Introduction

Complex regional pain syndrome (CRPS) is a debilitating and painful condition characterized by sensory, neurologic, autonomic, and/or trophic impairments [1]. CRPS is a clinical diagnosis based on accepted criteria including the Veldman criteria, the International Association for the Study of Pain (IASP) criteria for CRPS, Valencia criteria, and most commonly the Budapest criteria [2,3,4]. The global prevalence of CRPS is estimated between 5.5 and 26.2 per 100,000 persons per year [5•] affecting more females than males between the ages of 40 and 70 years old [6]. The total annual healthcare costs and prescription costs after a diagnosis of CRPS are estimated to be 2.17-fold and 2.56-fold higher compared to baseline, respectively [7].

The exact pathophysiologic mechanism of CRPS is unknown. It has been postulated that CRPS is a result of a multifactorial derangement in the inflammatory system, immune system, peripheral and central pain signaling pathways, and autonomic nervous system [1, 2]. The primary objective of this scoping review is to describe the available treatment modalities for CRPS. In addition, we appraise the level of evidence and degree of recommendation for each treatment modality, as well as propose an updated treatment algorithm for CRPS.

Methods

Search Strategy and Study Selection

An entry was registered in the International Prospective Register of Systematic Reviews (PROSPERO) database before study initiation. A systematic search was performed by an experienced librarian (L.C.H.) after search terms were previously determined by the study authors. Relevant keywords included the following: “physical therapy,” “acupuncture,” “transcutaneous electrical nerve stimulation,” “amitriptyline,” “nortriptyline,” “ketamine,” “gabapentin,” pregabalin,” “tricyclic antidepressant,” “opioid,””neuromodulation,” “spinal cord stimulation,” “dorsal root ganglion stimulation,” “peripheral nerve stimulation,” “intrathecal systemic steroid,” “scrambler therapy,” “cannabis,” “bisphosphonates,” “topical diclofenac,” “topical lidocaine,” “sympathetic nerve block,” “sympathetic neural lysis,” “surgery,” and “treatment of complex regional pain syndrome.” Databases Ovid MEDLINE(R), Ovid Embase, Ovid Cochrane Central Register of Controlled Trials, Ovid Cochrane Database of Systematic Reviews, and Scopus were queried from 1990 to the search date (April 26, 2023) for the selected search terms and synonyms. The actual strategies listing all search terms, Boolean operators, and how they were combined are available in eAppendix 1. All articles were screened by two authors (Y.F.H. and E.K.) independently including titles, abstracts, and full texts. Inclusion criteria included observational studies and randomized controlled trials (RCTs) in the English language that reported change in pain intensity after implementation of a treatment modality (conservative, pharmacologic, or interventional) for CRPS. Exclusion criteria included case reports or case series and animal studies. Discrepancies were resolved by a third author (R.S.D.).

Data Extraction

Study characteristics (study design, funding source, treatment modality, and sample size) and outcomes of interest were extracted. Primary outcome was change in pain intensity from baseline. Secondary outcomes included change in physical functioning and mental health. For each treatment modality, RCTs were selected for further data extraction and analysis. If no RCTs were available, observational or retrospective studies were selected.

Evidence Appraisal

We utilized the United States Preventive Services Task Force (USPSTF) Criteria to appraise the evidence level and degree of recommendation for each treatment modality (eAppendix 2 and eAppendix 3). A degree of recommendation A correlates to the highest level of recommendation where there is good evidence that the measure is effective, and benefits outweigh the harms. Conversely, a degree of recommendation D correlates the lowest level of recommendation where there is at least moderate evidence that the measure is ineffective. A degree of recommendation I correlates to insufficient, low-quality, or contradictory evidence and that recommendation cannot be determined.

Results

Search Strategy

Of 3027 studies that were screened after deduplication, 322 studies underwent full-text review (Fig. 1). Of studies that underwent full-text review, sixty-five studies were included: 53 RCTs (9–19, 22–27, 31–56, 60–67, 71, 82, 83, 85, 86), 6 prospective studies (20, 21, 28, 29, 74, 81), and 6 retrospective studies (30, 72, 73, 78–80). The key study characteristics are reported in Table 1, 2, 3, and 4.

Fig. 1
figure 1

PRISMA flow diagram of study selection process

Full size image
Table 1 Study characteristics for conservative treatments
Full size table
Table 2 Study characteristics for topical pharmacologic treatment
Full size table
Table 3 Treatment of CRPS with oral, intravenous, or intramuscular pharmacologic formulations
Full size table
Table 4 Interventional treatments for CRPS
Full size table

Clinical Presentation

History

Patients with a suspected diagnosis of CRPS typically present with pain that is disproportionate to a recent trauma in a non-dermatomal distribution [2]. CRPS typically manifests with abnormalities in sensory, vasomotor, sudomotor, and/or motor/tropic signs and symptoms. The pain may migrate from the site of initial injury towards the torso and even to the contralateral limb [4].

Physical Exam

The physical examination should be conducted by exposing the affected limb and non-affected limb to assess for evidence of sensory, vasomotor, sudomotor/oedema, and motor/trophic abnormalities as listed by the Budapest Criteria [4]. Neurological and vascular exams should be performed to make sure that the presenting symptoms are not explained by a central/peripheral nerve injury or vascular pathology.

Diagnostic Tests

There is no single diagnostic test or lab for CRPS. CRPS is a clinical diagnosis. However, diagnostic workup should be considered in patients presenting with a suspected diagnosis of CRPS to rule out other pathologies. Labs include a complete blood count, C-reactive protein, erythrocyte sedimentation rate, and/or anti-nuclear antibody to rule out infectious, inflammatory, and/or auto-immune pathologies [3, 75]. Nerve conduction tests and electromyography may also be conducted to assess for peripheral nerve lesions. Thermography, sweat test, bone scintigraphy in the acute phase, triple-phase bone scintigraphy within the first 5 months, plain radiographs in the chronic phase, bone mineral density, and sympathetic block may also support a diagnosis of CRPS.

Conservative Treatment Options

This section describes conservative treatments for CRPS including physical therapy (PT), mirror therapy (MT), acupuncture, and transcutaneous electrical stimulation (TENS) (Table 1). Six RCTs comprising 317 participants were identified that investigated the effects of PT on patients with CRPS [8,9,10,11,12,13]. Lee et al. reported that PT reduced pain and improved gait impairment [8]. Adding aerobic exercise to PT was superior to PT alone [9]. Moreover, pain exposure PT resulted in improved disability and function compared to protective pain contingent treatment [10, 11]. Oerlemans et al. demonstrated that PT was superior to occupational therapy (OT) in reducing pain and improving active range of motion [12]. PT was more cost-effective [13]. Overall, there was favorable improvement in pain intensity (level I, degree B) and physical functioning outcomes (level I, degree B) following PT for CRPS.

Three RCTs comprising of 106 patients evaluated the effectiveness of MT on patients with CRPS [14,15,16]. Cacchio et al. showed that MT improved pain and motor function compared to controls [14]. Adding MT to conventional therapy resulted in higher functional independence measure-motor and -pain scores [15]. For the maximal impact on pain and disability, MT needed to be performed in a sequential order such as hand laterality recognition, imagined movements, and mirror movements [16]. Overall, the literature supports the association of MT and improvement in pain intensity (level I, degree B) and physical functioning outcomes (level 1, degree B) in CRPS.

The efficacy of acupuncture in treating CRPS was investigated in two RCTs [17, 18] and two prospective case–control studies [18] which collectively comprised of 372 patients. Korpan et al. observed no statistical difference in pain intensity in CRPS patients treated with acupuncture or sham [17]. However, electric acupuncture was demonstrated to improve pain and functionality [18, 19]. Additionally, Zheng et al. (2018) showed that combining acupuncture with rehabilitation led to improvement in upper extremity motor function and enhancement of quality of life [20]. Overall, the association of acupuncture in improvement of pain intensity (level I, degree C) and functionality (level I, degree C) in CRPS varied among the studies and its benefit remains unclear.

One RCT comprising 30 patients reported on the use of TENS in the management of CRPS [21]. It showed significant improvements in pain scores, edema, range of motion (ROM), and functionality among both subjects in the TENS and Sham-TENS groups with greater improvements favoring the TENS group. This study presents evidence supporting TENS providing modest pain relief and functional improvement (Level I, degree B), although the number of high-quality studies is limited.

Pharmacological Treatment—Topical Treatments

Three RCTs comprising of 64 patients compared topical treatments for CRPS, including eutectic mixture of local anesthetics (EMLA) cream [22], ketamine [23], and fatty cream with dimethyl sulfoxide (DMSO) to placebo [24] (Table 2). After 2 months of topical DMSO-containing cream along with PT, patients’ RSD-scores improved (median improvement 4 vs 3, P < 0.01) as well as VAS pain scores (2.9 vs 1, P > 0.1) [22]. Finch et al. reported improved allodynia in both treated limbs as well as in the ipsilateral forehead and hyperalgesia (30 min after application). Interestingly, Strauss et al. found that EMLA cream improved tactile resolution (t(11) = 3.98, P < 0.01) and motor function (t(10) = 2.57, P < 0.05) in the hand but not pain [24]. Overall, the efficacy of topical medications on pain intensity (Level I, degree C) and functional improvement (level I, degree C) in subjects with CRPS remain unclear with some studies showing inconsistent results of improvement versus no change when compared to placebo.

Pharmacologic Treatment—Oral, Intravenous, and Intramuscular Formulations

This section describes the treatment of CRPS with oral, intravenous (IV), and intramuscular (IM) pharmacologic formations including gabapentinoids, opioids, bisphosphonates, systemic steroids, and ketamine (Table 3). The benefits of gabapentinoids have been studied on patients (n = 586) with CRPS in two RCTs [25, 26], two prospective studies [27, 28], and one retrospective study [29]. In a RCT comparing gabapentin and placebo, van De Vusse et al. observed that gabapentin did not alleviate pain. Further, patients who were treated with gabapentin reported side effects of dizziness, somnolence, and lethargy [25]. In another RCT, Brown et al. compared gabapentin and amitriptyline. Both medications provided significant pain relief without any differences in side effects [26]. In the two prospective studies, both gabapentin and pregabalin provided pain relief but no significant improvement in function [27, 28]. Lastly, Lee et al. reported that combining IV mannitol, steroids, and oral gabapentin improved pain intensity, finger ROM, swelling, and grip strength in CRPS. Overall, participants with CRPS who received gabapentinoids achieved inconsistent pain relief (level I, degree C) and functionality (level I, degree C), with some studies reporting no benefit compared to placebo (or control arms) and some studies reporting modest benefit.

Two RCTs assessed the administration of systemic opioids in 63 patients with CRPS. One study reported that morphine (30 mg) with memantine (40 mg) was associated with greater improvement in pain and disability (habitual pain: 5.47 to 1.40; P < 0.001; movement pain: 8.03 to 2.84; P < 0.001) compared to morphine alone [30]. Harke et al. observed that after spinal cord stimulation (SCS) cessation, 8-day therapy with carbamazepine but not morphine delayed the recurrence of pain compared to placebo (P = 0.038) [31]. Overall, these studies suggest it is unclear if opioid medications offer benefit to patients with CRPS compared to placebo (level I, degree D).

Seven RCTs examined bisphosphonates for pain control in 249 patients with CRPS. Young et al. compared IV pamidronate to oral prednisolone and found significant reduction in VAS scores at 4 weeks in both groups compared to baseline, but only prednisolone improved swelling [32].

Multiple RCTs compared various IV bisphosphonates to placebo with different doses and durations [33,34,35, 36•]. Adam et al. showed that IV alendronate led to significant improvements in spontaneous pain, tenderness, swelling, and motion compared to placebo [33]. Varenna et al. demonstrated that IV clondronate significantly reduced pain and clinical global assessment compared to placebo [34]. Robinson et al. found that a single infusion of pamidronate significantly improved pain and physical function [35]. In another study [36•], four doses of IV neridronate led to significant reduction in pain intensity compared to the placebo group. The patients in the neridronate-treated group were able to discontinue NSAIDs and acetaminophen within 2 weeks compared to only 45% in the placebo group.

Manicourt et al. examined the effect of oral bisphosphonate in patients with CRPS. They showed that 8 weeks of oral alendronate improved pain, pressure tolerance, and joint mobility [37]. During an extension of this study where all patients could use oral alendronate, the patients who were previously in the placebo group showed significant improvement in the same outcomes that were achieved by the patients that were in the alendronate cohort (P < 0.05).

Varenna et al. was the only RCT examining IM bisphosphonate therapy for CRPS. They showed that the IM neridronate group achieved significantly higher pain reduction compared to placebo [38]. In an extension phase of the study, where patients initially receiving placebo treatment were treated with IV neridronate, both groups had high rates of patients with greater than 50% pain relief and significant improvement in function at 12 months [39]. Overall, these studies suggest that both oral and IV bisphosphates are effective in reducing pain intensity (level I, degree B) and improving functionality (level I, degree B) in patients with CRPS.

There were three RCTs conducted by Kalita et al. that evaluated the effects of prednisolone in CRPS [40,41,42]. These studies comprised of 154 patients. In the first study, prednisolone was superior to piroxicam with 83.3% vs 16.7% of the subjects achieving significant improvement in pain [40]. The second study divided the patients that previously responded to high-dose steroid into two groups: continue steroid or discontinue steroid. The group that discontinued steroid had a higher occurrence of symptoms compared to those who continued steroid [41]. The third study showed that CRPS patients receiving a higher dose of steroid (prednisolone 20 mg vs 40 mg) experienced significantly higher pain reduction [42]. Overall, these studies demonstrated that a short-course of systemic steroid reduced CRPS pain (level I, degree B). Functionality did not improve with systemic steroid treatment (level I, degree D).

Two RCTs comprising 79 patients evaluated the effectiveness of IV ketamine on pain from CRPS [43, 44]. Both RCTs showed that IV ketamine infusion for 4 or 10 days resulted in significant improvement in pain at the 3-month follow-up. A sub-analysis of one of the RCTs showed that there was an inverse relationship between pain intensity and motor movement parameters in the affected limb such as velocity, amplitude, and frequency [45]. These studies reported that patients receiving ketamine infusion developed higher percentages of nausea, vomiting, and psychomimetic effects compared to placebo. Overall, IV ketamine has been shown to reduced pain intensity but may manifest with intolerable side effects (level I, degree C). However, it does not appear to affect functionality (level I, degree D).

Interventional Treatments

Interventional treatments for CRPS include sympathetic plexus blocks, spinal cord stimulation (SCS), dorsal root ganglion stimulation (DRG-S), peripheral nerve stimulation (PNS), and intrathecal drug delivery system (IDDS; Table 4).

Eight RCTs assessed the efficacy of sympathetic plexus blocks for CRPS [46,47,48,49,50,51,52]. Four of the eight RCTs focused on stellate ganglion block (SGB) for upper extremity CRPS [46,47,48, 53]. Naskar et al. showed that there was no significant difference between SGB injectate with ropivacaine plus clonidine and methylprednisolone [47]. When comparing SGB to T2 paravertebral block (PVB), PVB resulted in higher success rates of reducing pain intensity, duration of pain relief, and patients’ satisfaction [48]. Toshniwal et al. demonstrated that continuous SGB and continuous infraclavicular brachial plexus block were equivocal in reducing pain and edema and improving range of motion [53].

One study investigated the benefits of thoracic sympathetic block (TSB) [49]. Rocha et al. found that TSB resulted in significant improvement in pain relief as assessed by the McGill Pain Questionnaire, Neuropathic Pain Symptom Inventory, and depression scores compared to sham procedure.

Lastly, three RCTs evaluated the effectiveness of lumbar sympathetic ganglion block [50, 51]. Meier et al. compared IV lidocaine plus lumbar sympathetic ganglion injection with normal saline versus IV normal saline plus lumbar sympathetic ganglion injection with lidocaine [50]. The authors showed that lumbar sympathetic ganglion injection with lidocaine was more effective in reducing allodynia and pain scores. Yoo et al. showed that botulotoxin type A injection into the lumbar sympathetic ganglion decreased more pain than local anesthetic injection [51]. Freitas et al. demonstrated that pulsed radiofrequency ablation of the sympathetic lumbar plexus was equally effective in reducing pain as a lumbar sympathetic ganglion block [51]. Overall, the current evidence suggests that sympathetic ganglion block is associated with meaningful pain relief in CRPS (level I, degree B).

Only one RCT was found evaluating sympathectomy for CRPS [54]. The study randomized 20 patients into two arms: radiofrequency or phenol lumbar sympathectomy. There was no statistically significant difference between groups but both had significant pain relief up to 4 months after the procedure [54]. Overall, this study showed that either radiofrequency or phenol lumbar sympathectomy may reduce pain in CRPS, although future studies supporting this association are warranted (level I, degree C).

Neuromodulation with SCS has been demonstrated to be effective in treating CPRS [76••, 77]. Six RCTs comprising of 342 patients were identified [55,56,57,58,59,60,61,62]. Kemler et al. showed that SCS improved pain and health-related quality of life compared to PT only over a 2-year period [56, 57]. However, these benefits diminished over a 5-year period [58] due to neural habituation [78] or device-related issues [79]. Interestingly, the effects on CRPS pain did not differ between different SCS settings [55, 61]. Additionally, SCS restored the quantitative sensory testing thresholds for the affected limb to that of the non-affected limb [59, 60]. Overall, dorsal-column SCS is effective in treating pain from CRPS (level I, degree B) and is also approved by the FDA for this indication.

DRG-S also carries FDA labeling for treatment of pain from CRPS and may provide substantial pain relief [80]. One RCT and one prospective study comprising of 164 patients were identified [62,63,64,65,66]. Deer et al. demonstrated that DRG-S provided a greater percentage of participants achieving ≥ 50% pain relief (PPR), quality of life, and psychological measures compared to dorsal-column SCS at the 3-month follow-up [63]. DRG-S was associated with higher cost than dorsal-column SCS [64, 65], and more subjects preferred DRG-S over dorsal-column SCS [66]. Overall, these studies highlighted that DRG-S substantially improves pain from CRPS (level I, degree B).

PNS may alleviate pain by disrupting the nociceptive afferent fibers, downregulating inflammatory mediators, and changing the local microenvironment [81,82,83]. One prospective study and three retrospective studies comprising 207 patients were identified [67,68,69,70]. These studies showed that PNS reduced pain and opioid consumption. Functional improvement varied between studies. Overall, these studies showed that PNS reduces pain intensity (level II-2, degree B) with variation in improvement in functionality (level II-2, degree C).

Two RCTs by Munts et al. compared intrathecal medications to placebo [71, 72]. Intrathecal glycine failed to show superiority over saline for pain, dystonia or global impression scores, and a single dose of intrathecal methylprednisolone did not provide pain relief in CRPS at 4 weeks of follow-up [71, 72]. Intrathecal baclofen (ITB) was administered in either a slow or rapid infusion (4-times faster) with no difference on pain [73]. Finally, when comparing intrathecal clonidine to adenosine, both treatments showed improvements in hyperalgesia and allodynia with clonidine having a threefold higher effect for pain relief [74]. Overall, the current evidence showed that only intrathecal clonidine or adenosine may improve pain intensity although future studies are warranted to support this association (level I, degree C). There was no report of functional improvement (level I, degree I).

Discussion

In this systematic review, we described modest to moderate improvement in pain intensity from PT, OT, MT, acupuncture, and TENS therapy, although changes in functionality were inconsistent. Topical medications such as EMLA and Ketamine cream were associated with decreased allodynia and hyperalgesia. There were mixed results supporting gabapentinoids for reduction of pain in CRPS. It was unclear whether systemic opioid alone was able to provide pain relief in CRPS because the study design of included studies reported adjuvant therapy such as memantine or carbamazepine added to systemic morphine. There was no study comparing morphine to placebo. There were positive outcomes associated with IM and IV bisphosphonates. Patients showed modest to moderate improvement in pain intensity and functionality. Systemic steroid provided moderate short-term pain reduction, although pain recurred after 1 month of discontinuation. Similarly, ketamine was able to provide short-term pain relief, although adverse effects were commonly reported. In terms of interventional therapy, no benefit was reported for intrathecal drug delivery system. Sympathetic ganglion block, sympathectomy, SCS, DRG-S, and PNS were associated with modest to moderate improvement in CRPS pain intensity.

The findings in this systematic review are consistent with the recommendations from prior reviews in the literature [84]. From a therapeutic approach, we recommend starting with conservative therapies such as PT with aerobic exercises/pain exposure, OT, MT, acupuncture, and TENS. These therapies may need to be implemented concurrently alongside other treatment modalities to decrease pain and restore function (Fig. 2). In refractory cases of CRPS, clinicians may consider treating the patient with conservative therapies, medications, and interventional therapy at the same time. Neuromodulation and sympathectomy are considered last-resort therapies given its invasive nature, high cost, and potential adverse events. Opportunities for future research exist for treatment modalities with limited RCT data such as PNS therapy and neuropathic pain medications (e.g., gabapentinoids, tricyclic antidepressants, and serotonin-norepinephrine reuptake inhibitors).

Fig. 2
figure 2

Treatment recommendations for complex regional pain syndrome. TENS, transcutaneous electrical stimulation; EMLA, eutectic mixture of local anesthetics; PO, enteral; IM, intramuscular; IV, intravenous; SCS, spinal cord stimulation; DRG, dorsal root ganglion; PNS, peripheral nerve stimulation

Full size image

Conclusion

In conclusion, this systematic review equips the clinician with important updates on conservative, pharmacologic, and interventional treatment modalities for CRPS-related pain. There is level I evidence supporting modest to moderate improvement in pain intensity from physical therapy, occupational therapy, massage therapy, acupuncture, and TENS. EMLA and ketamine cream were associated with decreased allodynia and hyperalgesia. Intramuscular or intravenous bisphosphonate therapy may achieve modest to moderate improvement in pain intensity and functionality. Systemic steroid and ketamine provided clinically significant pain reduction, although it was of short duration. Interventional therapy, including sympathetic ganglion block, sympathectomy, dorsal column spinal cord stimulation, dorsal root ganglion stimulation, and peripheral nerve stimulation, may provide modest to moderate improvement in pain with although the level of evidence was limited.