This was an epidemiological observational, cross-sectional, national study carried out among French rheumatologists between 2006 and 2007, after obtaining approval of french national ethical committee (CCTIRS). Each participating rheumatologist was asked to record data for all consecutive adult patients who underwent a diagnostic or therapeutic procedure (puncture, or infiltration, or both) at the time of consultation, irrespective of localisation. The observation period extended over 1 month of visits.

Rheumatologists were selected from the French rheumatologist' representative TVF database (source CEGEDIM group). In order to guarantee a high level of respondents, details of the study and a request to participate were sent out to all rheumatologists across France in two mail-shots. This enabled some adjustment in the second mailing according to the level and regionality of the response to the first. The rheumatologists in the first mail-shot were chosen by drawing lots stratified according to administrative region.

The experience of each rheumatologist was determined indirectly from their age, assuming that they had started practising professionally when they were 27-years-old. The study was proposed to 1800 French rheumatologists with private practice. After 2 contacts and propositions, 339 rheumatologists accepted to participate. Finally, of these 339 rheumatologists, only 252 rheumatologists have participated actively in the study and have recruited patients. The demographics of these rheumatologists were compared to the global demographics of French rheumatologists and did not demonstrate significant difference, according to age, sex-ratio, number of years of practicing, location.

All consecutive adult patients (>18 years of age) suffering from a non-malignant musculoskeletal pathology for which a diagnostic and/or therapeutic procedure (puncture, or infiltration, or both) was carried out at the time of consultation were recruited (maximum of 40 patients per rheumatologist). As this was an observational study, the only exclusion criteria were mesotherapy or acupuncture procedures. All patients provided informed consent before procedure.

The following data were recorded for each patient: age; sex; underlying pathology for which the procedure was required; previous experience of the patient regarding a similar type of procedure; details of the procedure 20 performed (type, localisation, aim, planned or not); analgesic protocols (preventative, post-procedure) used; evaluation of pain before procedure and during procedure, and of the patients' experience post-procedure.

Pain intensity was evaluated using verbal scale (none, mild, moderate or severe). The verbal rating scale is a commonly used scale, easier to understand than Visual Analog Scale. It has been validated in many studies and is one of the three most commonly used scale 21.

Patient's pain intensity was assessed immediately before the procedure, considering mean pain intensity during the previous week before procedure. Then, after the procedure, patients had to report to the physician their level of pain induced by the procedure,.

The primary objectives of the study were to evaluate: (i) the respective frequency of different articular and peri-articula r procedures in rheumatological primary practice; (ii) the frequency and intensity of pain linked to these different procedures; (iii) the demographic factors for procedural pain. (iv) the type and frequency of analgesic treatment (preventative and post-procedural) administered for procedural pain;

Statistical analysis was carried out in order to compare patients who had suffered from pain whilst undergoing the procedure to those who had not. Explanatory analyses were conducted with the characteristics of the patients to explain the existence of pain. When patients had more than one procedure during the visit, each procedure was assessed separately.

For descriptive data, quantitative variables are presented as number (n), means and standard deviation (SD), and qualitative variables as n and % of available data. For the descriptive and univariate analysis, quantitative variables were analysed using the Wilcoxon or Mann-Whitney test, and qualitative variables were compared using the Chi²-Pearson or Fisher's exact test. The level of significance selected was 0.05.

Multivariate analysis was carried out by stepwise logistic regression with the dependent variable being the presence or absence of pain. Significant factors to the threshold of 20% in univariate analysis were included in multivariate analysis. Entry threshold was set-up at 10% and conservation threshold to 15%.

Statistical analysis was carried out using SAS^® (version 8.2; SAS Institute, North Carolina, USA).

A total of 240 French rheumatologists took part in this study (57.7% male, mean age 49 years). 68.1% had exclusive private practice, 31.4% had a mixed hospital/surgery practice and the rest were practicing exclusively in hospital. The rheumatologists each had an average of 22 years experience.

Data were analysed for 8446 patients who underwent a puncture, or infiltration, or both, during visit to 240 rheumatologists across France. The mean age of the patients was 62 ± 14 years and the majority were female (64.4%). The demographic characteristics of the patients are summarised in Table 1.

In 31.3% of the cases patients had joint osteoarthritis, tendonitis in 18.1%, radiculalgia in 9.1%, back pain in 7.2%, inflammatory rheumatological disorder in 3.8% (mostly rheumatoid arthritis), and a tunnel syndrome in 4.4%. In 66.8% of the cases, patients were prescribed analgesics for the pain related to the underlying pathology.

In total, 91.8% of patients underwent only one procedure. The knee was the main site of intervention (45.4%), followed by spine (19.1%), shoulder (15.0%) and small joints (12.0%). The procedure was unilateral in 91.6% of patients, intra-articular in 57.0%.

The procedure was carried out for therapeutic purposes in the majority of patients (91.8%), and was diagnostic and therapeutic in 6.8%. Over half of the patients (56.1%) were given a corticosteroid injection, while 26.0% received an injection of hyaluronic acid. In the remaining cases, procedure was a synovial fluid aspiration or combination of aspiration and injections (Table 2). Approximately half (53.9%) of the interventions were planned before the consultation.

The majority of patients (81.7%) experienced some degree of pain induced by the procedure. Pain was considered to be severe by 5.3% of patients, moderate by 26.6%, and mild by 49.8%. A small proportion of patients (18.3%) experienced no pain (Figure 1).

Univariate analysis demonstrated that the intensity of pain varied significantly associated with the site of the injection/aspiration. Pain was considered to be more intense (moderate to severe) in 41% of cases involving the small joints, 32% involving the head and spine, 30% for the shoulders and 28% for the knees (Figure 2). Other factors like age (p < 0.001), sex of the patient (p < 0.001), type of underlying rheumatological pathology (p < 0.001), pain in the previous week related to this underlying pathology (p < 0.001), analgesic treatment for pain related to this underlying pathology (p < 0.001), procedure planned before the visit (p = 0.001) and patient's procedural perception evaluated by the physician (p < 0.001) were significant factors associated with procedural pain.

On the contrary, univariate analysis demonstrated that actual flare of the disease, history of previous procedure, number of procedures performed on the same visit, body-side of the procedure, type of injection (intra vs extra-articular), aim of the procedure (therapeutic vs diagnostic procedure) were not significant factors for the occurrence of procedural pain.

From this univariate analysis, fourteen demographic and pain ratings explanatory variables linked to the occurrence of pain during the procedure (with significant threshold of 20% in univariate analysis) were included in the multivariate analysis (Table 3). It demonstrated significant differences for intensity of pain linked to the underlying pathology during the week preceding the procedure (p < 0.0001), site on which the intervention was required (p = 0.0006) and the patient's procedural perception evaluated by the physician (p < 0.0001). Thus, pain was considered to be significantly and independently more intense in patients with more severe pain linked to their underlying pathology, for procedures performed in small joints, globally in accordance with the perception of the procedure by the rheumatologists. There was a trend towards more intense pain in women, but not significant. The use of preventative analgesic protocol was also significantly associated with the presence of procedural pain in univariate analysis (p = 0.005), demonstrating that physicians were taking into account the possibility of that induced pain.

On the whole, rheumatologists thought that the majority of their patients had tolerated the procedure very (53.1%) or quite (39.8%) well, and that only 0.7% of patients had suffered badly during the procedure (Figure 3). This was in accordance with the intensity of pain experienced during the procedure by their patients (kappa score: 0.44).

Less than half of the patients received any analgesia (either preventative or curative) for their pain. Only 5.7% of patients were given preventative analgesic medication, usually local anaesthesia (34.8%), xylocaïne patches (20.5%), acetaminophen alone (14.3%; Step 1 analgesics I) and Step 2 analgesics (11.2%). In most patients this was initiated by the physician (82.6% of cases) rather than at the request of the patient. The mean duration between receiving preventative analgesia and the procedure was 38 ± 74 min. There was no significant difference in the use of preventative analgesia between patients undergoing a single procedure and those undergoing two (p = 0.683).

Post-procedure analgesia was given to 36.3% of patients and was initiated by the physician in 94.6% of the cases. The most common analgesics prescribed were step 2 analgesics (weak opioids combined to acetaminophen) (35.8%), non-steroidal anti-inflammatory drugs (NSAIDs) (34.8%) and acetaminophen alone (34.2%).