Article Text

other Versions

Download PDFPDF

Effectiveness of intravenous acetaminophen for postoperative pain management in hip and knee arthroplasties: a population-based study
  1. Ottokar Stundner1,
  2. Jashvant Poeran2,3,
  3. Hannah Noemi Ladenhauf4,
  4. Marc Moritz Berger1,
  5. Steven B Levy5,
  6. Nicole Zubizarreta2,3,
  7. Madhu Mazumdar2,
  8. Janis Bekeris1,6,
  9. Jiabin Liu6,
  10. Leesa M Galatz3,
  11. Calin S Moucha3 and
  12. Stavros Memtsoudis1,6,7
  1. 1Department of Anesthesiology, Perioperative Medicine and Intensive Care Medicine, Paracelsus Medizinische Privatuniversitat, Salzburg, Austria
  2. 2Department of Population Health Science and Policy, Institute for Healthcare Delivery Science, Icahn School of Medicine at Mount Sinai, New York, New York, USA
  3. 3Department of Orthopedic Surgery, Icahn School of Medicine at Mount Sinai, New York, New York, USA
  4. 4Department of Pediatric and Adolescent Surgery, Paracelsus Medizinische Privatuniversitat, Salzburg, Austria
  5. 5Department of Pharmacy, Mount Sinai St Luke's, Mount Sinai Health System, New York, New York, USA
  6. 6Department of Anesthesiology, Hospital for Special Surgery, New York, New York, USA
  7. 7Department of Healthcare Policy and Research, Weill Cornell Medical College, New York, New York, USA
  1. Correspondence to Dr Jashvant Poeran, Department of Population Health Science and Policy, Institute for Healthcare Delivery Science, Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA; jashvant.poeran{at}


Background and objectives The significance of intravenous over oral acetaminophen (APAP) as part of multimodal analgesic protocols is contested, particularly when considering its relatively high price and use in a surgical cohort such as total hip or knee arthroplasty (THA/TKA), which generally tolerates oral medications. This study aims to elucidate APAP’s effectiveness in a large, population-based patient sample.

Methods 1 039 647 THA/TKA procedures were sampled from the Premier Healthcare claims database 2011–2016. APAP use was categorized by intravenous/oral and use on the day of surgery, postoperative day 1 and thereafter. Outcomes were opioid utilization (in oral morphine equivalents), length and cost of hospitalization, and opioid-related adverse effects (respiratory, gastrointestinal, and naloxone use as a proxy). Mixed-effects models measured the associations between intravenous/oral APAP use and outcomes. Percent (%) change and 95% CIs are reported.

Results Overall, 23.6% (n=245 454) of patients received intravenous APAP; of these, 56.3% (n=138 180) received just one dose on the day of surgery. After adjustment for relevant covariates, particularly use of >1 dose of intravenous APAP (compared with no use) on postoperative day 1 was associated with −6.0% (CI −7.2% to −4.7%) reduced opioid utilization; this was −10.7% (CI −11.4% to -9.9%) for use of > 1 dose oral APAP on postoperative day 1. Further comparisons regarding other outcomes also favored oral (over intravenous) APAP.

Conclusions These results do not support the routine use of intravenous APAP in patients undergoing lower joint arthroplasty, especially since oral APAP shows more beneficial outcome patterns.

  • acute Pain
  • pain
  • postoperative
  • acetaminophen
  • arthroplasty
  • hip
  • arthroplasty
  • knee
  • analgesia

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Major total lower extremity joint arthroplasty (TJA) is a painful procedure, necessitating a robust postoperative analgesia strategy.1 Multimodal analgesic regimens aim at lowering opioid consumption along with reducing the incidence of potentially hazardous opioid-related side effects, by adding non-opioid modalities such as non-steroidal anti-inflammatory drug (NSAIDs) or gabapentinoids to the protocol. Intravenous acetaminophen (APAP), a relatively new option for multimodal analgesic regimens in the USA, has been available in most of Europe since the early 2000s,2 while approval for the US market only occurred in late 2010. Although countless studies have corroborated APAP’s efficacy in the non-operative setting,3 benefits of the intravenous route in postoperative analgesic protocols are debated, particularly among surgical patients such as those undergoing TJA that may tolerate oral medications.4–6 Moreover, meta-analyses demonstrating some benefit of intravenous APAP acknowledge the limited quality of evidence, while they also may not reflect clinical practice as comparisons are made with placebo.7 8 Of additional concern is the relatively high cost of intravenous APAP in the USA, compared with the oral formulation,9 which when used in high-volume surgeries such as TJA would amount to a substantial cost burden.10 Therefore, in the absence of robust, pragmatic data on this subject, using US national claims data we aimed to (1) provide an overview of intravenous APAP utilization patterns after TJA in the USA, (2) evaluate the impact of intravenous APAP on opioid utilization, length and cost of hospitalization, and opioid-related complications, and (3) compare how this relates to oral APAP. Based on the known dose–response relationship between morphine and opioid-related adverse effects, we postulated a reduction of 25% in opioid utilization to be clinically significant.11


Patient population

We used data from the Premier Healthcare database (Premier Healthcare Solutions, Charlotte, North Carolina).12 This all-payer database contains data on nationwide hospitalizations including International Classification of Diseases-9th Revision Clinical Modification (ICD-9) codes, Current Procedural Terminology codes, as well as billed items.

Inclusion/Exclusion criteria

The study sample included patients from January 2011 to December 2016 who underwent lower extremity TJA (ICD-9 code 81.51 and 81.54, for hip and knee arthroplasties) and received at least one of the seven most commonly prescribed opioids (fentanyl, hydrocodone, hydromorphone, meperidine, morphine, propoxyphene, and oxycodone).13 Exclusion criteria were unknown gender or discharge status (n=591), cases classified as an outpatient procedure (n=3615), and cases with opioid utilization greater than the 95th percentile opioid utilization (to address outliers, n=32 008) or were treated at a hospital performing <30 TJAs (to ensure sufficient sample size per cluster, n=346).14

Study variables

An analysis plan was created a priori to define all study variables, the main analyses and sensitivity analyses. The main effect of interest was the use of intravenous APAP, which was defined from inpatient billing data. This was categorized into 1 (1000 mg), or >1 doses, on the day of surgery (POD 0, which may also include preoperative use as we did not have specific timestamps, just the day of billing), postoperative day 1 (POD 1), or later (POD 1+). The same categorization was used for oral APAP separately. This was represented by six separate variables as there can be overlap and breaks in intravenous/oral APAP administration regimens which prohibit the creation of a single (mutually exclusive) variable. The contrast between 1 dose and >1 dose of APAP was determined based on the observed patterns of use where most patients received either one dose or four doses per day, with a minority of patients receiving two or three APAP doses. The main outcome of interest was opioid utilization over the entire hospital stay. This was additionally categorized into opioid use on POD 0, 1 and 1+. Opioid utilization was based on billing for opioids and was converted into oral morphine equivalents (OME) calculated using the Lexicomp ‘opioid agonist conversion’15 and the GlobalRPH ‘opioid analgesic converter’.16 The secondary outcomes were length and cost of hospitalization, and opioid-related adverse effects (categorized into respiratory and gastrointestinal complications).13 In addition, we used billing for intravenous naloxone as a marker for an opioid-related complication.

Patient demographic variables included age, gender, and race (white, black, other). Healthcare-related variables were insurance type (commercial, Medicaid, Medicare, uninsured, other), hospital location (urban, rural), hospital size (<300, 300–499, ≥500 beds), hospital teaching status, and the median annual number of TJAs performed per hospital. Procedure-related variables included year of procedure and type of procedure (hip or knee arthroplasty). Anesthesia/analgesia-related variables included use of neuraxial anesthesia, patient-controlled analgesia (PCA), peripheral nerve blocks, and non-opioid analgesics (intravenous/oral APAP, gabapentin/pregabalin, NSAIDs, cyclo-oxygenase-2 [COX-2] inhibitors, and ketamine). Overall comorbidity burden was assessed using the Quan adaptation of the Charlson Comorbidity Index.17 We additionally included substance use/abuse (including smoking), chronic pain conditions, psychiatric comorbidity variables, and opioid use disorder given their association with opioid utilization.18 19


First, univariable associations between intravenous APAP use and all study variables were assessed using standardized differences instead of p values as group differences easily reach statistical significance in large sample sizes. A standardized difference of 0.1 (or 10%) generally indicates a meaningful difference in covariate distribution between groups.20 21 In addition, trends in opioid utilization were assessed before and after the introduction of intravenous APAP to the US market, as well as the interhospital variation in the utilization of intravenous APAP. Mixed-effects models measured associations between the use of intravenous APAP and outcomes; effect estimates were then compared with those for the use of oral APAP. Models were adjusted for variables based on clinical significance and/or imbalance between groups; given the large study sample size and decreased need for parsimony, in practice this resulted in adjustment for all available variables. Models included a random intercept term that varies at the level of each hospital, accounting for correlation of patients within hospitals as they are likely to experience similar care. Two-sided statistical significance was defined with an alpha set at 0.05. Due to the skewed nature of the continuous outcome variables (opioid utilization, length of stay, and cost of hospitalization), these outcomes were modeled using the gamma distribution with a log link function using PROC GLIMMIX in SAS V.9.4 statistical software.22 23 For these variables, instead of ORs, percent (%) change compared with the reference is reported along with 95% CIs. Bonferroni adjustments were applied to adjust for multiple testing (96 hypotheses) and reduce the risk of type I errors (false positive associations).

Sensitivity analyses

Two sensitivity analyses were performed to assess robustness of our results. First, we compared effect estimates for intravenous APAP use with estimates from a control non-opioid analgesic, NSAIDs, as we would theoretically expect both to be associated with improvements in opioid utilization, opioid-related adverse effects, and length and cost of hospitalization. The second sensitivity analysis addressed the potential for confounding by indication: intravenous APAP potentially used in patients with more pain, thus requiring more opioids. This bias would theoretically result in intravenous APAP being associated with increased use of opioids, thus increasing the threshold to demonstrate any potential beneficial effect of intravenous APAP. Therefore, we restricted our sensitivity cohort to hospitals in which intravenous APAP was used in ≥50% of TJA cases under the assumption that hospitals with a high intravenous APAP use did so as part of standard postoperative pain protocols, and thus would be less susceptible to confounding by indication.


Records for 1 039 647 patients undergoing total hip (n=356 654) or knee arthroplasty (n=682 993) at 607 hospitals between 2011 and 2016 were extracted. Table 1 lists all study variables by intravenous APAP use. Overall, 23.6% (n=245 454) of patients received intravenous APAP; this increased from 3.7% in 2011 to 29.3% in 2016 (from row percentages instead of column percentages depicted in table 1). Among this cohort 56.3% (n=138 180) received just one single dose of intravenous APAP on POD 0, while 41.1% (n=100 897) received >1 dose; these percentages do not add up to 100% as there can be overlap and breaks in APAP administration regimens. On POD 1, 18.5% (n=45 528) of the intravenous APAP cohort received a single dose and 19.3% (n=47 452) multiple doses. On POD 1+, 2.8% (n=6891) received a single dose and 2.6% (n=6326) multiple doses. Use of intravenous APAP was more common in whites, in higher volume hospitals, and patients receiving other non-opioid analgesics. Interestingly, use of neuraxial anesthesia and PCA was lower in those receiving intravenous APAP, all with standardized differences >0.1.

Table 1

Study variables by intravenous APAP use

Figure 1 demonstrates substantial interhospital variations in intravenous APAP utilization among TJA recipients (median 10.3%, IQR 0%–38.9%) with a significant number of hospitals not using intravenous APAP at all. Figure 2 displays trends in median per-patient opioid utilization stratified by intravenous APAP utilization categories. Overall, a decreasing trend in opioid utilization is seen, with the majority of intravenous APAP categories below the general opioid utilization trend line only after 2014.

Figure 1

Interhospital variation in intravenous (IV) APAP administration to hip and knee arthroplasty recipients. APAP, acetaminophen.

Figure 2

Trends in unadjusted median per-patient opioid utilization (in oral morphine equivalents), stratified by intravenous (IV) APAP use categories. APAP, acetaminophen; POD, postoperative day.

Table 2 lists unadjusted outcomes by intravenous APAP (upper part) or oral APAP (lower part) use. For both intravenous and oral APAP, no clear consistent patterns emerge regarding expected reduced opioid utilization coinciding with equal reductions in opioid-related adverse effects or length and cost of hospitalization. In addition, unadjusted effects on opioid utilization (compared with no intravenous or oral APAP use) appear minimal except for patients receiving one dose of intravenous APAP on POD 1. This appears largely driven by a reduction in opioid utilization on POD 0, suggesting that patients who receive intravenous APAP on POD 1 are those receiving less opioids on the day of surgery.

Table 2

Unadjusted outcomes by intravenous APAP (upper) or oral APAP (lower) use

Table 3 lists the results from our mixed-effects models. In general, relative adjusted effects of intravenous APAP on opioid utilization appeared minimal. After adjustment for relevant covariates, particularly use of >1 dose of intravenous APAP (compared with no use) on POD 1 was associated with −6.0% (CI −7.2% to −4.7%) reduced opioid utilization; this was −10.7% (CI −11.4% to −9.9%) for use of >1 dose oral APAP on POD 1. These effects extended to length and cost of hospitalization. Interestingly, while not in line with equally great reductions in opioid utilization (and thus likely not connected), patients receiving >1 dose of intravenous APAP on POD 0 showed reduced odds of respiratory (OR 0.80, CI 0.68 to 0.94) and gastrointestinal (OR 0.77, CI 0.67 to 0.89) complications and naloxone use (OR 0.76, CI 0.65 to 0.89).

Table 3

Adjusted associations between intravenous APAP (upper) or oral APAP (lower) and outcomes

Oral (compared with intravenous) APAP showed more pronounced and consistent patterns in opioid reduction coinciding with reduced odds for opioid-related adverse effects, particularly in those receiving >1 dose on POD 1. Counterintuitively, both intravenous and oral APAP use on POD 1+ were associated with consistently increased opioid utilization, length of stay, costs, and adverse events, suggesting that intravenous and oral APAP use beyond POD 1 may be selectively used in patients with more pain, thus requiring more opioids.

When stratifying results by total hip and knee arthroplasty, patterns similar to those observed in tables 1–3 emerged (data not shown).

Results from the sensitivity analyses are displayed in the online supplementary material. Compared with intravenous APAP, the use of NSAIDs on particularly the day of surgery showed a more consistent pattern of reductions in opioid utilization paired with lower odds for opioid-related adverse effects. Even after addressing the potential for confounding by indication, minimal opioid-sparing effects of intravenous APAP were seen. Moreover, oral APAP demonstrated slightly more pronounced opioid-sparing effects.

Supplemental material


The results of this analysis including more than one million patients undergoing total hip or knee arthroplasty in the USA from 2011 to 2016 show that approximately one in five patients received intravenous APAP; more than half of these were only given a single intravenous APAP dose on POD 0. Since its market introduction in the USA in late 2010, intravenous APAP use has increased more than eightfold in the presence of generally decreasing trends in opioid utilization among TJA patients. Crucially, compared with oral APAP, we found no superior benefit of intravenous administration, neither in terms of opioid utilization, opioid-related complication risk, length of hospital stay, nor cost of hospitalization. These results were further corroborated in sensitivity analyses looking at a control non-opioid analgesic and addressing confounding by indication. Overall, APAP was associated with a weak and inconsistent effect on opioid utilization, and a moderate impact on the incidence of opioid-related complications. Importantly, no clinically significant reduction in OMEs (>25%) could be observed in any of the groups. Seemingly paradoxically, APAP administration beyond POD 1 was uniformly associated with a significant reverse effect: higher opioid utilization, higher opioid-related complication risk, longer hospital stay, as well as higher cost of hospitalization. While speculative, this finding may represent administration in patients with higher levels of pain, which is known to disproportionately complicate convalescence after major total lower extremity joint replacement.1

APAP is a widely used, first-line analgesic-antipyretic with very low potential for complications and side effects, if dosed correctly and adjusted to patients’ liver function3; its efficacy in the postoperative setting is well documented, with a long-standing safety record after oral administration.24 Due to extensive first-pass effects, APAP’s oral bioavailability is limited to 60%–70%, dependent on dose and formulation; intravenous APAP has been suggested as a more effective mode of administration.24 However, owing to high cost (particularly in the USA) and unclear comparative advantage, the utility of intravenous APAP has been challenged from both clinical and economic standpoints. Indeed, several initiatives to reduce intravenous APAP utilization have been proposed; Vincent et al for example reported savings close to half a million dollars at one urban medical center, invoked particularly by avoidance of intravenous APAP in the perioperative setting.10 25 Even in light of these controversies, in the perioperative setting intravenous APAP appears to be an appealing addition to multimodal analgesic strategies particularly in those who cannot tolerate oral medication. Moreover, given APAP’s long history of use, practitioners are universally familiar with this substance.

Multimodal analgesia has been shown to improve patient comfort, expedite postoperative mobilization and recovery, and prevent development of chronic pain after TJA (1). Moreover, the incidence of immediate postoperative adverse events can readily be reduced by the use of multiple analgesic modalities. In patients undergoing TJA, Memtsoudis et al26 found a stepwise beneficial effect for each non-opioid analgesic modality that is added to the strategy, including regional anesthesia, APAP, steroids, gabapentinoids, NSAIDs, COX-2 inhibitors, or ketamine. Patients who received more than two additional modalities (as compared with ‘opioids only’) were found to have significantly lower risk for respiratory (−19%) and gastrointestinal (−26%) complications, lower opioid utilization (−18.5% OME), and shorter length of stay (−12.5% days). However, while the authors noted that the most profound effects were found for NSAIDs or COX-2 inhibitors when added to the multimodal strategy, the comparative role of both intravenous and oral APAP was unclear. In our data set, NSAIDs were administered in 55% of patients receiving intravenous APAP and in 46% of patients not receiving intravenous APAP on the day of surgery.

Reduction in the prescription of opioids is another increasingly important goal of multimodal analgesia. This is especially the case as the overprescription of this drug class has contributed to the creation of a national crisis in the USA and other parts of the world.27 Non-opioid constituents of multimodal analgesia ideally avoid opioid-related side effects, development of tolerance, and necessity for distribution of outpatient opioid prescriptions. Therefore, they are highly sought after. Regional anesthesia is one candidate, demonstrating efficacious results with regard to pain and complication outcomes28; however, its use is associated with increased complexity, requirements for training, and potential complications. Widespread application of regional anesthetic techniques is not deemed feasible in all settings and is contraindicated in a number of settings, particularly when coagulation disorders are present.29 Hence, a pharmacologic solution to the problem seems more desirable to many practitioners. Multiple smaller studies demonstrate beneficial effects of intravenous APAP as such a solution, including lower pain scores,30 31 shorter hospital length of stay,32 and better patient comfort. On the other hand, O’Neal et al33 found no additional analgesic or opioid-sparing effects of oral or intravenous APAP when administered to patients immediately after TJA in 174 patients. Meta-analyses and larger population-based studies yield similarly equivocal results. Blank et al34 summarized 17 prospective studies where intravenous APAP’s ability to reduce opioid consumption over 24 hours after abdominal surgery was analyzed. While no significant impact of intravenous APAP on 12-hour or 24-hour narcotic requirements could be detected overall, subgroup analysis suggested that patients undergoing open laparotomy would benefit most from intravenous APAP, and that other analgesic agents showed a more pronounced effect with regard to the main outcome. In line with this, Wasserman et al35 conducted a claims-based study involving 181 640 patients who underwent open colectomies. The authors found a reduction of opioid utilization predominantly when more than one dose of intravenous APAP was used; however, the demonstrated effect was marginal and did not reach clinical significance. Importantly, no meaningful difference between oral and intravenous APAP administration could be detected. Population-based studies in other surgical cohorts have additionally questioned the utility of intravenous APAP. Mörwald et al36 analyzed analgesic prescription patterns in 117 269 patients undergoing lumbosacral spinal fusion and were not able to detect a reduction of opioid utilization or opioid-associated adverse events when intravenous APAP was administered. Similar to the current study, the authors found that most patients (73.6%) only received one single dose of APAP on POD 0. Ambiguous results were also documented after TJA. Yang et al8 conducted a meta-analysis of four studies involving 865 patients who underwent total hip or knee arthroplasty. The authors found significantly lower weighted median pain scores in patients who received additional intravenous APAP as part of a multimodal analgesic strategy as well as lower opioid requirements. Sun et al6 conducted a meta-analysis of two randomized controlled studies including 236 patients, on whom TJA was performed. No significant benefit of intravenous APAP, neither on pain scores nor on opioid consumption, could be detected over oral APAP. The authors of all mentioned meta-analyses acknowledge that wide heterogeneity prevails in the analyzed studies, and that higher quality, larger studies would be warranted to elucidate the suggested effects.

The validity of results gained from smaller studies were frequently doubted as they are thought to be underpowered for small effects and for those with a high interindividual variability.37 In contrast to the existing studies on APAP’s effects in patients undergoing major lower extremity joint replacement, our study is based on a large, real-world population sample, thus representing actual care and minimizing a potential underpowering bias. Moreover, use of intravenous APAP in trials may not translate to use in actual clinical practice as we found that most used just one dose. Nevertheless, several limiting factors must be mentioned. First, the analysis is retrospective in nature and the data have not primarily been sampled for scientific analysis. Pain scores as a direct comparator of analgesic impact have therefore not been sampled. However, this disadvantage is partially offset by the size of the data set. Opioid utilization, opioid-related complications, and surrogate parameters (eg, administration of naloxone) were used instead of pain scores as more robust, comparable, and clinically relevant outcome parameters. Appropriate sensitivity analyses were conducted to address common biases of population-based analytics. Moreover, data reflect individual clinical practices rather than strict protocols and thus contain considerable heterogeneity in practice. This heterogeneity is however characteristic of a realistic sample and does not reduce the validity of the analysis. Particularly for studies with a focus on variables with high interindividual variability, the population-based set-up has been repeatedly advocated.38

Future research should evaluate whether patients could profit from an intensified APAP dosing regimen beyond the immediate postoperative period. Moreover, the comparative impact of NSAIDs and other adjuncts to a multimodal analgesic regimen should be evaluated more definitively and exclusively, as many of these have a comparatively good safety profile and low complication risk as APAP does.

In conclusion, using a large population sample, our study found that intravenous APAP shows only marginal effectiveness in reducing opioid utilization and opioid-related complications in patients undergoing lower extremity TJA. The maximum recommended daily dose is 4 g in healthy adults,3 thus raising the question whether APAP was used in the most efficacious dosage. The intravenous administration route showed no additional benefits; whenever feasible, APAP can be administered orally, producing similar, if not slightly better effects.



  • Contributors OS contributed to the generation of the research plan, helped interpret the analyzed data, and drafted the manuscript. JP contributed to the generation of the research plan, analyzed the data, and contributed to drafting the manuscript. HNL helped interpret the raw and analyzed data, and reviewed the manuscript. MMB contributed to the generation of the research plan, helped interpret the analyzed data, and assisted in drafting the manuscript. SBL provided field-specific guidance, helped to generate relevant variables, and interpret the analyzed data. NZ contributed to the generation of the research plan, helped in analyzing the data, and reviewed the manuscript. MM contributed to the generation of the research plan, provided senior statistical advisory, generated sensitivity analyses, audited the statistical methodology, and reviewed the manuscript. JB contributed to the generation of the research plan, helped in analyzing the data, and reviewed the manuscript. JL provided field-specific guidance, and helped to generate relevant variables and interpret the analyzed data. LMG contributed to analyzing the data and reviewed the manuscript. CSM provided senior field-specific guidance, helped analyze the data, and reviewed the manuscript. SM provided senior clinical advisory, assisted in the generation of the research plan, helped interpret the analyzed data, and assisted in drafting and reviewing the manuscript.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Ethics approval This retrospective cohort study was approved by the Mount Sinai Hospital Institutional Review Board (project #14–00647).

  • Provenance and peer review Not commissioned; externally peer reviewed.