Comparative Effectiveness of Pain Management Interventions for Hip Fracture: A Systematic Review

Background: Pain management is integral to the management of hip fracture. of nonpharmacologic for managing pain after effective for relieving acute pain and reducing delirium. Low-level evidence suggests that preoperative traction does not reduce acute pain. Evidence was insufficient on the benefits and harms of most interventions, including spinal anesthesia, systemic analgesia, multimodal pain management, acupressure, relaxation therapy, transcutaneous electrical neurostimulation, and physical therapy regimens, in managing acute pain. Limitations: No studies evaluated outcomes of chronic pain or exclusively examined participants from nursing homes or with cognitive impairment. Systemic analgesics (narcotics, nonsteroidal anti-inflammatory drugs) were understudied during the search period. Conclusion: Nerve blockade seems to be effective in reducing acute pain after hip fracture. Sparse data preclude firm conclusions about the relative benefits or harms of many other pain management interventions for patients with hip fracture. Agency for Healthcare Research and Quality.

methods. This review provides a summary of the methods; details are available in the full report (13).

Data Sources and Searches
We searched electronic databases and clinical trial registries, with no language restrictions, from January 1990 to December 2010 and hand-searched scientific meeting proceedings and reference lists (Appendix Table 1, available at www.annals.org). Search terms were selected by scanning search strategies of systematic reviews on similar topics and by examining index terms of potentially relevant studies. A combination of subject headings and text words were adapted for each electronic resource. This included terms for hip fracture (fracture* and [hip or intertrochanter* or petrochanter* or subtrochanter* or intracapsular or extracapsular or petrochant* or trochant* or hip or "femoral neck"]) and pain (pain* or heal or healing or therap* or recover* or "quality of life" or rehabilitat* or "drug therapy" or pharmacological or acupunct* or acupress* or traction or "electrical stimulation" or "passive motion" or morphine or acetaminophen or paracetamol or tylenol or anesth* or analges*). We excluded research published before 1990 because surgical procedures and medical care for this patient population have changed substantially since then. We examined all pain management interventions used for hip fracture, including (but not limited to) nerve blockade, systemic analgesics (for example, opiates and nonsteroidal anti-inflammatory drugs [NSAIDs]), anesthetics, and traction.

Study Selection
Two reviewers independently screened titles, abstracts, and the full text of identified articles; disagreements were resolved by consensus or third-party adjudication. Randomized, controlled trials (RCTs); nonrandomized, controlled trials (non-RCTs); cohort studies (prospective or retrospective); and case-control studies were included if they were published in 1990 or later; included adults aged 50 years or older who were hospitalized with acute hip fracture because of low-energy trauma; and examined any pain management intervention, regardless of method of administration or time point during the care pathway.

Data Extraction, Quality Assessment, and Rating the Body of Evidence
Two reviewers independently extracted data, assessed methodological quality, and rated the body of evidence; discrepancies were resolved by consensus or third-party adjudication. Extracted data included study characteristics, inclusion and exclusion criteria, participant characteristics, interventions, and outcomes. We used the Cochrane Collaboration's Risk of Bias tool (14) to assess RCTs and non-RCTs (Appendix Table 2, available at www.annals.org) and the Newcastle-Ottawa Scale (15) to assess observational studies.
We assessed the strength of evidence (rated by using the GRADE [Grading of Recommendations Assessment, Development and Evaluation] approach) (16) for outcomes that the clinical investigators had identified as the most clinically important: acute pain (Յ30 days), chronic pain (Յ1 year), 30-day mortality, and incidence of serious adverse events (delirium, myocardial infarction [MI], renal failure, and stroke) (Table). We assessed study design (experimental or observational), risk for bias (low, medium, or high), consistency (no inconsistency, inconsistency present, or unknown), directness (direct or indirect), and precision (precise or imprecise).

Data Synthesis and Analysis
We pooled the studies if the study design, interventions being compared, and outcomes were considered to be similar. Odds ratios (ORs) for dichotomous outcomes were combined by using the DerSimonian and Laird randomeffects model (17), except for rare events (Ͻ1% incidence), in which case the Peto method was used (18). Continuous outcomes were combined by using the mean difference (MD) or standardized mean difference (SMD) when different scales were used across studies (19). The SMD allows data transformation to a uniform scale by dividing the mean by its SD. Statistical heterogeneity was quantified by using the I 2 statistic (20). When heterogeneity was substantial (I 2 Ͼ 75%), we investigated sources of heterogeneity by determining the effect of important modifiers: intervention details (type and quantity), study design and risk for bias, and effect of imputed data. A priori subgroup analyses included sex, age, race, body mass index, marital status, comorbid conditions, prefracture functional ability, and family distress. All analyses were performed with Review Manager, version 5.01 (The Cochrane Collaboration, Copenhagen, Denmark).

Context
Multiple pain management strategies are available for patients after hip fracture.

Contribution
This systematic review found moderate-level evidence that nerve blockades reduced acute pain and delirium after hip fracture and low-level evidence that preoperative traction did not affect pain. Benefits and harms of such interventions as spinal anesthesia, systemic analgesia, multimodal pain management, acupressure, relaxation therapy, transcutaneous electrical neurostimulation, and physical therapy were unclear.

Caution
No or few studies evaluated outcomes of chronic pain or systemic analgesics.

Role of the Funding Source
The Agency for Healthcare Research and Quality suggested the initial questions and approved copyright assertion for the manuscript but did not participate in the literature search, data analysis, or interpretation of the results.

RESULTS
We identified 9357 citations and included 83 unique studies (64 RCTs, 5 non-RCTs, and 14 cohort studies) with 14 to 1333 participants (median, 60 participants; interquartile range, 40 to 90 participants]) (Figure 1). The full report (13) includes the search strategies and details of the included and excluded studies.
Studies were published between 1990 and 2010 (median year, 2003; interquartile range of years, 1998 to 2007), and nearly all were published in peer-reviewed journals. Most were RCTs performed in single-university settings in Europe, which investigated pre-or intraoperative pain management interventions in the acute care setting, and none was from institutional settings (such as nursing homes). Five of the 83 studies were from North America, and 55 were from Europe. The mean age of study participants ranged from 59 to 86 years. Most participants were female (74%). Almost one half of the studies (n ϭ 31) excluded participants with cognitive impairment or delirium. Pain, the primary outcome, was measured with an array of validated visual and numerical pain scales (for example, visual analogue scale).
Of the RCTs, only 2 (21,22) were considered to be at low risk for bias (Appendix Table 3, available at www .annals.org). The methodological quality of the cohort studies was moderate (median, 7; interquartile range, 6 to 8). Interventions were divided into 8 groups (Table and  Appendix Table 4 Most of the evidence for key outcomes and serious adverse events came from single trials and cohort studies. The strength of evidence was low to moderate for some interven-tions in alleviating acute pain, preventing delirium, and decreasing the 30-day mortality rate (Table). The strength of evidence for the remaining outcomes was insufficient because of inadequate numbers of studies and study power.
Acute pain was reported in studies comparing various nerve blockades with no blockade (Figure 2), but pooled results were not reported because of significant heterogeneity (I 2 ϭ 92%). Sensitivity analyses showed statistically significant results in favor of each type of blockade except 3-in-1 and fascia iliaca blockades (Table). When blockades were used, pain on movement and at rest were inconsistent and dependent on the type of blockade (Table), but less use of  The incidence of delirium, a common complication of hip fracture, was statistically significantly decreased with nerve blockades compared with no blockade. Other than  reducing the incidence of delirium, nerve blockades offered no advantage in preventing cardiac complications (22,28,50), deep venous thrombosis (25,26), MI (Table), nausea or vomiting (21,24,29,35,40,47), pulmonary embolism (22,28), respiratory infection (22,25,26,28,49), stroke (Table), surgical-wound infection (22,26), urine retention (24,47,52), or urinary tract infection (26,52). Nerve blockade reduced the length of acute hospitalization (50,52); however, pooled results were not reported because of marked heterogeneity (I 2 ϭ 93%). Mortality rates were unchanged at 30 days or 1 year (24, 25, 52) ( Table) with the use of nerve blockades. Similar mortality results were found in participants with preexisting heart disease (OR, 0.10 [CI, 0.01 to 1.90]) or who were living independently before hip fracture (OR, 1.00 [CI, 0.06 to 16.76]). Nerve blockade and neuraxial anesthesia had similar results in terms of acute pain, use of additional pain medications (48), and delirium (Table). No outcome differences were attributable to any nerve blockade in trials that assessed the use of bupivacaine versus ropivacaine (46) or the addition of clonidine to the injectate (43).

Systemic Analgesia
Of 3 RCTs comparing systematic analgesics (Table), 1 (83) reported that intravenous parecoxib is more effective than the combination of intramuscular diclofenac and meperidine. A second RCT (84) reported that intrathecal isotonic clonidine was more effective than hypertonic clonidine.

Multimodal Pain Management
Two cohort studies described the use of multiple pain management strategies (sequential or in parallel) as part of a clinical pathway. One study (86) compared a postoperative protocol of tramadol plus acetaminophen compared with standard care. The other study (87) compared a preoperative protocol of skin traction, morphine, and acetaminophen with standard care and reported no difference between groups for 30-day or 1-year mortality or for adverse events. Both studies reported no difference in the incidence of delirium (Table).

Physical Modalities and Physical Therapy Regimens
Yielded acute pain, length of acute hospitalization (88,90), or use of additional analgesic medication (88, 95) did not statistically significantly differ between preoperative skin traction and no traction (Table and Figure 3). Thirtyday mortality did not statistically significantly differ between skin and skeletal traction (Table).
Transcutaneous electrical nerve stimulation (TENS) was compared with a sham control in 2 studies, 1 before (100) and 1 after (99) surgery. Both studies reported pain relief with TENS (Table). When TENS was applied after surgery, pain on movement was reduced (Table) (Table).

Complementary and Alternative Medicine
Acupressure provided short-term pain relief compared with a sham intervention (Table). The Jacobson relaxation technique (a 2-step process of contracting and relaxing specific muscles) also provided short-term pain relief (Table) and reduced the need for additional pain medication (MD, Ϫ8.43 [CI, Ϫ15.11 to Ϫ1.75]).

DISCUSSION
Hip fracture is associated with substantial pain, but the evidence on pain management after hip fracture was surprisingly sparse. We identified only 83 studies that met the inclusion criteria, precluding firm conclusions for any single approach or for optimal overall pain management. Most studies were RCTs conducted in single-university settings in Europe, with few studies from North America. Studies primarily evaluated pharmacologic management of postoperative pain in the acute care setting, including documentation of adverse events and short-term mortality. Most studies focused on a single discipline (for example, anesthesiology or rehabilitation), despite supporting evidence that optimal pain management is multidisciplinary (104,105). No included study examined pain beyond 30 days. More often than not, functional recovery, health-related quality of life, and health services utilization were underreported. Despite these limitations, at least 5 general conclusions can be made.
First, nerve blockade seems to be more effective than standard care. It reduced the need for supplemental systemic analgesia and lowered the risk for delirium, although most studies did not report on how nerve blockades may affect ambulation or rehabilitation. Several technically different blockade approaches have been studied in this context, and most were found to be beneficial. The evidence does not support a preference for more complex techniques (for example, catheter-based, continuous epidural infusions) over simpler, safer approaches (for example, single- shot femoral blockade). Nerve blockade is within the repertoire of most practicing anesthesiologists, but many clinicians do not perform it routinely because they believe the additional time, effort, and supervision required may outweigh the benefits (106). On the contrary, the evidence would suggest that blockade may be beneficial. Second, spinal anesthesia, although effective and safe, does not demonstrably differ from general anesthesia in rates of mortality, delirium, or other medical complications. This is in contrast to other recent reviews (107,108), which show a greater benefit for regional anesthesia than general anesthesia. The discrepancy may reflect improvements in the safety and adverse-effect profile of general anesthesia or in the medical care of older participants in the past 2 decades (because the aforementioned reviews included studies dating back to 1977).
Adding other agents to plain local anesthesia for spinal anesthesia does not seem to affect outcomes outside the operating room. Furthermore, larger doses of a spinal anesthetic may cause more hypotension without improving pain control or other outcomes (109). Evidence was insufficient to show that multimodal analgesia yields improvements over single modalities.
Third, surprisingly few studies focused on systemic analgesics, opiates, or traditional NSAIDs. Because most NSAID trials are for short-term treatment of chronic pain and given the risks associated with NSAIDs (for example, gastrointestinal damage, bleeding, and drug interactions), NSAID use in this population has not been aggressively pursued. However, it is surprising that given the regular daily use of NSAIDs for acute, surgical, and arthritis pain, we identified only 3 trials of systemic analgesia in hip fracture published since 1990. Of these, 1 small trial of intravenous parecoxib (an agent not available in North America) compared with diclofenac and meperidine reported a significant finding. Because meperidine use is discouraged among older populations owing to its association with delirium (110), this study may have even further limited clinical utility.
Fourth, preoperative traction does not reduce pain or complications compared with no traction. These results are consistent with those of a Cochrane review (111).
Finally, only 4 nonpharmacologic interventions were identified: acupressure, relaxation therapy, TENS, and physical therapy. Although these interventions were based on limited evidence, they seem to be safe and may result in clinically meaningful reductions in pain. A consideration of nonpharmacologic interventions is that, similar to nerve blockade, proper application and instruction may require additional time and training of allied health professionals.
A strength of this review is the comprehensiveness of the search, which included 25 electronic databases, conference proceedings, and forward-searching of included studies. Language was not an exclusion criterion, and we included studies published in 8 languages. The main limitations relate to the quantity, quality, and external validity of the available evidence. First, important subgroups of patients with hip fracture were frequently excluded. Almost one half of the studies (n ϭ 31) excluded patients with cognitive impairment or who could not cooperate. With approximately 30% to 60% (11,12) of the elderly population with hip fracture having some degree of cognitive impairment, this represents a significant selection bias. Furthermore, approximately 25% of patients with hip fracture are from nursing homes (112, 113), and none of the included studies examined this population exclusively.
Second, because our literature search included studies only from 1990 onward, we may not have identified older studies of systemic analgesics and anesthetics. Even so, with the changes in hip fracture care over the past 20 years, evidence found before 1990 may not apply to current care standards that focus on early patient mobility.
Third, the lack of standardized outcomes limits the interpretation and applicability of the results. Although pain and function are correlated (3), most studies focused on pain relief and did not evaluate an intervention's effects on the patient's ability to mobilize after surgery, a factor linked to recovery levels after hip fracture (114).
Fourth, we found minimal evidence about managing pain after hospital discharge or assessment of the long-term effects of early postoperative pain management on subsequent recovery. Finally, because of the low incidence of postoperative complications, no individual study had adequate power to detect adverse events.
For future studies, we suggest a conditional set of minimal scientific criteria to guide the next generation of trials on pain management of hip fracture to provide more valid and clinically useful information. Adequately powered, multicenter studies that assess safety, effectiveness, and appropriateness of pain management after hip fracture should include, at the least, subgroup analyses comparing patients with and without cognitive impairment and community-dwelling versus nursing home residents. Study follow-up should extend beyond the acute care setting and should evaluate patients at least 6 months after fracture (3). Standardized and validated outcome measures should be selected to allow meaningful comparisons across interventions and studies. Relevant outcomes include validated pain scores, prescription of analgesics, and adverse events. Given how common impaired cognition occurs in this patient population, nonverbal pain assessment scales are recommended (107,115,116). Outcome measures should also include functional status, health-related quality of life, and health service utilization.
In summary, the evidence showed that most pain management interventions improved short-term pain, but few studies of long-term clinical importance were available. Until more rigorous evidence is generated, the review summarizes what should be considered the current state of the evidence for managing hip fracture pain.