Local infiltration analgesia: a 2-year follow-up of patients undergoing total hip arthroplasty

Total hip arthroplasty (THA) is a common and standardized surgical procedure. Postoperative mortality after hip joint replacement is low, but a number of complications remain, including persistent post-surgical pain (PPSP) [1], hip dislocation [2], infection [3], and deep vein thrombosis [4]. Strategies to reduce morbidity and mortality include the posterior surgical approach, mechanical and pharmacological prophylaxis of deep vein thrombosis, and the use of spinal anesthesia [5]. One of the important factors affecting patient satisfaction with THA is good postoperative pain management [6]. Poorly managed postoperative pain can lead to chronic post-surgical pain and, therefore, aggressive postoperative pain management is important [7]. Local infiltration analgesia (LIA) using a combination of large-volume local anesthetics (LA) and non-steroidal anti-inflammatory drugs (NSAIDs) injected systematically peri-articularly has been used for pain management with variable success [8]. In two earlier studies, we found improved analgesia when using LIA as compared to intrathecal morphine (ITM) during the first few days postoperatively after knee and hip arthroplasty [9, 10].

The use of NSAIDs administered orally or systemically during orthopedic surgery has been controversial due to the risk of delayed healing from the inhibition of osteoclast and osteoblast formation [11], specifically during fracture surgery. LA injected intra-articularly have also been suspected to cause chondroplasia in the presence of intact cartilage [12]. Thus, delayed healing and functional recovery, local anesthetic toxicity, PPSP, and infection may be some of the risks associated with the use of LIA. We hypothesized that LIA has no negative effects on post-discharge and chronic post-surgical pain, mobility, recovery of function, quality of life, or major postoperative complications. Therefore, the primary aim of our study was to assess hip function and the secondary aims were to assess quality of life, post-discharge and PPSP at 6 months after surgery and all major complications at up to 2 years after surgery.

Approval for this study was obtained from the Regional Ethics committee in Uppsala as well as from the Swedish agency on drugs prior to patient recruitment for the initial larger study. The initial larger study was registered in Clinicaltrials.gov (Identification number: NCT01281891) and monitored by an independent organization based in the hospital for quality control; it was conducted in accordance with Good Clinical Practice. Patients were recruited from the Department of Orthopedic Surgery, University Hospital, Örebro, Sweden between 2011 and 2012. The study reported here is a secondary analysis of the data from the initial study on post-discharge events, including hip function and quality of life. These data are therefore part of the larger study but have never been published; in contrast, the primary data on early postoperative recovery and postoperative pain management have been published previously [10].

A total of 80 patients in the age group 50–85 years, with ASA physical status I–II, who were undergoing elective THA were randomized in this prospective, double blind, parallel group, longitudinal study. Written and verbal informed consent was obtained from all patients prior to inclusion. The exclusion criteria were chronic pain requiring opioid medication, known allergy to local anesthetics, contraindications for using NSAIDs or receiving regional anesthesia, and inability to follow verbal or written instructions. Patients were allocated to one of two groups, namely, Group ITM (standard of care at our hospital) and Group LIA, according to a computer-generated randomization sequence by the hospital pharmacy. Thus, double-blinded syringes were sent to the operating room with the drugs/placebo according to the randomization sequence.

A total of 80 patients were randomized, but two patients were subsequently excluded from the analysis, one in each group: one patient decided not to continue the study after randomization (Group ITM) and the other patient left the study because the spinal anesthetic did not achieve adequate surgical analgesia (Group LIA) (this patient subsequently received general anesthesia). Three additional patients were excluded from the analysis of the post-discharge follow-up: two patients in Group ITM (one with suspected postoperative infection and the other patient did not wish to continue with the study) and one patient in Group LIA (postoperative re-operation due to loosening of the cup) (Fig. 1). The age (66 ± 9 and 68 ± 8), weight (83 ± 23 and 86 ± 20), and ASA I/II physical status (12/25 and 14/24) were comparable between patients in Groups ITM and LIA, respectively.

No statistical differences were found in the NRS (pain now and average pain during the preceding week) between the groups during postoperative weeks 1–4 (Fig. 2a, b). In general, average pain scores as assessed with the NRS were low (<3) in both groups at all measurement points.

No significant differences were found in rescue analgesic consumption between Group ITM and Group LIA during postoperative days 8–14 [median (range); paracetamol: 15 (range 0–34) vs. 14 (0–28) g, respectively (p = 0.417); tramadol: 250 (0–2100) mg vs. 0 (0–2100) mg, respectively (p = 0.314)]. The incidence of persistent post-surgical pain (NRS > 3 or HOOS > 30) in the two groups is shown in Table 1. No statistically significant differences were found in the incidence of PPSP between the groups at either 3 or 6 months (p = 0.115). Two patients (6%) in group ITM had pain (NRS > 0) during walking at 6 months following surgery.

One patient in the LIA group had a complete luxation of the hip joint on postoperative day 35 and one patient had sub-luxation on postoperative day 15. Two patients in the ITM group were re-admitted because of pain and difficulty in mobilization on postoperative day 10 and after 4.5 months, respectively, and one had sub-luxation at postoperative day 14. Additionally, one patient in the ITM group had deep vein thrombosis after 4.5 months, and one in Group LIA was admitted to the hospital due to a suspected allergic reaction to diclofenac after 3 months. No patient in either group had surgical site infection. Other significant side effects during the 2-year follow-up are presented in Table 2.

There was an overall improvement with time in the HRQoL as assessed using the EQ-5D values (p < 0.001), with a significant interaction (group × time) (p = 0.036); the latter indicated somewhat higher scores (= better health) in Group LIA (Table 3). There was an overall improvement in the summary measure of the physical component of the SF-36 (PCS) over time (p < 0.001) (Table 3), but no interaction was found between group × time. No difference in the MCS of the SF-36 was found over time, and no group × time interaction was seen in the MCS (Table 3).

A significant improvement was found over time in all components of HOOS (pain, symptoms including stiffness and range of motion, activity limitation in daily living and hip-related quality of life) (p < 0.001), with the exception of the sport and recreation function (Table 4). A significant overall interaction (group × time) was seen only in HOOS-Pain (p = 0.004), with somewhat lower HOOS-Pain in Group ITM, specifically at 3 months, but there was no statistical difference between the groups.

In this study, we found no differences in pain intensity or analgesic consumption after home discharge between patients receiving LIA and those receiving ITM after THA. Functional recovery was similar between the two treatment groups for up to 6 months after surgery. No difference in PPSP was found between the groups at 3–6 months following surgery, and the incidence of post-surgical complications was similar between the groups up to 2 years later. A significant improvement in HRQoL was found using the EQ-5D tool but not with the SF-36 in the LIA group, postoperatively. The small difference seen between the groups in these endpoints has minimal clinical importance, and we therefore believe that LIA is a safe technique when used for the management of postoperative pain following THA.

We used a specific instrument for the assessment of hip function, i.e., the HOOS. This questionnaire has been used in patients with hip diseases with or without arthritis [13]. It is used to assess the impact of treatment of hip diseases as well as to describe the consequence of exposure of a population of patients to a treatment or management strategy. The Western Ontario and McMaster Universities Arthritis Index score has also been used in this setting, but we found that the HOOS is a simple and easily understood questionnaire, and it has been widely used in the Swedish population. In the present study, patients in both groups showed a steady improvement over time in all components of the HOOS score, with the exception of the dimension sports and recreation. All patients were back to normal function and activities related to the hip after 6 months. No inter-group differences were found, except in the pain component of HOOS for which a significantly greater number of patients in the LIA group had a pain score of > 30. This aspect is discussed in greater detail in the following text.

Orthopedic surgeons have often been concerned about the use of NSAIDs in patients undergoing orthopedic procedures [14]. Since prostaglandins play an important role in the regulation of osteoblast and osteoclast formation, inhibition of prostaglandin production due to the use of NSAIDs may retard bone formation [11] or lead to renal insufficiency, specifically in elderly patients. Therefore, NSAIDs may be expected to have significant consequences when used in several clinical situations where bone formation or remodeling is an important factor in healing, as during bone fractures [15] or non-cemented hip prostheses. However, it is important that the advantages of NSAIDs for pain relief should be weighed against the possible risk for delayed healing, and caution should be exercised during the first two postoperative weeks, specifically after fracture surgery. In a study on rats, celecoxib therapy was found to worsen the mechanical properties of callus formation during the early phase of fracture repair [16], possibly due to inhibition of angiogenesis [17]. However, no clinical evidence of prosthesis migration or differences in pain scores, range of motion, and subjective outcome were found after 2 years when comparing celecoxib with placebo in patients undergoing total knee replacement [18]. In the present study, two patients in the LIA group (5.4%) who were administered peri-articular injections of ketorolac experienced prosthetic dislocation during the 2-year post-discharge follow-up. In contrast, in a multi-centre study from France that involved > 2000 primary hip arthroplasties, 10.4% of all patients who underwent revision arthroplasty had hip dislocation [2]. This is similar to the results obtained from the Swedish arthroplasty register where 8.7% of patients underwent revision arthroplasty due to dislocation. In our study, a total of three patients had hip dislocation (4%), with no inter-group difference.

The incidence of PPSP, as measured by NRS, and rescue pain requirement was similar between our two study groups during the 6-month follow-up. We used pain intensity exceeding 3 on the NRS or 30 on the pain component of HOOS when defining PPSP. There is no clear international definition of pain intensity when assessing PPSP and therefore the reported incidence varies considerably (27–38%) in patients undergoing THA [1]. The low incidence of PPSP in our study (<10% on NRS) could be due to our use of a more strict definition, i.e., pain intensity > 3 on NRS or > 30 on the pain component of HOOS. PPSP has been a matter of concern, and several studies have recently highlighted this problem [1, 7]. Prior to the start of our study, we were concerned that the use of NSAIDs in the LIA group may manifest as PPSP syndrome due to delayed healing. However, PPSP does not seem to have been a clinical problem in the small group of patients we studied over 2 years. Thus, despite theoretical arguments as well as results from animal studies suggesting that NSAIDs should not be administered during orthopaedic surgery, we found that two doses of ketorolac given peri-articularly had no clinically relevant adverse effects while offering good postoperative analgesia.

Local anaesthetics have been used for the management of postoperative pain and have been injected in different tissue planes with good effect [19]. Although a single dose of LA appears to have only a short effect-duration [20], the use of catheters and the intermittent injection of LA peri-neurally, subcutaneously, intra-fascially, as well as intra-abdominally have prolonged the duration of analgesia [9, 21]. Despite the fear of systemic absorption when using large doses of LA, as during LIA, toxic plasma concentrations have not been measured, and only rarely have adverse effects been described [22].

We used two different questionnaires to assess HRQoL. The EQ-5D is a standardized method to assess quality of life in five dimensions. The questionnaire is well validated, but it is non-specific and can therefore be used in different situations [23]. The average value on the EQ-5D provides an assessment of patient-assessed health and quality of life, with high scores suggesting good health. We found that there was a slight deterioration in the scores on the EQ-5D at 7 days postoperatively compared to preoperative values, with higher scores in the LIA group. Thereafter, there was a steady improvement, and at 6 months, patients had reached a high quality of life suggesting a significant improvement in perceived health, compared to preoperative values. Similarly, we also found an improvement over time in patient-assessed HRQoL using the physical component score (PCS) of the SF-36 in both groups, but not in the mental component score (MCS), suggesting that physical improvement in health may not necessarily translate into mental improvement.

Functional recovery after THA did not differ between patients receiving LIA and those receiving ITM at 6 months following surgery. Post-discharge pain, analgesic consumption, and PPSP were similar between the groups. HRQoL at 6 months and the incidence of adverse events up to 2 years following the surgery were also similar between the groups. LIA would appear to be a good alternative to ITM for postoperative pain management following THA and to be associated with mild post-discharge pain, satisfactory functional recovery, a good HRQoL, a low risk for PPSP at 6 months and a low incidence of adverse events up to 2 years after surgery.