9.6 Non-opioid analgesics

Non-opioid analgesics encompass the non-steroidal anti-inflammatory drugs (NSAIDs) and paracetamol (acetaminophen). All these drugs have analgesic and antipyretic properties. The NSAIDs include acetylsalicylic acid (ASA, aspirin), dipyrone (metamizole), and numerous other drugs in diverse classes. The advantages of non-opioid analgesics include their wide availability, familiarity to patients, effectiveness for milder pain conditions, ease of administration, additive analgesia when combined with other analgesics, and relatively low cost. The disadvantages include a ceiling effect for pain relief and the risk of side effects, including the potential for serious gastrointestinal (GI), renal, and cardiovascular toxicity (NSAIDs), and hepatotoxicity (paracetamol).

Non-opioid analgesics are widely used to manage mildtomoderate pain. In most Western societies, patients use over-the-counter formulations to treat pain related to headache and musculoskeletal ailments. In palliative medicine, they represent the first step of the analgesic ladder when used alone, or with adjuvant drugs, for mild pain and are an important supplement to opioids and adjuvant drugs at higher steps of the ladder (World Health Organization, 1996). The NSAIDs seem to be particularly useful in inflammatory pain and specific conditions such as bone pain. Paracetamol has limited anti-inflammatory effects, but is safer than the NSAIDs for long-term use. Dipyrone (metamizole) is used in some countries to treat pain, but has been removed from the market in others because of an association with life-threatening agranulocytosis.

NSAIDs and paracetamol inhibit the production of prostaglandins (PGs). PGs are lipid-soluble molecules that are produced by enzymatic breakdown of arachidonic acid, which is itself produced from cell-membrane phospholipids. PGs are not stored in the body, but are produced constitutively as mediators of physiological effects in many tissues, or induced as part of the inflammatory cascade in in response to various noxious stimuli. Only 44 years ago, Vane postulated that inhibition of PG synthesis was the main effect of analgesics such as ASA (Vane, 1971). Subsequently, this led to the development of different types of drugs aiming at inhibiting the synthesis of PGs. A large number of NSAIDs were designed, and after the discovery of different isoforms of cyclooxygenase (COX), more selective COX-inhibitors became the focus of NSAID development. The selective COX-2 NSAIDs (coxibs) are the latest group to emerge from these activities.

The two isoforms of COX that have been best characterized are COX-1 and COX-2 (Fig. 9.6.1). COX-1 is largely a constitutive enzyme that produces PGs in many tissues. These PGs protect gastric mucosa, maintain normal kidney function, and promote platelet aggregation. Most COX-2 is induced as part of the complex processes that lead to inflammation, fever, and pain. All NSAIDs inhibit both COX-1 and COX-2, but there is a large variation in the relative effects on the two isozymes. Compared to conventional ‘non-selective COX-1/COX-2’ NSAIDs, coxibs inhibit COX-2 to a much greater extent.

The potential positive effects produced when NSAIDs inhibit PG synthesis include reduced inflammation and less pain. Serious adverse effects are possible, however, if the normal physiological effects of the PGs are disrupted (Abramson et al., 1985; McCormack, 1994; Richardson and Emery, 1995). Managing the latter risk is key to the use of these drugs in palliative care. The therapeutic benefit of the NSAIDs may be compromised if those PGs that are involved in nociception and inflammation cannot be reduced sufficiently, or adverse effects on normal physiology related to PG inhibition outweigh the therapeutic effects.

The tissue release of PGs is complex and is mediated by many types of noxious stimuli, and also by the actions of other substances like bradykinin and cytokines. The latter compounds induce nociception themselves and stimulate the synthesis of PGs and other mediators involved in nociception, such as substance P and calcitonin gene-related peptide (Burke et al., 2006).

Some of the functions affected by PGs also are mediated by leukotrienes. Leukotrienes are responsible for and involved in anaphylactic reactions, broncho-constriction, and chemotaxis, as well as vascular permeability and inflammation (Burke et al., 2006). PG synthesis inhibitors, such as NSAIDs and ASA, do not inhibit the production of leukotrienes, which to some extent may explain the limited efficacy of NSAIDs and ASA when treating pain of inflammatory origin.

The specific mechanisms underlying the beneficial antipyretic and analgesic actions of NSAIDs are not fully understood. Mediators of the inflammatory process—especially the cytokines—induce the production of COX-2 within hours. The induction of COX-2 does not only take place in injured tissues, but also in the central nervous system (CNS). In the CNS, high concentrations of COX-2 cause increasing concentrations of PGs, which, in turn, induce central sensitization and consequently more pain (Baba et al., 2001; Samad et al., 2001). However, under normal physiological conditions, COX-2 is found only at very low concentrations in the body, including brain tissue (Breder et al., 1995; Seibert et al., 1997). In animal studies, endothelial cells express COX-2 mRNA in response to systemic interleukin-1 beta and the resulting production of PGs in the hypothalamus area may play an important role in producing fever (Cao et al., 1996).

In addition to COX-1 and COX-2, a COX-3 has been proposed as a variation expressed by the COX-1 gene (Chandrasekharan et al., 2002). The existence of COX-3 in the CNS of humans has been controversial, and discussions regarding COX variants and their impact on antipyretic and analgesic effects are ongoing (Warner and Mitchell, 2002; Kis et al., 2003, 2004; Snipes et al., 2005).

COX-1 is constitutive, which means that it exists and acts continuously in many different tissues (Fig. 9.6.1). In gastric epithelial cells, COX-1 predominates and is responsible for the production of the protective PGI₂ (prostacyclin). Multiple actions of PGI₂ are involved in the protection of gastric mucosa: maintenance of mucosal blood flow, mucus production, secretion of bicarbonate, and also a positive influence on epithelial cell regeneration in order to provide a well-protected gastric mucosa. Thus, the inhibition of the production of PGI₂ by NSAIDs can upset this protective equilibrium (Scarpignato, 1995).

COX-1 is also responsible for the synthesis of PGE₂, which in addition to vasodilatation during inflammatory processes counteracts vasoconstriction in the kidneys. Under normal physiological conditions, the PG synthesis activity in the kidney is low and its role in modifying renal blood flow (RBF) is not of major importance (Ruilope et al., 1986). However, if RBF is critically lowered, glomerular filtration rate (GFR) may be partly restored by vasodilatatory effects of PGE₂. If PGE₂ production is reduced due to the use of NSAIDs, volume depletion of different aetiologies may aggravate the reduction of RBF. Moreover, inhibition of PGs may also result in a higher extracellular concentration of electrolytes such as sodium, which may cause water retention and oedema. In 10–25% of the patients treated with NSAIDs, sodium retention can be found (Burke et al., 2006).

There is growing evidence that nephrotoxicity is not only related to the influence of COX-1, but also to COX-2 activity. COX-2 is also constitutively expressed in the kidney and is regulated in relation to changes in intravascular volume. Metabolites of COX-2 have also been discussed in the regulation of sodium excretion and renin release and the maintenance of RBF (Harris, 2006).

COX inhibition by NSAIDs also reduces platelet aggregation. Thromboxane A₂ (TXA₂) and PGI₂ are produced in platelets. PGI₂ prevents aggregation of platelets, but does not influence endothelial adherence, whereas TXA₂ is mainly responsible for coagulation. NSAIDs temporarily inhibit platelet aggregation by decreasing TXA₂ synthesis via incomplete and reversible inactivation of COX-1. The antithrombotic action of ASA is mainly due to inhibition of TXA₂-mediated platelet aggregation related to a complete and irreversible inactivation of COX-1 in platelets by acetylation of the enzyme (Patrono et al., 2004).

In an effort to minimize side-effects such as GI complications, renal impairment, and increased cardiovascular risk, efforts have been made to develop NSAIDs that selectively suppress the inducible form of COX, that is, COX-2, and also reduce the action of microsomal prostaglandin E₂ synthase (mPGES-1)-derived PGE_2. The coxibs have these effects. PGES-1 was identified as the most important PGE₂ synthase and its inhibition may lead to a relevant reduction of inflammation, fever, and pain in animal models. However, even if mPGES-1 seems to be a paramount target for inhibition resulting in a therapeutic use for any diseases related to inflammation and cancer, current knowledge about its role in clinical settings is limited (Koeberle and Werz, 2009).

NSAID-induced reduction in prostaglandin formation could potentially have other therapeutic effects, but supporting data are yet very limited. Epidemiological surveys of users and non-users of NSAIDs and ASA have shown that NSAIDs may alter the development and growth of malignancies. Clinical trials in patients with familial adenomatosis polyposis have shown the efficacy of NSAIDs in reducing the number as well as the size of colorectal polyps. However, a primary preventive effect has not yet been demonstrated. NSAIDs are also supposed to have a preventive and growth inhibitory effect in extracolonic epithelial malignancies. Pre-clinical studies show promising results with combination treatments of either chemotherapy or radiotherapy with NSAIDs. These NSAID effects in cancer cells may be mediated not only by COX enzymes, but also by interactions with downstream effectors of inflammation (Meric et al., 2006).

NSAIDS also have been proposed to be effective in the treatment of cancer-induced cachexia. The effect can be explained through the ‘cytokine pathway’ in the development of cachexia. Consequently, the blocking effect of NSAIDS on cytokines may slow down the development of cancer cachexia.

In general, the inhibition of cyclooxygenase isoforms is the main mechanism of the non-opioid analgesics, especially the NSAIDs. However, there are several other effects of non-opioid analgesics beyond COX inhibition, and these too may ultimately be shown to be important in the analgesic and anti-inflammatory effects. Interactions with cholinergic, monoaminergic, and endocannabinoid systems may be responsible for anti-inflammatory effects, as well as for the reduction of pain and fever; they also may be responsible for some adverse effects (Hamza and Dionne 2009).

Most NSAIDs are administered orally, and absorption takes place mainly in the upper GI tract; stomach mucosa may also absorb a substantial proportion of the dose, especially at low pH. If administered as suppositories, most of the NSAID crosses the mucosal membrane easily. Recommended maximum doses and pharmacokinetic data of widely used NSAIDs are shown in Table 9.6.1.

With oral administration, effects usually begin within 30 minutes and peak effects, which mirror maximum concentration, occur within 120 minutes. The rate at which effects decline depends on the half-life of each NSAID.

There is evidence that topical administration of NSAIDs can be an effective way of treating pain in special indications. The pharmacology of topical-applied NSAIDs is not well understood (Mason et al., 2004).

Several NSAIDs, such as ketorolac and ibuprofen, are available in injectable formulations, which may be administered intravenously or intramuscularly. There are only a few publications which compare possible advantages and disadvantages of the parenteral and oral routes (Tramer et al., 1998).

All systemically administered NSAIDs are highly protein-bound. Plasma protein binding ranges from 90% to 99%. NSAIDs are predominantly metabolized in the liver by conjugation to sulphate and glucuronide compounds. Hepatic metabolism, which is by the cytochrome P450 system, is subject to many sources of individual variation, including genomic factors and the influence of other drugs. There may be circadian differences as well, but the potential clinical relevance of this phenomenon remains unclear (Burke et al., 2006). Conjugation of NSAIDs in other tissues of the body occurs, but accounts for a small and insignificant percentage of the metabolism.

Most of the NSAIDs are rapidly distributed in all tissues of the body. The more lipid-soluble a NSAID is, the more it will distribute into the CNS. However, the very high plasma protein binding will retain most of the drug within the plasma compartment (Burke et al., 2006).

NSAIDs are eliminated in the urine in free and conjugated forms. The relative amounts of free and conjugated compounds are highly dependent on urinary pH. Higher pH favours more acidic forms of elimination products. A small percentage of the drug can be found in the bile as well, which indicates that excretion also takes place via the intestines into the faeces (Burke et al., 2006).

Although data from populations with musculoskeletal pain or pain due to inflammatory disorders suggest that NSAIDs are likely to be efficacious in the treatment of pain associated with cancer and other advanced medical illnesses, the evidence of efficacy is very limited. A useful measure of relative efficacy is the number needed to treat (NNT), which can be calculated from data aggregated through systematic reviews. The NNT is the number of patients who need to be treated before one patient benefits at a specified level—such as 30% or 50% pain relief for a specified period of time—when compared to placebo. A systematic review of NSAID efficacy in cancer pain, alone or in combination with opioids, was unable to provide a NNT for the analysed studies because few studies reported the percentage of responders as dichotomous data (McNicol et al., 2005). Based on limited data, the authors cautiously concluded that NSAIDs seem to be more effective than placebo for cancer-related pain. The evidence is not adequate to determine comparative efficacy or safety across NSAIDs, and the relative efficacy of NSAIDs in different cancer-related pain mechanisms cannot be established. There are no data specific to other severe illness. Given the short duration of studies, the long-term safety of NSAIDs for cancer pain also has not been established.

NSAIDs administered alone or in combination with other analgesics should be used carefully. Potential adverse effects must be weighed carefully against potential advantages on a case-by-case basis, especially during chronic use. Given the broad role that PGs have in modifying physiological mechanisms, potential adverse effects vary widely.

The most common adverse effects associated with the intake of NSAIDs occur in the GI tract (Table 9.6.2). The effects vary from symptoms without evident end-organ injury, such as nausea, pyrosis, pain, or bloating, to more serious problems, such as gastric and/or intestinal erosions followed by ulcerations. Erosions or ulceration, which can be observed in up to 30% of NSAIDs users, produce a range of associated symptoms of varying severity, and haemorrhage may range from harmless to life-threatening. Ulcers are often asymptomatic until haemorrhage occurs. Compared to non-users, patients who receive long-term treatment with a non-selective COX-1/COX-2 NSAID have an approximate fivefold higher risk of peptic ulcer disease and a fourfold higher risk of upper GI bleeding.

Patients receiving COX-2 selective inhibitors have significantly lower risk of GI toxicity than those receiving a non-selective COX inhibitor (Lanas, 2010). Although COX-1 inhibition is not the only mechanism involved in NSAID-induced GI toxicity, a systematic review of randomized controlled trials has shown that COX-2 selective inhibitors produced significantly fewer gastroduodenal ulcers and clinically important ulcer complications than non-selective NSAIDs (Rostom et al., 2007). Whether the outcome included bleeding events, symptomatic ulcers, or both, the GI risks with coxibs was consistently about half of that with traditional NSAIDs. However, there is evidence, both from this systematic review and from large controlled trials such as CLASS (Silverstein et al., 2000) and SUCCESS-1 (Singh et al., 2006), that this safety advantage is reduced in patients receiving concomitant low-dose ASA treatment. For example, in the SUCCESS-1 study, the risk of ulcer complications in patients receiving naproxen or diclofenac was significantly higher than in those receiving the COX-2 selective inhibitor celecoxib in the absence of ASA; in contrast, there was no significant difference in risk between the two groups of ASA users (Singh et al., 2006). A recent meta-analysis of all available trials including patients taking low-dose ASA combined with either non-selective NSAIDs or COX-2 selective inhibitors indicates a 28% reduction of GI risk in patients taking the combination of ASA + COX-2 (Rostom et al., 2009).

Celecoxib is the only COX-2 inhibitor currently available in the United States. Rofecoxib and valdecoxib were marketed but were withdrawn in 2004 and 2005, respectively, because of concern related to prothrombotic effects and other toxicities associated with long-term use.

The past decade has seen major advances in the prevention and management of ulcer complications, such as a decrease in the prevalence of Helicobacter pylori infection and improved treatment of acute ulcer bleeding. Recent evidence suggests that these developments have been reflected in a change in the pattern of NSAID-related GI complications seen in clinical practice (Lanas et al., 2009). Thus, while the incidence of complications involving the upper GI tract has decreased steadily during the last decade, perforations and bleeding in the lower GI tract have increased. Such findings suggest that, whereas attention has traditionally focused on NSAID-related complications in the stomach or duodenum, a broader perspective regarding the potential adverse effects of NSAIDs in the GI tract as a whole is needed.

Although there is a connection between NSAID use and lower GI haemorrhage, the evidence of NSAID-associated risk of lower GI bleeding is weaker than the evidence of upper GI bleeding. A systematic literature review reported that mucosal breaks or small intestinal injuries were present in up to 71% of NSAID users, and that up to 88% of patients with lower GI bleeding were NSAID users (Laine et al., 2006); the odds ratios (ORs) for upper GI and lower GI bleeding or perforation associated ranged from 1.9 to 18.4 and from 2.5 to 8.1, respectively. The risk of such problems was lower in patients receiving COX-2 selective inhibitors than in those receiving non-selective COX-1/COX-2 inhibitors (Laine et al., 2006).

The risk of GI toxicity varies among the non-selective COX inhibitors as well. Ibuprofen seems to be the least harmful drug in terms of the risk of upper GI bleeding. In comparison to non-users of NSAIDs, patients taking coxibs seem to have a low relative risk of GI bleeding followed by ibuprofen and diclofenac. A medium risk seems to be present for flurbiprofen, indomethacin, and naproxen, and a high risk has been proposed for ASA and ketorolac (Hernandez and Rodriguez 2000).

The risk of GI toxicity also varies with many other factors. The duration of the use of NSAIDs seem to be weakly associated with the risk of developing GI bleeding, whereas the incidence of haemorrhage is highly associated with high doses (Hawkins and Hanks, 2000). Specific characteristics and comorbidities also strongly influence the likelihood of adverse GI effects. Risk is related to age above 65 years; Helicobacter pylori infection; peptic ulceration within the last year; simultaneous intake of corticosteroids, low-dose ASA, and/or anticoagulants; far-advanced disease; cardiovascular disease, renal and hepatic impairment, and diabetes mellitus; smoking; and excessive alcohol use (Hernandez and Rodriguez, 2000).

Due to the high risks of GI side effects of traditional NSAIDs in frail patients, the use of gastroprotective strategies prevails, although the evidence is still limited. Proton pump inhibitors (PPIs) and misoprostol both reduce the incidence of gastric and duodenal ulcers, as well as recurrence in patients receiving traditional NSAIDs. Although misoprostol is slightly more effective in preventing gastric ulcers, PPIs are better tolerated, as misoprostol often causes diarrhoea (Rostom et al., 2002; Dubois et al., 2004). Patients at increased risk of GI complications should receive either a non-selective NSAID with an appropriate gastroprotective agent, such as a PPI, or a COX-2 selective inhibitor alone. These two strategies were compared in a randomized, double-blind trial involving 287 arthritis patients who received either diclofenac plus omeprazole, or celecoxib, 200 mg twice daily, for 6 months (Chan et al., 2002). The risk of recurrent ulcer bleeding did not differ significantly in the two groups. More recently, the combined use of a COX-2 selective inhibitor and a PPI to prevent recurrent ulcer bleeding in high-risk patients was studied. In this randomized, double-blind trial, 441 patients with upper GI bleeding received celecoxib, 200 mg twice daily, alone or in combination with esomeprazole, 20 mg twice daily, for 12 months (Chan et al., 2007)⁾. The incidence of recurrent bleeding within 12 months was significantly lower with combination therapy than with celecoxib alone, and there were no differences in discontinuation rate or the incidence of adverse events between the two groups.

Non-selective NSAIDs inhibit both constitutive COX-1 and inducible COX-2, the rate-limiting enzymes involved in the production of PGs and thromboxane. In addition to their role in inflammation and pain, PGs are important mediators of vascular tone, salt and water balance, and renin release. PGE₂ is a mediator of sodium reabsorption in the distal renal tubule and acts as a counter-regulatory factor under conditions of increased sodium reabsorption by limiting salt and water reabsorption (Whelton, 2000). PGI₂ and PGE₂ increase potassium secretion, primarily by stimulating the secretion of renin and activating the renin–angiotensin system, which leads to increased aldosterone secretion. These vasodilatory PGs also increase RBF and GFR under conditions associated with decreased actual or effective circulating volume, resulting in greater tubular flow and secretion of potassium. In healthy hydrated individuals, renal PGs do not play a major role in sodium and water homeostasis (Whelton, 2002). Under conditions of decreased renal perfusion, however, the production of renal PGs serves as an important compensatory mechanism. Numerous conditions are associated with a decrease in renal perfusion, including dehydration, blood loss, congestive heart failure, cirrhosis, diuretic use, and restricted sodium intake. Under these conditions, non-selective NSAIDs may produce adverse effects, including a decrease in GFR. Conversely, in conditions of high salt intake and/or volume expansion, non-selective NSAIDs may induce salt retention, which may elevate blood pressure or make pre-existing hypertension worse.

These effects produced by the non-selective COX-1/COX-2 inhibitors also occur with the selective COX-2 inhibitors. Both types of NSAIDs can cause acute renal failure and hypertension (Cheng and Harris, 2004) during short-term or long-term use. All these drugs are especially likely to reinforce renal impairment in patients with pre-existing renal insufficiency. Patients with chronic kidney diseases, like those with heart failure, cirrhosis, dehydration, or any other co-morbidity causing activation of the sympatho-adrenergic and/or renin–angiotensin system, depend very much on a normal PG status for maintaining adequate function of the kidneys. Because the production of PGE₂ and PGI₂ depends on the activity of COX-2 (Qi et al., 2002), there is no difference in this risk between the non-selective COX inhibitors and the selective COX-2 drugs. Long-term treatment with either non-selective COX-1/COX-2 inhibitors or the selective COX-2 inhibitors also may significantly increase the risk of developing and maintaining renal failure, as may the use of NSAIDs with a long half-life (Henry et al., 1997).

Renal papillary necrosis induced by the non-selective COX-1/COX-2 inhibitors is well recognized and case reports also have described this complication with the selective COX-2 inhibitors (Whelton, 1999; Akhund et al., 2003). Recently, tubulointerstitial injury has been reported with COX-2 inhibitors (Ortiz et al., 2005). The vascular supply within the renal papillae is dependent on local renal PG production. Clinical conditions resulting in volume depletion or decreased RBF in association with NSAID ingestion may lead to elevated concentrations of NSAIDs and their metabolites within the papillae, which inhibit the vasodilatatory effect of the PGs and lead to renal papillary necrosis.

Congestive heart failure also may result from treatment with NSAIDs, particularly in medically frail and elderly patients. Patients admitted to the hospital diagnosed with congestive heart failure had a higher prevalence (17%) of recent NSAID use than patients admitted with other diagnoses (12%) (Page and Henry, 2000; Merlo et al., 2001). The risk of congestive heart failure increased with an additional history of heart disease

As mentioned earlier, TXA₂ is a potent vasoconstrictor and promotes platelet aggregation. TXA₂ is induced by the activity of COX-1 and, in effect, acts as a prothrombotic agent. On the other hand, PGI₂ causes vasodilatation, and thereby inhibits the aggregation of platelets. Selective inhibition of COX-2 increases the relative activity of TXA₂, which subsequently may facilitate the formation of thromboses. The literature relating to the risk of prothrombotic complications from NSAIDs, including myocardial infarction and stroke, is complicated, but supports the conclusion that all NSAIDs—including non-selective COX-1/COX-2 inhibitors and selective COX-2 inhibitors—could pose a risk of these complications as a result of their effects on COX-2. There appears to be important variation across drugs, however. In patients with coronary heart disease, a reduction of COX-2 activity produced by treatment with selective COX-2 inhibitors increase the risks for acute cardiovascular events such as strokes, thromboses, and myocardial infarction (Fitzgerald, 2004; Bresalier et al., 2005; McGettigan and Henry, 2006). However, some studies do not observe these outcomes during treatment with celecoxib (McGettigan and Henry, 2006). A meta-analysis of randomized trials including coxibs and non-selective NSAIDs taken for more than 4 weeks showed that coxibs and high-dose regimens of ibuprofen and diclofenac were associated with a moderate increase in the risk of vascular events (defined as myocardial infarction, stroke, or vascular death) (Kearney et al., 2006). However, high-dose naproxen was not associated with such events. Another recent systematic review confirmed that the relative risk of myocardial infarction varied with the individual NSAIDs. An increased risk was observed for diclofenac and rofecoxib; the latter drug had a dose–response trend, with risk higher for doses greater than 25 mg/day (78%) than for lower doses (18%) (Hernandez-Diaz et al., 2006). Rofecoxib was withdrawn from the market in September 2004.

Notwithstanding the limitations in the literature, it is clinically prudent to include prothrombotic cardiovascular risk as among the potential adverse effects produced by both the selective COX-2 inhibitors (Flavahan, 2007) and the non-selective COX-1/COX-2 inhibitors (Kearney et al., 2006; Warner and Mitchell, 2008). The ‘imbalance’ theory alluded to previously offer a reasonable explanation: any NSAID that reduces COX-2-dependent PGI₂ in endothelium without a commensurate effect on COX-1-dependent TXA₂ in platelets will predispose to a prothrombotic state by increasing the relative activity of TXA_2. All NSAIDs may be prothrombotic because all have an inhibitory effect on COX-2 (Cheng et al., 2002).

Given the risk profile of the NSAIDs there are important relative contraindications to their use in palliative care. Treatment should be undertaken cautiously, if at all, in patients at high risk for GI haemorrhage, or those who would be unlikely to cope with bleeding should it occur. The former group includes those with a haemorrhage within the last year, a history of peptic ulcer disease or NSAID-induced gastroduodenopathy in the past, advanced age, severe medical frailty, or concurrent treatment with a corticosteroid.

Patients with clinically significant renal insufficiency usually are considered to have too high a risk for NSAID therapy, and patients with liver disease and those predisposed to adverse cardiovascular outcomes should be considered to have a relative contraindication. The latter group includes those with a history of symptomatic atherothrombotic disease in the past (myocardial infarction, angina, stroke, transient ischaemic attacks, or symptomatic peripheral vascular disease) and those with significant risk factors, such as a history of poorly controlled hypertension, hyperlipidaemia, or smoking. Patients with a history of NSAID-induced asthma or a documented allergic reaction to any NSAID also should not generally be offered NSAID therapy. When the European Agency for the Evaluation of Medicinal Products reviewed a recent addition to the coxibs available in Europe, etoricoxib, it noted that the drug is contraindicated in cases of hypertension, severe hepatic dysfunction, inflammatory bowel disease, and congestive heart failure (European Medicines Agency, 2008).

NSAIDs also are associated with drug–drug interactions that may be particularly important in the medically ill. For example, they may reduce renal function during concomitant lithium, methotrexate, and amino-glycoside therapy, which can give rise to increased plasma concentrations. High plasma protein binding of NSAIDs can increase the concentration of free, active drug in some cases, such as phenytoin, and lead to unanticipated toxicity as a result. Patients who have a coagulopathy or are receiving anticoagulant therapy should be considered to have a strong relative contraindication to NSAID therapy. NSAIDs also may attenuate the metabolism of warfarin and thus increase its effect.

One should take into consideration that most of the studies of patients with advanced disease and short life expectancy typically involve small numbers of patients. Therefore, the studies investigating side effects are probably underestimating the frequency of such symptoms and signs. As a general rule in old and/or frail patients, polypharmacy should be limited if possible. Indications for the use of all drugs should be carefully considered by evaluating the probability of beneficial effects versus the probability of side effects. These considerations seem to be of utmost importance when using NSAIDS.

Topical NSAIDs, which may be applied via a rub-on solution, gel, or adhesive skin patch, offer obvious theoretical advantages by minimizing the systemic complications of these drugs. They have a relatively low bioavailability compared with oral NSAIDs and this may account for their favourable safety profile. Topical patches provide a fixed constant dose and local action. Gel formulations can be used on parts of the body that are not easily accessible by a patch, such as the fingers. Side effects tend to be localized to the site of application, such as itching and rashes.

Topical NSAIDs have been recommended for use in osteoarthritis (National Institute for Health and Clinical Excellence, 2010). After topical application, therapeutic levels of NSAIDs can be demonstrated in synovial fluid, muscles, and fasciae, and this finding suggests that they may have their pharmacological effects on both intra- and extra-articular structures. Although it is assumed that their mechanism of action is similar to that of oral NSAIDs, the topical drugs produce a maximal plasma NSAID concentration of only 15% of that achieved following oral administration of a similar dose (Lin et al., 2004). There have been no studies of topical NSAID therapy in palliative care and their efficacy for localized pain associated with non-musculoskeletal causes is unknown. The potential role of topical NSAIDs needs to be explored in palliative care.

ASA and its derivatives are the prototypes of NSAIDs and have been known since ancient times. In the mid-eighteenth century, Edward Stone, an English clergyman, wrote the first scientific description of the effects of willow bark.

ASA covalently and irreversibly inhibits both COX-1 and COX-2. This is an important feature of ASA, as the duration of the effects are related to the turnover rate of COX in the different tissues. Platelets are particularly susceptible to ASA-mediated irreversible inhibition of COX, because they have limited capacity for protein synthesis and thus cannot regenerate COX enzymes. This means, that a single and small dose of ASA will inhibit the COX enzymes for the lifetime of the platelets (Patrono et al., 2004).

The prevalent use of low-dose ASA for cardioprotection must be considered when NSAIDs are administered for pain. The combination of a non-selective COX-1/COX-2 inhibitor and low-dose ASA increases the risk of GI toxicity. The combination of a selective COX-2 inhibitor and low-dose ASA reduces the relative benefit of the coxib on GI risk, particularly if treatment lasts more than 6 months. Some NSAIDs, such as ibuprofen, may attenuate the antithrombotic effects of ASA and reduce its cardioprotective effects. Because the latter interaction would be less likely to occur if the NSAID is taken after (perhaps 2 or more hours) the dose of ASA, patients should be instructed to do so. Paracetamol has no adverse interaction with low-dose ASA therapy (Gaziano and Gibson, 2006).

The results of systematic reviews of single-dose studies confirm ASA’s efficacy as an analgesic. Analysis of the analgesic dose–response tells us that higher doses give more analgesia, and comparing ASA with paracetamol the analgesia produced by the two drugs are very similar. One gram of ASA gives the same analgesia as 1 g of paracetamol (Gaziano and Gibson, 2006; McQuay and Moore, 2007). Furthermore, significant benefit of ASA over placebo was shown for ASA 600–650 mg, 1000 mg, and 1200 mg, with NNT values for at least 50% pain relief of 4.4, 4.0, and 2.4, respectively. However, in single-dose studies even low doses of ASA 600/650 mg produced significantly more gastric irritation and drowsiness than placebo (Edwards et al., 1999). The GI side effects of ASA are the main reason for the limited use of ASA in palliative care.

Paracetamol (acetaminophen; N-acetyl-p-aminophenol) is one of the most commonly used analgesic and antipyretic drugs worldwide, and it is widely available by prescription and over the counter. Paracetamol, a so-called coal-tar analgesic, was discovered by accident as an active metabolite of phenacetin. The site of action seems to be in the brain; however, the mechanism of action is still poorly understood. In a double-blind, placebo-controlled study in healthy volunteers, Piletta and colleagues obtained some evidence for a central analgesic action of paracetamol. Application of a transcutaneous electrical stimulus to the sural nerve caused a flexion reflex and a subjective sensation of pain. In contrast to ASA, paracetamol raised the threshold to both types of pain, indicating an analgesic action both at the spinal cord level and in higher centres (Piletta et al., 1991).

Although there is considerable overlap in the actions of paracetamol, NSAIDS, and ASA regarding inhibition of COX, in vitro studies show that paracetamol is not a highly potent inhibitor of COX-1 and COX-2 (Ouellet and Percival, 2001). As mentioned earlier, a third form of COX has been proposed, which has been referred to as COX-3 (Ayoub et al., 2006). However, cloning studies have failed to confirm the existence of COX-3 in humans and the search for a convincing mechanistic explanation of the therapeutic activity of paracetamol continues (Qin et al., 2005).

Paracetamol also possesses antipyretic activity and the brain is likely to be the site of its antipyretic effects. The antipyretic activity of the compound resides in its aminobenzene structure and the effect seems to be caused by inhibition of COX-2 or a variant of this enzyme (Simmons, 2003). Because of the association between ASA and Reye’s syndrome in children, paracetamol is the antipyretic drug of choice in children, and when used in recommended doses, has few side effects, and is remarkably well tolerated.

Paracetamol is commercially available in a considerable number of products, both alone and in combination with other drugs. It can be administered orally as tablets (conventional, sustained release, effervescent), capsules, powders, and elixirs, and can be given rectally as suppositories. Paracetamol is rapidly and almost completely absorbed from the GI tract. Gastric emptying rate rather than the diffusion across the intestinal mucosa is the rate-limiting step in paracetamol absorption after oral administration. Therefore, any drug, disease, or other condition that alters the rate of gastric emptying will influence the rate of absorption. Although paracetamol is rapidly absorbed from the GI tract, it is incompletely available to the systemic circulation due to hepatic first-pass metabolism accounting for a 10% loss in therapeutic doses. Intestinal metabolism also contributes to decreased bioavailability (Tone et al., 1990). In adults, the bioavailability of paracetamol after administration of suppositories is approximately 60% (Beck et al., 2000). After oral administration of therapeutic doses, the concentration in plasma reaches a peak in 30–60 minutes, and the half-life in plasma is about 2 hours. Paracetamol is relatively uniformly distributed throughout most body fluids (Prescott 1996). The proportion of paracetamol bound to plasma proteins is small and varies from 5% to 20% (Milligan et al., 1994). Biotransformation takes place primarily in the liver and the oxidative reactions via the cytochrome P450 system are followed by conjugation. After therapeutic doses, 90–100% of the drug may be recovered in urine within the first day, primarily after hepatic conjugation with glucuronic acid (about 60%), sulphuric acid (about 35%), or cysteine (about 3%); small amounts of hydroxylated and deacetylated metabolites also have been detected (Steventon et al., 1996).

The lack of dichotomous data in randomized trials, and a lack of trials in medically ill populations, largely precludes calculation of a NNT for paracetamol use in patients with serious illness. Studies also have not evaluated long-term efficacy. Nonetheless, evidence suggests the use of non-opioids alone is superior to placebo for mild pain in cancer and medical illness, at least during short-term treatment (McNicol et al., 2005).

The efficacy of paracetamol compared to NSAIDS has not been thoroughly investigated (Ventafridda et al., 1990) and the combination of paracetamol and NSAIDS has not been investigated in cancer-related pain. Most studies have assessed paracetamol in combination with other analgesics, and these studies do not give any information on the analgesic effect that paracetamol provides on its own (Carlson et al., 1990; Chary et al., 1994). Only one randomized, double-blind, placebo controlled study indicated that the addition of paracetamol to ongoing oral opioid therapy improved pain relief and general well-being in cancer patients (Stockler et al., 2004).

At therapeutic doses of paracetamol, side effects are rare, and the clinical advantage during long-term treatment is that the side effects are less severe than with the NSAIDS. Allergic reactions have been described, and during long-term treatment chronic headache may occur (Meskunas et al., 2006). Toxicity from paracetamol usually is due to either accidental or deliberate overdose. A small proportion of paracetamol undergoes P450-mediated N-hydroxylation to form N-acetyl-benzoquinoneimine, a highly reactive intermediate metabolite. This metabolite normally reacts with sulfhydryl groups in glutathione. At large doses of paracetamol (usually considered in those without liver disease to be a single dose > 10 g), the metabolite is formed in sufficient amounts to deplete liver cells completely of glutathione, which seems to trigger hepatotoxicity and a prolonged rise in liver-derived transaminase and alkaline phosphatase levels in the serum. Intervention to sustain hepatic glutathione is an effective treatment for paracetamol overdose and administration of N-acetyl-L-cysteine, which replenishes glutathione stores, remains the treatment of choice (Josephy, 2005).

Dipyrone is a popular medicine for pain relief in many countries and is used to treat postoperative pain, colic pain, cancer pain and migraine. In other countries (e.g. United States, United Kingdom, and Japan), however, the drug has been removed from the market, or not approved, due to concerns about serious adverse effects. Although the data are insufficient to draw any conclusions about the influence of dose or route of administration, dipyrone has been associated with potentially life-threatening blood disorders, such as agranulocytosis.

A single 500 mg oral dose of dipyrone provides at least 50% pain relief to adults with moderate or severe postoperative pain, with efficacy similar to ibuprofen 400 mg. A single 2.5 g intravenous dose is equivalent to 100 mg intravenous tramadol for at least 50% pain relief. In efficacy studies, no serious events or adverse event withdrawals were reported (Edwards et al., 2010). A small controlled trial indicated that dipyrone adds significantly to the analgesic effect of morphine in patients with cancer-related pain (Duarte et al., 2007).

NSAIDs have important risks and should be used cautiously in the medically ill. There are many factors that increase the risk of GI, renal, or cardiovascular toxicity, and populations with serious or life-threatening illness are likely to have one or more of these factors. Caution must be exercised in elderly or medically frail patients and those with significant disorders affecting liver or kidney. Risk factors for GI haemorrhage, such as peptic ulcer disease or Helicobacter pylori infection, and risk factors for cardiovascular disease, such as prior coronary artery disease, uncontrolled hypertension, diabetes, and others, present relative contraindications to the long-term use of NSAIDs. The risk of these drugs increases when the potential for pharmacokinetic or pharmacodynamic drug–drug interactions occur, and co-treatment with corticosteroids, low-dose ASA, or anticoagulants, for example, should further increase caution in the use of these drugs.

Patients at high risk for adverse effects during long-term oral NSAID therapy should be considered for other strategies, such as paracetamol, topical NSAIDs, adjuvant analgesics, and opioids. When treatment with NSAIDs is indicated and effective, it is strongly recommended to use the lowest effective dose for the shortest period of time, and to combine the treatment with a gastroprotective agent (usually a PPI; misoprostol and higher-dose H₂ antagonists are alternatives). Unless otherwise contraindicated for other reasons, a COX-2-inhibitor, such as celecoxib, should be strongly considered. Patients with advanced illnesses are often at high risk and either lowest effective dose of an NSAID combined with PPI or the lowest effective dose of a coxib should be considered.

Cardiovascular risks are associated with COX-2 inhibition, and consequently, are concerns during treatment with all NSAIDs. Although the data are inconclusive, there do appear to be important drug-selective effects; naproxen, for example, may pose a lesser risk than other NSAIDs, and some studies suggest a low risk for celecoxib. Both relatively higher doses and longer treatment durations seem to increase the cardiovascular risks. Based on current evidence, patients at high cardiovascular risk should be treated with either naproxen (plus PPI) or celecoxib, and the use of the lowest effective dose for the shortest possible duration of time is recommended.

Generally, the efficacy and side effects of NSAIDs should be monitored and evaluated during long-term treatment. If effectiveness is doubtful, the drug should be discontinued.

There remains a need for a substantial increase in the number of high-quality trials of non-opioid analgesics in patients with serious or life-threatening illnesses, such as cancer. Studies that specifically address the question of whether addition of a non-opioid to an opioid analgesic regimen actually increases efficacy and/or reduces side effects are required. In addition, the safety and efficacy of chronic use of non-selective COX-1/COX-2 inhibitors versus the coxibs needs to be established. Emerging questions about the selective COX-2 inhibitors, such as their potential anti-angiogenic properties, remain unanswered (Steinbach et al., 2000). Translation of emerging preclinical insights into distinct mechanisms for pain of different aetiologies (e.g. bone metastases) would be another step towards more rational pharmacotherapy (Davar, 2002).

Due to the potential for adverse effects encountered during NSAID therapy, combined with a very economically lucrative market, this is an area of continuing major investment in research. The COX-inhibiting nitric oxide donors (CINODs) are a new class of agents designed for the treatment of pain and inflammation. CINODs have a multi-pathway mechanism of action that involves COX-inhibition and nitric oxide donation. The anti-inflammatory and analgesic effects of COX-inhibition are reinforced through inhibition of caspase-1 regulated cytokine production, while nitric oxide donation provides multi-organ protection. The CINODs are devoid of hypertensive effects in animal models and their mechanism of action suggests that they may not cause oedema. CINODs also have other renal-sparing effects, being better tolerated than NSAIDs in models of kidney failure. CINODs have been shown to prevent platelet activation in vitro and exhibit antithrombotic activity in vivo. In animal models of ischaemia/reperfusion treatment with CINODs results in improved recovery of heart contractility and reduced left ventricular end-diastolic pressure, in contrast to the effects of NSAIDs and ASA. Naproxcinod has been the first, and so far the only, CINOD investigated in clinical trials. These studies have shown a slight improvement in GI tolerability in comparison to naproxen in surrogate endpoints (number of gastric and duodenal ulcers) and a significant reduction in the risk of destabilization of blood pressure control in patients with osteoarthosis taking antihypertensive medications in comparison to either naproxen and rofecoxib (Baerwald et al., 2010; White et al., 2011). The lack of outcome studies, however, has precluded the approval of naproxcinod by the US Food and Drug Administration.

NSAIDs that release H₂S as a mechanism may have greater GI and cardiovascular safety, and are being investigated in preclinical models. Both naproxen and diclofenac hybrids have been reported to cause less GI injury than parent NSAIDs. These novel chemical entities exert a variety of beneficial effects in rodent models of cardiovascular and metabolic disorders through a mechanism that might involve the release of H₂S and/or by exerting antioxidant effects. The beneficial role these mechanisms in clinical settings await a proof-of-concept study (Fiorucci and Distrutti, 2011).

The short-term use and benefits of topical NSAIDs have been reflected in guidelines for the treatment of osteoarthritis. The evidence with regards to efficacy of these drugs is predominantly limited to the osteoarthritis and rheumatoid arthritis populations (National Institute for Health and Clinical Excellence, 2010). Their efficacy needs to be further investigated in chronic cancer pain conditions of musculoskeletal origin.