Non-steroidal anti-inflammatory drugs in the perioperative period

Learning objectives.

By reading this article you should be able to:

• •Develop a safe plan for incorporating non-steroidal anti-inflammatory drugs (NSAIDs) into a multimodal approach to pain management.
• •Differentiate between NSAIDs based on their efficacy, duration of action and risk profile.
• •Understand the key factors that increase the risk of adverse events associated with NSAIDs.
• •Apply this knowledge in clinical practice to optimise analgesia while minimising the risks.

Key points.

• •The use of NSAIDs requires individualised risk-benefit analysis.
• •High doses and prolonged use increase all the risks associated with NSAIDs.
• •The risk of renal injury is increased in hypoperfused states, pre-existing renal dysfunction and drug interactions.
• •The cyclooxygenase (COX)-2 selectivity of an NSAID indicates its cardiovascular risk.
• •NSAID hypersensitivity reactions are related to COX-1 inhibition.

NSAIDs are highly effective analgesics. Multimodal regimens for analgesia have been emphasised recently to limit the use of opioids, after controversies over their potentially harmful effects. Despite being an essential component of multimodal analgesia, the prescription use of NSAIDs is often met with hesitation. Although caution is advisable in the perioperative period, completely avoiding NSAIDs based on potential risks is unwarranted. The prescription of NSAIDs should be guided by careful consideration of known risks, which can be minimised by selecting the most appropriate NSAID, at a specific dose, over a specified timeframe, together with an adequate consent process. In this review we examine the different NSAIDs available and factors affecting their analgesic efficacy. We then offer guidance on their use by known patient-specific factors, with mitigation strategies to help minimise risk. These recommendations may change practice by increasing the use of short courses of NSAIDs for patients undergoing surgery, to optimise multimodal analgesia.

The role of prostaglandins in nociception

Cyclooxygenase (COX) converts free arachidonic acid found in the cell membrane to prostaglandins D₂, E₂, F_2⍺ and I₂, and thromboxane-A₂ (TxA₂). Individual prostaglandins each preferentially interact with their specific G protein-coupled receptors. Prostaglandins are synthesised as part of the wider local inflammatory process. Local somatosensory neurones contain prostaglandin receptors, which when activated sensitise the neurone through various intracellular mechanisms such as phosphorylation of sodium channels.¹ This lowers the threshold at which these sodium channels open, thereby increasing neuronal excitability and promoting the transmission of noxious stimuli from the peripheries to the CNS. Figure 1 gives an overview of the mechanisms through which COX-1 and COX-2 perform their respective functions.

Fig 1.

Mechanisms of action for COX-1 and COX-2: a visual representation of their distinct functions within the body. Ang II, angiotensin II; COX, cyclooxygenase; GI, gastrointestinal tract; HCO₃⁻, bicarbonate; Na Ch, sodium channels; PG, prostaglandin; PKA, protein kinase A; TRPV, transient receptor potential vanilloid.

Sub-types of cyclooxygenase: COX-1 and COX-2

The COX-1 isoform is primarily responsible for the formation of TxA₂ and prostaglandins that fulfil many physiological roles such as gastric mucosal protection. The COX-2 isoform is induced at sites of inflammation leading to synthesis of prostaglandins that result in inflammation, peripheral sensitisation of nociceptors and pyrexia. For instance, COX-2 expression increases in the hypothalamus in response to circulating cytokines, triggering PGE₂ synthesis, which leads to the response of pyrexia. However, COX-2 is also ‘constitutive’ in some critical areas such as the kidneys and vascular endothelium, where it catalyses the formation of prostacyclin (PGI₂).

Non-selective NSAIDs inhibit both COX-1 and COX-2, whereas COX-2-selective drugs such as ‘coxibs’ (a class of drugs that selectively inhibit COX-2; examples include celecoxib and etoricoxib) and diclofenac have a higher affinity for COX-2 than COX-1. However, the COX-2 selectivity of these drugs can be variable over time. Initially, diclofenac inhibits both isoforms of COX, but as plasma concentration decreases, its effect on COX-1 diminishes while continuing to inhibit COX-2. Similarly, celecoxib also has a small inhibitory effect on COX-1 initially, which corresponds with the increased peak plasma concentration of the drug.

Clinical efficacy

Table 1 compares commonly used NSAIDs; their efficacy is evaluated using the number needed to treat (NNT). The number needed to treat refers to the number of patients with moderate to severe postoperative pain that would need to be treated for one patient to experience at least 50% pain relief with an analgesic compared with placebo. The number needed to treat of intramuscular morphine 10 mg is 2.9, similar to most NSAIDs.² NSAIDs have been shown to reduce the use of opioids by ∼40%.⁵

Table 1.

Comparison between NSAIDs used commonly for acute postoperative pain. NNT refers to the number of patients with moderate to severe post-surgical pain that would need to be treated for one patient to experience at least 50% pain relief with an analgesic compared with placebo. Duration of action is inferred from the median time to rescue medication. These data are obtained from the Cochrane collaboration.² ∗When known, the COX-1/COX-2 IC₅₀ ratio is quoted.^3†Data quoted from oral diclofenac potassium. P.R., per rectum.

NSAID	Single dose (mg)	Routes	NNT	Duration of action (h)	Time to peak plasma concentration (h)4		COX-1: COX-2 selectivity∗	Notes
					Fed	Fasted
Ibuprofen	200–400	P.O., i.v., topical	2.7–2.5	4.7–5.6	2	1.3	Not selective	Doses above 400 mg do not enhance analgesia
Naproxen	250–500	P.O.	3.4–2.7	8.9	1.9	1.9	Not selective	Higher GI ulceration risk compared with ibuprofen
Diclofenac†	25–50	P.O., p.r., i.v., i.m., topical	2.4–2.1	3.1–4.5	1.5	0.5	Similar to coxibs	Slow-acting (enteric-coated) formulations have vastly inferior efficacy for acute pain
Celecoxib	200	P.O.	4.2	6.6	3.2	2.7	1:30	Inhibits CYP2D6
Etoricoxib	90–120	P.O.	1.7–1.8	20	3	1	1:344	Metabolism is reduced in hepatic impairment
Parecoxib	20–40	I.V., i.m.	2.4–2.2	6.9–10.6	30 min, i.v.		1:62	Prodrug for valdecoxib Precipitates in Hartmann's solution Reduce dose to 20 mg in patients with moderate hepatic impairment or weight <50 kg Avoid if history of sulfonamide allergy
Ketorolac	10–30	I.V., i.m.	2.5	4.5	5 min, i.v. 50 min, i.m.		Not selective	Limit use to 2 days Contraindicated during breastfeeding
Indomethacin	50–100	P.O., p.r.	Not effective				Not selective	Use for as little as 7 days can result in serious GI toxicity
Meloxicam	7.5–15	P.O., i.v.	No data		5–6		Similar to coxibs	Recent development of i.v. formulations Half-life of ∼20 h

NICE guidelines on NSAIDs

NICE guidelines (NG180) affirm many of the principles detailed in this article.⁶ The guidelines endorse oral ibuprofen for most postoperative pain, with the exception of hip fracture cases. If the oral route is unavailable, then intravenous non-selective NSAIDs are preferred over coxibs. NICE acknowledges that concerns over cardiac and renal complications have previously constrained NSAID use after surgery. Despite these concerns, the guidelines assert that traditional oral NSAIDs can be safely incorporated into postoperative care with judicious management. This represents a shift towards increased use of these drugs in the postoperative setting.

Combination therapy

There is substantial value in using pharmacological synergism by combining multiple analgesics that have different mechanisms of action. This strategy aims to increase efficacy and allow dose reduction, thereby reducing the risk of adverse effects without compromising analgesia. A prime example of this is a regimen involving a combination of ibuprofen 200 mg with paracetamol 500 mg. This combination has proved to be particularly effective with an NNT of 1.5.² Similarly, the combination of ibuprofen 200 mg with caffeine 100 mg also has greater efficacy, as indicated by an NNT of 2.1. Therefore, this approach presents a viable strategy to enhance both patients' safety and analgesic efficacy.

Basic pharmacokinetics

In general, NSAIDs have a high bioavailability (80–100%) and reach peak plasma concentration within 2–3 h. These acidic drugs have a pK_a of 4–5 and are extensively bound to plasma proteins, which can lead to competition for the same binding sites with drugs such as warfarin. The liver metabolises NSAIDs, and their metabolites are excreted by the kidneys.

Interpatient variability in response

At the individual level, there may be significant variations in efficacy between different NSAIDs. According to the British National Formulary (BNF), only 60% of patients will respond to a single NSAID. Reasons for this include differences between NSAIDs in non-prostaglandin mechanisms of action, and individual differences in drug metabolism such as CYP2C9 polymorphisms, leading to the significant variations in the half-lives of ibuprofen and celecoxib.⁷ Patients whose pain does not respond to an NSAID from one class may derive analgesia from a different class of NSAID.

The role of NSAIDs via non-systemic routes

Topical NSAIDs have been used effectively for acute musculoskeletal pain with an adverse effect profile similar to that of placebo.⁸ However, there is limited research on the use of adding NSAIDs as an adjuvant to local anaesthetic agents for wound infiltration, and the potential benefits and risks of NSAIDs via this route remain unclear.⁹

Pre-emptive use of NSAIDs in adult and paediatric perioperative practice

Using NSAIDs before the surgical incision (pre-emptive analgesia) is postulated to reduce central sensitisation and increase analgesic efficacy compared with NSAIDs initiated after surgery. However, there are no clinically significant differences in pain or opioid use between pre- and post-incision NSAIDs.¹⁰ Furthermore, pre-incision NSAIDs will affect platelet function and therefore potentially increase blood loss (see NSAIDs and the haematological system.)

The impact of food on NSAID absorption and peak concentrations

Taking NSAIDs with food is a common practice, which may delay the time to peak plasma concentrations of the drug. However, the time to peak plasma concentrations is an important determinant of efficacy.¹¹ A lower dose of 200 mg of ‘fast-acting’ ibuprofen has a lower NNT (2.1) than that of a standard 200 mg dose (2.7). Fast-acting formulations of ibuprofen reach peak plasma concentrations in about 20–40 min, whereas standard ibuprofen takes around 90–120 min. Higher peak plasma concentrations of NSAIDs may result in increased concentrations within key sites of action such as synovial fluid, from which the drug diffuses out relatively slowly.¹¹ Taking NSAIDs with food may reduce dyspepsia and GI ulceration, but this has never been formally studied in humans. Evidence from animal studies does not support this practice.¹²

NSAIDs and the renal system

Haemodynamically-mediated acute kidney injury

Both COX-1 and COX-2 enzymes are involved in the synthesis of prostaglandins in key areas of the kidney, including the glomerular endothelium, medullary interstitial cells and collecting tubules. COX-2 expression is always present to some degree in the kidneys, hence selective COX-2 inhibitors still confer a similar renal risk profile to non-selective NSAIDs. Prostaglandins act to preserve glomerular filtration rate through a vasodilator action on the renal afferent arterioles. Renal blood flow is normally independent of prostaglandins, and at present there is no proved association between acute kidney injury (AKI) and NSAID use in patients with normal kidney function.¹³ However, during times of physiological stress when vasoconstrictors are present, prostaglandin-mediated vasodilation helps counteract any potential reduction in perfusion to the kidney and maintains the glomerular hydraulic pressure, which is the main driving force for glomerular filtration. Therefore, the use of NSAIDs should be avoided when significant fluid shifts or blood loss is expected, especially when other analgesic techniques provide sufficient analgesia for the initial perioperative period, such as intrathecal opioids for major colorectal surgery.

Drug interactions, chronic kidney disease and kidney injury

Renal blood flow autoregulation is altered by the combination of NSAIDs and renin–angiotensin system inhibitors (RAS-I). Renin–angiotensin system inhibitor treatment does not alter the risk of NSAID-associated AKI when NSAIDs are used for ≤3 days.¹⁴ Beyond this duration, harm may become apparent. The number needed to harm for one episode of AKI after a year's treatment with an NSAID is 347.¹⁵ Loop diuretics in particular increase the risk of AKI and concurrent use should be avoided.¹⁵

NSAIDs do not appear to have an effect on the decline of glomerular filtration rate (GFR) in patients with chronic kidney disease (CKD) stages 1–3.¹⁶ This is in contrast to CKD stages 4 and 5, where a decline in estimated GFR (eGFR) is significantly faster with NSAID therapy. It is reasonable that CKD stage 1–2 patients can use NSAIDs for up to 7 days.¹⁷ It should be noted that this does not apply to high doses of NSAIDs, which can significantly increase the risk of accelerated decline in renal function.¹⁸ For the purpose of this discussion, ‘high-dose’ is considered as per the guidelines provided by the Agency for Healthcare Research and Quality, although it is important to note that there is no universally accepted definition for ‘high-dose’ NSAIDs.¹⁹ For instance, doses of ibuprofen ≥2.4 g per day or naproxen doses >1 g per day are considered high.

NSAIDs and the cardiovascular system

Selective COX-2 inhibitors, such as diclofenac and the coxibs, can lead to a prothrombotic state through inhibition of (mostly COX-2 mediated) prostacyclin synthesis without inhibiting (COX-1-mediated) TxA₂ synthesis. Patients at high risk of cardiovascular disease (CVD) are particularly vulnerable.²⁰ Coxibs and diclofenac have a place in acute pain management, but should only be used for a short period of time and should be avoided in patients who have (or are likely to have) CVD, including single perioperative doses.^21,22 Non-selective NSAIDs such as naproxen and ibuprofen have a more favourable risk profile in this regard: naproxen does not seem to be associated with an increased risk of major vascular events, and there is no increase in risk of myocardial infarction (MI) with ibuprofen in doses <1.2 g daily.^20,23

All NSAIDs enhance sodium reabsorption and diminish free water clearance. This can lead to peripheral oedema and frank heart failure (HF). Patients with a previous history of HF have an increased risk of a recurrent episode of HF, with increasing doses being associated with increased risk of HF-related death.²⁴ All NSAIDs can increase blood pressure with an average increase in mean arterial pressure of 5 mmHg, but with considerable variability.²⁵ The National Institute for Health and Care Excellence in the UK recommends avoiding NSAIDs in patients with uncontrolled hypertension.

NSAIDs and the gastrointestinal system

The incidence of serious gastrointestinal (GI) complications such as bleeding, perforation or death is ∼0.76% at 6 months and increases to 1.5% at 1 yr for patients taking NSAIDs continuously.²⁶ Proton pump inhibitors (PPIs) are commonly used to reduce this risk: PPIs reduce the risk of endoscopically detected ulcers by 65%, and the risk of clinically detectable ulcers by >90%.^27,28 PPIs should be used when any risk factors for GI ulceration are present. These include age >65 yrs, a previous history of GI ulceration, concurrent use of aspirin (including low dose), corticosteroids or anticoagulants.²⁹ These risk factors were identified by the American College of Gastroenterology (ACG). According to the ACG, the presence of three or more of these risk factors categorises a patient as high risk, and so a coxib and PPI are cautiously recommended in such cases. It is important to note that any patient with a prior history of complicated peptic ulceration is automatically considered high risk.

High-dose NSAIDs are associated with two to three times increased risk of GI complications, and some NSAIDs are more ulcerogenic than others (indomethacin>naproxen>ibuprofen>coxibs). Duration of NSAID therapy is also a key determinant of NSAID-mediated GI risk, with NSAIDs inducing GI effects normally after 2–3 months.³⁰ For patients who are at high risk of both GI and cardiovascular NSAID-related adverse events, it is recommended to avoid using NSAIDs altogether. In this cohort of patients coxibs carry an unacceptable high risk of cardiac events, whereas non-selective NSAIDs carry an unacceptable risk of adverse GI events, even when taken with PPI cover.

In summary, the perioperative risk of GI ulceration and bleeding can be mitigated by (i) using NSAIDs with lower ulcerogenic risk profiles; (ii) avoiding high doses; (iii) utilising short courses; and (iv) giving with a PPI when GI risk factors are present. NSAIDs should be avoided in patients at high risk of both CVD and GI ulceration.

NSAIDs and bowel anastomotic leakage

The use of NSAIDs in the setting of colorectal cancer surgery is still a subject of debate with most studies being observational and retrospective. Randomised control trials are needed to clarify this issue. The Enhanced Recovery After Surgery (ERAS) society colorectal guideline considers the evidence inconclusive and recommends using NSAIDs, citing evidence that patients receiving NSAIDs have faster return of flatus, stool, oral diet and significantly lower opioid consumption.³¹

NSAIDs and inflammatory bowel disease

It is estimated that NSAIDs may cause clinical relapse in ∼20% of patients with quiescent inflammatory bowel disease (IBD).³² Coxibs do not appear to be associated with relapse of IBD, but caution should still be exercised.³³ The decision to use NSAIDs in patients with IBD should be in consultation with a gastroenterologist.

NSAIDs and the haematological system

Platelets only express the COX-1 isoform, which catalyses the formation of TxA₂. Coxibs at therapeutic concentrations thus do not impair platelet function or do not cause adverse perioperative bleeding.³⁴ In general, non-selective NSAIDs do not increase the risk of bleeding, especially when given only after surgery. However, when given before orthopaedic surgery (hip fracture patients), NSAIDs do increase the rates of perioperative transfusion.³⁵ It is important to note that this study looked at pre-admission use of NSAIDs, not a short course of NSAIDs given after surgery for acute pain management. NSAIDs before surgery do not seem to increase bleeding in children undergoing tonsillectomy, whereas in adults there is a five-fold increased risk of bleeding.³⁶ If there are concerns regarding any potential contribution of NSAIDs to perioperative bleeding, then they should generally be withheld before surgery for at least four to five times the drug half-life, which can be 3 days for naproxen, or 24 h for ibuprofen. In patients with pre-existing platelet defects, such as von Willebrand disease or thrombocytopenia, NSAIDs should be avoided.

The concurrent use of non-selective NSAIDs with other medications that affect the clotting cascade, such as clopidogrel or heparin, may increase the risk of vertebral canal haematoma with central neuraxial block.³⁷

NSAIDs and the musculoskeletal system

Although COX-2 inhibition has been shown to impair fracture healing in rodent studies, there is no clear evidence to support this concern in humans. Initial poor quality studies have given way to more recent high-quality evidence, showing no effect on clinical or radiographic bone healing in adults or children.³⁸ Similar concerns have been raised with regard to cartilage repair and the impact of NSAIDs on chondrocytes, especially on chondrocyte differentiation in the early phases of chondral repair.³⁹ It is currently unclear what effect NSAIDs have because of the lack of human studies. The known positive impacts that NSAIDs have on early rehabilitation after surgery need to be weighed up against the theoretical negative effect on chondrocyte proliferation.

NSAIDs and hypersensitivity reactions

Hypersensitivity reactions can be either immunological (allergic) or non-immunological (idiosyncratic, pseudoallergic). Non-immunological reactions to NSAIDs are caused by COX-1 inhibition alone, therefore any NSAID causing COX-1 inhibition can induce a non-immunological reaction. Coxibs very rarely produce non-immunological reactions, given their lack of action on COX-1. In contrast, aspirin with its strong COX-1 action is the archetypical NSAID causing non-immunological reactions, hence older terms such as aspirin-exacerbated respiratory disease, which has now been superseded with NSAID-exacerbated respiratory disease (N-ERD), reflecting the ability of all COX-1 inhibiting NSAIDs to evoke a hypersensitivity response. N-ERD is defined as a chronic eosinophilic, inflammatory disorder of the respiratory tract occurring in patients with both asthma and chronic rhinosinusitis with nasal polyps, symptoms of which are exacerbated by NSAIDs, including aspirin. NSAID-exacerbated respiratory disease is estimated to have a prevalence of 1.4% in the general population and 6.9% in asthmatics. Table 2 categorises the major non-immunological reactions.

Table 2.

Classification of non-immunological reactions relating to COX-1 inhibition.⁴⁰

Type	Name	Time to onset (h)	Clinical features	Associated underlying disease	Other terms
1	NSAID-induced asthma and rhinitis	1–3	Bronchospasm, rhinorrhoea, nasal congestion, periorbital oedema	10–30% incidence in patients with asthma and chronic rhinosinusitis	Aspirin-exacerbated respiratory disease, NSAID-induced bronchospasm
2	NSAID-induced urticaria/angioedema	0.5–1.5	Both acute urticaria and angioedema or worsening of chronic urticaria	10–30% incidence in patients with chronic urticaria	Aspirin-induced urticaria and aspirin-exacerbated cutaneous disease
3	Blended reaction	0.5–3	Respiratory and cutaneous features

Diagnosis of non-immunological reactions can be made if the patient gives a history of repeated previous reactions to aspirin or other COX-1 inhibiting NSAIDs. Management involves the avoidance of aspirin and other COX-1 inhibiting NSAIDs.

The vast majority of patients with NSAID-induced asthma will tolerate coxibs. High-quality evidence suggests that coxibs are safe, but manufacturers' labels continue to advise against their use in patients with asthma. Very sensitive patients may react to a minor degree of COX-1 inhibition seen with paracetamol or high doses of coxibs.

Use of NSAIDs during the postpartum period

NSAIDs have an important role in regimens for postpartum analgesia. However, concerns have been raised with their use in certain situations, specifically severe pre-eclampsia. Concerns regarding NSAIDs exacerbating hypertension are in general unfounded and NSAIDs should be used preferentially over opioids.⁴¹ Use of NSAIDs in this group follows the same principles in other patients, in that high doses and prolonged courses should be avoided. If these patients have unstable or increasing blood pressure beyond a 24-h period despite medical management then NSAIDs should be discontinued. In addition, the clinician should also be mindful of the impact of NSAIDs on renal perfusion and platelet function in critically unwell women. NSAIDs should therefore generally be avoided in women with pre-eclampsia-related renal impairment, or HELLP syndrome. Pregnant patients should avoid NSAIDs because of the risk of premature closure of the ductus arteriosus in utero with resulting pulmonary hypertension of the fetus.

NSAIDs in breastfeeding mothers undergoing surgery

Most NSAIDs (excluding aspirin in doses >150 mg per day) are excreted in breast milk in very small amounts. Ibuprofen concentrations have been shown to be unmeasurable in the breast milk of women taking 400 mg of ibuprofen every 6 h for pain after Caesarean section. In contrast, the breastfed infant is exposed to 9–21% of the maternal dose of aspirin, which is found in breast milk for up to 24 h after the initial dose.⁴² This may adversely affect the infant's platelet function and carry a theoretical risk of Reye's syndrome.

Use of NSAIDs for children

The use of NSAIDs in children follows similar principles to those in adults. High doses and prolonged courses should be avoided. Similar contraindications exist, with NSAIDs being avoided in children with a history of IBD, coagulopathy, renal dysfunction, drug interactions and states of hypoperfusion. There are age limitations to using NSAIDs, with a general recommendation to avoid them in children aged <6 months, although in the UK ibuprofen is now available over the counter in formulations for infants aged ≥3 months. The use of NSAIDs in preterm infants for the treatment of patent ductus arteriosus is associated with GI bleeding and renal injury. Recommended doses of NSAIDs in children can be found in the BNF.

Conclusions

The use of NSAIDs can be challenging because of patient-related risks and drug interactions, which may lead to clinicians avoiding them altogether. However, minimising the use of opioids may outweigh their associated risks, and NSAIDs can be incorporated into most multimodal analgesic plans. Healthcare professionals need appropriate information to make a balanced risk–benefit analysis of NSAIDs. This review quantifies the risks of NSAID use in different clinical subgroups, providing a framework for joint decision-making. After assessing a patient's gastrointestinal, cardiovascular and renal risks, adjusting the NSAID type, dose, duration, and incorporating adjuvant medications can minimise but not eliminate the risks. A decision-making tool (Fig. 2) based on key evidence from this article can guide this process.

Fig 2.

A guide to NSAID decision-making for the short-term postoperative use of NSAIDs. Excludes patients with hip fracture.

Declaration of interests

The authors declare that they have no conflicts of interest.

MCQs

The associated MCQs (to support CME/CPD activity) will be accessible at www.bjaed.org/cme/home by subscribers to BJA Education.

Biographies

Edward Halvey BSc FRCA is a fellow in anaesthesia at Sir Charles Gairdner Hospital in Perth, Australia. He was previously a consultant anaesthetist at Northumbria Healthcare NHS, UK, and has a special interest in regional anaesthesia and acute pain medicine.

Edward R. Mariano MD MAS FASA is a professor at Stanford University School of Medicine and chief of anaesthesiology at the Veterans Affairs Palo Alto Health Care System. He was previously a director of the American Society of Regional Anaesthesia and Pain Medicine.

Nat Haslam FRCA is a consultant anaesthetist at South Tyneside and Sunderland Foundation Trust. He specialises in regional anaesthesia and acute pain medicine, and is a board member and Honorary Secretary of Regional Anaesthesia-UK.

Matrix codes: 1D02, 2E01, 3E00