Hip fractures are a global public health issue. In Canada, over 30,000 hip fractures occur annually, and approximately two million people suffer a hip fracture each year worldwide.1 The term hip fracture refers to fractures spanning from the femoral head–neck junction to 5 cm below the lesser trochanter. This injury occurs commonly among the elderly population, and its incidence is increasing as the proportion of elderly individuals rises.2 This is concerning because hip fractures have an extensive morbidity profile that substantially affects the functional status and post-injury independence of patients.3 Elderly hip fracture patients are also at a higher risk of developing postoperative complications including delirium, chronic pain, ambulation difficulties, and death.3 Effective analgesia is critical for patients to participate in the postoperative rehabilitation needed to prevent disability. Although opioids are irrefutably effective analgesics, their use in the frail geriatric population is potentially hazardous because of severe adverse effects, including nausea, vomiting, constipation, delirium, and respiratory depression.4 To help address this issue, multimodal analgesia has been recommended with various nonopioid pharmacologic treatments to manage pain optimally.5

Acetaminophen (also known internationally as paracetamol) is an effective adjunct to opioids because of its well-established analgesic effects and safety profile.6 This medication is critical in the geriatric population, who have a high incidence of comorbidities. Acetaminophen is a cornerstone of multimodal postoperative analgesia, and its scheduled oral formulation is widely used. Nevertheless, oral acetaminophen has limited bioavailability. The systemic concentration of an oral formulation is reduced by up to 40% by first-pass hepatic metabolism.7 If the patient is also receiving concomitant opioids, absorption can be limited further.8 Acetaminophen has a central site of action, so its systemic concentration is critical for uptake across the blood–brain barrier. As such, as an intravenous (iv) formulation, acetaminophen has a higher and more reliable bioavailability.8 In addition, different studies have analyzed the pharmacokinetics of iv acetaminophen and showed that its maximum observed concentration (Cmax) and area under the curve were significantly higher than oral acetaminophen.8,9,10,11,12 Nevertheless, a drawback of the iv formulation is that it is more expensive than the oral formulation. At the time of article submission, the current cost in Canada of a 1-g dose of iv acetaminophen ranged from CAD 10 to CAD 19. This is compared against the average cost of oral acetaminophen at CAD 0.03 for a 1-g dose.13

The efficacy of iv acetaminophen, compared with oral acetaminophen for postoperative analgesia, is contentious. Recent studies in patients undergoing elective hip and knee arthroplasties and colorectal surgery (both elective and emergent) have shown no significant improvement in postoperative analgesia when comparing iv acetaminophen to oral acetaminophen.6,14,15,16,17 Previous studies have not reported adverse effects with the use of iv acetaminophen.6,14,15,16,17,18 As emergency hip fracture patients are vulnerable to adverse effects of opioids, it is critical to determine whether this formulation improves analgesia and other outcomes, justifying the higher cost. In this systematic review, we sought to assess the efficacy and safety of iv acetaminophen in elderly hip fracture patients in reducing pain or preventing postoperative complications. Our research question was, “In older hip fracture patients, does the use of perioperative intravenous acetaminophen, compared against oral acetaminophen, affect postoperative outcomes?”

Methods

Inclusion and exclusion criteria

The inclusion criteria were as follows: 1) general human population ≥ 50 yr of age with no geographic restrictions; 2) known diagnosis of a hip fracture undergoing emergency surgical repair/replacement of the hip; 3) iv acetaminophen used as an intervention; 4) randomized controlled trials, prospective and retrospective cohort studies, or case–control and nested case–control studies; 5) studies in English or French; and 6) studies reporting any of the following assessments and outcome measures: pain, opioid use, delirium, cognitive impairment, ambulatory ability, length of stay, discharge location, readmission, quality of life, depression, overall health, costs, all-cause morbidity, or all-cause mortality.

The exclusion criteria were as follows: 1) pathologic hip fractures secondary to a malignancy or metastasis; 2) elective arthroplasty; 3) qualitative studies; 4) nonhuman studies (in vivo, in vitro, other); and 5) case reports.

Search strategy

This systematic review protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) (registration number: CRD42021198174; registered 15 August 2021). The following databases were searched from inception via the Ovid search interface: Medline, Medline ePubs/In-Process (daily), Embase, Cochrane Central Register of Controlled Trials, and the Cochrane Database of Systematic Reviews. The Web of Science (Clarivate™; London, UK), Scopus (Elsevier; Amsterdam, The Netherlands), and the Cumulative Index to Nursing and Allied Health (EbscoHost, 1982–present) databases were also searched from inception. ClinicalTrials.gov and the World Health Organization’s International Clinical Trials Registry Platform (ICTRP) were also searched. All databases and trial registries were searched on 21 September 2021.

The search process followed the Cochrane Handbook19 and the Cochrane Methodological Expectations of Cochrane Intervention Reviews20 for conducting the search, the PRISMA 2020 guideline21 for reporting the search, and the Peer Review of Electronic Search Strategies (PRESS) guidelines,22 drawing on its 2015 Guideline Evidence-Based Checklist to avoid potential search errors.

Preliminary searches were conducted and full-text literature was mined for potential keywords and appropriate controlled vocabulary terms (such as Medical Subject Headings [MeSH] for Medline and Emtree descriptors for Embase) using the databases, target articles, and the Yale MeSH Analyzer.23 The search strategy used the following terms: “Acetaminophen” AND “Hip Fractures OR Hip Surgery,” with each component being fleshed out with controlled vocabularies, text word terms, and synonyms.

To supplement the search, references and citations were searched on the Web of Science (Clarivate) for one target citation: Tsang KS, Page J, Mackenney P. Can intravenous paracetamol reduce opioid use in preoperative hip fracture patients? Orthopedics. 2013 Feb;36(2 Suppl):20–4.

https://doi.org/10.3928/01477447-20130122-53. PMID: 23379572. For details, please see the Medline search strategy provided in the Electronic Supplementary Material, eAppendix.

Study selection

Duplicated studies were detected and removed using Covidence (Melbourne, Australia). Next, two reviewers (J. C. and K. M.) independently screened the titles and abstracts according to the eligibility criteria. Disagreements were resolved by consensus or by consulting a third author (J. W. or S. S.).

The approved articles underwent a full-text review by the same two reviewers (J. C. and K. M.) to ensure they met all the eligibility criteria. Reasons for excluding articles at this stage were recorded. The resulting list of included studies for full review was circulated to the entire systematic review team to identify any potentially missing studies.

Data extraction and quality assessment

In duplicate, two authors (J. C. and K. M.) extracted study characteristics from the approved articles. For each included study, the following data were extracted: year, study location, study design, demographics, sample size, intervention, details on the primary outcome, details on the secondary outcome, and follow-up period. All the statistics, except for confidence intervals (CIs) of differences, came from the original four articles. Two authors (J. C. and K. M.) independently rated each article’s quality using the Newcastle-Ottawa Quality Assessment Scale.24 Each study was judged on nine items. These items were the representativeness of the exposed cohort, the selection of the nonexposed cohort, the ascertainment of exposure, the demonstration that the outcome of interest was not present at the start of the study, whether the study controlled for type of hip fractures, whether the study controlled for age and sex, the assessment of the outcome, whether the follow-up was long enough for the outcome to occur, and the adequacy of the follow-up. These nine items were categorized into the following groups: study group selection, study group comparability, and ascertainment of the outcomes of interest and scored accordingly. If disagreements occurred, they were resolved by consensus or by consulting a third author (J. W. or S. S.). We initially planned to complete a meta-analysis; however, this was not possible because study design and outcomes were not consistent between the studies.

Calculation of 95% confidence intervals of differences

To calculate the 95% CIs of the differences in means for studies that did not report these, we used the following formula: μ1 − μ2 = (M1 − M2) ± ts(M1 − M2), where M1 and M2 = sample means; t = t statistic determined by confidence level; and s(M1 − M2) = standard error.25

Morphine milligram equivalent conversion

To ensure consistency and accurate comparison, all opioid medications described in the studies were converted to morphine milligram equivalents (MME) as outlined by the 2016 Centers for Disease Control and Prevention guidelines.26 For intravenous opioids, the values were converted to their oral equivalent using the following ratio: 1 mg of iv morphine = 3 mg oral morphine.

Results

Search results

The literature search yielded 3,510 studies, with the search in the Cochrane database yielding 187 articles (Figure 1). After the duplicates were removed, 2,355 studies remained. Once titles and abstracts were screened, 106 articles were deemed eligible for full-text assessment. Of these articles, four studies met the inclusion criteria and were subsequently included in the final review.3,4,27,28 Additionally, we contacted Hansen et al. to obtain the hip fracture-specific data that were acquired during their study.28

Study characteristics

Of the four studies, one study was a prospective cohort study4 and the rest were retrospective cohort studies (see Table 1).3,27,28 Study locations included the UK (n = 1)4 and USA (n = 3).3,27,28 Three of the four studies, excluding the study conducted by Hansen et al., directly compared the control group of oral acetaminophen with the intervention group of iv acetaminophen.3,4,27 These three studies included 255 control patients and 279 intervention patients in total. Hansen et al. compared 64,395 control patients prescribed iv opioids to 17,928 intervention patients prescribed iv acetaminophen.28 Hansen et al. informed us that all patients received standard of care (personal correspondence), so there was no restriction on oral acetaminophen in both the control and intervention groups of this study.28 Nevertheless, Hansen et al. did not describe the dosing regimens of oral acetaminophen. Excluding the study by Hansen et al. (age and sex characteristics of fracture patients were not provided), the average age of participants was 81 yr and 29% were male. All studies reported the mean opioid dose used,3,4,27,28 and three used the visual analogue scale as a subjective measure of pain.3,4,27 Three studies reported the length of hospital stay3,27,28 and two studies reported the percentage of patients discharged home.3,27 Only one study analyzed each of the remaining outcomes: total cost of hospitalization,28 missed physical therapy sessions,27 chart-based delirium identification,3 the need for one-to-one supervision as a surrogate outcome of delirium,3 and readmission rate.3

Table 1 Study characteristics
Full size table

Study outcomes

Pain

All four studies compared the mean opioid dose administered to the intervention and control groups (Table 2).3,4,27,28 Three studies reported a significantly lower mean opioid dose with the administration of iv acetaminophen.3,4,27 One of these studies, by Connolly et al., reported the dose for postoperative days 0 to 3.3 Compared against the oral acetaminophen group, the iv acetaminophen group required significantly less iv opioid administration on postoperative day 1 (7.14 mg vs 2.22 mg; P = 0.01). There were no significant differences between the two groups on postoperative day 2 (4.56 mg vs 2.22 mg, P = 0.11) and postoperative day 3 (2.7 mg vs 1.2 mg, P = 0.16). The study by Hansen et al. also compared the opioid doses in the postoperative period from postoperative days 0–2.28 In this study, the use of iv acetaminophen was associated with a difference of 0.6 mg compared with the iv opioid group; however, this difference was not significant (P > 0.01).

Table 2 Study outcomes
Full size table

One study by Tsang et al.4 reported the preoperative opioid dose from hospital admission to the anesthetic room. They found that the iv acetaminophen group required a significantly lower total opioid dose (mean, 65.4 mg vs 19.5 mg P < 0.005) and significantly fewer opioids per day (mean 27.0 mg vs 11.4 mg, P < 0.005). Lastly, Bollinger et al. reported the opioid dose from the time of admission to discharge.27 Comparing oral vs iv acetaminophen, Bollinger et al. reported a total mean opioid consumption of 41.3 mg vs 28.3 mg (95% CI of the difference, -8.195 to 34.195;Footnote 1P < 0.001), and a daily mean opioid consumption of 9.6 mg vs 7.8 mg (95% CI of the difference, -0.004 to 3.604;A P = 0.05).27

Three studies used a visual analogue scale (VAS) to assess patients’ postoperative pain levels, directly comparing the control oral acetaminophen group to the intervention iv acetaminophen group.3,4,27 Two studies were retrospective cohort studies,3,27 and the remaining study was a prospective cohort study.4 These three studies included 255 control patients and 279 intervention patients. With all patients receiving appropriate standardized analgesic medications, with the reduction in MME not at the expense of the patient’s pain, only Bollinger et al. reported a significant VAS pain score reduction in the iv acetaminophen group vs the oral acetaminophen group (mean, 2.8 vs 4.2; 95% CI of the difference, 0.98 to 1.82;A P < 0.001).27

Physiotherapy sessions

Bollinger et al. reported the percentage of missed in-hospital physical therapy (PT) sessions as a surrogate marker for postoperative pain.27 This study included 169 control patients and 157 intervention patients, and 27.5% were male. Comparing oral to iv acetaminophen, the latter group missed 10.4% sessions compared with 21.8% in the control group (95% CI of the difference, -7.232 to 30.032;A P < 0.001).27

Length of stay

Three studies reported the length of hospital stay.3,27,28 All of these studies reported a significant decrease in the length of stay for patients administered iv acetaminophen (Table 2). Bollinger et al. and Connolly et al. directly compared oral and iv acetaminophen, with the length of stay at 4.4 days vs 3.8 days (95% CI of the difference, -0.033 to 1.233;A P < 0.001) and 8.47 days vs 6.37 days (P = 0.04), respectively.3,27 Hansen et al. reported a mean difference of -0.55 days in the length of stay for the iv acetaminophen group (P < 0.01).28

Discharge home

Two studies comparing 227 control patients with 232 intervention patients reported the percentage of patients discharged home.3,27 During hospitalization, patients who experienced adverse outcomes were discharged to a secondary care facility, such as an acute rehabilitation facility or a nursing institution.27 Only Bollinger et al. described a significant difference between the two groups. They noted that 19% of patients were discharged home in the iv acetaminophen group compared with 7% in the oral acetaminophen group (P < 0.001).27

Readmission rate

In one study, Connolly et al. reported the readmission rate, comparing 58 control patients and 65 intervention patients.3 The readmission rate was 18.5% in the iv acetaminophen group vs 27.6% in the control group; nevertheless, the difference was not statistically significant (P = 0.23).

Costs

One study by Hansen et al. compared the total cost of hospitalizations (USD); the authors compared 64,395 control patients on iv opioids with 17,928 intervention patients on iv acetaminophen. Calculating the systems-level cost of hospitalizations among 14 different departments, the authors reported a decrease of USD 879.60 for hip fracture patients administered iv acetaminophen vs those who received iv opioids. Nevertheless, this decrease was not statistically significant (P > 0.01).28 The authors of this study predominantly attributed the decrease in hospital costs to the difference in length of stay between the groups.28

Delirium

Only one study (by Connolly et al.) reported delirium incidence by reviewing medical records, including various healthcare professionals’ notes and the administered treatments.3 Using a Chart-based Delirium Identification Instrument (CHART-DEL), the authors selected a “possible” level of delirium as a cutoff for a positive diagnosis.29 Significantly fewer patients developed delirium when managed with iv acetaminophen (15.4%) than patients managed with oral acetaminophen (32.8%; P = 0.02).3

The same study also included the need for one-to-one supervision as a secondary outcome measure. The need for one-to-one supervision was lower in the iv acetaminophen group than in the oral acetaminophen group (9.2% vs 24.1%; P = 0.03).3

Safety/adverse effects

None of the studies reported adverse effects of iv acetaminophen compared with oral acetaminophen.

Quality assessment

Table 3 shows the quality scores of the studies. Two studies scored 8 out of 9,3,27 while the other two studies scored 5 out of 9.4,28 The main concerns with the latter studies were the lack of standardization regarding sex distribution between groups and the inadequacy of follow-up of the cohorts. Neither article by Tsang et al. nor Hansen et al. mentioned the exclusion criteria of pathologic fractures secondary to malignancy or metastasis. These studies also failed to match patients for age and sex in the two groups. Lastly, the study by Hansen et al. solely used billing codes to define analgesic exposures but provided little reference assessing the accuracy of this surrogate marker.

Table 3 Quality assessment: Newcastle-Ottawa Scale
Full size table

Discussion

Although the evidence is very limited, our results suggest that iv acetaminophen may improve analgesia, both pre- and postoperatively, as shown by the reduction of opioid medications consumed. Interestingly, despite the significant reduction in opioid consumption with iv acetaminophen, the VAS pain scores were similar or lower, suggesting that iv acetaminophen may provide sufficient analgesia to maintain the patients’ pain relief. Additionally, the length of hospital stay was reduced when patients used iv acetaminophen. Nevertheless, the effect of iv acetaminophen on hospital discharge is unclear since the two studies reporting this outcome had conflicting results. In addition, the number of PT sessions missed, readmission rates, delirium, and costs were only reported in one study each.

The effects of iv acetaminophen on opioid reduction in other types of surgery, such as general surgery, have also been mixed.17 Studies have found that the iv formulation is not more effective than oral acetaminophen in these patient populations. In elective joint arthroplasties, Studner et al. showed that oral acetaminophen produced more consistent results in opioid reduction, along with fewer reported opioid-related adverse effects than iv acetaminophen did.14 The discrepancy in results between elective arthroplasties and nonelective fracture repair may arise for multiple reasons.30 The most notable factor is the older patient population that undergoes hip fracture repair surgeries compared with elective arthroplasty. The average age of the patients included in our studies was 81.1 yr. As such, these patients have a higher burden of perioperative comorbidities and frailty. For instance, an elderly patient may have renal impairment, which prevents the use of nonopioid analgesic medications, such as nonsteroidal anti-inflammatory drugs. This patient population also has a much higher risk of postoperative morbidity and mortality than elective joint arthroplasty patients.31

Of the two studies reporting the percentage of patients discharged home, only Bollinger et al. showed a significant increase with iv acetaminophen.3,27 The correlation between analgesia and subsequent functional outcomes has been studied in other settings, including elective hip arthroplasty. For instance, Erlenwein et al. reported that maximal preoperative pain intensity independently influenced daily function six months after surgery.31 The authors indicated that those patients with more persistent hip pain intensity had more limitations in daily activities.31 One study in our review reported that patients using iv acetaminophen missed fewer PT sessions than those using oral acetaminophen did.27 They concluded that postoperative pain, associated with prolonged bed rest, could disrupt PT sessions. This result may be clinically significant as increased participation in PT sessions suggests improvements in pain, potentially leading to the earlier recovery of their functional status. Further investigation on whether better analgesia improves subsequent functional outcomes, specifically in elderly hip fracture patients, is warranted. Thus, future studies investigating medication-induced side effects should also include clinical outcomes, such as the patients’ ability to complete activities of daily living independently following orthopedic surgery.

Although the acquisition cost of both iv and oral acetaminophen is relatively low, widespread use of iv acetaminophen can lead to substantially higher costs. The price for iv acetaminophen varies depending on the country, region, and institution. Nevertheless, this cost may be mitigated if the iv formulation provides better analgesia, enabling quicker rehabilitation and reducing adverse outcomes, such as delirium, which may reduce the overall length of hospital stay.32 Furthermore, one study showed that iv acetaminophen reduces postoperative delirium and the subsequent need for one-to-one supervision, but this study did not compare the costs associated with hospitalization.3 In a study at our institution, delirium was correlated with a mean incremental episode-of-care cost of CAD 8,286 compared with patients who experienced no delirium.33 A recent study from the USA reported that the cumulative healthcare costs attributable to delirium after major elective surgery were USD 44,291 per patient per year.34 Thus, iv acetaminophen may reduce overall hospital costs when compared with the oral formulation.

Elderly hip fracture patients are at very high risk for delirium. Connolly et al. reported that delirium was reduced in hip fracture patients who received iv acetaminophen.3 Although there are no randomized trials on iv acetaminophen in hip fracture patients, the effect of iv acetaminophen on reducing delirium was recently examined in cardiac surgery patients.35 Cardiac surgery has a high incidence of postoperative delirium and postoperative neurocognitive disorders. A pilot randomized study (DEXACET) and the subsequent multicentre protocol proposal (PANDORA) explored the mitigating effect of iv acetaminophen on delirium in the cardiac surgical population.35,36 The DEXACET study was the first randomized controlled trial to analyze the effects of iv acetaminophen on delirium outcomes. With a sample size of 120 patients, the authors showed that the use of iv acetaminophen significantly reduced the incidence of postoperative delirium (-18%; P = 0.01).35 Although these results are encouraging, one of the major weaknesses of the DEXACET study’s methodology is the use of a placebo rather than an active comparator such as oral acetaminophen.37

Thus, the study by Connolly et al. on hip fracture patients showing “significant reduction of delirium and hospital length of stay” provides preliminary evidence supporting the use of iv acetaminophen in reducing postoperative delirium.3 Nevertheless, as the authors acknowledged, the retrospective methodology, the lack of follow-up, the “chart-based” diagnosis of delirium, and its “likelihood,” instead of an actual diagnosis based on validated clinical tools, along with the use of a “before and after” study design with historical cohorts, limit the quality of this study. These factors could undermine the finding that delirium is reduced.

Limitations

The present systematic review has several limitations. First, our literature search yielded only four articles and lacked high-level evidence publications, such as prospective randomized controlled trials, which emphasizes the gap in the literature on this subject. Second, the analysis of many postoperative outcomes, such as delirium, was reported by only one study. Third, we could not conduct a meta-analysis because of heterogeneity, variable outcome definitions, inconsistencies between comparison groups, and different timelines for administration of iv acetaminophen (before vs after surgery). Fourth, the sample sizes of the studies included in our review were small, with only one study including more than 350 people. None of the studies reported 95% CIs for the pairwise difference in outcomes. Only one study provided aggregate data for us to calculate a 95% CI; however, the distribution assumption used when calculating the 95% CI from aggregate data may not match the actual distribution of the data very well and the calculated CI may be more conservative than the actual CI. In addition, none of the studies included pharmacokinetic/pharmacodynamic analyses. Finally, one study received funding from the company that manufactures iv acetaminophen.27

Conclusion

There is some very limited low-level evidence that iv acetaminophen may decrease the use of preoperative and postoperative opioids and shorten the length of hospital stay. Nevertheless, there were few studies in this patient population, and all of the studies used historical controls. As such, our results should be interpreted with caution. Our review highlights a prominent gap in the literature regarding the efficacy of iv acetaminophen and its impact on postoperative outcomes. With the aging population and increasing numbers of hip fracture patients, improving the perioperative care of these vulnerable patients is essential. Therefore, adequately powered randomized controlled trials are needed to determine iv acetaminophen’s efficacy in improving mortality, pain, activities of daily living, mobility, and health-related quality of life in elderly hip fracture patients.

Figure 1
figure 1

Prisma flow diagram of literature search

Full size image