these cases. Second, patients who fit the high-risk profile may be targeted for risk reduction interventions, such as administration of vitamin C or regional anesthetic blocks. However, before we fully advocate for these interventions, more robust evidence supporting preventive efficacy is needed.

Limitations of this study relate to the constraints inherent in retrospective review of a claims database. We cannot distinguish among affected digits or rays (eg, index vs middle finger, central vs ulnar ray) due to lack of CPT code specificity. In addition, in the event that patients experienced CRPS after treatment but were not diagnosed formally and designated with an ICD diagnosis code, they would not be captured in this study. In our attempt to identify potential uncoded cases of CRPS by including patients who underwent stellate ganglion block without an ICD CRPS diagnosis in our second regression model, we found that significant predictors were largely consistent with the initial regression based only on patients with an ICD diagnosis of CRPS. However, due to potential undercoding, it remains possible that the true incidence of CRPS following treatment of Dupuytren contracture is higher than that reported here. Conversely, it is possible that patients who did not have true CRPS postoperatively were labeled with a CRPS diagnosis based on clinicians’ assessment of postoperative pain, without adherence to standard diagnostic criteria for CRPS. ²⁶ This would lead to an overestimate of CRPS diagnosis and skew epidemiologic accuracy.

In addition, our analysis is limited to a commercial insurance database, and thus may not be representative of all patients within the United States. This study does not contain patients with Medicaid as payer, which implies that the studied population is of higher socioeconomic standing. Although we included patients with a Medicare supplement, this database does not include many of the patients in the general Medicare population with Dupuytren contracture. As Dupuytren disease is a problem known to affect people in the later decades of life, future studies could focus on the older population through use of a dedicated Medicare database. Furthermore, while our study identifies risks of CRPS, future studies should continue to investigate modalities of perioperative risk reduction that may be deployed to limit the development of this debilitating and recalcitrant condition.

In conclusion, in this commercial claims database, the incidence of CRPS following treatment of Dupuytren contracture was 0.31%, which is lower than that previously reported. Risk factors included female sex, long-term disability status, and more extensive operative procedures. Patients with these characteristics should be considered for preoperative and postoperative risk reduction measures to limit the development of CRPS.