In France, for example, approximately 68,000 total knee replacement (TKR) operations were carried out in 2012 and the number of operations is rising every year. The majority of the patients involved are women with gonarthrosis and the average age is around 70. A number of these patients also have a body mass index which is higher than 30. The chronic pain that occurs after this type of surgery (CPSP) has only been described and studied over about the last 10 years, mostly because it impairs the overall functional outcome and the patient’s perception of the quality of the operation, as well as the corresponding change in quality of life [1].

On the basis of a transverse study carried out in almost 500 patients who were surveyed using a questionnaire 3 or 4 years after TKR, Wylde et al. showed that 44 % of them presented with CPSP, of which 15 % had a severe form [2]. Liu et al. reported more recently, based on a transverse study in more than 1,000 patients, that the incidence of CPSP was 53 % after TKR, measured 1 year after the operation [3]. Also 1 year after surgery, a pharmaco-epidemiological study in almost 2,000 patients after TKR showed a high frequency of analgesic use (47 %), including medications for neuropathic pain (8.6 %) and strong opioids (5.6 %) [4].

Like the clinical aspects, the mechanisms of CPSP after TKR have only been studied by a very small number of published works: of necessity these are referred to here indirectly, together with the risk factors. Based on the retrospective study carried out by Rousset [5] and the orthopaedic literature [6, 7], certain mechanisms can, however, be highlighted which are commonly implicated:

While the mechanical and non-mechanical aetiologies which are orthopaedic in nature result in replacement of the prosthesis in the short or longer term, which in most cases correspondingly leads to resolution of the pain [6–8], a number of patients whose prosthesis is satisfactory from an orthopaedic point of view still have chronic pain: in these cases the mechanism causing the pain may be neuropathic or may not be determined. Taking into account the importance of proprioceptive afferents from the knee and on the basis of certain clinical descriptors [2, 5], it might be suggested that a “phantom knee” phenomenon occurs in some patients, particularly those with specific preoperative risk factors: pain at rest, catastrophising and anxiety.

The risk factors for the occurrence of CPSP after TKR are the clinical aspects that have been most closely studied: these are no different from the risk factors for CPSP that have been identified in relation to other types of operation but they have been described particularly well in the case of TKR. On the individual level these essentially appear to be linked to somatic and psychiatric comorbidities both before and after surgery and more broadly to those associated with the operation itself. The degree of CPSP risk that depends on the factors that have been identified does, however, appear to be only modest according to the literature (odds ratios between 0.9 and 1.4), while it is higher in terms of pre- and postoperative pain (odds ratios between 1.4 and 2).

Very few studies have addressed the clinical description of CPSP after TKR and most of these have used validated diagnostic tools to assess the neuropathic nature of the chronic pain.

Wylde et al. reported that in patients with CPSP 3 to 4 years after TKR the descriptors most commonly reported in the simplified Mac Gill questionnaire (aching, tender, tingling), and after using the PainDetect questionnaire they stated that neuropathic pain is probably present in 13 % of patients with pain or 6 % of patients who have undergone operations [2]. Nevertheless, while the descriptors identified using the Mac Gill questionnaire do not suggest neuropathic pain, the PainDetect questionnaire was developed for use in chronic lumbo-sciatic pain. Buvanendran et al. used the S–LANS questionnaire to show the preventative effect of pregabalin given perioperatively on the incidence of neuropathic pain at 3 and 6 months [35], although this validated scale is still moderately sensitive. Finally, Rousset showed using a descriptive questionnaire distributed 3 months after knee arthroplasty that analysis of patient records based on the DN4 questionnaire indicated a prevalence of CPSP of 48 %, with a neuropathic component in 42 % of cases, which was responsible for high medication use, an increase in sleep disorders and prolonged periods of time off work [5]. Recent studies in other types of operation suggest that the search for signs of nerve damage and symptoms suggestive of neuropathic pain during the early and late postoperative period (3–6 weeks), particularly using the DN4 questionnaire [36], should be done routinely to allow early recognition and appropriate treatment and therefore limitation of the harmful effects of chronic pain [37, 38].