Optimal pain management is important after hip fracture treatment to ensure patient comfort and to facilitate rehabilitation. However, postoperative pain is frequently undertreated in elderly patients1, and inadequate pain control after hip fracture is associated with prolonged bed rest, interruptions in physical therapy, and delays in walking, potentially increasing the risks of thromboembolism, functional impairment, and longer hospitalization2-4.
Multimodal pain management with preemptive medication and intra-articular and periarticular injections during total joint arthroplasty involves the use of multiple agents that act on different regions of the pain pathway and provide appropriate pain relief with less reliance on opioids5. Multimodal pain management has the potential to minimize postoperative pain while reducing total opioid consumption with its potential associated adverse effects6-9. However, previous studies of multimodal pain management have been performed on patients who had undergone elective surgery6-9.
Elderly patients sometimes respond poorly to questionnaires, including the visual analog scale (VAS) for pain. Delirium after hip fracture in elderly patients is common, and 28% to 61% are unable to respond appropriately to questionaires10-14. In the present study, we developed a computerized method to record the frequency of use of patient-controlled analgesia to provide an objective method to measure severity of pain in patients who could not respond appropriately.
This prospective, randomized, controlled study was conducted to determine whether multimodal pain management provides substantial clinical benefits with regard to pain level, functional recovery, total amount of fentanyl used, frequency of use of patient-controlled analgesia, patient satisfaction at discharge, and perioperative complications. In addition, we compared the correlation between frequency of use of patient-controlled analgesia and pain level.
The design and protocol of this prospective study were approved by the institutional review board at our hospital, and all patients provided informed consent preoperatively. The study was registered in the ClinicalTrials.gov Protocol Registration System (trial number NCT01112436).
One hundred and ten consecutive patients (age range, sixty-five to ninety-two years) who had undergone bipolar hemiarthroplasty after hip fracture from March 2010 to April 2011 were considered eligible for this study. The exclusion criteria applied included mental illness, intolerance or allergy to any of the drugs used in this study, and refusal to participate. Decisions to enroll or exclude patients were made by an investigator who did not otherwise participate in the study.
In this randomized, single-blinded study, randomization into one of the two study groups was performed with use of Microsoft Excel (Microsoft, Redmond, Washington) to generate random numbers. Group allocations were made by a statistician, who also did not otherwise participate in the study, and were unknown to the investigators and patients. Allocations were kept in a set of sealed envelopes. One hour prior to surgery, the appropriately numbered envelope was opened and the card inside determined the group allocation.
The patients were divided into two groups: Group I, in which they received preemptive analgesic medication and intraoperative periarticular injections, and Group II, in which they did not. The patients assigned to Group I received oxycodone SR (10 mg) and celecoxib (200 mg) with 10 mL of water one hour before surgery as well as intraoperative periarticular injections, whereas those in Group II did not receive medication or intraoperative injections.
For Group I, a 100-mL periarticular injection cocktail (300 mg of ropivacaine [40 mL], 10 mg of morphine sulfate [10 mL], 30 mg of ketorolac [Troloc] [1 mL], 300 μg of 1:1000 epinephrine [0.3 mL], 1000 mg of cefmetazole [10 mL], and normal saline solution [38.7 mL]) was prepared in two 50-mL syringes, and 20 mL of the cocktail in the first syringe was injected into the anterior aspect of the capsule before femoral stem insertion. Before joint capsule closure, the remaining 30 mL was injected into the synovium, capsule, and gluteus muscles. After joint capsule closure, 50 mL of the periarticular injection cocktail in the second syringe was infiltrated into the subcutaneous tissue, the capsule, the fascia lata muscle, and the periosteal area of the fracture site for intertrochanteric fractures. All surgical procedures were performed with the patient under spinal or general anesthesia.
All surgical procedures were performed by a single surgeon (Y.-C.H.) using a posterior surgical approach, and a cementless bipolar hemiarthroplasty prosthesis (Bencox; Corentec, Seoul, South Korea) was implanted in all cases.
All patients were managed with the same perioperative protocol. Oral aspirin or rivaroxaban was prescribed for at least ten days and up to thirty-five days postoperatively. Above-the-knee elastic stockings were placed on all patients perioperatively and worn for a minimum of two weeks, and a calf sequential compression device was used for each patient15. All patients were allowed to walk or stand as tolerated using a walker or crutches on the second day, and a standard postoperative rehabilitation protocol was adopted. To control postoperative pain, intravenous fentanyl was administered with use of a computerized intravenous patient-controlled analgesia system (AUTOMED 3300; ACE Medical, Seoul, South Korea). The patient-controlled analgesia was set at a bolus of 0.1 μg/kg with a lockout time of fifteen minutes and continuous infusion of 0.1 μg/kg/hr (total regimen, 10 μg/kg/100 mL). The patients were instructed to push the button of the patient-controlled analgesia system each time they felt pain. When a patient complained of persistent pain with a VAS pain score of >3, an additional 50 μg of fentanyl was injected intravenously by the investigator or duty nurses until a VAS pain score of <3 was achieved. The patients and a clinical investigator (J.-S.L.), who prospectively collected demographic data and preoperative clinical details using predesigned datasheets and entered them into a database with all clinical information, were unaware of group identities until the final data analysis.
Preoperative clinical status was evaluated with the surgical risk American Society of Anesthesiologists (ASA) rating scale16, and postoperative walking ability was assessed with the categories described by Koval et al.17. The presence of preoperative cognitive impairment/dementia was determined within twenty-four hours after admission with use of the Mini-Mental State Examination (MMSE)18 and Delirium Rating Scale (DRS)19. The determination of a cognitive impairment was defined as an MMSE score of <24 or a DRS score of >10.
The primary outcome variables were pain levels on postoperative days one, four, and seven and at discharge. Before surgery, the patients were instructed to use a 10-cm VAS with end points labeled “no pain” and “the worst possible pain” to assess maximum pain level (VASmax) during the assessment day and minimum pain level (VASmin) during the assessment day. At a fixed time (5 p.m.) on each assessment day, VAS scores were collected by one blinded investigator with more than two years of experience interviewing patients about postoperative pain. In addition, the VAS score after fentanyl administration was collected by a duty nurse or the patient’s caregivers thirty minutes after injection.
The secondary outcome variables were frequency of use of patient-controlled analgesia, total fentanyl use, delirium, patient satisfaction, wound complications, drug-related side effects, and functional recovery. The frequency of use of patient-controlled analgesia and total fentanyl use (summed total of fentanyl delivered by patient-controlled analgesia plus additional intravenous injections of fentanyl) were assessed at twelve hours, between twelve and twenty-four hours, between twenty-four and thirty-six hours, between thirty-six and forty-eight hours, and between forty-eight and sixty hours after surgery.
Levels of overall patient satisfaction were determined on the day of discharge with use of a 10-point scale. The wound complications noted were delayed wound-healing, prolonged wound drainage, and wound infections. Patients were closely observed postoperatively for narcotic and ropivacaine-related side effects. The narcotic-related side effects noted were nausea, vomiting, pruritus, urinary retention, and respiratory depression. The ropivacaine-related side effects were blurred vision, hearing problems, peripheral paresthesia, dizziness, uncontrolled muscle contraction, convulsion, hypotension, bradycardia, headache, and itching.
Statistical Analysis
To estimate the required cohort size, a pilot study was conducted by measuring the VAS pain scores on postoperative day one for ten patients, half of whom had received multimodal pain management and half of whom had not. The standard deviations of the VAS scores at postoperative day one were 3.1 and 2.5 for these patients. Setting a requirement to detect a VAS-score difference of 1 point between the two study groups, with α = 0.05 (one-tailed) and a power of 80%, we determined that ninety-eight patients were required per group.
Two interim analyses were prospectively planned for when 25% and 55% of the patients were enrolled. The first interim analysis was done in August 2010, with a stopping-guideline p value of ≤0.001. The data safety monitoring committee reviewed the results and recommended that the study continue. The second interim analysis was done in April 2011, with a stopping-guideline p value of <0.02 based on the Lan-DeMets error spending function with O’Brien-Fleming-type boundaries. The data safety monitoring committee reviewed the data and noted that the primary outcomes, including the postoperative pain scale score, had reached significance.
We used an intention-to-treat strategy, with all participants included in the analysis irrespective of whether they had completed the study. Two options were considered for dealing with missing values: baseline values carried forward for missing data and missing data replaced with the average of the other group at that time point.
Koval categories for walking levels were dichotomized as ambulatory (activity levels I, II, and III) or housebound (activity levels IV, V, VI, and VII). For intergroup comparisons, we first tested data distributions for normality using the Shapiro-Wilk test. Normally distributed results are presented as means and standard deviations, and the Student t test was used to perform comparisons. Non-normally distributed results were expressed as medians and interquartile ranges and were analyzed with use of the Mann-Whitney U test. Spearman correlation analysis was used to examine the relationship between the total number of times that the patient-controlled-analgesia button was pushed by the patient and the VASmax values and between the total number of times that the patient-controlled button was pushed and the VASmin values.
The chi-square or Fisher exact test was used to analyze descriptive variables, as appropriate. Results are reported as means and standard errors. P values of <0.05 were considered to indicate significance, and the analysis was performed with use of PASW Statistics-18 software (IBM, Armonk, New York).
Source of Funding
Financial support for this study was obtained from Kolon Pharmaceuticals, South Korea.
Although the planned sample was 196 patients, patient recruitment was stopped at 110 as the primary outcomes (postoperative pain scores) had reached significance.
Of the 110 patients eligible for the study, twenty-eight were excluded: twelve because of preoperative mental illness, four because they refused to participate, and twelve because they could not complete the questionnaire. Of the eighty-two patients included, forty-three were randomized to Group I and thirty-nine, to Group II (Fig. 1).
Age, sex, height, weight, body mass index (BMI), ASA class, and operative time did not differ between the two groups (see Appendix). The mean VASmax score on postoperative day one was 4.8 in Group I and 5.8 in Group II (p = 0.008). Despite the administration of rescue analgesic, the mean VASmax score in Group II was >5 on postoperative day one. The highest pain levels in both groups were experienced on postoperative day one and gradually diminished with time. The mean VASmax score on postoperative day four was 3.8 in Group I and 4.8 in Group II (p = 0.014). However, no significant difference was observed between the two groups on postoperative day seven or at discharge (Fig. 2). The VASmin scores on postoperative day one and postoperative day four were significantly lower in Group I (p = 0.008 and p = 0.023, respectively). However, no intergroup difference was observed on postoperative day seven or at discharge (see Appendix). Fentanyl consumption was higher in Group II at all time points (see Appendix). The mean total amount of fentanyl consumption in Group I (665.5 μg; range, 602.5 to 756.8 μg) was 29% less than that in Group II (935.0 μg; range, 743 to 1100.0 μg) (p < 0.001).
The frequency of use of patient-controlled analgesia was higher in Group II at all time points (see Appendix). The mean number of times that the Group-I patients pushed the patient-controlled-analgesia button (13.5; range, 7.3 to 17.8) was 59% less than the mean number of times that the Group-II patients pushed it (33.0; range, 21.0 to 40.0) (p < 0.001). The frequency of use of patient-controlled analgesia had a fair correlation with VASmax and VASmin values (ρ = 0.338, p = 0.003 and ρ = 0.317, p = 0.006, respectively).
Walking ability at hospital discharge did not differ between the two groups (p = 0.864). The time until the patient started walking or performing standing exercises was marginally, but not significantly, different between the two groups (mean, 2.9 days in Group I and 3.9 days in Group II; p = 0.067).
Neither the incidences of nausea and vomiting and postoperative delirium nor the length of hospital stay differed between the two groups. Patient satisfaction was significantly higher in Group I (p = 0.016) (see Appendix).
There has been increasing interest in multimodal pain management for postoperative pain control, early rehabilitation, and reduction of opioid consumption following elective surgery. This prospective randomized study demonstrates that preemptive analgesic medication and intraoperative periarticular injections in elderly patients undergoing hemiarthroplasty for a hip fracture significantly reduce postoperative pain and opioid consumption.
Although numerous intervention studies have been performed to find ways to reduce postoperative pain and improve effective rehabilitation after hip fracture surgery over the past two decades, evidence about pain management after this surgery is sparse. Abou-Setta et al.20 performed a systemic review of eighty-three studies, including two cohort studies on multimodal pain management for patients with a hip fracture, but they focused on the development of clinical pathways for pain management and could not recommend evidence-based guidelines for pain management associated with hip fractures.
Benefits of multimodal pain management have been reported with elective orthopaedic surgery. Busch et al.8 performed a prospective randomized study of sixty-four patients who received an intraoperative periarticular multimodal drug injection during hip arthroplasty. They concluded that such an injection reduced postoperative patient-controlled analgesia requirements and pain during activity, findings that concur with our results. In the present study, multimodal pain management reduced maximum and minimum pain scores until the fourth postoperative day and fentanyl consumption until sixty hours after surgery.
Pain severity after hip fracture is a significant predictor of functional outcome and is associated with longer hospital stays, delayed walking, and postoperative complications2-4,6,21. Several studies have shown that pain reduction improves activity. Busch et al.8 reported that VAS scores for pain on activity were lower for patients who had received an intra-articular injection. Peters et al.22 reported that pain reduction had a limited effect on activity during the second postoperative day. In the present study, although there was no significant difference in early walking activity between the two groups, the patients started walking or performing standing exercises one day earlier in Group I.
In this study, the total amount of fentanyl administered was 29% less in Group I, a finding similar to that in other studies. Murphy et al.23 performed a prospective randomized study of the use of periarticular injection in ninety-one primary total hip arthroplasties. They reported that mean morphine consumption was 45.8% less in the periarticular injection group. Peters et al.22 observed that an intraoperative periarticular injection during total hip arthroplasty reduced postoperative narcotic requirements.
Postoperative pain levels determined with a VAS are subjective and might be influenced by psychological and cultural factors, personal experiences, and operative technique. Le Resche et al.24 reported that participants given the same pain stimulus rated the severity of pain differently from one another and that the same participants assessed pain produced by the same level of stimulus quite differently at different times. In the present study, elderly patients with a hip fracture had difficulty expressing pain using the VAS. To overcome these limitations, we noted the frequency of use of patient-controlled analgesia, which provided an indicator of total pain intensity. Total use of patient-controlled analgesia, determined with a computerized patient-controlled analgesia system, was 59% less in Group I, and the total amount of fentanyl used was 29% less in that group. These findings suggest that frequency of use of patient-controlled analgesia can be employed as an indicator of pain intensity as well as to calculate the amount of fentanyl used.
We could not find any difference between groups with regard to opioid related-complications such as nausea, vomiting, and dizziness, but this may be related to the goals of this study. The primary outcome variables in this study were pain levels on postoperative days. As the sample size of this study was calculated on the basis of the primary outcome measure, the sample size may have been insufficient to detect differences except in the primary outcome measure. In this study, we found that multimodal pain management did not increase the incidence of drug-related side effects or wound complications. However, some investigators have reported that injection of analgesics caused chondrolysis and complications associated with the systemic absorption of drugs25-27.
This study has several limitations. First, it was a prospective study with a single-blind design. Although the patients and their families were not told their group assignments, it is possible that the patients in Group II were aware that they were in that group because they did not receive preemptive analgesic medication. This could be a source of bias similar to a placebo effect. However, to minimize the potential influence of bias, one investigator who was completely blinded conducted all assessments of patients’ pain scores during the study period.
Second, we could not standardize the anesthesia type because of patient fear of spinal or general anesthesia and because there were some contraindications to regional anesthesia such as impaired coagulation and previous spinal surgery. However, the types of anesthesia were similarly distributed in the two groups (see Appendix). Systematic reviews of anesthesia for hip fracture surgery have identified modest evidence that regional anesthesia decreases the risk of postoperative confusion, but there was not sufficient evidence to draw conclusions regarding effectiveness of pain management20,28.
Third, we focused on the early postoperative period. Although we would not expect differences between the groups at four weeks, an additional study is required to evaluate whether reduced pain scores and fentanyl consumption translate into a lower incidence of chronic pain development.
Our data suggest that the use of preoperative preemptive analgesia combined with periarticular intraoperative injections of multimodal drugs around hips undergoing bipolar hemiarthroplasty for treatment of a fracture reduces postoperative pain and patient-controlled analgesia requirements.
Disclosure: One or more of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of an aspect of this work. None of the authors, or their institution(s), have had any financial relationship, in the thirty-six months prior to submission of this work, with any entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. Also, no author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.