A prerequisite for enhanced recovery after hip and knee arthroplasty is optimization of the patient with comorbidities, and several factors have to be considered. The patient has to be prepared properly for the forthcoming anesthesia and surgery. Preoperative optimization of comorbidities may reduce pain from other locations than the operated joint and hence improve overall function (Ayers et al. 2013). A number of medical and psychological comorbidities have been identified to be associated with increased postoperative pain in TKA; among these are low back pain, pain in the contralateral joint, heart disease, depression, and anxiety (Singh and Lewallen 2013).

Also, patients with comorbidities have been shown to stay longer in hospital (to take longer to fulfill functional discharge criteria) and to be less satisfied with pain treatment – indicating a need for special attention (Husted et al. 2008; Williams et al. 2007). However, patients with comorbidities should not be excluded from fast-track surgery as these patients cannot only follow a fast track, but this large subgroup of patients has been shown to benefit the most from optimized treatment (Dowsey et al. 1999; Jørgensen et al. 2013).

Preoperative information should cover specific topics, i.e., pain, to ensure that the patients know what to expect and what is expected of them. Provision of preoperative information can alleviate anxiety regarding the upcoming anesthesia (Jlala et al. 2010) and surgical procedure but not regarding postoperative pain in general (McDonald et al. 2004). Also, in support of this, a recent review found no consistency in the association between patients’ preoperative expectations and treatment outcomes for TKA and THA including pain, function, and satisfaction (Haanstra et al. 2012). However, another review concluded that pain catastrophizing and low preoperative mental health have an influence on outcome after TKA: in follow-ups shorter than 1 year, patients with pain catastrophizing reported more pain postoperatively, and in long-term follow-up, lower preoperative mental health was associated with lower scores on function and pain (SF 12 and SF 36). For THA, only limited, conflicting, or no evidence was found (Vissers et al. 2012). As also the most recent review on the potential influence of preoperative education on postoperative pain found only 1 of 13 studies with a reduction in postoperative pain following preoperative education compared to the noneducation group, it seems that preoperative education is of little if any value to reduce postoperative pain (Louw et al. 2013). Preoperative identification of pain catastrophizers/high-pain responders could allow for a more aggressive pain treatment protocol in these specific cases, but efforts so far have failed to identify these patients by simple means (Lunn et al. 2013a).

The effect of physical prehabilitation on pain has been studied in a few studies. For patients waiting to have a TKA, there was no effect on pain, function, walking speed, or muscle strength, whereas patients waiting for a THA had less pain and better function following an exercise program for 3–8 weeks (Gill and McBurney 2013). The effect of prehabilitation on postoperative pain following TKA was not significant (and neither was there any effect on stiffness, ROM, strength, or function) (Silkman Baker and McKeon 2012). There are no studies specifically addressing prehabilitation in THA and postoperative pain, but although a study found hip extensor and flexor muscle strength and endurance to be associated with postoperative limping, there was no association between the severity of limping and pain – indicating little chance of a pain-reducing effect of prehabilitation in THA (Horstmann et al. 2013).

Fast-track anesthesia should be safe, minimize the impact of the surgical stress response, and facilitate the transition from surgical anesthesia to fast postoperative ambulation (within 1–2 h). However, all safety studies so far are performed with a conventional setup. Studies comparing neuraxial block (spinal, epidural) versus general anesthesia (GA) have found regional analgesia to result in lower 30-day mortality, less in-hospital complications, less bleeding, and less deep venous thromboembolism (DVT) and may also give a reduction in operating room (OR) time (Mauermann et al. 2006; Memtsoudis et al. 2013). Only GA and spinal analgesia are recommended for TKA, whereas both spinal and epidural are recommended for THA by the PROSPECT group (www.​postoppain.​org) and is found to be safe in a large consecutive series (Pumberger et al. 2013).

There is only one published study in fast-track TKA comparing GA and spinal anesthesia focusing on early functional recovery. GA resulted in shorter LOS, less nausea and vomiting and dizziness, less pain after 6 h, and less use of opioids, and patients were able to walk earlier and were also more preferred by patients (Harsten et al. 2013). The best choice of anesthesia for fast-track THA and TKA is thus debatable: GA may facilitate improved early functional recovery including less pain and opioid use, whereas spinal analgesia may have a better safety profile – studies on safety in fast-track THA and TKA using GA are lacking.

Surgical pain originates from the surgical wound and a rational approach to perioperative pain treatment has been directed toward the use of local infiltration analgesia (LIA) at the site of surgery. LIA has gained popularity due to the simplicity, low price, effectiveness of the technique, and the fact that the quadriceps strength remains unaffected thus facilitating early ambulation. Most of the more than 25 published randomized studies, of which only few are placebo controlled, have insufficient design hindering interpretation and do not allow comparison with other analgesic techniques. There is little evidence to support the use of LIA in THA either intraoperatively (Lunn et al. 2011a) or with a continuous postoperative wound infusion (Andersen et al. 2011; Solovyova et al. 2013) – when used in conjunction with multimodal opioid-sparing analgesia. In TKA, the data support the intraoperative use of the LIA technique (Andersen et al. 2008a), but there are limited data supporting the use of wound catheter administration. The analgesic effect is prolonged by using a compression bandage (Andersen et al. 2008b). LIA is effective in providing immediate pain relief – without motor blocking – allowing immediate mobilization postoperatively. This could result in less thromboembolic episodes and allow for early functional recovery (Husted et al. 2010). Various drugs are used for LIA, mostly bupivacaine and ropivacaine providing low pain scores for 6–8 h, but a new liposomal form of bupivacaine may provide this effect for up to 72 h which could further facilitate early postoperative recovery.

Peripheral nerve blocks may also like LIA reduce postoperative pain effectively but with accompanying motor problems making early mobilization difficult and risky (7 % falls) (Ilfeld et al. 2010). Also, pain and function is not improved beyond 24 h by a femoral block (Ilfeld et al. 2011) and hence will not ease early functional recovery beyond the first postoperative day. A new modality in the form of an adductor canal block may be promising in preserving quad control better than a femoral nerve block with equivalent pain control – but also technically demanding and time consuming (Jæger et al. 2013).

Steroids in the form of methylprednisolone (125 mg) have been found to be extremely potent in reducing postoperative pain in both THA and especially in TKA for 24–32 h following a single shot given during the operation (Lunn et al. 2011b, 2013b). Early functional recovery is also facilitated by less opioid use, less nausea, and less fatigue.

Postoperative pain following fast-track THA (LOS <3 days) with a standardized multimodal opioid-sparing regime is at acceptable levels with concomitant low use of opioids in >95 % of patients before day 10 (Andersen et al. 2009). Pain following TKA is more pronounced: 52 % of patients reported moderate pain, and 16 % severe pain when walking 1 month after surgery with a concomitant increase in the use of strong opioids (56 % used weak or strong opioids on day 30) (Andersen et al. 2009). In a fast-track setting, it has been shown that 90 % of the TKA patients are able to walk independently on the first postoperative day and all managed to walk on postoperative day 2 with a maximum of moderate pain intensity (Holm et al. 2010). Still, pain is a challenge also in the early postoperative phase as a study found pain to be one of the three main clinical reasons (the others being dizziness and muscle weakness) for patients to be hospitalized following fast-track THA and TKA after 24 and 48 h (Husted et al. 2011).

The key to successful pain treatment in fast-track THA and TKA is multimodal opioid-sparing analgesia, where different drugs are combined to produce a synergistic pain-reducing effect with few side effects and with a low need for opioids (and as rescue medication) only. This has lead to the avoidance of delirium (Krenk et al. 2012) and a reduction of postoperative cognitive dysfunction with a reduction of postoperative pain to preoperative levels by day 5–9 (Krenk et al. 2013). No or only reduced use of opioids are the goal as opioids may slow recovery by its side effects (dizziness, sedation, nausea). A number of different drugs can be used to reduce both pain and the need for opioids, spanning from paracetamol and NSAIDs/COX-2 inhibitors over ketamine, alpha-2 agonists, serotonin-specific reuptake inhibitors (SSRI), to gabapentinoids, and more will come. Paracetamol and a COX-2 inhibitor have been shown to reduce pain and the need for opioids (Maund et al. 2011) and COX-2 inhibitors also in addition to reduce inflammation and increase range of motion (ROM) (Schroer et al. 2011) without compromising prosthetic fixation (Meunier et al. 2009). Gabapentinoids have been shown efficient in reducing postoperative pain across specialties (Tiippana et al. 2007) but in a recent study in TKA failed to reduce pain or opioid use (Paul et al. 2013). Procedure-specific dose-response studies are warranted.