Complex Regional Pain Syndrome: Analgesic Outcome

STUDY MODEL Retrospective observational study in the form of a chart review. STUDIED POPULATION AND PERIOD Patients followed at the Maisonneuve-Rosemont Hospital Pain Management Center (PMC) for complex regional pain syndrome (CRPS) who received at least one analgesic intervention between January 1, 2020 and October 2025.

This period was chosen to obtain a sufficient number of patients and to best reflect current practice.

SAMPLE SIZE A sample size calculation will not be performed for this study. The PMC is a tertiary pain clinic that treats a large number of patients, and it is estimated that the chosen period will yield a sample of approximately 200 patients.

GENERAL PROCEDURE Data will be collected from digitized records in the Clinibase and Oacis clinical applications.

An initial sampling of charts will be performed in Clinibase by the planning, programming, and research department of the CIUSSS de l'Est-de-l'Île-de-Montréal. A search of the Clinibase database using the type of appointment at the Pain Management Center between January 1, 2020, and October 23, 2025, will provide the chart numbers of potentially eligible patients. The appointment types searched will include: block, ketamine infusion, pamidronate, stellate block, and lumbar sympathetic block.

Each chart will then be reviewed in Oacis to locate the first consultation note at the Pain Management Center in order to verify inclusion criteria-specifically, a diagnosis of complex regional pain syndrome as determined by the Budapest criteria. The use of an eligible analgesic intervention (plexus block, pamidronate infusion, ketamine infusion, stellate or lumbar sympathetic block) for the treatment of CRPS pain will be confirmed. The analgesic technique does not necessarily need to have been performed at the first visit to be eligible. Charts meeting the inclusion criteria will be retained for the research project and analyzed in more detail to collect the required data.

DATA COLLECTION

Patient data:

• Demographic data
• Age
• Sex
• Association of the chart with a worker's compensation agency (CNESST in Quebec) (yes/no) (This information will be collected because affiliation with a workers' compensation program is an independent factor for poorer response to analgesic interventions. It is therefore a confounding factor for the primary endpoint-intervention success) (Clin J Pain 1995;11:94-102);
• Patient's medication (SSRI antidepressants (yes/no), tricyclic antidepressants (yes/no), gabapentinoids (yes/no), NSAIDs (yes/no), opioids (yes/no))
• Initial trauma (fracture, sprain, surgery, stroke, other);
• Post-traumatic stress disorder (yes/no). This variable will be collected for the same reason as CNESST affiliation. A PTSD diagnosis significantly affects pain intensity and treatment response, particularly if inadequately addressed (J Pain Res 2018;11:527-36).

CRPS data:

• Upper or lower limb involvement;
• Left or right;
• Duration of symptoms at the time of first analgesic intervention (measured from the initiating event) (<12 months, 12-18 months, >18 months);
• Symptoms reported by the patient:
• Sensory: hyperalgesia (yes/no), allodynia (yes/no)
• Vasomotor: temperature asymmetry (yes/no), color asymmetry (yes/no), skin color change (yes/no)
• Sudomotor: edema (yes/no), sweating change (yes/no)
• Motor/trophic: decreased range of motion (yes/no), tremor (yes/no), weakness (yes/no), dystonia (yes/no), trophic changes in skin, nails and/or hair (yes/no).
• Signs observed by the physician:
• Sensory: hyperalgesia (yes/no), allodynia (yes/no)
• Vasomotor: temperature asymmetry (yes/no), color asymmetry (yes/no), skin color change (yes/no)
• Sudomotor: edema (yes/no), sweating change (yes/no)
• Motor/trophic: decreased range of motion (yes/no), tremor (yes/no), weakness (yes/no), dystonia (yes/no), trophic changes in skin, nails and/or hair (yes/no).
• CRPS category (type I, type II, ROS, NOS)
• CRPS subtype, as determined by the main symptom reported at the first consultation:
• Vasomotor: predominance of vasomotor and sudomotor signs/symptoms
• Sensory: predominance of sensory signs/symptoms (hyperalgesia, allodynia)
• Florid: most symptoms present and/or motor symptoms predominant
• Undetermined (chart notes do not allow subtype determination)

First intervention performed:

• Plexus block
• Infraclavicular, axillary, interscalene, popliteal sciatic
• Dose of agent used (mL)
• Adjuvant (yes/no)
• Type of adjuvant: dexamethasone (yes/no), dexmedetomidine (yes/no)
• Sympathetic block
• Stellate, lumbar
• Ultrasound-guided (yes/no), fluoroscopy (yes/no), anatomical landmarks (yes/no)
• Dose of agent used (mL)
• Ketamine infusion
• Dose (mg)
• Duration (min)
• Concomitant use of midazolam (yes/no) If yes, dose (mg)
• Pamidronate infusion o Dose (mg)

Follow-up visit data:

• Analgesic success (yes/no), defined as significant pain reduction that allowed improved function, according to patient report.
• Report of adverse effects (yes/no)
• Type of adverse effect reported: pain increase, severe pain at puncture site, hematoma, infection, nausea/vomiting, nightmares, general malaise, other)

Data from other follow-up visits:

• Number of different analgesic interventions required to achieve significant relief: 1, 2, 3, 4, or no analgesia
• Type of intervention that was beneficial (if the first intervention did not provide relief) STATISTICAL ANALYSES Data will be stored in an Excel file and analyzed with SPSS Statistics software. Continuous data will be analyzed using the Shapiro-Wilk test to assess normality. Normally and non-normally distributed values will be reported as mean (standard deviation) and median (interquartile range), respectively. Categorical data will be reported as number (percentage). The primary objective-the frequency distribution of first-line analgesic interventions-will be reported as number (percentage), as will the analgesic success rate for each intervention. The success rate of each technique by CRPS subtype and symptom duration will then be analyzed using the Chi-square or Fisher's exact test. A p < 0.05 will be considered statistically significant.

To better compare the different techniques, a propensity score approach will be used. The probability of receiving each treatment will be calculated based on the following characteristics: age, sex, symptom duration, limb type, CNESST affiliation, medications used, treating physician, and CRPS subgroup. These probabilities will be calculated using multinomial logistic regression and will define the propensity scores. Each patient's characteristics will be weighted using the Inverse Probability of Treatment Weighting (IPTW) method. After weighting, group balance will be verified by ensuring that standardized mean differences (SMD) for each covariate are below 0.1. Finally, the weighted success rate for each technique will be calculated. Weighted Odds Ratios (OR) with 95% confidence intervals will also be computed using weighted logistic regression for each treatment compared to the reference treatment. Sensitivity analyses will be performed using other weighting methods such as Overlap Weighting, 1:1 Matching, and Entropy Balancing.

To identify the best treatment for each subgroup, the probability of success as a function of treatment, subgroup, and a "treatment × subgroup" interaction term will be calculated using logistic regression. If the interaction term is significantly associated with the probability of success, stratified analyses will be conducted by subgroup. For each subgroup, a propensity score will be calculated using the same method described above, excluding the subgroup variable from adjustment. Weighted success probabilities will be calculated for each treatment and each subgroup, as well as ORs with 95% CIs compared to the reference treatment, to identify the optimal treatment for each subgroup.